I'i,~ic Di~fIDA~~,hani'$ ()·w ~ MA:RG VICTO,lI ,,"',-.i,"" a THE '!' ~ i '. BASIC BIOMECHANICS of the MUSCULOSKELETAL SYSTEM Margareta Nordin, P.T., Dr. Sci. Director. Occupational and Industrial Orthopaedic Center (OIOC) Hospital fOf Joint Diseases Orthopaedic Institute Mt. Sinai NYU Health Program of Ergonomics and Biomechanics, New York University Research Professor Department of Orthopaedi" and Environmental Heallh Science School of Medicine. New York University New York. New York Victor H. Frankel, M.D., Ph.D., KNO President Emeritus Hospital for Joint Diseases OrthopaedIC Institute Professor of Orthopaedic Surgery New York University School of Medicine New York, New York Dawn Leger, Ph.D., Developmental Editor Kajsa Forssen, Illustrator Angela Lis, P.T., M.A.. Editorial Assistant $~~ LipPINCOIT WILLIAMS & WILKINS • A wohl'r! Klllwtr (ornp.lny l'hil.ldclphi'l • B"llimore • New"ork . london [lucllOS Ailes' Hong Kong· Sydney· T{lkyo / foreword Mechanics and biology have always fascinated humankind, The irnportance or understanding the biomechanics of the musculoskeletal system cannot be underestimated, Much a([entian has been paid in recent years to genetic and biomoleClilar research, but the stlld~' of the mechanics of structure and of the whole body s}'stcm is still of immense importance. Musculoskeletal ailments arc among the most prevalent disorders in the wodd and will continue to grow as the population ages. Since the days when I first studied biomechanics in Sweden with Carl Hirsch, through my years as an orthopaedic surgeon, teacher, and researcher, I have alway's emphasized combining basic and applied research with clinical experiencc, This text represents my fifth effort to integrate biomechanical knowledge into clinical training for patient carc. It is not a simple task but by relating the basic concepts of biomechanics to everyday life, rehabilitation. orthopaedics. traumatology, and patient care are greatly enhanced. Biomechanics is a multidisciplinary specialty, and so we have made a special effort to invite contributors from many disciplines so that individuals from dilTerent fields may feel comFortable reading this book. Together with an invaluable team, Margareta Nordin and I have produced Ihis third edition of Basic Biol1lechanics oFthe A'lusclt!o.\'keletal Systelll, The new edition is shall1cncc! and improved thanks to the input from the students and resi· dents in orthopaedics Ihal during the past 10 years have llsed the texl. This book is written for students and with a major input from students and will hopefully be used to edunHe students and residents for many ~Iears to come. Although the basic information contained in the book remains largely unchanged, a considerable amount of extra information has been provided throughoul. ~Ve have also made a special point to document with the key references any significant changes in the field of biomechanics and rehabilitation. It has always been m)' interest lO bridge the gap between engineering knowledge and clinical carc and praclice, This book is written primarily for clinicians such as orthopaedists, physical and occupational therapists, clinical ergonomisls, chiropractors, and olher health professionals who arc acquil-ing a working knowledge of biomechanical principles for use in the evaluation and treatment of musculoskeletal dysfunction. We only hope that if you find this book int~resting. you will seek more in-depth study in the field of biomechanics. Enjo\' it. discuss it, and become a beller clinician and/or researchcl: Vve are extremely proud that Basic BiomeclUluics oj" the !\tlllscliioskeic/lli Sysle111 has been designated "A Classic" by the publishers, Lippincott Williams & Wilkins. We Ihank the readers, students, professors, and all who acquire thc tcxt and lise it. VielO,. H. Frallkel, M.D., Ph.D., KNO vii l'~ e"!,-g,! !,•.-~--~"".'!~~~.""~"""c""',:::~::""'-'''''''';'-:'''"''''.''''''·'''''--'''''--''''''''''':~'''i~!1'···"'-"-"'·E"'.r,;>d"._""."" .....,. ~"-h_,,:_/>;:~B:>~'¥,,..'13iA:s"' *w.c-:;·_!l , _~_iIl~!il i.iIl=i1!~i1!i!i!lWl;41.4~l'l!~$~rm!1il·, ~·" "'1l i&~.l\l.l!l~t!%_"t81; . ., Preface ever possible. We retained the selected examples to illustrate lhe concepts needed for basic knowledge of the musculoskeletal biomechanics; we also have kept the important engineering concepts throughout the volume. We have added four chapters on applied biomechanics topics. Patient case studies ancl calculation bo:'\cs have been added to each chapter. \Ne incorporated flowcharts throughout the book as teaching tools. The text will serve as guide to a deeper understanding of musculoskeletal biomcchanics gained through funher reading and independent research. The information presented should also guide the rt.'ader in assessing the literature on biomechanics. "Vc have attemptcd to provide therapcutic examples but it was not our purpose to cover this area; instead, \ve have described the undel'lying basis for rational therapcutic or exercise programs. An introductory chapter describes the inlporlance of the study of biomechanics, and an appendix on the international system of measurements serves as an introduction to the physical measurements used throughout the book. The reader needs no more (han basic knowledge of mathematics to fully comprehend the material in the book, but it is important to review the appendix on the Sl System and its application to biomechanics. The body of the third edition is then divided into three sections. The first section is the Biomechanics of Tissues and Stnlcturcs of the Musculoskeletal System and covers the basic biomechanics of bone, ligaments, cartilage. tendons, muscles, and nenres. The second section covers the Biomechanics-of Joints, including every joint system in the human body. Chapters range from the foot and ankle through the cervical spine, and co\'er eve I:" joint in between. The third sec- Biomechanics uses physics and engineering concepts to describe the motion undergone by the various body segments and forces acting on these body parts during normal activities. The inter-relationship of force and motion is important and must be understood if rational treatment programs are to be applied to musculoskeletal disorders. Deleterious effects may be produced if the forces acting on the areas with disorders rise 10 high levels during exercise or other activities of daily living. The purpose of this text is to acquaint the readers with the force-motion relationship \vithin the musculoskeletal system and the various tech~ niqucs llsed to understand these relationships. The third edition of Basic Biol1/eclwllics of rite iHllScliloskeleral System is intended for use as a textbook either in conjunction with an introductory biomechanics course or for independent study. The third edition has been changed in many ways, but it is still a book that is designed for use by students who are interested in and want to learn about biomechanics. It is primarily written for students who do not have an engineering background but who want to understand the most basic concepts in biomechanics and physics and how these apply to the human body Input from students has greatly improved this third edition. We have used the book for 10 years in the Program of Ergonomics and Biomechanics at New York University', and it is the students and residents who have suggested the changes and who have continuously shown an interest in developing and irnproving this book. This edition has been further strengthened by the contribution or the students over the past year. \Vc formed focus groups to understand better what thc students wanted and applied their suggestions wher- e .-.. ---~-_.-- . _--.-._.. _------- .-.- . ----.. -.-.... -'------------~--- .. ,~".-- . . -.. ..- ...-.--..-------.. ~.~ ~ -.,,"------~----- -",,--.-.---.------ ~ tion cover~ some topics in Applied Biomechanics, including chapters on fracture fixation; arthroplasty; silting, standing and lying; and gait. These arc basic chapters that sl:l\r e to intra· c1uce topics in applied biomechanics: they arc not in-depth explorations of the subject. Finally. we hope that the revision and expan- oFthe kluscu/oskelela/ Syslel1l will bring about an increased awareness of the imparlance of biomechanics. II has never been our intention to complL'tely cover the subject, but instead provide a basic introduction to the field that will lead to further study or this important lopic. sion of this third edition of" Basic 13io11leclulJIics Margarela NOf(!;11 alld Viclor H. Frankel ""at ~.,,;,~~~~ :,,~,~t, Acknowledgments _--.. ---------~ I i ! !Ii I ! ! I I ; This book was made possible through the outstanding contributions of many individuals. The chapter authors' knowledge and understanding of the basic concepts of biomechanics and their wealth of experience have brought both breadth and depth to this work. Over the past 10 years. questions raised by students and residents have made this book a better teaching tool. The Third Edition could not have been done without the students who have shared their cornmen(s and really sCnItinizcd thc Second Edition. There arc too many names LO list here, but we thank each student who asked a question or made a suggestion during the course of his or her studies. Special thanks to the students who panicipated in several focus groups. whose input was invaluable in finalizing the contents and design of the text. Vve are honored and grateful for the contributions of everyone who has worked to prepare this new edition. 'vVe can honestly say that this third edition is written ror the sludent and by students and residents who leave the classroom with the knowledge to enhance our life and existence. A book of this size with its large number of figures, legends, and references cannot be produced without a strong editorial team. As project editor, Dawn Leger's continuous effort and perseverance and thoughtfulness shines through the entire book. She has contributed not just to the editing but also to logistics, and as a stylist, as an innovator, and a friend. Our editorial assistant, Angela Lis, is a physical thcrapisl and recent recipient of the MA degree in Ergonomics and Biomechanics from NYU. As a recent graduate, Angela was also a recent USCI' of the book, and she devoted several months to help finalize this edition. She created the flowcharts and scrutinized all the figures, patient cases, and caku- -1.'- • JUlian boxes. Angela look this book to her hear and \ve arc all the bettcr for her passion an attention to detail. The illustrator: Kajsa Forssen. has now worke on all three editions of this text. Her never-failin grasp of hiomechanical illustrations, her simp city and exactness of figures, is always appr ciated. In drawing all the figures and graphs, sh considers how they would translate into a slide into a computer-generated presentation. Kaj Forssen is one of the top iIIustralOrs that we hav ever worked wilh, and she has been an importa member of the publication (cam. This book was also made when publicatio companies I11ergcd and merged again, and in th end we are deeply grateful to Ulila Lushnyck who has with her team at Lippincott \·Villiams Wilkins been responsible ror the production. Sh has worked with tremendous energy and posiliv thinking, put the book together in record spce and we fonvard our sincerest gratitude to he \Ne are also thankful for a development gra provided by Lippincott Williams & Wilkins finance this effort. Our colleagues al the Occupational an Industrial Orthopaedic Cenler and the Depar ment or Orthopaedics of the Hospital ror Joi Diseases Orthopaedic Institule functioned critical reviewers and contributors to th chapters. Special thanks is extended to Dav Goldsheydcl" for assislance in reviewing lh biomechanical calculation boxes. to Marc Campello as a contributor and reviewer, and Shira Schccter-vVeiner for contributing to th spine chapteI: Much thanks to Dr. Mark Pitma 1'01' supplying vital x-rays ror the new edition. \,V are parlicularly grateful to DI: Markus Pielr for contributing with the latcst on intr abdominal pressure. to Dr. Ali Sheikhzadeh r reviewing chapters and contributing ne • references, to Dr. Tobias Lorenz for his work on the first section, and to all other staff at the Occupational and Industrial Orthopaedic Center who have been managing the center while we are absorbed wilh the book. \'\Fe arc most grateful to Drs. Bejjani, Lindh, Pitman, Peterson, and Stuchin for their COI1l1·j· bUlions to the second cdition which sen'cd as a framework for the updated third edition. The third edition of Basic Biomechallics orEiIe iHuscllloskeletal System was supported throughout its production by the Research and Development Foundation of the Hospital for Joint Diseases Orthopaedic Institute and the hospital administration, to whom we forward our sincere gratitude. To all who helped, we say' again, thank yOLi and TACK SA MYCKET. klargareta Nordin and Victor fl. Frankel Contributors Gunnar B. J. Andersson, M.D., Ph.D. Professor and Chairman Department of Orthopaedic Surgery Rush-Presbyterian-SI. Luke's Medical Center Chicago, IL Thomas P. Andriacchi, Ph.D. Biomechanical Engineering Division Stanford University Stanford, CA Victor H. Frankel, M.D., Ph.D., KNO President Emeritus Sherry I. Backus, M.D., P.T. Senior Research Physical Therapist and Research Associate Motion Analysis laboratory Hospital for Special Surgery New York, NY Ann E. Barr, Ph.D., P.T. Physical Therapy Department College of Allied Health Professionals Temple University Philadelpllla, PA Fadi Joseph Bejjani. M.D .. Ph.D. Director of Occupational Musculoskeletal Diseases Department University Rehabilitation Association Newark, NJ Maureen Gallagher Birdzell, Ph.D. Departmenl of Orthopaedic Surgery Hospital for Joint DiseasesiMI. Sinai NYU Health New York, NY Marco Campello, P.T., M.A. Associate Clinical Director Occupational and Industrial Orthopaedic Center Hospital for Joint DiseasesiMI. Sinai NYU Health New York. NY Dennis R. Carter. Ph.D. Professor Biomechanical Engineering Program Stanford University Stanford, CA -_. __._---._._-_._--_._ Hospital for Joint Diseases Orthopaedic Institute Professor of Orthopaedic Surgery New York University School of Medicine NevI York, NY Ross Todd Hockenbury, M.D. River City Ortl1opaedic Surgeons LouiSVille, KY Assistant Professor ~.- Craig J, Della Valle, M.D. NYU-HJD Department of Orthopaedic Surgery Hospital for Joint Diseases School of Medicine New York University New York, NY _._._.- Clark T. Hung, Ph.D. Assistant Professor Department of Mechanical Engineering and Center for Biomedical Engineering Columbia University New York, NY Debra E. Hurwitz. Ph.D. Assistant Professor Department of Orthopaedics Rush·Presbyterian-St. Luke's lvIedical Center Chicago, IL Laith M. Jazrawi. M.D. NYU·HJD Department of Orthopaedic Surgery Hospital for Joint Diseases School of Medicine New York University New York, NY Frederick J, Kummer, Ph.D. Associate Director, Musculoskeletal Research Center Hospital for Joint DiseasesiMt. Sinai NYU Health Research Professor, NYU-HJD Department of Orthopaedic Surgery Scl100l of Medicine New York University New York, NY _._-_._._._-_. _.-.._.-._-_ _._._-_._ _ _-_._-- _ __ _._. xiii Dawn Leger, Ph.D. Adjunct Assistant Professor NYU-HJD Department of Orthopaedics School of Medicine New York University New York, NY Jane Bear-Lehman, Ph.D., OTR, FAOTA Assistant Professor of Clinical Occupational Therapy Department of Occupational Therapy Columbia University College of Physicians and Surgeons New York. NY Margareta Lindh, M.D., Ph.D. Associate Professor Department of Physical MeeJicine and Rehabilitation Sahlgren Hospital Gothenburg University Gothenburg, Sweden Angela Lis, M.A., P.T. Research Physical Therapist Occupational and Industrial Orthopaedic Center Hospital for Joint DiseasesiMt. Sinai NYU Health New York, NY Associate Professor Physical Therapy Program (orporacion Universitaria Iberoamericana Bogota, COLOMBIA Tobias Lorenz, M.D. Fellow Occupational and Industrial Orthopaedic Center Hospital for Joint Diseases/Me Sinai NYU Health New York, NY Goran Lundborg, M.D. Professor Department of Hand Surgery Lunds University Malmo Allmanna Sjukhus Malmo, Sweden Ronald Moskovich, M.D. Associate Chief Spine Surgery NYU-HJD Department oj Orthopaedic Surgery Hospital for Joint Diseases School of Medicine New York University New York, NY Van C. Mow, Ph.D. Director Orthopaedic Research Laboratory Department of Orthopaedic Surgery Columbia University New York, NY Robert R. Myers, Ph.D. Associate Professor Department of Anesthesiology University of California San Diego La Jolla, CA Margareta Nordin, P.T., Dr. Sci. Director, Occupational and Industrial Orthopaedic Center (OIOC) Hospital for Joint Diseases Orthopaedic Institute !vlt. Sinal NYU Health Program of Ergonomics and Biomechanics New York University Research Professor Department of Orthopaedics and Environmental Health Science School of Medicine, New York University New York, NY Kjell Olmarker, M.D., Ph.D. Associate Professor Department of Orthopaedics Sahlgren Hospital Gothenburg University Gothenburg, Sweden Nihat bzkaya (deceased) Associate Professor Occupational and Industrial Orthopaedic Center Hospital for Joint Diseases Research Associate Professor Department of Environmental Medicine New York University NelN York, NY Lars Peterson, M.D., Ph.D. Gruvgat 6 Vastra Frolunda Sweden Mark I. Pitman, M.D. Clinical Associate Professor NYU-HJD Department of Orthopaedic Surgery School of lvIedicine New York University New York, NY Andrew S. Rokito, M.D. Associate Chief. Spons Medicine Service ASSistant Professor NYU-HJD Department 'Of Ortllopaedic Surgery School of Medicine New York University New York, NY Bjorn Rydevik, M.D., Ph.D. Professor and Chairman Department of Orthopaedics Sahlgren Hospital Gothenburg University Gothenburg. Sweden G. James Sammarco, M.D. Program Director fellowship in Adult Reconstructive Surgery foot and Ankle Orthopaedic Surgery Program The Center for Orthopaedic Care, Inc. Volunteer Professor of Orthopaedic Surgery Department of Orthopaedics University of Cincinnati Medical Center Cincinnati, OH Chris J. Snijders, Ph.D. Professor Biomedical Physics and Technology faculty of Medicine Erasmus University Rotterdam, The Netherlands Steven Stuchin, M,D. Director Clinical Orthopaedic Services Director Arthritis SefYice Associale Professor NYU-HJD Department of Orthopaedics School of Medicine New York University New York, NY Shira Schecter Weiner, M.A., P.T. Research Physical Therapist Occupational and Industrial Orthopaedic Center Hospital for Joint Diseases/Mt. Sinai NYU Health New York. NY Joseph D. Zuckerman, M.D. Professor and Chairman NYU-HJD Department of Orthopaedic Surgery Hospital for Joint Diseases School of Medicine New York University New York, NY Contents Introduction to Biomechanics: Basic Terminology and Concepts 2 Biomechanics of the Foot and Ankle 222 Niha! bzkaya, Dawn Leger G. James Sammarco, Ross Todd Hockenbury Appendix 1: The System International d'Unites (SI) 18 (11) Biomechanics of Bone 2 Margareta Nordin, Shira Schecter Weiner, adapted from Margareta Lindh Dennis R. Carter trw ~ Biomechanics of Tissues and Structures of the Musculoskeletal System Biomechanics of the Lumbar Spine Biomechanics of the Cervical Spine 2 Ronald Moskovich (f} Biomechanics of the Shoulder 318 Craig J. Della Valle. Andrew S. Rokito. Mauree Gallagher Birdzell, Joseph D. Zuckerman Biomechanics of the Elbow 26 340 Laith M. Jazrawi, Andrew S. Rokito, Maureen Gallagher Birdzell, Joseph D. Zuckerman Victor H. Frankel, Margareta Nordin Biomechanics of Articular Cartilage 60 Biomechanics of the Wrist and Hand 358 Van C. Mow, Clark T. Hung Ann E. Barr, Jane Bear-lehman adapted from Steven Stuchin, Fadi J. Bejjani Biomechanics of Tendons and Ligaments 102 Margareta Nordin, Tobias Lorenz, Marco Campello e Applied Biomechanics Biomechanics of Peripheral Nerves and Spinal Nerve Roots 126 Bjorn Rydevik, Goran lundborg, Kjell Olmarker, Robert R. Myers Biomechanics of Skeletal Muscle ~ Introduction to the Biomechanics of Fracture Fixation 390 Frederick J. Kummer 148 Biomechanics of Arthroplasty Tobias lorenz, Marco Campell0 adapted from Mark 1. Pitman, Lars Peterson ([$ ~.. , Biomechanics of Joints e o Biomechanics of the Knee 400 Debra E. Hurwitz, Thomas P. Andriacchi, Gunn B.J. Andersson Engineering Approaches to Standing, Sitting, and Lying 420 Chris J. Snijders 176 ED Biomechanics of Gait Margareta Nordin, Vietor H. Frankel Ann E. Barr, Sherry l. Backus Biomechanics of the Hip Index 202 438 459 Margareta Nordin, Victor H. Frankel x BASIC BIOMECHANIC of the MUSCULOSKELETA SYSTE -~ '" I Introduction to Biomechanics: Basic Terminology and Concepts Nihat OZkaya, Dawn Leger Introduction Basic Concepts Scalars, VeCtors, and Tensors Force Vector Torque and Moment Vectors Newton '$ l.aws Free-Body Diagrams Conditions for Equilibrium Statics Modes of Deformation Normal and Shear Stresses Normal and Shear Strains Shl?ar·5train Diagrams Elastic and Plastic Deformations Viscoelasticity Material Properties Based 011 Stress-Strain Diagrams Principal Stresses Fatigue and Endurance Basic Biomechanics of the Musculoskeletal System Part I: Biomechanics of Tissues and Structures Part 11: Biomechanics of Joints Part III: Applied Biomechanics Summary Suggested Reading Introduction Biomechanics is considered a branch of bioengineering and biomedical engineering. Bioengineering is an interdisciplinary field in which the principles and methods from engineering. basic sciences. and technology arc applied to design. test, and manufacture equipment for use in medicine and to understand, define, and solve problems in physiology and biology!. Bioengineering is one of several specialty areas that corne under the general field of biomedical engineering. Biomechanics considers the applications of classical mechanics 10 the analysis of biological and physiological svstems. Different aspects of biomechanics utilize different parts or applied mechanics. For example, the principles of statics havc been applied to analyze the magnitude and nature of forces involved in various joints and muscles of the nUlsculoskeletal system. The principles of dynamics have been utilized for motion description, gait analysis, and segmental motion analysis and have many applications in sports mechanics. Thc mc~ chanics of solids provides the necessary tools for developing the field constitutive equations For biological systems that are used to evaluate their functional behavior under dilTerent load conditions. The principles of fluid mechanics have been used to investigate blood flow in the circulatory system, air flow in the lung, and joint lubl'ication. Research in biomechanics is aimed at improving our knowledge of a vcry complex structure-the human body. Research activities in biomechanics can be divided into three areas: experimcntal studies, model analyscs, and applied research. Experimental studies in biomechanics arc done to determine the mechanical properties of biological materials, in~ eluding the bone, cartilage, muscle, tendon, ligament. skin, and blood as a whole or as parts constituting them. Theoretical studies involving mathematical model analy1ses have also been an im~ ponant component of research in biomechanics. In general, a model that is based on experimental findings can be used to predict the efrect of environmental and operational factors without resorting to laboratory experiments. Applied research in biomechanics is the application of scientific knowledge to bcnefit human bcings. vVe know that musculoskeletal injury and illness is one of the primary occupational hazards in industrialized countries. By learning how the musculoskeletal system adjusts to common work conclitions and by developing guidelines to assure that manual work conforms more closely to rhe physica limitations of the human body and to natural bod rnO\'cmCnlS, these injuries rnay be combatlcd. Basic Concepts Biomechanics of the musculoskeletal system r quires a good understanding of basic mechanic The basic terminology and concepts from mechan ics and physics arc utilized to clcscribe intcrn forces of the human body. The objective of studyin thcs~ forces is to understand the loading conditio of soft tissues and their mechanical responses. Th purpose of this section is to rC\'jew the basic con cepts of applied mechanics that are used in biome chanical literature and throughout this book. SCALARS, VECTORS, AND TENSORS Most of the concepts in mechanics arc either scal or vector. A scalar quanlity has a magnitude onl Concepts such as mass, energy', power, mechanic work, and temperalure are scalar quantities. For e ample, it is suffkicnt to say that an object has 8 kilograms (kg) of mass. A vector quanlity, con versely, has both a magnitude and a direction ass ciated \vith it. Force, moment, velOcity, and accele ation arc exall'lples of vector quanlities. To describ a force fully. one must state how much force is a plied and in which direction it is applied. The ma nitude of a vector is also a scalar quantity. The ma nitude of any quanlity (scalar or vector) is always positi\'c number corresponding to the numeric measure of that quantity_ Graphically, a vector is represented by an arrow The orientation of the alTow indicates the line of a tion and the arrowhead denotes the direction an sensc of the vectm: 'If 1110re than one vector must b shown in a single drawing, the length of each arro must be proportional to the Inagnitude of the vect it represents. Both scalars and vectors arc speci forms of a more general category of all quantities mechanics called tensors. Scalars arc also known "zero-01"(Ic1· tensors," whereas vectors aJ'e "firs order tensors." Concepts such as stress and strai conversely, are "second-order tensors." FORCE VECTOR Force can be defined as· mechanical disturbance load. Whcn an object is pushed or pulled, a force applied on it. A force is also applied when a ball thrown or kicked. A force <:\Cling on an object may deform the objecl, change its slale of m~lion, 0"1' both. Forces may be c1assif-lcd in variolls ways according to their effects on the objects to \vhicf'l they arc applied or according to their orientation as con~­ pared with one another. For example, a force may be internal or external, normal (perpendicular) 0'1' tangential; tensile. compressive. 01" shear; gravitational (weight); or frictional. Any two or more forces acting on (\ single body 111ay be coplanar (acting on a lwo~dimcnsional plane surface); collinear (have a common line of action); concurrent (lines of action intersecting at a single point); or parallel. Note that weight is a special form of Force. The weight of an object on Earth is the gravitational force exerted bv Earlh on the mass of that object. Thc magnitude ~'r the wcighl of an object on Ea~·t11 is equal the mass of the object times the magnitude of the gravitational acceleration, \vhich is approximately 9.8 mt> ters pCI' second squared (111/s 1). For exampl~, a 10-k<J object weighs approximately 98 newtons (N) o~ Earth. The direction of weight is always vertically do\vl1\vard. t; TORQUE AND MOMENT VECTORS The effect of a roree on the object it is applied upon depends on how the rorce is applied and how the object is suppo!"ted. For example, when pulled. an open door will swing about the edge along which it is hinged lO the wall. \-Vh'll eallses the door 10 swin cJ is the torque generated by the applied force abou~ an axis that passes through the hinges of the door. If one stands on the free-end of a diving board, the hoard will bend, What bends the board is the momel1l of the body weight about the fixed end of the board. In general. torque is associated with the ro~ tnlional and twisting action of applied forces, while moment is related to the bending action. However, the mathematical defmition of moment and torque is the same. Torque and moment arc vector quantities. The magnitude of the tonlue Of rnoment of a force about a point is equal to the mannitude of the force times the length of the shortc:t distance between the point and the line of action of the force which is known as the lever or moment arm. Con~ W)-_-I Definition of torque. Reprinred with permission from DZkaya, N. (998). Biomechanics. In w.N. Rom, Environmental and Occupationa( Medicine (3rd ed., pp, 1437~1454), New York: Lippincott·Raven, sider a person on an exercise apparatLls who is holding a handle that is attached to a cable (Fig. I-I), The cable is wrapped around a pulley and attached to a weight pan. The weight in the \veight pan stretches the cable such that the magnitude F of the tensile force in the cable is equal 10 thc weight of the weight pan. This force is transmitted to the person's hand through the handle, At this instant, if the cable allached to the handle makes an angle 0 with the horizontal. then the force E exerted by the cable on the person's hand also makes an angle 0 with the horizontal. Let 0 be a point on the axis of rolation of the elbow joint. To dctermine the magnitude of the moment due to force f about 0, extend the line of action of force f and drop a line from 0 that cuts the line of action of F at right angles. If the point of intersection or the twO lines is Q, then the dbtance d between 0 and Q is the lever arm, and thc magnitude of the rno~ ment M of force E about the clbow joint is M = dE The direction of the moment ,·cctor is perpendicular to thc plane defined by the line of action of E and line 00, or for thb two-dimensional case, it is counterclockwise. NEWTON'S LAWS Relatively few basic laws govern the relationship betwcen applied forces and corresponding motions, Among these, the laws of mechanics introduced by Sir Isaac Newton (1642-1727) are the most important. NeWLOn's first law states that an object at rest will remain at rest 01' an object in motion will move in a straight line with constant velocity if the net force acting on the object is zero. Newton's second law states that an object with a nonzero net force acting on it will accelerate in the direction of the net force and that the magnitude of the acceleration will be proportional to the magnitude of the net force. Newton's sccond law can be formulated as E = m ;), Here, E is the applied force. m is the mass of the object, and i! is the linear (translational) accelcration of the object on which the force is applied. If more than one force is acting on the object. then E represcnts the net or the resultant force (the vector sum of all forces), Another way of stating Newton's second law of motion is M = I Q., where M is the net or resultant moment of all forces acting on the objcct, I is thc mass moment of inertia of the object, and ~ is the angular (rota- , tional) acceleration of the object. The mass m and mass moment of incrtia I in these equations of motion arc measures of resistance to changes in 1110- lion. Th~ larger the inertia of an object, the m difficult it is to sel in motion or to SlOp if it is rend)' in motion. Newlon's third law states that to every act there is a reaction and that the forces of action a reaction between interacting objects are equal magnitude, oppositc in direction, and have same line of action. This law has important appli tions in constructing free· body diagrams. FREE· BODY DIAGRAMS Free-body diagrams are constructed to help iden the forces and moments acting on individual pa of a system and to ensure the correct use of equations of mechanics La analyze the system. this purpose. the parts constituting a system are i lated from their surroundings and the effects of s roundings arc replaced by proper forces and m ments. The human musculoskeletal system consists many parts that are connected to one anot through a cornplcx tendon. Iigamcnt, muscle, a joint SU·uctUfC. In somc analyses, the objective m be to investigate the forces involved at and arou val'ious joints of the human body for different p tural and load conditions. Such analyses can be c ried out by separating the body into two parts al joint of interest and drawing the free-body diagr of one of the parts. For example, consider the a illustrated ill Figure I ~2. Assume thal the forces volved at the elbow joint arc to be analyzed. As lustrated in Figure 1-2, lhe entire body is separa into two at the elbow joint and the free-body d gram of the forearm is drawn (Fig. 1-2B). Here, E is the force applied to the hand by the handle the cable attached to the weight in the weight pa \V is the total wcight of thc lower arm acting the center of gravity of the lower arm, £,\\1 is the force excrted by' the biceps on the dius, £.,,; is the force exerted bv the brachioradi muscles on the radius. £.\l~ is the force exerted by the brachialis musc on the ulna, and f1 is the resultarll reaction force at the hume ulnar and humeroradial joints of the elbow. N that the muscle and joint reaction forces repres the mechanical effects of the upper ann on lower arm. Also note that as illustrated in Fig 1-2;\ (which is not a complete free-body diagra equal magnitudc but opposite muscle and joint action forces act on the upper arm as wcll. " -, ' .~ ... dum implies that the body of concern is either at rest or moving with constant velocity. For a body to be in a slate of equilibrium, it has to be both in translational and rotational cquilibl-iul11. A body is in translational cquilibriun1 if the net force (vector sum of all forces) acting on it is zero. If the Ilt:t force is zero, then the linear acceleration (time rate of change of linear velocity) of the body is zero, or the linear velocity of the bod,Y is either constant or zero. A body is in rotational equilibrium if the net moment (vector sum of the moments of all forces) act· ing on it is zero. If the net moment is zero, then the angular acceleration (time rate of change of angular velocity) of the body is zero, or the angulal· yelocily of the body is either constant or zero. Therefore, for a body in a state of equilibrium, the equations of motion (Newton's second law) take the following special forms: A ~E = E B w Forces involved at and around the elbow joint and the free-body diagram of the lower arm. Reprinted with permission from dzkaya. N. (7998). Biomechanics. In W.N. Rom, Environmental and Occupational Medicine (31d ed., pp. 1437-1454). New York: Lippincott-Raven. CONDITIONS FOR EQUILIBRIUM Statics is an area within applied mechanics that is concerned with the anal~:sis of forces on rigid bodies in equilibrium. A rigid body is one that is assumed to undergo no deformations. [n reality, evcr)' object or matcrial may undergo deformat ion to an extent when acted on by forces. (n some cases, the amount of deformation may be so smalllhal il may not affect the desired analysis and the object is assumed to be rigid. In mechanics, the term cquilib- 0 and ~rvl =0 rt is important to remember that force and moment arc vector quantities. For example, with respect to a rectangular (Cartesian) coordinate system, force and moment vectors may have components in the .'\. y, and z directions. Therefore, if the net force acting on an object is zero, then the sum of forces acting in each direction must be equal lo zero (IF, = 0, IF, = 0, IF, = 0). Similarly, if the net moment on an object is zero. then the sum of moments in each direction must also be equal to zero (lM, = 0, lM,. = 0, lM, = 0). Thel'efore, for three-dimension force systems there arc six cOl1ciilions of equilibrium. For two-dimensional force sys[ems in lhe xy-plane, onl~: three of these conditions (IF, = 0, ~F, = 0, and ~M, = 0) need to be checked. STATICS The principles of slatics (equations of equilibrium) can be applied to investigate the muscle and joint forces involved at and around the joints for various postural positions of the human body and its segments. The immediate purpose of static analysis is to provide answers to questions such as: What tension must the neck extensor muscles exert on the head to support the head in a specined position? \OVhen a person bends, what would be the force ex~ ertcd by the erector spinae on the fifth lumbar vertebra? Ho\\'" does the con1pression at the elbow, knee, and ankle joints vmy with externally applied forces and with different segmental arrangements? How docs the force on the femoral head vary with loads carried in the hand? \,Vhat arc the forces in~ volved in various muscle groups and joints during different exercise conditions? In general. the unknowns in static problems involving the musculoskeletal s:'stcm arc thc magnitudes of joint reaction forces and muscle tensions. The mechanical analysis of a skelclal joint requires that we know the vector characteristics of tensions in the muscles, the proper locations of muscle attachments, the weights of body segmcnts, and the locations of the centers of gravity of the body segmems. Mechanical models are obviously simple representations of c0l11plex systems. Many models are limited by the assumptions that must be made to reduce the system under considcration to a st.atically determinate one. Anv model can be improved by ~onsidering the comril)utions of other muscles, but (hat will increase the number of unknowns and make the model a statically indeterminate one. To analyze the improved model. the researcher would need additional information related to the muscle forces. This inforrrlalion can be gathered through electromyography measurements of muscle signals or by applying certain optirnization techniques. A similar analysis can be made to investigate forces involved at and around other major joints of the musculoskeletal system. MODES OF DEFORMATION When acted on by externally applied forces. objects may translate in the direction of the net force and rotate in the direction of the net torque acting on them. If an object is subjected to externally applied forces but is in stalic equilibrium. then it is most likely that there is some local shape change within the objec!. Local shape change under the effect of applied forces is known as deformation. The extent of deformation an object may undergo depends on many' factors, including the material properties. size, and shape of the object; environmental factors such as heat and humidity; and the nlagnitudc, direction, and duration of applied forces. One way of distinguishing forces is by observing their tendencv to deform the object they are applied upon. For example. the object is said to be in tension if the body tends to elongate and in compression if it tends to shrink in the direction of the applied forces. Shear loading differs from tension and compression in that it is caused b:! forces acting in directions tangent to the area resisting the forces ~ causing shear, whereas both tension and compression are caused by collinear forces applied perpendicular to the areas on which they act. It is common ':, to call tensile and compressive forces normal o ial forces: shearing forces are tangential forces jects also deform when they are subjecled to f that cause bending and torsion, which are relat the moment and torque actions of applied forc A matel"ial nwv respond differently to diff loading configurations. For a given material. may be different physical properties that mu considered while analyzing the response of tha terial to tensile loading as compared with com sive or sheai' loading, The mechanical propert mnterials are established through stress analys subjecting them to various experiments such as axial tension and compression, torsion, and ing tests. NORMAL AND SHEAR STRESSES Consider the whole bone in Figure \-3;,\ that is jected to a pair of tensile forces of magnitude F bone is in static equilibriulll. To analyze the f induced within the bone, the method of section be applied by hypothetically cutting the bone two pieces through a plane perpendicular to the axis or the bone. Because the bone as a whole equilibrium, the two pieces must individually equilibrium as well. This requires that at the cu tion of each piece there is an internal force t equal in magnitude but opposite in direction t externally applied force (Fig. 1-38). The int force is distributed over the entire cross-sec area of the cut section. and E represents the resu of the distributed force (Fig. 1-3C). The intens this distributed force (force per unit area) is k as stress. For the case shown in Figure 1-3. be the force resultant at the cut section is perpendi to the plane of the cut. the cOITesponciing str called a normal or axial stress. It is customar:y t the symbol (T (sigma) to refer to normal stresse suming that the intensity of the distributed fo the Cllt section is uniform over the cross-sec area A of the bone, then u::::: FlA. Normal stresse are caused by forces that tend to stretch (elon matcl"ials aJ"C marc specincally known as t stresses; those that tend to shrink them are kno compressive stresses. According to the Standar ternational (SO unit system (see Appendix), str are measured in newton per square meter (N which is also known as pascal (Pa). There is another form of stress, shear s which is a measure of the intensity of internal f acting tangent (parallel) to a plane of cut ~i H' r '.i~ ,..-__..J-_=- - F .....~--I .. ~- F A A ~-G?- F ...... =' ~, B c Definition of normal stress. Reprinted wirh permission from OZkaya. N. (1998j. Biomechanics. In W.N. Rom, Environmental and Occupatiol1<11 r..,ledicine (3rd ed., pp. i437-145r+). New York: Lippincorr·RiNen. , d; !: example. consider the whole bone in Figure 1-4A. The bone is subject to a number of parallel forces that act in planes perpendicular to the long a,is of the bone. Assume that the bone is cut into two parts through a plane perpendicular to the long axis of the bone (Fig. 1-48). If the bone as a whole is in equilibrium, its individual parts must be in equilibrillm as well. This requires that there must be an internal force at the cut section that ,lets in a direction tangent to the cut surFace. If the magnitudes of the external forces arc known, then the magnitude F of the internal force can be calculated by considering the translational and rotational equilibrium of onc of the parts constituting the bone. The intensity of the internal force tangent to the Clit section is known as the shear stress. It is customary to usc the symbol T (tau) to refer to shear stresses (Fig. 1-4C). Assuming that the intensity of the force tangent to the cut section is uniform over the cross-sectional area A of the bone, then T = FlA. NORMAL AND SHEAR STRAINS Strain is a measure of the degree of deformation. As in the case of stress, two types of strains can be distinguished. A norm~l'l strain is deflnecl as the ratio of the change (increase or decrease) in length to the original (undeformed) length, and is commonly denoted with the symbol € (epsilon). Consider the whole bone in Figure \-5. The total length of the bone is I. If the bone is subjected to a pair of tensile forces. the length of the -bone may increase to I' or by an amount .1i\ = I' -I. The normal strain is the ratio of the amount of elongation to the original F., F, A Shear strains are related to distortions caused shear stresses and arc cornmonly denoted with t symbol y (gamma). Consider the rectangle (ABC shown in Figure 1-6 thm is acted on by a pair of ta gential forces that deform the rectangle into a p allelogram (AB'C '0). 'If the relative horizontal d placement of the top and the bOllom of t rectangle is d and the height of the rectangle is then the average shear strain is the ratio of d and which is equal to the tangent of angle y. The angle is lIsllall~" vcry small. For small angles. the tange of the angle is approximately equal to the angle self measured in radians. Therefore, the avera shear strain is "y = cllh. Strains arc calculated by dividing two quantit measured in units of length. For most application the deformations and consequently the strains volved may be very small (c,g" 0,001), Strains c also be gi\'en in percemages (e.g.. O.l%). STRESS-STRAIN DIAGRAMS B F, Different I11mcrials may demonstrate differe stress-strain relationships. Consid(~r the stre strain diagrarn shown in Figure 1-7. There arc distinct points on the curve, which arc labeled as P, E, Y, U, and R. Point 0 is the origin of the Sli'e strain diagram, which corresponds to the initial ( load, no deformation) state. Point P represents t proportionality limit. Between 0 and P. stress a strain are linearly proportional and the stre strain diagram is a straight line. Point E represen the clastic limit. Point Y is the .\"ield point, and t stress (T.. corresponding to the yield point is call the yield slrength of the material. At this Slr level, considerable elongation (yielding) can occ without a corresponding increase of load. U is t highest stress point on the stress-strain diagra The stress (r is the ultimate strength of the mat ial. The last point on the stress-strain diagram is \vhich represents the nq)ture or failure poinl. T stress at which lhe failure occurs is called the ru ture strength of the material. For some materials may not be easy to distinguish the elastic limit a the yield point. The yield strength of sLieh materi is determined by the offset method, which is a plied b.y drawing a line parallel to the linear secti of the stress-strain diagram that passes through strain level or approximately 0.2 % • The intersecti of this line with the stress-strain ClWVC is taken be the vielel point, and the stress corresponding this po-int is called the ~\pparent yield strength the material. ll Definition of shear stress. Reprintecl with permission {rom Ozkaya, N, (I 99B). Biomechanics. /11 W.N. Rom, Environmenlal and Occupational Medicine (3rd ed" pp. 1437-/454). New York: LippincorrRaven. I length, or E = c,11 1. If the length of the bone increases in the direction in which the strain is calculated. then the strain is tensile and positive. If the length of the bone decreases in the direction in which the strain is calculated, then the strain is compressive and negative. "nfj , ----,- I' F ~ , "'" I -- .; .:.\ ~ - ): -. I '.l/ ...... F 1 Definition of normal strain. Reprinted with permission from 6zkaya. N. (/998). Biomechanics. In W.N. Rom, Environmental (1nd Occupational !v1edione (lrd ed., pp. /437-1454). New York: Lippi(lCOfl-R(l~'efl. • Note that a given material may behave dilTcr~ ently under different load and environmental con~ ditions. If the curve shown in F'igurc 1~7 represents the stress"strain relationship for a material under tensile loading, there ma).o' be a similar but ELASTIC AND PLASTIC DEFORMATIONS or Elasticit:-.· is defined as lhe ability a material to resume its original (stress-free) size and shape on removal of applied loads. 1n other words, if a load different curve representing the stress-strain relationship for the same material under compressive or shear loading. Also. temperature is known 10411leI' the relationship between stress and strain. For some materials, the stress-strain relationship may also depend on the rate at which the load is applied on the material. u A y I_ 'I d S' B ~C / - - - - - - - - - 7 C' / / h / r1.t / / / / / ,l- , / AL----;-.-------';0 F Stress-strain diagrams. Reprinted with permission Definition of shear strain. Reprinted wirh permission from OZkaya, N. (1998). Biomechanics. In W.N. Rom, Environmental and Occupational Medicine (3rd ed.. pp. /437-1454). New York: Lippincou-Raven. • from Ozkaya, N. (1998). Biomechanics. In W.N. Rom, Environmental and Occupational Medicine (3rd ed.. pp. 1437-1454)., New York: LippincouRaven . ; is applied on a material such that the Stress generated in the material is equal to or less than th~ elastic limit, the deformations that took place in the material will be cOlllpletcl.v recovered once the applied lands arc removed. An elastic material \vhose stress·strain diagram is a straight line is called a linearly clastic material. For such a matcthe stress is linearly proportional to strain. slope of the stress-strain diagram in the e1asregion is called the elastic or Young's rnodulus of the material. which is commonly denoted by E. ,Therefore, the relationship between stress and strain for linearly elastic materials is a := E€. This equation that relates normal stress and strain is called a material function. For a given material. different material functions may exist for different modes or derormation. For example, SOme materials may exhibit linearly elastic belHwior under shear loading. For such materials, the shear stress T is linearly proportional to the shear strain y, and the constant of proportionality is called the shear modulus, or the modulus of rigidity. If G represents the modulus of rigidity, then ,. = Gy. Combinations of all possible material functions for a given material form the constitutive equations for that material. Plnsticity implies permanent deformations. Materials may undergo plastic deformations follo\ving elastic deformations when they are loaded beyond their elastic limits. Consider the stress-strain diagram of a material under tensile loading (Fig.I-7). Assume that the stresses in the specimen arc brought to a level greater than the yield strength of the material. On removal of lhe applied load. lhe material will recover the elastic deformation that had taken place by following an unloading path parallel to the initial linearly elastic region. The point where this path cuts the strain axis is called the plastic strain. which signifies the extent of perrl1a~ nent (unrecoverable) shape change that has taken place in the material. Viscoelasticity is the characteristic of a material that has both fluid and solid properties. Most materials arc classified as eilher fluid or solid. A solid material will deform to a ccrLain extent when an exlernal force is applied. A continuously applied force on a Ouid body will cause a continuous deformation (also known as flow). Viscosity is n fluid property thut is a quantitative measure of rcsis· tance to flow. Viscoelasticity is an example of how areas in applied mechanics can overlap, because it ulilizes the principles of both fluid and solid mechanics. G E linearly elastic material behavior. Reprinted wirh permission from OZkclycl, N. (1998)_ Biomecll<lflics. In W.N. Rom. Environmental and Occupattonal MecHcme (3rd cd., pp J437-1Li54.J. Ne....,; York: Lippincott- Raven VISCOELASTICITY \·Vhcn they are subjected to relatively low stress els, many materials such as metals exhibit ela material behavior. They undergo plastic defor tions at high stress levels. Elastic materials defo instantaneously when they are subjected to ex nally applied loads and resume their original sha almost instantly when the applied loads are mo\·cd. For an elastic material, stress is a function strain only, and the strcss-strain relationship unique (Fig. 1-8). Elastic materials clo not exh time-dependent behavior. A different gl'OUp of m rials, such as polymer plastics, metals at high t peratures, and almost all biological materials, hibits gradual deformation and recovery w subjected to loading: and unloading. Such mater are called viscoelastic. The response of viscoela materials is dependent on how quickly (he loa applied or removed. The extent of deformation viscoelastic materials undergo is dependent all rate at which the deformation-causing loads are plied. The stress-strain relationship for a viscoela material is not unique but is a f1.lI1ction of time or rate at which the stresse.s and strains are develo in the material (Fig. 1·9). The word "viscoelastic made of two words, Viscosity is a fluid property ,--; ->". I.'· " r Increasing suain rale Ii) < Strain rate~dependent viscoelastic with an instantaneous strain that would remain at a constant level until the load is removed (Fig. I-lOB). At the instant when the load is removecl, the deformation will instantl)' and completely recover. To the same constant loading condition, a viscoelastic material will respond with a strain increasing and ciCCI-casing graduall)r. If the material is viscoelastic solid, the recovery will eventually be complete (Fig. 1-.1 DC). If the material is viscoelastic fluid, complete recovery will never be achicved and there will be a residuc of defOl'mation lerr in the material (Fig. 1-IOD). As illustrated in Figure l-11A, a stress·relaxation experiment is conducted material be- havior. Reprinted with permission /rom 6zkaya. N. (1998). Biomedlc1llic5. In WN. Rom, Environmental and Occupational Medicine (3rd ed., P.o. 1.137-/454). New York: Lippincorr·Raven. is a measure of resistance to now. Elasticity is" solid material property. Therefore, viscoelastic materials possess both nuid- and solid-like properties. For an elastic material, the energy supplied to deform the material (strain energy) is stored in the material as potential energy. This energy is available to return the material to its original (unstressed) size and shape once the applied load is removed. The loading and unloading paths for an elastic material coincide. indicating no loss of energy. Most elastic materials exhibit plastic behavior at high stress levels. For e1asto-plastic materials, some of the strain energy is dissipated as heat during plastic defat-mat ions. For viscoelastic materials, some or the strain energy is stored in the material as potential energy and some of it is dissipated as heat regardless of whether the stress levels are small or large. Because viscoelastic materials exhibit lime-dependent material behavior. the differences between elastic and viscoelastic material responses are most evident under time-dependent loading conditions. Several experimental techniques have been designed to analyze the time-dependent aspects of material behaviOl: As illustrated in Figure 1-1004, a creep and recovery test is conducted by applying a load on the matcl¥ial, maintaining the loael at a constant level for a while, suddenly removing the load, and obsen;jng the material response. Under a creep and recovery test. an elastic material will respond -~ o Creep and recovery test. Reprinred wirh permission from Ozkay,l, N. (1998). Biomechanics. In W.N. Rom, Environmental and Occupational Medicine (3rd ed., pp. 1437-745/1). New York: LippincorrRaven. , 'by straining the Olalcriallo a level and maintaining the constant strain while observing the stress response of the material. Under a stress-relaxation lcst, an elastic mater-ial will respond with a stress developed insw.ndy and maintained at a consWnl level (Fig. I-II B). That is, an elastic malcrial will not exhibit a stress-relaxation behavior. t\ viscoelaslie material. conversely', will respond with an initial high stress level that will decrease over time. If the m;terial is a viscoelastic solid, the stress level will nevcr rcduce to zcro (Fig, I-lie), As illuSlrated in Fiourc I-II D, the stress will evct11uallv o . reduce to zero for a viscoelastic nuid. A 'U '0 cr o·u a MATERIAL PROPERTIES BASED ON STRESS-STRAIN DIAGRAMS PRINCIPAL STRESSES There are infinitely many possibilities of constructing elements around a given point wilhin a structure. Among these possibilities, there may be = E, to B The stress-strain diagrams of two or Il"wrc materials can be compared to determine \vhich m<:ucrial is rei· atively stiffer, l1C:lrdcl~ tougher, more ductile, or more brittle. For example, the slope of the stress-strain di~ agram in the clastic region represents the clastic modulus that is a measure of the relative stiffness of materials. The higher the elastic modulus, the stiffer the material and the higher its resistance to deformation. A ductile material is one that exhibits a large plastic deformation prior to failure. A britlie material, such as glass, shows a sudden failure (rupture) without undergoing a considerable plastic deformation. Toughness is a measure of the capacity of a material to sustain permanent defonllation. The toughness of a matedal is measured b~: considering the total area under its stress-strain diagram. The larger this area, the tougher the malerial. The ability of a material to store or absorb energy without permanent deformation is called lhe resilience of the material. The resilience of a material is measured by its modulus of resilience, which. is equal to the area under the stress-strain curve in the elastic region. Although thcy arc not directl\' rclated to the stress-strain diagrams, other important concepls are used to describe material properties. For cxam~ pie, a material is called homogeneous if its properties do not vary from location to location within the material. A material is called isotropic if its properlies are independent of direction. A material is called incompressible if it has a constant denSity. 0"0 C U: l110 G D 10 Stress-relaxation experiment. Reprinted with permission from Ozkaya, N. (1998). Biomechanics. In WN. Rom. Environmental and Occupational Medicine (Jrd ed.• P.o. 1437-1454). Ne~··1 York: Lippincott-Raven. one: element for which the normal stresses maximum and minimum. These maxin1lrm minimum normal stresses arc called the princi stresses, and the planes whose normals are in directions of the maximum and minimum stJ"e are called the principal plancs, On a princi plane, (he normal stress is either maximum minimum. and the sheal" stress is zero. It is kno that fracture or material failure occurs along planes of maximum stresses, and structures m be designed by taking into consideration the m / imulll stresses involved. Failure by yielding cessive deformation) n.lay occur whenever largest principal stress is equal to the y strength of the material or failure by rupture m occur whenever the largest principal stress ,is equal to the ultimate strength of the material. For a given structure and loading condition. the principal stresses ma~" be within the limits of operational safely. However, the structure must also be checked for critical shearing stress. called the maximum shear stress. The maximum shear SlI-es$ occurs on a material element for which the normal stresses are equal. A a I - Tension - ] o min 1 cycle FATIGUE AND ENDURANCE - - (j" am· - - - ':'.~ (j Principal and maximum shear stresses are useful in predicting the response of materials (0 static loading configurations. Loads that Illay not cause the failure of a structure in a single application may cause fracture when applied repeatedly. Failure may occur aher a few or many cycles of loading and unloading, depending on factors such as the amplitude of the applied load, mechanical properties of the material, sIze of the structlire, and operational conditions. Fracture resulting from repeated loading is called fatigue. Several experimental techniques have been developed to understand the fatIgue behavior of materials. Consider the bar shown in Figure 1-12;:1. Assume that the bar is made of a material whose ultimate strength is U'w This bar is first stressed to a mean stress level (1m and then subjected to a stress fluctuating over time, sometimes tensile and other times compressive (Fig. 1-128). The amplitude (T:, of the stress is such that the bar is subjected to a maxImum tensile stress less than the ultimate strength of the material. This reo versible and periodic stress is applied until the bar fractures and the number of cycles N to fracture is recorded. This experiment is repeated all specimens having the same material properties by applying stresses or varying amplitude. A typical result of a fatigue test is plotted in Figure 1-12C on a diagram showing stress amplitude versus numbct· of cycles to failure. For a given N. the corresponding stress value is called the fatigue strength of the material at that nun1ber of cycles. For a given stress level, N represents the fatigue life of the material. For some matel"ials, the stress amplitude versus number or cycles curve levels off. The stress CT, at which the fatigue curve levels off is called the endurance limit of the material. Below the endurance limit, the material has a high probability of not failing in fatigue, regardless of how many cycles of stress are imposed on the material. :· · . . . t o max - - ..... "7" - - .......: ..' - - - - - - ••••• - ~:.;. - - - ";"~ - - ~:~ •• ••••• •• ••••• lime Compression B °0- - - - - - - - - - - - - L-_,---'---,---,--N 10' 10' 10' c Fatigue and endurance. Reprinred with permission from Olkaya, N. (1998). Biomechanics. In W.N. Rom, Environmeniat and Occupational Medicine (3rd ed., pp. 1437-1454). New York: Lippincou-Raven. The fatiguc behavior or a material depends on several factors. The higher the temperature in which the material is used, thc lower the fatigue strength. The fatigue behavior is sensitive to surface imperfections nnd the presence of discontinuities within the material that can cause stress concentrations. The fatigue failure starts \vith the creation of a small crack on the surface of the material. which can propagate under the effect of repeated loads, resulting in [he rupture of [he material. Orlhopaedic devices lII)dergo repeated loading and unloading as a result of the activities of the patients and the actions of their 111uscles. Over a pe- riod of vears. a weight-bearing prosthetic dc\·icc or fi;.;:ati~n device can be subjected to a consiclerabk number of cycles of stress reversals as a result or noHnal daily activity. This cyclic loading and un~~ .•l;"N can cause faLigue failure of the device. a' b~" clinicians (Q provide an introducLor~: level knowh:dge about each joint s~'stem. PART III: APPLIED BIOMECHANICS A new section in the third edition of this book Biomechanics the Musculoskeletal System even a simple task c.'\ecuted b.v the musculoskeletal svstcm requires a broad. in-depth :~;' knowledge of various fields that ma~' include 1110tor control, neurophysiology, physiology. physics. and biomechanics. For example, based on the purpose ancl intention or a task and the sensOl'~' information gathered from the ph~'sical cndronmcnl and orielllatioll of tilL' body and joints, the central nervous system plans a strategy for a task execu:;.: lion. According to the strategy adoptc:d. Illuscles / . '.,' will be recruited lO provide Lh<..' forces and 1110mcnts required for the movement and I.Jalance of the s.)'slem. Consequently, the internal forces will be changed and soft tissues will experience different load conditions. The purpose this book is to present a \\'cll~ balanced synthesis of information gatllCred frorn various disciplines. pro\'iding a basic understanding of biomechanics of the musculoskeletal system. The material presented here is organized (0 cover three areas of musculoskeletal biomechanics. ;~; or PART I: BIOMECHANICS OF TISSUES AND STRUCTURES The material presented throughout this textbook provides an introduction to basic biolllechanics of the musculoskeletal system. Part I includes chapters on the biomechanics of bone. articular cartilage, tendons and ligaments, periphcral nerves, and skeletal muscle. These are augmcnted wilh case studies to illustrate the imponarll concepts for understanding the biomechanics of biological tissues. PART II: BIOMECHANICS OF JOINTS Part II of this textbook covers the major joiots of the human body, from the spine to the ankle. Each chapter contains information about the structure and functioning of the joint. along with case studies illuminating the clinical di~lgnosis and management of joint injlll)' and illness. The chnpters are written troduces important issues in applied biomechani These include the biomechanics of fracture fixati arlhroplasty; sitting, standing. and lying; and gait is important for the beginning studenl to und stand the application or biomechanical principles clirfcrcnt clinical areas. Summarv 1 Biomechanics is a young and dynamic fidd study based on the recognition thaI conventio cllginccl'ing thcorks and methods can be useful understanding and solving problems in physiolo and medicine. Biomcclwnics considers the appli tions of classical rncchanics to biological problem The flcld of biomechanil:s flourishes from the co eration among life scientists, physicians, enginee and basic scientists. Such cooperation require certain amount of common vocabulary: an engin must learn some anatom~: and ph)'siology, nnclm ical personnel need to understand some basic c cepts of physics and mathematics. 2 The information presented throughout t textbook is drawn from a large scholarship. The thors aim to introduce some of the basic concepts biolllechanics related to biological tissues a joints. The book does nOl intend to provide a co prehensive review of the literature, and readers encouraged to consult the list of suggcsted read below to supplcmcnt theil' knowledge. Some ba textbooks arc listed here, and studcnts should c sult peer-revicwed journals for in-depth presen Lions of the latest research in specialty arcas, SUGGESTED READING Black. J. (19SS). Onhop'lcdic Bionmleriuls in R,'sl,.'ardl and P til:c. New York: Churchill Li\·in!!stonc. Brollzino. J.D. (Ed.) (1995). The -Biom(.'dic<ll Engincl.'ring H. book. 80(:01 R:llOn. rL: CRC Press. B(lrst('in, A.H., & \Vright. T.~'I.( 1993). Fundamellt;lls of Or1hop< Biolllc:dmnics. Ballirnorl': Williams & Wilkins. Chaffin. 0.8.. & All{krsson. G.B.J. (1991). O<::cupational Bio ch~lllics (2nd l'<'I.). Nl'\\' York: John Wiley & Sons. Fung. Y.c. (1981). Biolllec!wnics: Mechanical Properties of Li Tissul.·s. New York: Springl.'r·\\:rJag. FUll!!. yc. (1990). Biomcch~l1lics: ;\·Iotioll, Flo\\" Slrcss, :llld Gro New York: Springl'r,Vl'rlOIg. ... :':. H~l\'. J.G .. 6:.. RI..'id. J.G. (1988). .'\nalOllw. '\lcchanil:s .md Human Mo-lioll (2nd cd.). Englewood Cliffs, NJ: Pn:llIkc-Hall. Kelly, O.L. (1971). Kinl.'siology: FLllldaml.'llWls of \Iolion D~·StTip· lion. En!.:.lcw()od Cliffs, NJ: Prelltice-Hall. \Iow, Vc.. &: Hayes, w.e. (1997). Basic Orthopaedic Biorncch.mics (2nd cd.). New York: Raven Press. ,\low. \I.e., Ratdiff, A".& Woo. S.L.·Y. (Eds.). (1990). BicHlll.:chanics or {)i~tl"(llrodial Joinls. Nr.:\\' York: Sp.-jnga-Verlag. NahulIl. A.M .. &. Md\'in. J. (Etls.). (19S5). The BiOlllcch:lI1il:s of TI'UUll'l. Norwalk. CT: l\ppl('ton-Ct:llhll~'.Crofts. Nordin, M .. Andersson. G.B.J., & Po\X", M.H. (Eds.). (1997). ~plllSCll­ 1()sk..:k'IJI Disonk'J1; in the \VorkpbcL'. Philadclphin: :\'1osby,Ycilr Book. Nordin. ~L & Franb,:1. \l.H. (Eds.l. (1989). Basit- OiolUl.'chanics of lhe :\hlst.:uloskdctal S,\':o;'lClll (2nd t:d,), Philaddphi:l: 1...:<.\ & F.:bi"'.:r OzkaYil,eN.·( 1998). Bi(Hn~d1iJnics. III W.N. Rom, EnvirorH1H:nlal and OCCllp<:lliollal :\kdicill~' Ord cd., pp. 1437-1454). New York: Li ppi m:olt -RaVCll. OZk:IY;\. N.. & Nordin. M. (1999). Fundamelltals of Biolll<.'ch'lllics: Equilihrium, :\lotioll. <lnd Dcfonlwtioll (2nd I..'d,). Nt:\\" York Sprin!.!(,I~Vl.'rlag. Schmid.Schonbl'i;l, G.\\'., \\'00. 5.1...·).... & Z\\"eifach. B.\V. (Eds.). (1985). Frontiers in Biomechanics. ~·h.'\\" York: Springer.Verlag:, Skal<:lk, R" & Chkn, S. (Eds.). (!98i). Hllndbonk or BiOi..'ngincering. New York: McGraw·Hill. Thompsoll. C. W. (1989). :\·tanu'll of Stlllt'lur<ll cd,). 51. Louis. MO: ·fill1l..'s Mirror/:\·Iosbv. Kinl~si()log~' (11th Williams. M., & lissner, H.R. (1992). Bion;l.'chanics of !·luman ;\'tolion (3r<l cd.). Phibddphin: Sallll(k:rs, Wintcl'. D.!\. (1990). Biol1lcchnnics :Illd ~:lolor Control or Hum"n Behavior (2nd I..'d.). New York: John Wiky &: Sons. Willlel'S. J,!lil.. &. Woo. S.L-Y. (Eds.), (1990). Multiple Muscle SysIl'IHS. New York: Springcr.verktg. The System International d'Unites (51) Dennis R. Carter The 51 Metric System Base Units Supplementary Units Derived UnitS Specially Named Units Standard Units Named for Scientists Converting to 51 From Other Units of Measurement fined in terms of the \V~\'c1cngth of radiation emitted from the krvpton-S6 atom. The''S! Metric System ,,;The System Intemational d'Unites (SI), the metric system, has evolved into the most exacting system of 'measures devised. In this section. the 51 units of lllcasurcmcnl used in the science of mechanics arc described. SI units used in electrical and light sciences have been omitted for the sake or simplicity. ~__ • fifo' <""'i,ilY ' BASE UNITS --~ The 51 units can be considered in three groups: I, tbe base units; 2. the supplementary unils; and 3. 'j;", the derived units (Fig. App-I). The base units are a ,', small group or standard measurements lhal have been arbitrarilv defined. The base unit for length is '-'rhe meler (Ill), "and the base Llnit for Illass is the kilogram (kg). The base units for time and temperature are the second (s) and the kelvin (Kl. respectively. Definitions 01" the base units have become.:: increas· .ingly sophisticated in response to the expanding needs Hnd capabilities of the scientific community (Table App-I). For example, the meter is now de- SUPPLEMENTARY UNITS The radian (rad) is a supplemental)' unit (() measure plane angles. This unit, like the base units, is arbitrarily defined (Table App-I). Although the radian is the 51 unit for plane angle. the unit of the degree has been retained for general use because il is firmly established and is widely Ltscd "round the world. A degree is equinllcnl to rrll80 rad. A" QUANTITIES EXPRESSED IN TERMS OF UNITS FROM WHICH THEY WERE DERIVED DERIVED UNITS ivlostunits of the 51 system are derived unils, meaning lhat they- are established from the base units in accordance with fundamcntnl physical principles. Some of these units are cxp['css~d in terms of the base units from which they are derived. Examples aloe area, speed, and acceleration. which arc expressed in the Sf units square meters (m~). meters per second (m/s), and meters pCI' second squared (In/s 2), respectively. or DERIVED UNITS WITH SPECIAL NAMES FORCE new Ion kg m/s 2 MOMENT OF FOACE ACCELERATION speED DENSITY VOLUME AREA qp ,- - . . PRESSURE & STRESS pascal N/m 2 ~\ . \l/~ \' Il/' _~r;) ~~ ~,I //'/~'" ~ ;~:::::<':~:~~J" J , ,~/~,:'!j ~DERIVED~ ,~m'?I~~'TS i~ ENERGY & WORK joule Nm POWER watt JIS TEMPERATURE degree Celsius K'- 273.15 radian (fad) PLANE ANGLE ----SUPPLEMENTARY UNIT meter (m) LENGTH kilogram (kg) MASS second (s) TIME kelvin (K) TEMPERATURE BASE UNITS The International System of Units. 19 The 51 unit of pressure, the pascal, is therefore defined in terms of the base 51 units as: Specially Nanzeel Units Other dedved units are similarl.v established from the base units but have been given special names (Fig App-I and Table App-I). These units are defined through the lise of fundarnental equations or physical laws in conjunction with the arbitrarily defined Sf base units. For example, Newton's second law of motion states that when a body that is free to Il10vC is subjected to a force. it will experience ~m ~lCcelcr­ alion proportional to thai force and inversely proportional to its own mass. i\Jlathcmatically, this principle can be expressed as: force = Illass X acceleration The Sf unit of force, the newton (Nl, is lherefore defined in terms of the base SI units as: 1N = 1 kg X I I11/S:! The Sf unit or pressure and stress is the pascal (Pa). Pressure is defined in hydroslaties as the force divided by the area of force application. Mathem4llicall~l, this can be expressed as: pressure = force/area ~._.m ..__ .._ I Pa = IN I I 111" Allhough the S[ base llnil of temperature is the kc.:lvin, the derived unit of degree Celsius (OC 01' c) is much marc commonly used. The degree Celsius is equivalent to the kelvin in magnitude. but the ab~ solute value of the Celsius scale differs frol11 that o the Kelvin scale such that °C = K - 273.15. ,",Vhen the 51 s~'stcm is used in a wide variely o measurements, the quantities expressed in terms of the base. supplemental. or derived units ma~t be either very large or very small. For example, the arca on the head of a pin is an extremely small number when expressed in terms of square meters Conversely, the weight of a whale is an extremely large number when expressed in terms of newtons. To accomrnodate the convenient representation o small or large quantities, a system of prefixes has been incorporated into the SI system (Table App-2), Each prefix has a fixed meaning and can be used with all 5Iunils. \!\Then used with the name or the unit, the prefb: indicates that the quanti!}! described is being expressed in some multiple o __.__._.__.. _ I ; Definitions of Sl Units Base 51 Units meter (01) The meIer is the length eqllal to 1,650,763.73 wavelengths in vacuum of the radiation corresponding to the transition belween the levels 2PHi and Sd" of the krypton-86 atom. kilogram (kg) The kilogram is Ihe unit of mass and is equallO the mass of the international proto· type of the kilogram. second (s) The second is the duration of 9.192,631,770 periods of Ihe radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium-133 atom. kelvin (k) The kelvin. a unit of thermodynamic temperature, is the fraction 1/273.16 of the thermodynamic temperature of the triple point of water. Supplementary SI Unit radian (rad) The radian is the plane angle between two radii of a circle that sub tend on the circum~ ference of an arc equal in length to the radills. Derived SI Units With Special Names newton (N) The newton is that force \tvhich, when applied to a rpass of 1 kilogram, gives it an aCceleration of 1 meter per second squared. 1 N= 1 kg rn/s / . pascal (Pa) The pascal is the pressure produced by a force of 1 newton applied. with uniform distribution, over an area of 1 square meter. 1 Pa = 1 N/m~. joule (J) The joule is the work done when the point of application of a force of 1 newton is displaced through a dist~nce of 1 meter in the direction of the force. 1 J = 1 Nm. wall (W) The watt is the power that in 1 second gives rise to the energy of 1 joule. 1 W ::: 1 J/s. degree Celsius (C) The degree Celsius is a unit of thermodynamic temperature and is equivalent to K - 273.15. Standard Units Nmned for Scientists Factors and Prefixes SI Prefix SI Symbol giga G mega Iv1 kilo k hecla h deka da deci d centi c rniJli rn micro I' nana n pica p One of the more interesting aspects of the SI is its lise of the names of famous scientists a dard units. In each case, the lInit was named scientist in recognition of his contribution Geld in which that unit plays a major role Reprinred wirh permission from Ozkaya. N.. & Nord,-n. M. (1999). Fundamentals oi Blomechani(!>: Equilibrium. Mo· tion. and Deiormatlon (2nd ed.) New York: Springer·lJer/ag. p, iO. ten times the unit used. For example. the millime· 'tel' (mm) is used to represent one thousandth (10"·1) of a meter and a gigapascal (Gpa) is uscd to denotc one billion (10') pascals. App-3 lists a number of Slullits and the scien which each was named. For example. the unit of force. the neWlo named in honor of the English scientist Si Ncwlon (1624-1717). Hc wa' cducalcd 'II College at Cambridge and later rclurned to College as a professor or mathematics. Early career, Newton made fundamental contribut malhcmalics Ihat formcd Ihc basis of diffc and integral calculus. His other major disc were in the fields of optics. astronomy. grav and mechanics. His work in gravitation wa portedly spurred by being hit on the head by ple falling from a tree. It is perhaps poetic that the SI unit of one newton is approxi equivalent to the weight of a medium-sized Newton W~lS knighted in t 705 by Qucen i\lt his monumental contributions to science. ---_._-------- , SI Units Named After Scientists Symbol Unit Quantity Scientist Country of Birth Oates A 1775-1 1736-1 1701-1 1791-1 1797-1 1857-1 1818-1 1824-1 1642-1 1787-1 1623-1 1823-1 1856-1 1745-1 1736 1804-1 ampere electric current Amphere, Andre·tvtarie France ( coulomb electric charge Coulomb, Charles Augustin de France O( degree celsius temperature (elsius. Anders Sweden F farad electric capacity Faraday, Michael England H Hz henry inductive resistance Henry, Joseph United States hertz frequency Hertz, Heinrich Rudolph Germany J joule energy Joule. James Prescott England K kelvin temperature Thomson, William (lord Kelvin) England N newton force Newton, Sir Isaac England fl ohm eleclrlc resistance Ohm. Georg Simon Germany Pa pascal pressure/stress Pascal, Blaise France 5 siemens electric conductance Siemens, Karl Wilhelm (Sir William) Germany (England) T testa magnetic flux density Testa. Nikola (roatia (US) volt electrical potential Volta, (ount Alessandro Italy walt power Watt, James Scotland weber magnetic flux Weber, Wilhelm Eduard Germany V W Wb Conversion of Units Moment (Torque) Length 1 centimeter (em) = 0.01 meter (01) 1 inch (in) = 0.0254 rn 1 foot (ft) = 0.3048 rn 1 dyn-cm = i Ibf-ft 1.356 N-m = i 0. 7 N-rn Work and Energy s' 1 yard (ydl = 0.9144 rn 1 kg-m' / 1 mile = 1609 m 1 dyn-cm = 1 erg = lO-} J 1 angstrom (A) = 10" rn 1 Ibi-It Time = 1 N-m = 1 Joule IJ) = 1.356 J Power = I minute (min) :::; 60 second (s) I kg· rn' / s' = 1 J/s 1 hour (h) = 3600 s I horsepower (hpl = 550 Ibl·IVs = 7461"1 1 day (dl = 86400 s 1 Watt (1"1) Plane Angle Mass J degree ("') .::;: 1</180 radian (fad) 1 pound mass (1 brn) = 0.4536 kilogr~lm (kg) 1 revolution (rev) 1 slug ::: 14.59 kg 1 rev;::: 211: raci -."-' 5.283 rad :=: 360" Temperature Force 1 kilogram force (kgf) = 9.807 Nevvton (f\J) <>( '" "r~ ~ 273.2 1 pound iorce IIbO = 4448 N I dyne (elyn) " I0 ~I Pressure and Stress 1 kgl rn·s:' = 1 N/m" = 1 Pascal (Pal I Ibi / in' (psi) = 6896 Pa I Ibl / ft' (psO = 92966 Pa I dyn / em' = O. I Pa Reprinted with pefFnisslon from OZkaya. 1'1., & No~dJn. l'1i. (1999). Fundamentals oi Biomechanics: EqUllib· rium. Motion. and DeformatIon (2nd ed.) New York: Spnnger-verlag. p. 11. The unit of pressure and stress. the pascal. was named after the French physicist, mathematician, and philosopher Blaise Pascal (1623-1662). Pascal conducted important investigations on the <.:harncteristics of vacuums and barometers and also invented a machine that would make mathematical calculations. His work in the area of hydrostatics helped lay the foundation for the later developmenl of these scientific ficlds_ In addition to his scicntific pursuits, Pascal was passionalely interested in religion and philosophy and tllllS wrote extensively on a wide range of subjects. The base unit of temperature, the kelvin. was named in honor of Lord Vv'illiam Thomson Kelvin (1824-1907). Named William Thomson, he was cd· ucated at the Universit.\· of Glasgo\\" and at C bridge University. Early in his career, Thol11son vcstigated the thermal propcnies of steam at a entific laboratory in Paris. At thc age of 32, returned to Glasgo\\" 10 accept the chair of Nalu Philosophy. His meeting with James JOllie in 1 stimulated interesting discussions on the natur heat, which eventually lecl to the establishmen Thomson's absolute scale of temperature, the Ke scale. In recognition of Thomson's contribution the field of thermodynamics, King Edward VIl c fen-cd on him the title of Lord Kelvin. The commonly llsed unit of temperature, the gree Celsius, \\ as named aftcr the Swedish lronomcl and lmenlO! Anders CelsiLl' (1701-17 Celsius was appointed professor or astronomy at the Inii"p-",i,\' of Uppsala at the age of 29 and remained the university until his death 14 years latec [n 1742, he described the centigrade thermometcl- in a paper prepared for the Swedish Academ!' of SciThe name of the centigrade temperature waS officially changed to Celsius in 1948. to Sf From Other of'Measurement Box App·t contains the formulae for the conversion ~f measurements expressed in English and non-Sl units into Sf units. One fundamental source of connJSlon in converting from one system to another is that (wo basic t~'Pes of Illeasun:menl systems exist. In the "ph\'sical" system (such as SI). the units of length. time, and nUlSS arc arbitrarily defined, and other units (including force) are derived fron1 these ,. base units. In "technical" or "gravitational" systems (slich as the English system). the units of len time, and force arc arbitrarily' defined, and o units (including mass) are derived from these b units. Because lhe units of force in gravitational tems are in fact the It'eights of stan(~"\rd masses. version to 51 is dependent on the acceleration mass due to the Earth's gravity, By' internati agreement. the acceleration due to gravity 9.806650 m/s'. This \'aille has been lIseel in estab ing some of the conversion factors in Box App-I. REFERENCES F~ircr, J.L. (19i7). SI .\leu;£' /-land/wok. Nt'w York: Ch StTibll('r'S Sons. Ozbva, N., ~ Nordin, ~'l. (1999). FWlfl(llllell1o!s or Bi c/;(wics: EquilihriulJI, .\lotioll, (lut! DejtmJ/(//ioll (2nd N<:w York: 5pringcr-Vl·r1ag. Pennychuick, C.J. {1974l. lIulldy .\!tllrif:l's oj" VI/il Crntl't' Pi/oon (or Riolof!..\' amI .\lcclulJI;cs. New York: John W &: Son~. \Vorld Health Org~lnizalion. jt's$iol1s. Gl.'llt'\"C WHO. (19i7). The SI /01' t!le fha/til Biomechanics of Tissue and Structures of the Musculoskeletal System Biomechanics of Bone Victor H. Frankel, Margareta Nordin Introduction Bone Composition and Structure Biomechanical Properties of Bone Biomechanical Behavior of Bone Bone Behavior Under Various loading Modes Tension Compression Shear Bending Torsion Combined l.oading Influence of Muscle Activity on Stress Distribution in Bone Strain Rate Dependency in Bone Fatigue of Bone Under RepetitivE: Loading !nfluence of Bone Geometry on Biomechanical Behavior Bone Remodeling Degenerative Changes in Bone Associated With Aging Summary References Flow Charts l _ ""''''''----------.--.:~.",----- lJitijoduction TbJ:i~~rpose or the skeletal system is to protect inte.r~.~V:"organs, provide rigid kinematic links and ~~ls:ctEtattachmcnt sites, and facilitate Illuscle ac· :·:::;:~"pJ:tf91iYli~9,body movement .. Bone has unique stl·UC· '-:/:::"'{:<'Y~lit~I.-.rI1d mechanical properties that allow it to ,,:~,<,:;;::,/,):',si;(rrY()llt;these roles. Bone is among the body's ,::)},:;;ll~lrci~~tstrllctllres; onl)' dentin and enamel in the :,:;:::';:,:::':i:,..'./t~¢th.>qreharder. It is one of the most dynamic and ::",';;::;.~/'i,1;i',.~t. ?9Pp'.cally active tissues in the body and re· >.";'\1.lAiri,S::~\ctiv0 throughollt life. A highl~; vascular tiss~e'.-.it has an excellent capacity for self-repair and ~lter its properties and configuration in rc. spOI~~e" to changes in mechanical demand. For example, changes in bone dcnsit~1 arc commonly obsei~·ed after pL:riods of disuse and of greatly increased use; changes in bone shapL' are noted dur· iog fracture healing and after certain operations. .;: Thus, bone adapts Lo the mechanical demands placed on i l. This chapter (iL'scribcs the composition and structure of bone tissue, the mechanical properties of bone, and the behavior or bone under different loading conditions. Various factors that affect the mechanical behavior of bone in vitro and in vivo also are disclissed. tan ·-'1 Bone Composition and Structure Bone tissue is a specialized connective tissue whose solid composition suits it for its supportive and proteclive roles. Like other connective tissues, it con· sisrs of cells and an organic extraccllular matrix of fibers and ground substance produced by the cells. The distinguishing reature of bone is its high con· tent of inorganic materials, ill the form of mineral salts, that combine intimately with the organic matrix (Buckwalter et al., 1995). The inorganic compo· nent of bone makes the tissue hard and rigid, while the organic component giv~s bone its flexil)ility and resilience. The composition of bone dirrers depending on site, animal age. dietar:.y histol)', and the pres· ence or disease (Kaplan et aI., 1993). In normal human bone, the mineral or inorganic portion of bone consists primarily of calcium and phosphate, mainly in the form of small crystals rc· sembling synthetic hydroxyapatite crystals \vith the composition Ca",(PO,)o(OI-l),. These minerals, which a~count for 60 to 70% of ilS dry weight, give bone its solid consistency. "Vatcr nccounts for 5 to SOk and the organic matrix makes tip the J'cmainder .. , ..~l.' of the tissue. Bone scn·cs <.lS a rcsen/Oir for esse minerals in the bod~·. particularly calcium. Bone minend is embedded in variousl~« orie fibers of the protein collagen, the fibrous po of the extracellular matrix-the inorganic ma Collagen ribers (type l) are tough and pliable they resist strelching ~lnd have lillie L'xtensib CollagL'1l composes ~\PPJ'().\.irnatdy 9(YYo of the tracellular matrix and accounts for approxima 25 lO 3W>'r' of the dr~! wl'ight of bone. A univ building block of the body, collagen also is chie!" fibrous component of other skddal s tures. (A detailed descriplion 01" the microstnlc and mechanical behavior of collagen is provide Chapters 3 and 4.) The gelatinous ground substance slIlToun thc mineralized collagcn fibers consists mainl protein polysaccharides, or gl~"cosarninogl (GAGs), primarily in the form of cOlllplt:x ma molecules called protcoglycans (PGs). The G s~rvc as a cementing substance between laye mineralized collagen fibers. Thcs~ GAGs, a wit.h various noncollagcI1ous glycoprotcins, co tute approximately 5% of the extracellular rna (The structure or PG:-:, whi,:h arc vital compon of artie, i:il' '_·;l~·{iL\gt·. is described in delail in C kr .) ..1 \Vater is fairl~: abundant in live bone. accoun for up to 25% of its total weight. Approxim 85°10 of the walt.'r is found in the organic ma around the collagen fibers and ground sllbsta and in the h.vdl·ation shells surrounding the crysl~ds. The other 15% is localed in the canals cavities that house bone cells and carry nutrien t h~ bone tissuc. At the microscopic level, the fllndamcmal S lllralunit of bone is the osteon, or haversian sy (Fig. 2·1). At the center of each osteon is a s channel, called a haversian canal, lIlat con blood vessels and nerVe fibers. The osteon itself sists of a concentric series of layers (lamella mineralized rnatrLx surrounding the central can configuration similar to growth ring~ in a trunk. Along the bOllndflries of each layel: or lamella small cavities known as lacunae, each cOl1la 011(.' bone cell. or ostcocyte (Fig. 2-1 C). Nume small channels, called canaliculi, radiate from lacuna, connecling the lacunae or adjacent lam and ultimately reaching the haversian canal. processes extend from the osteocytes into the ca culi, allowing nutrients '[Tom the blood vessels i haversian canal to reach the osteocyles. Lamellae OsteocYle~ Lacuna----- B A A. The fine structure of bone is illustrated schematically in a section of the shaft of a long bone depicted without inner marrow. The osteom. or haversian systems. are apparent as the structural units of bone. The haversian canals are in the center of the osteons, which form the main branches of the circulatory network in bone. Each osteon is bounded by a cement line. One osteon is shown extending from the bone (20x). Adapted from Basset!, CAL. (1965). £/ecrrical effects in bone. SCI Am, 213.18. B, Each osteon consists of lamellae. concentric haversian canal. Adapled irom Torrora G.J.. & Anagno takas. N.P. (198:1). Principles of Anatomy and Physiolog edJ. Ne~·v York: Harper gRow. C, Along the boundar the lamellae are small cavities known as lacunae, e of which contains a single bone cell, or osteocyte. ating from the lacunae are tiny canals, or canalicu into which the cytoplasmic processes of the osteoc extend. Adapted from Torrora G.;.. & AOclgflosrak05. N (1984). Principles of AnalOmy and Physiology (4th edJ. York: Harper & Ro~"/. rings composed of a mineral matrix surrounding the • At the pcriphcl)! of each osteon is a cement line, a nrtrrow area of cement-like ground SubSlrtllCe composed primarily of GAGs. The canaliculi of Ihc osteon do not pass this cement line. Like the canaliculi, the collagen fibers in the bOl1e matrix interconnect from one larndla to another within an osteon but do not cross the cement line. This intertwining of collagen fibers within the osteon undoubtedly in-, creases the bone's resistance to mechanical stress and probably explains _~vhy the cement line is the weakest portion of the bonc's microstructure. A typical osteon is approximately 200 micr tcrs (J..l) in (lin·rnctcr. Hence, evclY point in th teon is no more than 100 J..llll from the central catcd blood supply. In Ihe long bones, the os usually run longitudinally. but they branch quelltly and anaslOmose extensively with othcl: Jnterstitial lamellac span the regions bet complete O!"h:ons (Fig: 2-1..1). They arc contin with the ostcons and consist of the same mater a difTerent geol11ctric configuralion. As in th ! Frontal longitudinal section through the head. neck, greater trochanter, and proximal shaft of an adult femur. Cancellous bone, with its trabeculae oriented in a lattice, lies within the shell of cortical bone. Reprinted with permission from Gray, H. (T 985). Anatomy of the Human Body. (73(h American ed.J. Philadelphia: Lea & Febiger. 1 teons, no point in the interstitiallamellae is farther than 100 JLm from its blood supply. The interfaces between these lamellae contain an alTa)' of lacunae in which oSleocytes lie and from which canaliculi cXlcnd. !\l the macroscopic level, all Qonc,~ ..,!re.. c.Qrnposed of twotvpes, ,00~os'se-qus,~isslle: cortical. or compact, bone ~\n~d ci\'nc-c-il~·~IS.-Ol:"tl'abeclllal~"'iJ()nc'(FE~-"-i~·2). COl'tiC,,1 hoi1"Tili'ms the Otlt",:,11-"II:or' cortex: of the b()I"1~,tll~Cl.. has a ~l,~nse stt'llct,llre ,Si..11]i,ic.\I:t~)th~\i of ivory. C;ln~~Tlou's bo;;e within ihi~~I;ell is composed or-thin plates. o'j"lrabeclllae, in a loose mesll.-,~t1·.uc­ lUI~e; lhf([rlier'si-i·ccs--bctwe.en the lrabct:ulat:: are filled \\~hh red ;;all'O\';" (Fig. 2~j). C~I~~~ii~ll~-b~-~~~~e isaiTangeCi A, Reflected-light photomicrograph of cortical from a human tibia (40;<). B, Scanning electron p tomicrograph of cancellous bone from a human tibi (30)-:). Reprinted wirh permission from Carrer. D.R., & Na We. (1977). Compact bone fatigue damage. A microsco aminalion. Clin Onhop, 127, 265. in concentric lacunae-cQn'i"ainTilgJii!l)el- lae--ouiclocsn-ot em~l;-in haversian canals. The osteoc:\;tes re'ceive nUlricn-islhrough canaliculi from blood vessels passing through the red marrow. ~.Q.r­ lieal..bone always SUITOt.II).~~~ G!.pccl!.ous lJ.one. bUl· thi-rdative quantily of each lype varies' among bones and wilhin individual bones according to funclional requirements. On a rnicroscopic level, bone consists of .w and lamellar-bone (Fig. 2-4). Woven bOlle is-co erecl"immatllre bone. This type or bone is rOll t1 d enlbryo.-in the newborn, in the- fracturc'callus, lhe riietaphysial region of growing bone as wel tUIl-lOrS, ostcogcl]csis lmperfecla, and pagctic Lal-nellar bone begins to form 1 month arter bir activelv replaces woven bon~. Lamellar bone is therC'for~ a "marc" l11all~'c bOlle.--All bones are §urrounded by a dC.llse __ U_bJ~olls membrane called the periosteulll (Fig. 2-1.-\). Its ouler fa)~ei: (~ -pe-i::nl-eat-e-(Ch~);-'-GTood vessels (Fig. 2-5) and nerve fibers that pass into the cortex via Volk,llunn's cH"nalS: c()r\-i1c-EUn·g~\\;i.t~-1 thc--rl~i\.;e,:sian eLl m\ls '\Il'(Cc',~,~~,6~,lir~~Lt(? tll~-,s,a.·!1·~,~,1..1('-li~ ,I)one .. f\ n i n I'~<;-l~" o~_~,~~?g,~n .i..~,J,_~ver C()ll,t,,~1.'i Ils '" l?5)Il~ ~t:llsIY~ spo"nsible.....",'(or ,5:_".". ~(;nen~ting nc\v"" •.Gc;'n-c'''du'i-ing gro\Vtl~, . .",.. '-""'-"""~""""""" -""" . - which m":Cco\;'cr'ccT'witTl arlicufm- cartilage. I long borles, 'a- thin'llcr "nl'e'moi~a-ne:llle-cndo v:,\' line";; t heee n 1 ral (;;;";;clujIaty j ~,,",;il Ii icll;~ \VilTl'-y~cfl~~~~;~ fall'; -iiuirio\\':~ The' endostcum tai,'ls~~~s:l'~~;~l;ia~t; ,~,~d .alse),. giant n~-~,fli"'l-ll1-c' l.ioli"-e"lisctlTfec! osteoela-sfs: both of which i m pOI:f~li1i'I~(;fc's"'Ttl-dlc'l~cm odc ling aMi1~ir"(:cso I' of bone . - - -• • - and ,'cRail'. (ostcoblasls). The periosteum c th~cnli~ -b0I1::' except for the - joint-'Slld" - '-- - " - Lamellar ~ .f v~~, J /'i .\ } , l ,~. \. \ \) '("'1::.;, Woven Schematic drawing and photomicrographs of lamellar and woven bone, Adapted from Kaplan, F.5., Hayes, We., Keaveny. T.M., er a/. (1994). Form and funcrion of bone. In S.R. Simon (Ed.). Orthopaedic Basic Science (pp. 129, 130). Rosemonr, fL: AAOS. .~ Photomicrograph showing the vasculature of cortical bone. Adap,ed from K~lplan, w,e.. Keaven}~ rA1 .. et M (1994). Form and function of bone. In 5.R, Simon (Ed.). Orthopaedic Basic Science (p. /3i}. Rosemon£, It: AAOS. FS .. Nayes, • Biomechanical Properties of Bone Biomechanically, bone tissue may be regUl~ded a~ a t\v9~ph~,lse (biphasic) c01JlposiL(;~ I1),aterial;" with the mineral as onCjJhase and the collagen and ground substance as_.the o~her. In such matcl"'ials (n nonbio~­ logic-al exal,'lple is fiberglass) in which a strong. brittle material is embedded in a w~akcc more Oexible one, the combined substances ;:tre stronger for their weight than is either substance alone (Bassett, 1965), Functionally, the most important mechanical properties of bone are its strength and stiffness. These and other characteristics can best be understood for bone, or any other structure, by e:'\amining its behavior under .loading, that is, under the innuence of externally applied forces. Loading causes a deformation, or a change in the dimensions, of the~-strllctllre. \·Vhen a load in a known direction is imposed on a structure, the deformation 01" that structure can be rncasurcd and ploued 011 a loaddeformation curve. I\lluch information about the strength, stiflness, and other mechanical proper of the structure can be gained b~' examining curve. A hypothetical load-deformation curve for somewhat pliable fibrous structure, such as a l bone, is shown in Figure 2·6, The i1~~lial (strai line) portion of the curve, the c1asticJ'cgion. reve the elasticity bf the structure, th:u is, its capacity returning io its original shape after the load is moved, As· the load is applied, deformati911 occ btll is not permanent; th-e structure recovers its o inal sllape \\'hen unloaded. As loading c·ontinues. outcrni"ost fibers of the struCture begin to yield some point. This yidd poinl signals the elastic li of the structure, As, the load exceeds this limit, structure exhibits plastic behavior, I'cflecl~d in second (curved) portion of the eurvc, the p!as,tic gion. The structure \\~i11 no longer return to its or 11al dimensions when the load has been releas some residual deformation will be permanent loading-is progressively increased, the structure fail at'somc point (bonc' will fracturc), This poin indic~-iled by the ultimate failure point on the CU Plastic region / '" Yield point D ro S 1'- D 1 1 1 1 / 1 c i Ultimate failure point Energy 1 1 1 1 1 1 A D' Deformation Load-deformation curve for a structure composed of a somewhat pliable material. If a load is applied within the elastic range of the structure (A to B on the curve) and is then released, no permanent deformation occurs. If loading is continued past the yield point (B) and into the structure's plastic range (B to C on the curve) and the load is then released, permanent deformation results. The amount of permanent deformation that occurs if the structure is loaded to point 0 in the plastic region and then unloaded is represented by the distance between A and D. If loading continues within the plastic range, an ultimate failure point (C) is reached. • Three parameters for determining the strength of a structure are reflected on the load-deformation Clll-ve: 1, the load that the structure can sustain before failing; 2, the deformation ihat it cansustain before failing; and 3, the erler¥.Y that it CaIlstore be~ fore failing. The strel1gth in terms of loaclancLdefm:"mation, or ultimate strength, is in,clicatedOllJhe curve by the ultimate failure point. The streI,lgtl1 in terms of energy storage is indicated by the size of lhe area under the entire cun'e. The larger the area, the greater the energy that build~up in tfle structure as the load is applied. The stiffness of the structure is indicated by theslope of tl).<::.curve in the elastic region. Thesteepei::::the slope, the stifrer the material. the load-deformation curve is useful for deterInining the mechanical properties of whole struc~ turessuch as a whole bone, an entire ligament or tendqn, or a metai implant. This knowledge is helpful in the study of fracture behavior and repail~ the response of a struetlJre to physical stress, or the er~ fect of various treatment programs. However, char~ acterizing a bone or other structure in terms o material of \vhich it is composed, independent o geometry, requires standardization of the tes conditions and the size and shape of the test sp mens. Such standardized testing is useful for c paring the 111echanical properties of two or m materials, such as the relative strength of bone tendon tissue or the relative stiffness of various terials used in prosthetic implants. More pre units of measurement can be used when stand ized samples are tested-that is, the load per un area of the sample (stress) and the amount of de mation in terms of the percentage of change in sample's dimensions (strain). The curve generat a stress-strain curve. Stress)s:the load, or force, per unit area tha velops on a plane surface within a struc:lll['e.)1 sponse'iO'exteI~ililll)' applied loads. The three ~ most commonly used for measuring stress in s dardized samples of bone are ne\vtons per cent ter squared (N/cm:;); newtons per meter squared pascals (N/m 2 ,Pa); and megancwtons per m sqmii;cd; or mega pascals (MN/m 2 , MPa). Strain is the clef()l'mation (change in dimens that develoP?~yithiI1.~ls:tTuctureirl"I:t::sponse to ternallyapplied loads. The two basic types of st are ii'near strain, which causes "a- chang~ ill lengtl;"'6f'dlcspecimen, and shear strain, w causes }i"'~I~~i:~¥?i~.(ll~angulari'·clationships \v the structure. Linea-I'· strain is measured as amoliiltofflIlear de[ormati()n (lengthening or sh ening}"6fthesiilnple di\'ided by the sample's orig length. It is a nondim<;nsj()llal paramel<;r expre as a percentage (e.g:, centimeter per centime She'~1'1:'stt'~lTI1"is measured as the all"1Q.~lIlL.of ang ch~~!!.g,~. ,,,(Y)_i,l'-adglif _ ~~lj:~leI )'i l1gi"I"i.t}l<; pl.nne.o terest in the sample. It is expressedil"il'actians radian-'e(ill~llsai)proximately57.3°) (Internatio Society of Biomechanics, 1987). Stress and strain values can be obtained for b by placing a standardized specimen of bone ti in a testing jig and loading it to failure (Fig. 2 These values can then be plotted on a stress-st curve (Fig. 2-8). The regions of this curve are s lar to those of the load-deformation cUll/e. Load the elastic region do not cause perll1aD.<;nL~I~J( lion, buC6ncc:the yield point is excee.deeJ,s()ll1c Formation L5 permanent. The strengthoftheJl1~l inl in terms of energy' storage is repre~ent~elby area .~II1~1.<::I·theyntire curve. The stifFness is re sented h.ytheslope of th<? curve in the~Iasticreg A value for stiffness is obtained by dividing stress at ~ __point ir~_.thc clastic (straight line) porti()11 of trlC"cllI"\'e by i·he-~t~,~~·in ~t that point. This -is caITedtTle-nl0dlll~I~~ o!:-elastfcity (Young's modulus). Young'S modulus (E) is d~,T;'ed from Ihe relalionship betw';-en str"ss ('T) an~.strain(~): value E=<r/E The elasticity 01" a material or the Young's modulus E is equal to the slope of the stress (<r) and strain (E) diagram in the clastic linear region. E represents the stiffness of Ih" material, such Ihat Ihe higher the elastic modulus or Young's modulus, the stiffer the material (Ozkaya & Nordin, 1999). Mechanical propcrties differ in the two bone types. Cortical_bq__,!_c ~~ s~.Hfer_tlla!,! canccllQus_ bone, withstanding greater stress but less strain before falllll·c. C:-1-n~c~ITol-i-s -bolic-rn·-\'itl:-O--~~·~I~-~;jn lip to 50% of strains before yielding. \vhile cortical ~)(?nc viclds and fractures when the strain exccec!s 1.5 10 2.00/0. 8cc~\i.isc---<.)r ·ils~·I)O-'·()lis---stl~ucl·ure, ca..Q~J~I.~_ IOlls~~~=_.I:..'~_s a la-j";gc c~:.pa.::itS~_ (61~ enei·gy..sto,~ge (Ke"""ny & Hayes, 1993). The physical differen betwecn the two bone tissucs is quantified in ter of the apparent density of bone, which is ddined_ the mas~_.<?Ip·9ii~jI~~u£:i?,:·~~~ntin a unit of bon~ \ ~~f11e (gram per cubic ccn-titnctci:Tglcc]):-Frg"Llre depicts typical stress-strain qualities of cortical a trabecular bone with different bone densities tes under similar conditions. In general, it is enough lo describe bone strength with a sin number. A bctter way is to examine the stress-str curve for the bone tissue under the circumstan tested. To better understand th" relationship of bone other materials, schematic stress-strain curves bone, metal, and glass illustrate the differences mechanical behavior among these matcrials (F 2~10). The variations in stiffness are rcOecled in different slopes of the cun'cs in the clastic regi Metal has the steepest slope and is thLis the stiff material. C' .•• ---- .. -••• -------- ••••........••••........••••.... C PlastiC region B' A B" C" Stress-strain curve for a cortical bone sample teste in tension (pulled). Yield point (B): point past whic some permanent deformation of the bone sample curred. Yield stress (8'): load per unit area sustaine by the bone sample before plastic deformation too place. Yield strain (8"): amount of deformation wi stood by the sample before plastic deformation oc curred. The strain at any point in the elastic region of the curve is proportional to the stress at that point. Ultimate failure point (C): the point past which failure of the sample occurred. Ultimate stre (C'): load per unit area sustained by the sample be fore failure. Ultimate strain (C"): amount of defor mation sustained by the sample before failure. ~--Standardized bone specimen in a testing machine_ The strain in the segment of bone between the two gauge arms is measured with a strain gauge. The stress is calculated from the total load measured. Courtesy of Dennis R. Carter. Ph.D. • Apparent Densily ~ ~ ~ ..... 0.30 glee :::1 . - - . 0.90 glee CortIcal bone - - 1.85 glee 100 iii 50 / / -------------- / o.j.........,=.:...:. .:".+"..:..:. ".:....:..:.:. :".:." Trabecular bone :.c ..c.;.,:..:.":.. •.c.;":..:•.:.":...".f.:.c.;".:.":..".:.."---< 15 20 25 -'<":..:.":..:..:. . .:..:... o 5 10 scnce of a plastic region on the stl"L'ss-strain cur By contrasl, metal exhibits cxtensh·t: deformati before failing, as indicalC'd b~' a long plastic reg on the ClitYC. Bone also deforms before failing to a rnuch lesser extent than metal. The differen in the plastic behavior of metal and bone is the suit or differences in microlllcchanical events yield. Yielding in melal (tested in tension, pulled) is caused b~' plastic rIow and the fonrwti of plaslic slip lines; slip lines art: formed when molecules of lhe latticc structurc of mctal dis cate. Yielding in bone (tested in tcnsion) is caus Strain (%) Met Example of stress-strain curves of cortical and trabecular bone with different apparent densities, Testing was performed in compression. The figure depicts the difference in mechanical behavior for the two bone structures. Reprinted with permission from Ke,1'.'eny. T M., & Hc1yes, \tV. C. (1993). Mechanical properri(;.ls of cortical ancl rfDoecular bone, Bone. 7, 28S·]t]4, The clastic portion of the curve for glass und metal is a straight linc. indicating linearly ei<:\slic behavior; virtually no yielding lakes place before the yield point is reached. By comparison. precise testing of cortical bone hns shown thnt the elastic portion of the curve is not straight but instead slightly curved. indicating that bone is not linearly claslic in its bdlm'ior but yields somewhat during loading in the elastic rcgion (Bonefield & Li, 1967)" Table 2-1 depicts the mechanical propcrtics of selectcd bio· materials for comparison. Materials are classified as brittle or ductile depcnding on the extent of defor· mation before failure. Glass is a typical brittle material. and soft metal is a typical ductile material. The cUrrerence in the amount or deformation is reOected in the fracture surfaces or the two materials (Fig" 2-11)" When pieced togethcr aftcr fracture, lhe ductile material will not conform to its original shape whet'cas the brittle material will. Bone ex· hibits more brittle or ITIOI'C ductile behavior dt> pending on its age (younger bone being more ductile) and the rate at which it is loaded (bonc bcing more brittle at higher loading speeds). After the yield point is reached, glass deforms very little before failing, as indicated by the ab· Glass Bone Strain Schematic stress-strain curves for three materials. Metal has the steepest slope in the elastic region a is thus the stiffest material. The elastic portion of curve for metal is a straight line, indicating linearl elastic behavior. The fact that metal has a long pla tic region indicates that this typical ductile materia deforms extensively before failure. Glass, a brittle material, exhibits linearly elastic behavior but fails abruptly with little deformation, as indicated by th lack of a plastic region on the stress-strain curve. Bone possesses both ductile and brittle qualities demonstrated by a slight curve in the elastic region which indicates some yielding during loading with this region. • ~ by dcbonding of thL' ostC'ons 4\t the cement lin and micl'ofracturc (Fig. 2-12), while ~'idding bone as a result of compression is indicated I Mechanical Properties of Selected iI Biomaterials cracking of lhe osteons (Fig. 2-13). Ultimate Strength Modulus Elongation' (MPa) (GPa) (%) Metals Because the structure of bone is dissimilar in transverse and longillidinal. directions, it exhib diffcrerl'l mechanical propenics whc.I1 loaded alo differenl axes, a characleristic known as anisotro Co-Cr alloy Cast 600 220 forged 950 850 900 220 210 110 Stainless steel Titanium 15 10 15 Polymers Bone cement 20 2.0 2-4 ,.-; Ceramic Alumina 300 350 <2 100-150 8-50 10-15 1-3 20-35 2.0-4,0 10-25 Biological Cortical bone Trabecular bone Tendon, ligament 2-4 Adapted from Kummer, J,K, (1999) Impldn, biomillcrials In J.M. Spivak, P.E, DiCesare, Os. Fe[dman, K,!. Ko'''.1[, A.S. RokilO, & J.D. Zuckerrnan (Eds.). Orrhopaedic5.' A Study GlII'de (pp. 45-48). Ni:>,v Reflected-light photomicrograph of a human corti bone specimen tested in tension (30):). Arrows ind cate debonding at the cement lines and pulling ou of the osteoos. Courtesy 0; Dennis R. Carter. Ph.D. York: lvlcGraw-Hlil. 1t - - - - - - - II i Ductile fracture I I -. I Brittle fracture ~ I ,urface' of sample, of a ductile and a br;ttle material. The broken Jines on the ductile material indicate the original length of the sample. before it deformed. The brittle material deformed very little before fracture. Scanning electron photomicrograph of a human co tical bone specimen tested in compression (30X). A rows indicate oblique cracking of the osteons. Courtesy Dennis R. Carter, Ph.D. • 200 MPa ---- Strain Anisotropic behavior of cortical bone specimens from a human femoral shaft tested in tension (pulled) in four directions: longitudinal (L). tilted 30" with respect to the neutral axis of the bone, tilted 60", and transver (T). Data from Frankel, V.H., & Burstein, A.H. (1970). lhopaedic Biomech<1nlcs. Philadelphia: Lea & Febiger. • f·7jgure 2-14 shows the variations in strength and stiffness for cortical bone samples from a human remoral shah, tested in tension in four directions (Frankel & Burstein c 1970; Carterc 1978). The values for both parameters are highest for the samples loaded in the longitudinal direction. Figures 2·9 and 2·15 show trabecular bone slI-cngth and stiffness tesled in two directions: compression and tension. Trabecular or cancellous bone is approximately 25% as dense, 5 to look; as stiff, and five times as ductile as conical bone. Although lhe relationship bel ween loading patlerns and the mechanical properties of bone throughout the skeleton is extremely complex, it generally can be said that bone strength and stiffness are greatest in the dircction in which daily loads arc most commonly imposed. 10 B 6 4 2 0-1----1----1----1---+--"'"-1 o BiOlnechanical Behavior of Bone The mechanical behavior or bone-ils behavior under the innuence of forces and moments-is affected by its mechanical properties, its geometric characteristics, the loading mode applied, direclion of loading, rale of loading, and frequency of loading. 2 4 6 B Tensile strain (%) Example of tensile stress-strain behavior of trabec lar bone tested in the longitudinal axial direction the bone. Adapted from Gibson, L.1., & As/lby. M.F. (1988). Cellular Solids: Structure and Properties. New York: Pergamon, Press. pendicular to the applied load (Fig. 2-17). Under stnl.c:ture lengthens and narrows Clinically, fractures produced by tensilelqad are usually seen in Dones with a large proportio cancellous bone. Examples are fractures of the b or thCfifth metatarsal adjacent to the attachm of the pe-roneus brevis tendon and fractt,lres the calcaneus adjacent to the attachment of AchUlqs tendon. Figure 2-18 shows a tensile frac through the calcaneus; intense contraction of triceps surae muscle produces abnormall,Y high sile loads on the bone. sil~l(),l(ling, the .. Compression Tension • • Shear I'~. Compression I I i eli C..J Torsion ~L.-.- 1 Bending Combined loading _ Schematic representation of various loading modes. DUling compressive loading, equal ~nd opposite lO are applied toward the surface of the structure compressive stress and strain. result inside the st ture. Compressive stress can be thought of as m small forces directed into the surface of the stluct Maximal compressive stress occurs on a plane pendicular to the applied load (Fig. 2-19). Under c pressive loading, the structure shortens and widen Clinically, COin pression fractures are commo found in the vertebrae. which are subjected to h compressive loads. These fractures are most o seen in the elderly with osteoporotic bone tis Figure 2-20 shows the shortening and widen BONE BEHAVIOR UNDER VARIOUS LOADING MODES Forces and moments can be applied to a structure in various directions, producing tension, compression, bending, shear, torsion, and cornbinecl loading (Fig. 2-16). Bone in vivo is subjected to all of these loading modes. The folknving descriptions of these modes apply to structures in equilibrium (at rest or moving at a constant speed); loading produces an internal, deforming efFect on the structure. Tension During tensile loading, equal and opposite loads are applied olltward from the surface of the structure, ancCtensile stress and strai"n result inside the structure. Ten'sile stress can be thought of as manv small forces directed· <.nva:v from the -;'urface of th~ stlJIClure. Maximal tensile stress occurs on a plane per- Tensile loading. Tensile fracture through the calcaneus produced by strong contraction of the triceps surae muscle during a tennis match. Courtesy of Robert A. Winquisr, lA.D that takes place in a human vertebra subjected to a high compressive load. In a joint, compressive loading to failure can be produced by abnormally strong contraction of the surrounding muscles. An example of this effect is presented in Figure 2-2 t; bilateral subcapilal fractures of the Femoral neck Compression fracture of a human first lumbar ver bra. The vertebra has shortened and widened. • were sustained by a patient undergoing electroc vulsive therapy; strong contractions of the mus around the hip joint compressed the femoral h against the acetabulum. .. Compressive loading. Shear During shear loading, a load is applied parallel to surface of the structure, and shear,stress and st result inside the structure. Shear stress can thought of as many' small forces acting on the face of the structure on a plane parallel to the plied load (Fig. 2-22). J\ structure subjected t shear load deforms internaH.v in an angular-rllan right angles on a plane surface within the stnlc Force L Before loading ;~ 1Dfm I I _ Bilateral subcapital compression fractures of the femoral neck in a patient who underwent electrocon· vulsive therapy. e f become obtuse or acute (Fig. 2-23'). \'Vhel~e\'er a structure is subjected to tensile or compressivelqading, sflear stl:ess is produced. -Figli;·c 2-24 iIllls11:a'tcs angtiliii· deformatiorl in sfi~lctllrcs subjected (0 these loading modes. Clinically, sh<;ar fractures are most ohen seen in cancel lOlls bone. Human «clllit 'corticai bone exhibits different values for ultimate stress uncleI' compressive, tensile, and shear loading. Cortical bone can withstand greater stress in compression (approximately 190 Mpa) than in tension (approximately 130 !\Ilpa) and greater stress in tension than in shear (70 r\'lpa). The Under shear loading When a structure is loaded in shear, lines original at right angles on a plane surface within the struc ture change their orientation, and the angle beco obtuse or acute. This angular deformation indicat shear strain. .----------------- daslicily (Young's modulus) is approximately GPa in longitudinal or axial loading and appro 1l1alcly I I GPa in transverse loading. Human becular bone values for testing in compression approximately' 50 ivlpa and arc reduced lO appr *** D o <> 0 1 Unloaded I F. ~ '--------- _ Shear loading. 1 -~ Under lensile loading CJ <> *** Under compressive loading The presence of shear strain in a structure loaded tension and in compres.sion is indicated by angula deformation. Cross-section of a bone subjected to bending, showing distribution of stresses around the neutral axis. Tensile stresses act on the superior side, and compressive stresses act on the inferior side. The stresses are highest at the periphery of the bone and lowest near the neutral axis. The tensile and compressive stresses are unequal because the bone is asymmetrical. lateral roentgenogram of a "boot top" fracture duced by three-point bending. Courtesy of Robert \J1/inquist, M, D. mately 8 Mpa iF loaded in tension. The modulus or elasticity is low (0.0-0.4 GPa) and dependent on the apparent density of the trabecular bone and direc~ lion of loading. The clinical biomechanical consequence is that the direction of compression failure results in general in a stable fracture, while a fracture initiated by..' tension or shear ma!' have catastrophic consequences, .... 1_/ .. A_-------.JO .... .. B_-----.JO 1_/_ _ Two types of bending. A, Three-point bending. B, Four-point bending. Bending In bending, loads are applied to a structure manner that causes it to bend ~\bout an,axis. \V a bone 'is loaded in bending, it is subjected combination of tension and compression. Te stresses and strains act on one side of the ne axis!._'111d compressive stresses and strain?,Ic;t on oth~r side (Fig. 2-25); there are I}ostresse,:str~liI!?along tt;enelltral axis. The ~lagnitllde o stresses is proportional to their distance from neutral axis of the bone. The farther the stresse from the neutral axis, the higher their magnit Because a bone structure is asymmetrical, stresses may not be equally distributed. Bending may be produced by thl'qe forces (th point bendiilg) Ol:.Xour forces (roUI'~point bend (Fig. 2-26). Fractures produced b)' both type bending are commonly observed clinically, par lady in the long bones. Three-point bendirlgtakesplace when t forces acting on a structure pn)c!uce two equal mcnts, each being the product of one of the t\\' ripfle!~;;liforces and its perpcndicular distance. the axis of rotation (the point at which the mi Forceis applied) (Fig. 2-26;\). IF loading conti to the )-'iele! point, the structure, if homogene symmetrical, and with no structural or tissue ,"'''' ~ Fatigued muscle o , ~ ~~ _"~ ~ " ~ ~ ~ ~"_c feet, will break at the point of application of the middle force. A typict;~Jthree~pointl)eIl(ling fracture is tht: "boot top"Jracturesustainedb:vskiers. In the "boot top" fracture shown in Figure 2-27, one bending moment acted on the proximal tibia as the skier fell forward over the top of the ski boot. An equal moment, produced by the fixed foot and ski, acted on thc distal tibia. As the proximal tibia was bent fonvard, tensile stresses and strains acted on the posterior side of the bone and compressive stresses and strains acted on the anterior side. The tibia and fibula fractured at the top of the boot. Because adult bone is weaker in tension than in compression, failure begins on the side subjected to tension. Because immature hone is n10r~ ductile, it nUl)' fail first in compression, and a buckle fracture may result on the compressive side (Flowchart 2~ I). Four~point bending takes place when two fo couples acting on a stl'uclurc produce two eq moments. A force couple is formed when two par lei forces or equal magnitude but opposite direct are applied to a structure (Fig. 2-28;\). Because magnitude of the bending moment is the sa throughout the area between the two force coupl the structure breaks at its weakest point. An exa ple of a FourNpoint bending FraclUre is shown in F ure 2-28B. A stifT knee joint was manipulated inc rectl y' during rehabilitation of a palient with postsurgical infected femoral fracture, During A • A, During manipulation of a stiff knee joint during fracture rehabilitation, four-point bending caused the femur to refracture at its weakest point, the original fracture site. B, Lateral radiograph of the fractured femur. Courtesy of Kaj Lundborg, M.D. Cross-section of a cylinder loaded in torsion, showing the distribution of shear stresses around the neutral axis. The magnitude of the stresses is highest at the periphery of the cylinder and lowest near the neutral axis. manipulation. the posteJ"ior knee joint capsule and tibia Formed one force couple and the fClTlOral head and hip joint capsule formed the other. As a bending moment was applied to the femur, the bone failed at its \vcakest point, the original fracture site. Torsion In torsion, a load is applie~l to a strl1c~.lII·c _in a mannel' that causes i"l to l\\;j'st about an axis, and a lO~'gue--(or ornament) is produced within the -struclure.--\Vhen a structure is loaded in torsion, shear 61 , . , Ii' <J) ,,, ,!' ~~ @.- Shear ,: Compression ,i<? ....... 1 '-.-' stre:;ses arc distributed over the entire structure in b~n'ding, the magnitude of these stresses is ponional to their distance from the neutral (Fig. 2·29). The farther the stresses arc from neutral axis, the higher their magnitude. Under torsional loading, maximal shear stre act on plan~sp~~-;'alleland~perpen(ficular·to-then tl"al axis of the structure. 'In addition, max"fillal sile aild comp'ressi\'c stresses act on a plane dia nal to the neutral a.'\is of the structure. Figure illustrates these planes in a small segment of b loaded in torsion. The fracture pattern for bone loaded in tor suggests thal the bone fails first in shear, with formation of an initial crack parallel to the neu axis of the bone. A second crack usually fo along the plane or maximal tensile stress. Suc pattern can be seen in the experimentally produ torsional fracture of n canine femur shown in ure 2-31. Tension I ,; Experimentally produced torsional fracture of a c nine femur. The short crack (arrow) parallel to th neutral axis represents shear failure; the fracture at a 30'" angle to the neutral axis represents the plane of maximal tensile stress. Schematic representation of a small segment of bone loaded in torsion. Maximal shear stresses act on planes parallel and perpendicular to the neutral axis. Maximal tensile and compressive stresses act on planes diagonal to this axis. Combined loading, Allhough each loading mode has been conside separately, living bone is seldom loaded in one m only. Loading of bone in vivo is complex for principal reasons: bones .~.re ~onstantly subjecte multiple indeterminate loads and their geome structure is irregulm: In vivo mcasurement of strains on the antcrol11cdial surface of a hum adull tibia during walking and jogging dcm SlrateS the comph.:.xil.V 01" the loading patterns during these cOl11mon ph.'·siological activities (Lanyon el aI., 1975). Stress values calculated from these strain measurel~lents by Carter (1978) showed thal during normal walking, the strc.i.~es wcre COlllP.I"CSs.ivc dllring heel strike, tensile during- thc--s_t-~!.nce ph~y~·,·,indi1¥aill--c()il·~prcssive-.~Iu'ri,~,lg- pLlsl.1~off (Fig. 2-3"2A). VaI"ucs for shear stress \\'cn.:.~ relatively high in the later portion of the gait c)ldc, denoting significant torsional loading. This torsional loading \vas associated with external rotation of the tibia during stance ancl push-ofr. Dul"ing jogging. the stress pattern was quite difrerent (Fi-g. 2:328). The COlJlP.~·.~~.~~c_slrcssPFcdominatin!! at lac strike was followed bv high tensile stress~di.iring r>ll-sh-ofr. The sllenr slrcss~:vi~~--'o\V tfil'oughout 'tlle strich:\ denoting 111ini;lwl torsional loadhig"pl~o(lllced by slight e~'\tern~" and inlcrm~l 1'0tatiOll or-the tibia in an-ahernating pallcrn. The increase--fll specd from slow walking LO jogging in- crcas~d both the str~ss and the strain Oil the lib (Lan.'·on ct aI., 1975). This increase in strain w grt,,:atcr speed was confirmed in studies or locom lion in sheep, which ckmonslratcd a fivefold crcase in slrain values from siD\\" walking to f lrotting (Lanyon & Bourn, 1979). INFLUENCE OF MUSCLE ACTIVITY ON STRESS DISTRIBUTION IN BONE When bone is loaded in vivo, the contraction of t muscles attached to the bone alters dlC stress dist bution in [he bone. This muscle contraction d cr<.:ases 01· eliminates [ensile stress on the bone producing compressi\'c SlI-CSS thal neutralizes it ther partially or totall~·. The effect of muscle contraction can be ill trated in a tibia subjected to three-point bendin Figure 2~33A represents Lhe leg of a skier who falling: forward, subjecting the tibia Lo a bendi Jogging (2.2 m/sec) 12 10 Walking (1.4 m/sec) 4 Stress Tensile Compressive Shear (external rotation) 3 2 ~ z 1 ;;; 0 +''<,--+-!----o:..-.--+,__+ ~"' -', {jj ! \ .....;, 4· A ;i •••••••• \. 3 rn a. ~ I 1 i ~ ~ 6 4 ~ iii __.-;~'c...,...+- ' ..j Tensile Compressive Shear (external rotation) Shear (internal rotation) 8 2 ...i -~'.. .:; ... i "' \ .. i " r h. l,\ HS FF HO HO TO 2· (1975). Bone deformation recorded in vivo (rom srrain gauges c1tr<elched to the human tibial sharr. ACla Orlhop Scand, 46. 256. Figure coortesy of Dennis R. Carler, Ph.D "' i i "'\ ii \i s A. Calculated stresses on the anterolateral cortex of a human tibia during walking. HS, heel strike; FF, foot flat; HO, heel-off; TO. toe off; S, swing. Calwlated from Lanyon, L.E.. Hampson. W'.G.J., Goodship. A.E.. et af. ................ i O' , \·······1··../ / 1\ B 4 .l---=-E--=~'-:':::-----TS TS·TO 8, Calculated stresses on the anterolateral cortex of a human tibia during jogging. T5, toe strike; TO, toe o Calwfared [rom Lanyon. L.E.. Hampson. W.G.)., Goodship AE., et al. (1975). Bone deformation recorded in vivo fro strain gaoges aClacIJed ro the human tibial Shaft. Acta Orthop 5cand. 46. 256. Figure courtesy 01 Dennis R. Carte Pll.D. , ,-, , ,, • "" r:~ , , ,, '' : I B A. Distribution of compressive and tensile stresses in a tibia subjected to three-point bending. B, Contraction of the triceps surae muscle produces high compressive stress on the posterior aspect, neutralizing the high tensile stress. moment. High tensile stress is produced on the postcrior aspect of the tibia, and high compressive stress acts on the anterior aspect. Contraction of the triceps surae muscle produces great compressive stress on the posterior aspect (Fig. 2-338), neutralizing the great tensile stress and thereby' protecting the tibia from failure in tension. This muscle contraction may result in higher compressive stress on the anterior surface of the tibia and thus protect the bone from failure. Adult bone can usuall~/ withstand this stress, but immature bone, which is weaker, ma.y fail in compression. Muscle contraction produces a similar effect in the hip joint (Fig. 2-34). During locomotion, bending moments are applied to the femoral neck and tensile stress is produced on the superior cortex. Contraction of the gluteus medius muscle produces compressive stress that neutralizes this tensile stress, with the net result that neither compressive nor tensile stress acts on the superior cortex. Thus, the muscle contraction allows the femoral neck to sustain higher loads than would otherwise be possible. STRAIN RATE DEPENDENCY IN BONE Because bone is a viscoelastic material, its biomechanical behavior varies with the rate at which the bone is loaded (Le., the rate at which the load is plied and removed). Bone is stifTer and sustain higher load to failure when loads are applied higher rates. Bone also stores more energy be failure at higher loading rates, provided that th rates are within the physiological range. The in vivo claily' strain can vary considerab The calculated strain rate for slow walking is 0. per second, \vhile slow running displays a st rate of 0.03 per second. In general, when activities become more stre ous, the strain rate increases (Keaveny & Ha 1993). Figure 2-35 shows cortical bone behavio tensile testing at different physiological strain ra As can be seen from the figure, the same chang strain rate produces a larger change in ultim stress (strength) than in elasticit.\' (Young's mo lus). The data indicates that the bone is appr mately 30(/0 stronger for brisk walking than for s c"~,~ ,/ :, ,,: , , , , E , ,, , ;, , ' ' ' ' : " Stress distribution in a femoral neck subjected to bending. When the gluteus medius muscle is relax (top), tensile stress acts on the superior cortex an compressive stress acts on the inferior cortex. Con traction of this muscle (bottom) neutralizes the te sile stress. 400 1500/sec 300 300/sec " I~ l::l ~--- O.lIsec 200 'I" - - - - - O.Ollsec iii Brisk walking O.OOl/sec 100 Slow walking o'l'---+---+---+------i 0.0 0.5 1.5 2.0 Rate dependency of cortical bone is demonstrated at five strain rates. 80th stiffness (modulus) and strength increase considerably at increased strain rates. Adapted from McElhaney, J.H. (1966). Dynamic response of bone and muscle tissue. J Appl Physio1, 2}, I, /23/-/236. III walking. At very high strain rates (> I per second) representing impact trauma, the bone becomes more brittle. In a full range of cxperimentaltesting for ultimate tensile strength and elasticity of corti, cal bone, the strength increases by a factor of three and the modulus by a faclor of two (Keaveny & Hayes, 1993). The loading rale is clinically significant because it innuences both the fracture paUern and the amount of soft tissue damage at fracture. \Vhen a bone fractures. the stored energy is released. At a low loading rate, the energy can dissipate through the formation of a single crack; the bone and soft tissues remain relatively intact, with little or no displacement of the ,. bone fragments. At a high loading rate, however, the ;'·',greater energy stored cannOt dissipate rapidly :~ enough through a single crack, and comminution of bone and extensive soft tissue damage resull. Figure . 2-36 shows a human tibia tested in vitro in torsion at a high loading 1'4ue; numerou:j bone fragments ;;J' were produced, and displacement of the fragme was pronounced. Clinically, bone fractures fall into three gene categories based on the amount of energy releas at fracture: low-energy!, high-energy, and VCI)' hi energy. A low-energy fracture is exemplifled by simple torsional ski fracture; a high-energy fract is often sustained during automobile accidents; a a very high-energy fracture is produced by v high-muzzle velocity gunshot. FATIGUE OF BONE UNDER REPETITIVE LOADING Bone fractures can be produced b:v a single lo that exceeds the ultimate strength of the bone by repeated applications of a load of lower mag tude. A fracture caused by a repeated load appli tion is called a fatigue fracture and is lypically p duced either by few repetitions of a high load or many repetitions of a relatively normal load (C Study 2-1). The interplay of load and repetition for any m te!"inl can be plotted on a fatigue curve (Fig. 2-3 For some materials (some metals. for exampl thc fatiguc curve is asymptotic, indicating tha the load is kept below a certain level, theoretica the material will remain intact no matter h many repetitions. For bone tested in vitro, curve is not asymptotic. \,Vhen bone is subjected repetitive low loads, it may sustain microfractur Testing of bone in vitro also reveals that bone tigues rapidly when the load or deformation proaches its yield strength; that is, the number repetitions needed to produce a Fracture dim ishes rapidly. In repetitive loading of living bone, the fatig process is affected not only by the amount of lo and the number of repetitions but also bv number of applications of the load within a giv time (frequency of loading). Because living bon self-repairing. a fatigue fracture results only wh the remodeling process is outpaced by the fatig process-that is, when loading is so Frequent t it precludes the remodeling necessary to prev failure. Fatigue fractures are llsually sustained dur continuous strenuous physical activity, wh causes the muscles to become fatigued and redu .' their ability to contracl. As a result, the.v are l able to store energy aI~d thus to neutralize stresses imposed on the bone, The resulling al ation of the stress distribution in the bone cau Human tibia experimentally tested to failure in torsion at a high loading rate. Displacement of the numerous fragments was pronounced. B"ne Overloading vent failure. iVluscie fatigue occurred as a result of the abno mal loading pattern and the intensive uaining, It affected th 23-year-old military recruit was exposed to an intensive " , ' heavy physical training regime that included repetitive A continuous crawling in an awkward position for several weeks (Case Study Fig. 2-1-1 A). The repeated application of loads (high repetitions) and the number of applications of a load during a short period of time (high frequency of loading) muscle function in the neutralization of the stress imposed, leading to abnormal loading and altered stress distribution (Case Study Fig. 2-1-1 B). After 4 \-veeks of strenuous physical activity, the damage accumulation from fatigue at the femoral shaft lead to an oblique fracture. surpassed the time for the bone remodeling process to pre- Case Study Figure 2-1-1A. Abnormal loads at the femoral shaft occurred. ,'. ".3 m Injury Repetition The interplay of load and repetition is represented on a fatigue curve. • abnormally high loads to be imposed, and n fatigue damage accumulation occurs that Illav lead to a fracture. Bone may fail on the tensil~ side, on the compressive side, or on both sides. Failure on the tensile side resulls in a transverse crack, and the bone proceeds rapidly to complete fracture. Fatigue fnlctures on the compressive side appear to be pro- II ::1 'E 0006 I~ 0 004 l~ Ii g~ · I tI) 0.002 o Compression • TenSion Miles 10 100 1000 I • I ! • vlgorous,-:":::!::::""~~S~'b-'<:::::::-;:;-~ exercise Running -----------===--:::- Walking 0.0001+0-0--1---+---+----+---!------I 1~ 1~ 1~ Number 01 Cycles "i lam I -------- . I I I ' ,i i i Fatigue testing showing the number of cycles (x-axis) and strain range (y-axis) expressed as stress rangel modulus in human cortical bone specimens loaded in tension and compression. Typical strain ranges are shown for walking, running. and vigorous exercises. Note that resistance to fatigue fracture is greater in compressive loading. Ten miles represent approxj· mately 5,000 cycles, corresponding to the number of steps running during that distance. Aclapared from Carrer. D.R., Cater. W.E., Spengler, O.M.. Frankel, \l.H. (J 98 1). Fatigue behavior of adtilt cortical bone: the influence of mean srrc1ilJ anel strain range. Acra Orthop Seane!. 52.48/-490. J.----------------- duced more slowl~'; the remodeling is Icss casily o paced by the fatiguc process and the bone may proceed to complete fracture. This theol)1 of muscle fatigue as a cau~e of tigue fracture in the lower extremities is outlin in the schema in Flowchart 2-1 on p. 41. Figure 2-38 shows typical strain ranges for man femoral cortical bone during different act tics and distances. Resistance to fatiguc behavio great.er in compression than in tension (Keaveny Hayes, 1993). On average, approximately 5,000 clcs of experimental loading correspond to number of steps in to miles of running. One m lion cycles corresponds to approximately 1, miles. A total distance of less than 1,000 m could cause a fracture of the cortical bone tiss This is consistent with stress fractures repor among military recruits undergoing strenu training of up to 1,000 miles of nll1ning ove short period of timc (6 weeks). Fracturcs of in vidual trabeculae in cancellous bone have been served in postmortem hUlllan specimens and I be caused by fatigue accumulation. Common s arc the lumbar vcnebrae, the femoral head, a the proximal tibia. It has been suggested that th fractures may playa role in bone remodeling well as in age-related fractures, collapse of s chondral bone, degeneraLive joint diseases, a other bone disorders. INFLUENCE OF BONE GEOMETRY ON BIOMECHANICAL BEHAVIOR The geometrv of a bone greatl\' influences its m chanical bel~avior. In Le~lsion' and compressi the load to failure and the stiffness arc prop tional to the cross-sectional area of the bone. T larger the area, the stronger and stiffer the bo In bending, both the cross-sectional arca and distribution of bone tissue around a neutral a affcct the bone's mechanical behavior. The qu tity that takes inLO account these two factors bending is called the area moment of inertia larger moment of inertia results in a stron and stiffer bone. Figure 2-39 shows the in ence of the arc,) moment of inenia on the l to failure and the stiffness of three rectangu Slruclures thal have the same area but differ shapes. In bending, beam III is thc stillest of lhree and call withstand the highest load beca the greatest amount of material is distribuLed a distance from the neulral axis. For rectangu cross-sections, the formula for the area momcn - - - - ----,1-- 4 ){ 1 2x2 1x 4 I II III Three beams of equal area but different shapes subjected to bending. The area moment of inertia for beam I is 4/12; for beam II, 16112; and for beam 111,64/12. Adapted hom Franke', VH .. & Burstein, AH. (970). Orthopaedic Biomechanics. Philadelphia: Lea & Febiger. inertia is the width (8) multiplied by the cube of the height (1-1') divided by 12: B· H' 12 Because of its large area moment of inertia. bean"'! III can withstand four times more load in bending limn can beam I. A third factor, the length of the bone, influences the strength and stillness in bending. The longer the bone, the greater the magnitude of the bending moment caused by the application of a force. ,In a rectangular structure, the magniwde of the stresses produced al the point of application of Ihe bending moment is proponional to lhe length of the StI1.IClure. Figure 2-40 depicts the forces acting on two beams with the same width and height but different lengths: beam B is twice as long as beam A. The bending moment for the longer beatn is twice that for the shorter beam; consequently, the stress mag~ nitudc throughout the beam is twice as high. Because of their length, the long bones of the skeleton are subjected to high bending moments and. therefore, to high tensile and c01npressive stresses. Their tubular shape gives them the ability to resist bending moments in all directions. These bones have a large area moment of inertia because much of the bone tissue is distributed at a distance from the neutral axis. l The factors that affect bone strength and stiff in torsion are the same that operate in bending: cross-sectional area and the distribution of bone sue around a ncutral axis. The quantity that ta into account these two factol's In torsional load is the polar moment of inertia. The larger the p moment of inertia. the stronger LInd stiffer the b Figure 2-41 shows distal LInd pl'oximnl cr sections of a tibia subjected to torsional loading though the proximal section has a slightly sma bony area thun docs the distal section, it has a m higher polar moment of inertia because much o bone tissue is distributed at a distance from the tral axis. The distal section, while it has a la bony area. is subjected to much higher shear st bccause much of the bone tissue is distributed c to the neutral axis. The magnitude of the sh stress in the distal section is approximately do that in the proximal section. Clinically, torsi fractures of the tibia commonly occur distnlly. When bone begills to heal after fracture, b vessels and connective tissue from the periost migrate into the region of the fracture. formin cuff of dense fibrous lissue, or callus (woven bo around the fracture site. stabilizing that area ( 2-42A). The callus significantly increases the and polar moments of inertia. thereby increa the strength and stiffness of the bone in ben and torsion during the healing period. As the r Stress ~-"'~~-<:;magnilude -r-_I-L---... -L • S I Stress magnitude =, 25 2L _____ - - - - - - 2L ----.I as as bending moment. Hence, the stress magnitude throughout beam B is twice as high. Adapted from VH., & Burstein, A.H. (7970), Orthopaedic Bio· mechanics. Philadelphia: Lea & Febiger. A, Early callus formation in a femoral fracture fixe with an intramedullary nail. B, Nine months after jury, the fracture has healed and most of the callu cuff has been resorbed. Courtesy of Robert A. vVinqu MD ,, ,,, , ,,, ,, ,, ,, , ,,~l.,;_ Distribution of shear stress in two cross-sections of a tibia subjected to torsional loading. The proximal section (A) has a higher moment of inertia than does the distal section (B) because more bony material is distributed away from the neutral axis. Adap(ed from Franke!, VH., & Burstein, AH, (1970;' Orthopaedic Bio· mechanics. Philadelphia: Lea g Febiger. ture heals and the bone gradually regains its norm strength, the callus cuff is progressively resor and the bone returns to as ncar its normal size a shape as possible (Fig. 2-4213). Certain surgical procedures produce defects t greatly weaken the bone, particularly in torsi These defects fall into two categories: those wh length is less than the diameter of the bone (str raisers) and those whose length exceeds the bone ameter (open section defects). A stress raiser is produced surgically' when a sm piece of bone is removed or a screw is inserted. B strength is reduced because the stresses impo during loading are prevented From being distribu evenly throughout the bone and instead become c centrated around the defect. This defect is analog to a rock in a stream, which diverts the watel~ p ducing high water turbulence around it. The we ening effect of a stress rqiser is particularly mar under torsional loading; the total decrease in b strength in this loading mode can reach 60~o. c-::-<........) Empty screw hole _ -+-... Screw in place 125 _:.~rew removed ~l_te~..- 100 ~ ,2 / _.::....--- 75 [;; ~ 0> <;; c w 50 ~-. .,..---'y . /' 25 <I" raiser effect produced by the screws and by hol('s without screws had disappeared comple because the bone had ren1odclcd: bone had b laid down around the screws to stabilize them, the empty screw holes had been filled in with b In femora from which the screws had heen moved immediately before testing, however, energy storage capacity of the bone decreased 50 0hi, mainly because the bone tissue around screw sustained microdamage during sere\\' moval (Fig, 2-43), a An open seclion defect is a discontinuity in 345 2 6 7 B bone caused by the surgical removal of a piec Weeks Effect of ~crews and of empty screw holes on the en- ergy storage capacity of rabbit femora. The energy storage for experimental animals is expressed as a percentage of the total energy storage capacity for control animals. When screws were removed immedi- ately before testing, the energy storage capacity de creased by 50%. Adapted from Bur51!.>in. A.H., et al. 4 (1972), Bone strength: The effect of sere"v floles. J Bone Joint Surg. 54A, ] I i13. bone longer than the bone's diameter (e.g.• by clilling of a slot during a bone biopsy). Because ollter surface of the bonc's cross-section is no lo continuous. its abilit!1 to resist loads is allered, ticularly in torsion. In a normal bone subjected to tOl'sion, the s stress is distributed throughout the bone and ac resist the torque. This stress pattcl-n is illustrate the cross-section of a long bone shown in Fi 2-44A. (A cross-section with a continuous Ollter face is called a closed section.) In a bone wit Contro Burstein and associates (1972) showed the effect of stress rabel's produced by' screws and by empty screw holes on the energy storage capacity of rabbit bones tested in torsion at a high loading rate. The irnmediatc effect or drilling a hole and inserting a screw in a rabbit femur was a 74% decrease in energy stOl'age capacily. After 8 weeks, the stress Open section Deformation Stress pattern in an open and closed section under torsional loading. A, In the closed section, all the shear stress resists the applied torque. B, In the open section, only the shear stress at the periphery of the bone resists the applied torque. Load-deformation curves for human adult tibiae tested in vitro under torsional loading. The cont curve represents a tibia with no defect; the open tion curve represents a tibia with an open sectio fect. Adapted from Fr(lf)k~/. v.H., & Burstein, A.H. (1 Orthopaedic BiomechaniCS. Philadelphia: Leel & Febig graft was removed for use in an arthrodesis of the hi A fe\v weeks after operation, the patient tripped whi twisting and the bone fractured through the defect. Bone Remodeling Bone has the ability to remodel, b)' altering its siz shape, and structure, to meet the mechanical d mands placed on it (Buckwalter et aI., 1995). Th phenomenon, in which bone gains or loses cance lous and/or cortical bone in response to the level stress sustained, is summarized as \'VqJJr's la which states that the, remodeling of bone is infl enced and modulated by mechanical stress (Wolff, 1892): . Load on the~,keleton can b~,,,~lc.l.:(?I""l~plished by e tht:;,I.:,IX1~~_?(.:lc,activily or gravity. A positive correlatio exists b~twcen bOI1e mass and body we-ight. A gr~at body weight has been associated \vitha larger bo mass (Exner et al., 1979). Conversely, a prolong condition ohveightlcssness, such as that expedenc during space travel, has been found to result in d creased bone mass in weight-bearing bones. Astr nauts experience a fast loss of calcium and co sequent bone loss (Rambaut & Johnston, 197 \Vhedon, 1984). These changes are not complete reversible. ~'------I I I Normal A patient sustained a tibial fracture through a surgically produced open section defect when she tripped a few weeks after the biopsy. ..----------------open section defect, only the shear stress at the periphery of the bone resists the applied torque. As the shear stress encounters the discontinuity, it is forced to change direction (Fig. 2-44B), Throughout the interior of the bone, the stress nms parallel to the applied torque, and the amount of bone tissue resisting the load is greatly decreased. In torsion tests in vitro of human adult tibiae, an open section defect reduced the load to failure and energy storage to failure h,v as much as 90'10. The deformation to failure was diminished by! approximately 70% (Frankel & Burstein, 1970) (Fig. 2-45). Clinically, the surgical removal of a piece of bone can greatly \veaken the bone, particularly in torsion. Figure 2A6 is a radiograph of a tibia from which a Deformation Load-deformation curves for vertebral segments L5 to L7 from normal and immobilized Rhesus monkeys Note the extensive loss of strength and stiffness in the immobilized specimens. Adapted from Kazarian, L.L.. g Von Gierke, H.E. (1969) Bone loss as a result of immobilization and chelation, Preliminary results in ivlacaca mulatta. (lin Orthop. 65. 67. Bone Remodeling 30-year-old man who had a surgical removal of an A .. ulna plate after stabilization of a displaced ulnar fracture, Figure 2-48 shows anteroposterior (A) and lateral (8) roentgenograms of the ulna after late plate removal. The implant is used to stabilize the fracture for rapid healing. However, in situations such as this, the late plate removal decreased the amount of mechanical stresses necessary for bone remodeling. It is of concern when the plate carries most or all of the mechanical load and re· mains after fracture healing. Thus, according to Wolff's law, it will promote localized osseous resorption as a result of decreased mechanical stress and stimulus of the bone under the plate, resulting in a decrease in strength and stiffness of the bone. Disuse or inactivity has deleterious effects on the skeleton. Bcd rest induces a bone mass decrease of approximately' I (Ve- per week (Jenkins & Cochran, 1969; Krolner & Toft, 1983). In partial or total immobilization, bone is not subjected to the usual mechanical stresses, which leads to resorp- Roentgenogram of a fractured femoral neck to which a nail plate was applied. loads are transmitted from the plate to the bone via the screws. Bone has been laid down around the screws to bear these loads. Anteroposterior (A) and lateral (B) roentgenograms of an ulna after plate removal show a decreased bone diameter caused by resorption of the bone under the plate. Cancellization of the cortex and the presence of screw holes also weaken the bone. Courtesy of Marc Marrens, 1\.1. D. tion of the periosteal and subperiosteal bone and a decrease in the mechanical properties of bone (Le., strength and stiffness). This decrease in bone strength and stiffness was sho\vn b:v Kazarian and Von Gierke (1969), who immobilized Rhesus monkeys in full-body casts for 60 days. Subsequent compressive testing in vitro of the vertebrae from the immobilized monke.\!s and from controls showed up to a threefold decrease in load to failure and energy storage capacity in the vertebrae that had been immobilized; stiffness was also significantly decreased (Fig. 2-47). An implant that remains firmly" attached to a bone after a fracture has healed may also diminish the strength and stiFfness o( the bone. In the case of a plate fixed to the bone with screws, the plate and the bone share the load in proportions deter- o ~ nn •. c::7 UU 0 UW .... c nl '------------------ ,'- ." Vertebral cross-sections from autopsy specimens of young (A) and old (8) bone show a marked reduction in cancellous bone in the latter. Reprinted with permission ,I from Nordin. B,E.e. (1973). Metabolic Bone and Stone Disease. Edinburgh: Churchill Livingstone. C. Bone reduction I l__ P 9 9 W_i'_h_a__i_n__i'_'_'_h_e_m_a_t_i,_a_'_'Y_d_e__i'_t_e_d_._A_,_n_o_,_m_a_1_b_o_n_e mined by the geometry and n1mcrial properties of each structure (Case Stud)' 2-2). A large plate, carrying high loads, unloads the bone to a great extent; the bone then atrophies in response to this diminished load, (The bone may hypertrophy at the bone-screw interface in an altcmpt to reduce the rnicrol1lotion of the screws.) Bone resorption under a plate is illustrated in Figure 2-48. A compression plate made of a material approximately 10 times stifTer than the bone \Vas applied to a fractured ulna and remained after the fracture had healed, The bone under the plate earried a lower land than normal; it was partially resorbed, and the diameter of the diaphysis became markedly smaller. A reduction in the size of the bone diameter greatly decreases bone strength. particularly in bending and torsion, as it reduces the area and polar moments of inertia. A 20(M, decrease in bone diameter may reduce the strength in torsion by 60%. Changes in bone size and shape illustrated in Figure 2-48 suggest that rigid plates should be removed shortly after a fraeture has healed and before the bone has markedly diminished in size. Such a decrea.sc in bone size is usually accompanied by secondary osteoporosis, which' further weakens the bone (SI'itis et a!., 1980), An implant may cause bone hypertrophy at its attachment sites. An example of bone hypertrophy (top) is subjected to absorption (shaded area) durin the aging process, the longitudinal trabeculae beco thinner and some transverse trabeculae disappear tom). Adapted from Sifter!, R.S .. & Levy. R.N. (J98/j. T becular pacteffls and the internal architecture of bone. S_i_"_"_i_J_rv_lo_d_,_";_.S_,_2_2_'_. around scrcws is illustrated in Figure 2-49. A plate was applied to a femoral neck fracture and bone hypcrtrophied around the scrcws in resp to the increased load at thesc sites. H.~lpert may also result if bone is repeatcdly subjecte high mechanical stresses within the normal ph logical range. Hypertrophy of normal adult bon response to strenuous exercise has been obse (Dalen & Olsson, 1974; Huddleston et aI., 1 Jones et aI., 1977), as has an increase in bone sity (Nilsson & Wesllin, 1971), Degenerative Changes in Bone Associated With Aging A progressive loss of bone density has been sCI·veci as part of the normal aging process. The gitudinal trabeculae become thinner, and som the transverse trabeculae arc resorbed (SifTe Levy, 1981) (Fig. 2-50). The result is a marke duction in the amount of cancellous bone a thinning of conical bone. The relationship betw bone mass, age. and gender is shown in Figure The decrease ill bone Lissuc and the slight decr in the size or the bone reduce bone strength stillness. ro U o 1000 E w '" w m :;; 500 ID C o tIl 20 ~OWlng 40 60 BO Age (years) the relatIOnship between bone mass, age, and gender. On the top of the figure, a crosssection of the diaphysis of the femur and the bone mass configuration is shown. Reprinted ('lith permission from Kaplan, F.S., rlayes, W.c., Keaveny, T.M., er al (7994), Form and function of bone. In S.R. Simon (edJ Orthopaedic Basic Science (.0, 767). Rosemont, IL:AAOS Stress~strain curves for specimens from adult tibiae of t\VO widely differing ages te tension are shown in Figure 2~52. The u stress was approximately the sarne for the and the old bone. The old bone specimen withstand only half the strain that the youn could, indicating greater brittleness and a re in energy storage capacit},. The reduction density', strength, and stiffness results in in bone fragility!. Age~related bone loss depen number of factors, including genclel: age menopause, endocrine abnormality, inactiv use, and calcium deficiency. Over several d the skeletal mass ma.y be reduced to 5000 of trabecular and 25°/(; of cortical mass. In the decade, women lose approximately 1.5 to 2(1 while men lose only approximately half th (0.5 to 0.75(-/0) yearly. Regular physical activ exercise (Zetterbarg et aI., 1990), calcium, a sibly estrogen intake may decrease the rate mineral loss during aging. Summarv 1i Bone is a complex two-phase composit rial. One phase is composed of inorganic salts and the other is an organic matrix of c and ground substance. The inorganic com makes bone hard and rigid, whereas the component gives bone its flexibility and resi 2 tVlicroscopically, the fundamental str unit of bone is the osteon, or haversian composed of concentric layers of a mineraliz trix surrounding a central canal containin vessels and nellie fibers. Strain Stress-strain curves for samples of adult young and old human tibiae tested in tension, Note that the bone strength is comparable but that the old bone is more brittle and has lost its ability to deform. Adapted from Burs rein, A.H., Reilly, D. T, & Alartens, M. (976). Aging of bone tissue' Mechanical properties. Bone Joint Surg, 58A, 82. 3· rv(acroscopicall.\" the skeleton is comp cortical and cancellous (trabecular) bone. bone has high density' while trabecular bon in density over a wide range. Bone is an anisotropic material, exhibi ferent mechanical properties \vhen loaded in ent directions. iVlature bone is strongest and in compression. S Bone is subjected to complex loading during common physiological activities walking and jogging. Most bone fractures duced by a con1bination of several loading p s a m 6 Ivluscle contraction affects stress patt bone by producing compressive stress that p or totall).' neutralizcs thc tcnsilc stress acting on the bone. Bone is stiffer, sustains higher loads before failing, and stores more energy whcn loaded at higher physiological strain rates. ~i Living bone fatigues when the frequcnc y.' of loading precludes the remodeling necessary to prevent failure. The mechanical behavior of a bone is influenced by its geometry (length, cross-sectional area, and distribution of bone tissue around the neutral Bone remodels in response to the mechanical demands placed on it; it is laid down \vhere needed and resorbed where not needed. vVith aging comes a marked reduction in the amount of cancellous bone and a decrease in the thickness of cortical bone. These changes diminish bone strength and stiffness. REFERENCES Bassett. C.A.L. (1965). Electrical effects in bone. Sci .-\111,2/3. 18. BClIldicld, W.. 0.: Lt, e.H. (1967). AnisoLropy of nonelastic llow in bone. J App{ Physics, 38, 2450. Buckwaltcl~ lA.. Glimcher, \U., Coopel: R.R., et al. (1995). Bone biology. Part I: Stnlcture. blood supply, cells, matrix and mineralization. Pan II: Formation, form, remodelling and regubtion of cdl funcLion. ((nstructional Course Lectlln:). J BOlle JoiJl! Sllrg. 77A, 1256-1289. Burskin, A.H., Reilly, D.T.. 0.: \bnens, M. (1976). Aging of bone tissue: :vlechanical properties. J BOllI..' Joilll SI/1~!i. 58:\, 82 Burstein, A.H., L't al. (1972). Bone strength: The effecl of scre\\" holes. J BOlle Joillt Slil~!i, 54:\, 1143. n.R. (1978). Anisotropic analysis of strain roselle information from cortical bone. J BiolJlcch, II, 199. D.R.. 0.: I-!ayes, W.e. (1977). Compact bone fatigue damage: microscopic examination. Clill Of/hop, /27,265. N.. 0.: Olsson, K.E. (1974). Bone mineral content .md physiactivity. Acla Orthop Scall(/, 45, 170. G.U., et al. (1979). Bone densitometry using computed tomography. Part I: Selective determination of trabecular bone and oLher bone mineral paraml'lers. Normal values in children and adults. BrJ Radiol, 52, 14. Frankel, VH., &: Burslein. A.H. (1970). Ort/lOpacdic Biol/li.'clul Philadelphia: Lea 0.: Febigel: f-Iuddbaoll, A.t.. Rockwell, D.. Kulund, D.N., et a1. (1980). B mass in lifetime tennis athletes, hUlA, N4, 1107. !nternaLional Societv of Biomechanics (1987). QII{/lIlitics alld l oj" J1easllfi.'IIiCIIlS· iI/ Bio/lltxhal/ics (unpublished). Jenkins. D.P.. 0.: Cochran, T.H. (1969). Osteoporosis: The dram effecl of disuse of an extremity. Clill On/lOp, 64, 128. Jones, H., Priest . ./.. Hayes, \V.. et a1. (1977). f-lu!11l.'ral hYPl.'nr in response to exercise. J BOlle Joill/ SIiI;!5. 59:\, 204. Kaplan. ES .. Hayes. We., Kea\'eny. T.iVI., l'l al. (1994). Form funcLion of bone. In S.R. SiTllon (Ed.). Orlhopacdic Basic Sc (pp.I27-184). Rosemont. IL: AAOS. Kazarian, L.L., & \.lon Gil.'r'ke, I-I.E. (1969) Bone loss as a resu immobilization and chelation. Preliminary ri:sults in :\1< lllubtla. Clill Orthop. 65. 67. Ke.weny, T.M .. &: Hayes, \V.e. (1993). Mechanical propertil.'s of tical and trabecular bone. BOIIC, i, 285-344. Krolner, B.. 0.: Toft. B. (1983). Vertebral bone loss: An unheeded effecL of bedrest. C!ill Sci, 6e+. 537-540. Kummer, J.K. (1999). Implant Biomatl:.'rials. In: L\1. Sph'ak, DiCesare, D.S. Feldman, KJ. Ko\·al. A.S. Rokito. & J.D. Zu man (Eds.). Or/hop(i('dics:"\ S/Ildy Gllidc (pp. 45-48). New \IcGraw-I-lill. Lanyon. LE.. &. Bourn. S. (1979). The inlluence or lllechanicd Lion on the dc\'elopmcrH and remodeling of the tibia: An ex mental study in sh('ep. .1 BOlle Joinl SI/rg. 6/.4. 263. Lanyon, L.E., H.unpson. \\I.GJ .. Goodship, A.E., et al. (1975). deformaLion recorded in vin) from sLrain g'lllgl.'S attached t human tibial shaft. ..lew Onhop Sct/lul, "'6~ 256. Nilsson. B.E., & \\btlin, N.E. (1971). Bone density in athletes. On/lOp. 77, l79. 6zkaya, N., & Nordin, M. (1999). Fllndall/ell/tlls Biolllcc!ul EquilibriulII, ,HOlioll. alld Del(lnllatioll (2nd cd.) Ncw Springer-Verlag. Rarnballl~ P.c., 0.: ~Johnston. R.S. (1979). Prolonged weighLles and calcium loss in man. Ac!a ASll'ollal/lica, 6, 1113. Siffen, R.S .• & Levy, R.N. (1981). Tnlbccular patterns and the i nal architecture of bonl.'. ;\fl. Sinai J lIed, 48. 221. SltiLis, P., Paa\"{llainell, P., Karaharju, E.. ct al. (1980). Structura biolllechanical changes in bOl~e aher rigid plate fixation. C SlIIp., 23, 247. \Vhedon, G.D. (1984') Disuse osteoporosis: Physiological asp Cab!" 'fissile lilt, 36, 146-150. Wolff, J. (1892). Dlls Gt:set::. der halls!(JrlII(ltiOiI der Kllochcll. B I-lirschwald. Zetterberg e., Nordin. :\'1., Sko\Ton. M.L.. et al. (1990). SkeleL feClS of physical activity. Geri-7bpics. /3(4). 17-24. or C VASCULAR FACTO,RS J I MECHANICAL FACTORS I i i' SlruCturc Function ,ilJ ;j" w Protection, suppOrt. kinCffiJtic links ~~ Organic Component i Collagen Proteins 1 ;,N~~coliag~n'Proteln~; Type I ) ~1 1 Growth factor cell attachment proteins PG', \1 Inorganic Component Cry,Olh of HYDROXYAPATITE ' Calcium, Phosphate 1 1 : .j FLOW CHART 2·2 ~This flow chart is dl?~igncd Bone composition. structure, and functions.· (PG's, proteoglycans) for classroom or group discussion. Flow chart is not meant to be exhaustive. .1 '1 I fJ, ,,1'.j2ij4if@_~\M¥i"I:IIi:i ;:H%J! dl 'I);; )'MltII: i)() 1tWC' !~ ',~. •~ a. " > ~ "a0 E £ m x .."'" .s V '"'m" E 0 " 0 0 .0 B cm 0 E <; c m .- '"c0 {; "' u: V .0 em ~ 0 .~ C v ":;: ·u 'i3 0 0 m 2 E m ~ ~ § u .:" u .;;; C "5 x w 0 ~ 0 '"5 E 0 ~ m u .e " .,'"c • ". 0 ;;; "' ~ 0 '" ~ ,. ] I j ·(11 Gcnedc DIsturbance Blood Flow <.;,\' :..\;.:, \,.::,;:.\- ,~": Tumoral DIsturbance " J .;',: 'i.~,:('erjm"ry :j ,j i :\f~pvcloping' \i ,;,:::';;'Skclc~on 1 \:<·",-':::::::Y", I J 6~'~'~6'~i';~ndro$is /~\;~r~rth~s: N ,1 j .Ostco~h'~n'd~i;i'~ , disseC30S Kcil:, .Os,conccrosi~~~ "_fC~pr.'l1 head ~ 'Shcucrm3nn; of the femoral j:, ,,,,:,)::,::~,'.i:; ·Gauchcrs disease ·Cushing Syndrome condrlos' \0" , "';; .' J"f:,Vcrccbrac' IJ \' .' ,,-. ,:: ';:1 ~,i :j; 1/<1 .'~, ri PATHOLOGICAL OR FRAGILITY FRACTURE ij ~ t~: i~ o I'" FLOW CHART 2-4 Intrinsic factors associated with bone damage. Clinical examples.* "This flow chart is designed for d<lssroom or group discussion. Flow chart is not meant to be cxhamtivc. -''>-';;,i' ~C"' l' :,i,¥ (;~ " "-", ·,'i-".",,"';c.c -'~,,;.;':., ,: Biomechanics of Articular Cartilage Van C. Mow, Clark T Hung Introduction Composition and Structure of Articular Cartilage Collagen Proteoglycan WatE'r Structural and PhysicallniE'raction Among Cartilage Components Biomechanical Behavior of Articular Cartilage Nature of Articular Cartilage Viscoelasticity Confined Compression Explant Loading Configuration Biphasic Creep Response of Articular Cartilage in Compression Biphasic Stress-Relaxation ResponsE' of Articular Cartilage in Compression Permeability of Articular Cartilage Behavior of Articular Cartilage Under Uniaxial Tension Behavior of Articular Cartilage in Pure Shear Swelling Behavior of Articular Cartilage Lubrication of Articular Cartilage Fluid-Film Lubrication Boundary Lubrication Mixed Lubrication Role of Interstitial Fluid Pressurization in Joint Lubrication Wear of Articular Cartilage Hypotheses on Biomechanics of Cartilage Degeneration Role of Biomechanical Factors Implications on Chondrocyle Function Summary Acknowledgments References Flow Charts Composition and Structure Articular Cartilage Introduction Three types of joints exist in the human body: fib'ro~ls, cartilagin,olls. and synovial. Onl.y one of' these,;-,ollle syn.ovial. or diarthrodial, jo.int, allows a large degree of motion. In Y~_L1ng normal joints, the ,-~llticulalingJ)onecn~ls of diartl~~'odiaIj9_inls,!.re cov}--ered by a-thin (1-6 mill), den~c, tram~JJlcc_nt,~:-,'hile ,~~c6nnective-i.i.ssllc called hyaline arliCltlar. cartilage iCBox-3:!). Articul'IU~f\rtilage-;"s a highly s-p·ecia·lized . tissue precisely suited for withstanding the highly i~adcd joint environment without failure during an a\lerage individual's lifetime. PhXti.t~~l.Qgically, however, ,il is v51~~~I'.lIly an isolated li,ssuc, devQ.id of blood \;essds, 1).~-mphalic channels, and nelyrol.ogical inncr~alion~'-Flll'lllermoJ'c,its .c~l.llliar density is less than .·tK"t of any other tissue (Stoek\~c·il. 1979). [n (liartll-i'o(Ti~;fioints, articular cartilage has two - p-rimar'yfl~nctions': (1) to distl-ibute jo.int loads over ,a wiele al-ea, thus decreasing lhe strc:)ses sllst~!.!ned by the contacting join.t..~.sl~rraccs (Atcshian ct aI., 1995; Helminen et aI., 1987) and (2) to allow rel"tive movementoft!1cu:m,posing joinl surfaces \~:·,i.t"I,l""I~lin­ imal ft:'iction and wear (iVlow & Alcshian, 1997). In thi's"ch~l'I)iel','wc w'Hi describe how the biomcchanical pl'openics of articular cartilage, as determined by its composition and structure, allow for the optimal performance or these functions. ~rticular , I Cartilage A notable exception to the definition of hyaline articular G'lrtilage is the temporomandibular joint, a synovial joint in which fibrocartilage is found covering the bone ends. Fibrocartilage and a third type of cartilage, elastic cartilage. are closely related to hyaline cartilage embryologically and histologically but .'lre vastly different in mechanical and biochemical properties. Fibrocartilage represents a transitional cartilage found at the margins of some joint cavities, in the joint capsules, and at the insertions of ligaments and tendons into bone. Fibrocartilage also forms the menisci interposed between the articular cartilage of some joints and composes the outer covering of the interverlebral discs. the anulus fibrosus. Elastic cartilage is found in the external ear, in the cartilage of the eustachian tube, in the epiglouis. and in certain pans of the larynx. ~.~- ., of Chondrocytcs, the sparsely distributed cells in ar lllarc~\I:tiL\ge, account for less than 10% of the sue's volume (Stoekwell. 1979). Schematically. zonal arrangement of chondrocytcs is shown in ure 3-1. Despite their sparse distribution, chond cytcs manufacture, secrete, on!anil',---c, and rnain tile on!anicco'n;-I;~I~l-~f the~ extracellular ma (ECM) (Fos'lIlg & Hardingham, 1996; ·M~lir, 19 The org~nic ma~rix,is compo~~5:L~~[.:~,,(I(:nscp9.J~\ of fin:"-§?·I-(..1¥8~,,,(~'bril~ (n10stly t'ype-Jl~c-olh;~gen., miii6i' amounts .,;1' types V, VI, IX, and XI) that enrneshcd..Jn. a cOnCC.tl.traled ~olu.~i().n. of proteo cans (pcs) (l3ate;;'an et aI., I996;Eyre, 1980; M -~1983)_ 1n nOITn,)1 arLiculaL_cartilage the colla cOlllent ranges from 15 to 22% by wet weight the PG con-tent fronl4to 7% bv wet wei(Tht· the .' :::::> ' maining 60 to 85 % is walCI: inorganic salts. small amounts of other matrix proteins, glycop teins. and lipids (Mow & Ratcliffe, 1997). Colla flbrils and PGs, each being capablc of form structural nctworks of significant strength (Bro & Silyn-Roberts, 1990; Kempson et aI., 19 Schmidt el aI., 1990; Zhu et aI., 1991, 1993), arc structural components supporting the internal chanical stresses that result rrom loads being plied to the articular canilage, Moreover, th structural components, together with water, de mine the biomcchanical bchavior of this tis (Ateshian et aI., 1997; Maroudas, 1979; Mow et 1980, 1984; Mow & Aleshian, 1997). COLLAGEN Collagen is the mosl abundant protein in the b (Bateman et aI., 1996; Eyre, 1980). In articular ca lage. collagen has a high level of structural organ tion that providcs a fibrous ultrastructure (Cl 1985; Clarke, 1971; Mow & Ratcliffe, 1997). The sic biological unit of collagen is tropocollagen Sln.lcture composed of three procollagen polypep chains (alpha chains) coiled into left-handed hel (Fig. 3-2;\) thaI are furthcr coiled abollt each o inlO a right-handed triple helix (Fig. 3-28). Th rod-like tropocollagen molecules. 1.4 nanOJlle (nm) in diameter and 300 nm long (Fig. 3-2, C & polymerize into larger collagen fibrils (Batem ct al.. 1996; Eyre, 1980), In articular cartilage, th fibrils have an average' diamcter of 25 to 40 (Fig.3-2E, Box 3-2); howevec this is highly varia ...... Articular surface STZ (10.20%) ,,,,'•.,,•.,0' ., ,,~) •..,.> _,~ .~ ,., ," '. Middle zone (40-60<::0) '~,,~-' ,,-r:~ '~ ,~.-r ,'-' ~{~:f2jJ.~ Deep zone (30%) i~;~f Calcified zone ";·'5c..C_~,.,,1::;\,,_, Subchondral bone B " Tide mark Chondrocyte Photomicrograph (A) and schematic representation (8) of the chondrocyte arrangement throughout the depth of noncalcified articular cartilage. In the superficial tangential zone, chondrocytes are oblong with their long axes aligned parallel to the articular surface. In the middle zone, the chondrocytes are "round" and randomly distributed. Chondrocytes in the deep zone are arranged in a columnar fashion oriented perpendicular to the tidemark, the demarcation between the calcified and noncalcified tissue . • Scanning electron microscopic studies, for instance, have described fibers with diameters ranging up to 200 nm (Clarke, 1971). Covalent cross-links form between these tropocollagen molecules, adding to the flbrils high tensile strength (Bateman et aI., 1996). The collagen in articular cartilage is inhomogeneously distributed, giving the tissue a layered character (Lane & Weiss, 1975; Mow & Ratc1iITc, 1997). Numerous investigations Llsing light. transn1ission electron, and scanning electron microscopy have identifled three separate structural zones. For example, Mow et al. (1974) proposed a zonal arrangement for the collagen network sho\vn schematically' in Figure 3-3A. In the superficial tangential zone, which represents 10 to 20()0 of the total thickness, are sheets of fine, densely packed fibers randomly \voven in planes parallel to the articular surface (Clarke, 1971; Redler & Zimny, 1970; Weiss et aI., 1968). In the middle zone (40 to 60% of the total thickness), there are greater distances between the randomly oriented and homogeneousl::.. dispersed fibers. Below this, in the deep zone (approximatel.v 30% of the total thickness), the fibers come together, forming larger, radially oriented fiber bundles (Redler et aI., 1975). These bundles then cross the tidemark, the interface between articular cartilage and the calcified cartilage beneath it, to enter the calcified cartilage, thus forming an interlocking "root" sj'stem anchoring the cartilage to the uncler- lying bone (Bullough & Jagannath, 1983; Redl aI., 1975). This anisotropic fiber orientation is rored by the inhomogeneous zonal variations in collagen content, which is highest at the surface then remains relatively constant throughout deeper zones (Lipshitz et aI., 1975). This comp tionalla.vering appears to provide an important mechanical function by' distributing the stress m uniformly across the loaded regions the join sue (Selton et aI., 1995). Cartilage is composed primarily of type II c gen. [n addition, an array of difFerent coll (t)'Pes V, VI. IX, Xl) can be found in quantitati minor amounts within articular cartilage. Tvp collagen is present primarily in articular cm·ti the nasal septum, and sternal cartilage, as well or Differences in Coliagen Types Differences in tropocollagen alpha chains in various body tissues give rise to specific molecular species, or types of collagen. The collagen type in hyaline cartilage type II collagen, differs from type I collagen found in bone, ligament, and tendon. Type II collagen forms a thinner fibril than that of type I, permitting maximum dispersion of collagen throughout the cartilage tissue. (r A I l 1.4 nm B chain Triple helix Tropocollagen molecule o Collagen fibril with quarter-stagger array of molecules Fibril with repeated banding pattern seen under eleclron microscope Molecular features of collagen structure from the alpha chain ((I) to the fibril. The flexible amino acid sequence in the alpha chain (A) allows these chains to wind tightly into a righthanded triple helix configuration (8), thus forming the tropocollagen molecule (C). This tight triple helical arrangement of the chains contributes to the high tensile strength of the collagen fibril. The parallel alignment of the individual tropocollagen molecules, in which each molecule overlaps the other by about one quarter of its length (0), results in a repeating banded pattern of the collagen fibril seen by eledron microscopy (x20,OOO) (E). Reprinted wirh permission from Donohue, It'l1., Buss. D., Oegema. IR., et al. (19831- The effects of indireCl blunt trauma Oil adult canine arricuhu cartilage. J Bone Joint Surg, GSA. 948. • the inner regions of the intervertebral disc and meniscus. For reference, type I is the most abundant collagen in the human body and can be found in bone and soft tissues such as intervertebral discs (rnainl.y in the annulus fibrosis), skin. meniscus, tendons, and ligaments. The most important mechanical properties of collagen fibers nrc their tensile stiffness and their strength (Fig. 3-4;\). Although a Single collagen fibril has'-not be~n tested in len;ion. the tensile slI'ength of collagen can be inferred from tests on structures with high collagen contenl. Tendons, for example, are about 80% collagen (dry weight) and have a tensile stiffness of 10.1 !\'1Pa and a tensile strength of SO MPa (Akizuki et aI., 1986; Kempson, 1976, 1979; Wooet aI., 1987, 1997).Stec1, by comparison, has a tensile stillness of approximately 220 x 10" MPa. Although strong in tension. collagen fibrils offer little resis(ance lo ~olllpression because their large slendcnlcss ratio, the rati length to thickness, makes it easy for them to bu under cotl1prcssive loads (Fig. 3-48). Like bone, articular canilage is anisotropic material properties differ with the direction of l ing (Akizuki et aI., 1986; Kempson, 1979; Mow RalclifTc, 1997; Roth & Mow, 1980; Woo el 1987). Oft is thought that this anisotropy is relate the varying collagen fiber arrangements within planes parallel to the articular surface. It is thought, however, that variations in collagen f cross~link density, as well as variations in colla PG interactions. also contribute to articular c lage tensile anisotrop~", In tension, this anisotr is llsually described with respect to the direc of the articular surface split lines. These split l arc elongated fissures Jj·roduced by piercing the ticular surface with a small round awl (Fig. B STL A. Schematic representation, (Repriflleej \·virh permission from Mow v.c. el al. j1974j. Some surface dJc1riJcteristics of arriculaf cartilages. A scanning electron microscopy sludy and a theoretical mode! for the dynamic interaction of synovial fluid and articular car· tilage. J Bionll?{hanics, 7, 449), B. Photomicrographs (x3000; provided through the courtesy of Dr. T. Takei, Nagano. Japan) of the ultrastructural arrangement of the collagen network throughout the depth of articular cartilage. In the superficial tangential zone (STZ), collagen fibrils are tightly woven into sheets arranged parallel to the articular surface. In the middle Middle zone Deep zone zone, randomly arrayed fibrils are less densely packed to accommodate the high concentration of proteoglycans and water, The collagen fibrils of the deep zone form larger radially oriented fiber bundles that cross the tidemark, enter the calci· fied zone, and anchor the tissue to the underlying bone. Note the correspondence between this collagen fiber architecture and the spatial arrangement of the chondrocytes shown in Figure 3-1. In the above photomicrographs (B), the STZ is shown under compressive loading while the middle and deep zones are unloaded. Collagen Fibril <)0 A B 1m)) )))))) 1m)) )l»ll IllJli lJIlil I (> High tensile sliflness and strength Uttle resistance to compression Human femoral condyles Illustration of the mechanical properties of collagen fibrils: (A) stiff and strong in tension, but (B) weak and buckling easily with compression. Ad,lpted from Myers, E.R., Lai, WM., & MOV'I, VC (1984). A COfltinuum theory and all experiment for the ion-induced swelling be/Mvior carlilage. j Biomech Eng, Gelenkknorpel. Verhandlungen der Analomischen Gesellschaft, 106(21. 15/-/58. 12.248. Diagrammatic representation of a split line pattern on the surface of human femoral condyles. Reprinted wirh permission from Hu/rkrantz, W (1898). Ueber die Spafrrichwflgefl der Hultkrantz, 1898). The origin of the pattcrn is related to the directional variation 01" the tensile stilTness and strength characteristics of articular can ilage described above. To dale. ho\Vevel~ the exact rc;sons as to why articular canilage exhibits slIch pronounced anisolropies in tension is not known, nor is the functional significance anisotrop)'. or this tensile PROTEOGLYCAN l\llany types of PGs arc round in cartilage. Fundamentally, it is a large protein-polysaccharide mole- cule composed of a protein core to which one or more glycosaminoglycans (GAGs) are attached (Fosang & Hardingham, 1996; Muir, 1983; Ratcliffe & Mow, 1996). Even the smallest of these molecules, high'Can and deem-in, are quite large (approxim;t~ly I X 10'\ 111\\'). but they comprise less than 10~~ of all PGs present in the tissue. Aggrccans are much larger (1-4 X 10" mw), and they have the rcmarkable capability to attach to a hyaluronan mol· ccule (HA: 5 X 10' I11W) via a specific 1·-1A-binding rcgion (HABR). This binding is stabilized by a link protein (LP) (40-48 X 10" mw). Stabilization is crucial 10 the function of normal cartilage; without it, the components of the PG molecule would rapidly escape from the tissue (Hardingham & IVluir, 1974; H<lscall, 1977; Muil; 1983). Two types of GAGs comprise aggrecan: chondroitin sulfatc (CS) and kcmtan sulfate (KS). Each CS chain contains 25 to 30 disaccharide units, while the shorter KS chain contains 13 disaccharide units (MuiJ~ 1983). Aggrecans (previously rererred to as subunits in the American literature or as monomers in the UK and European literature) consist of an approximately 200.nanol11ctcr..long protein core to which approximately 150 GAG chains, and both 0linked and N-linked oligosaccharides, are covalently attached (Fosang & Hardingham, 1996; Muir, 1983). Furthermore, the distribution of GAGs along the protein core is heterogeneous; there is a region rich in KS and O-linked oligosaccharides and a rcgion rich in CS (Fig. 3-6M. Figure 3-611 dcpicts the famous "boule-brush" model for an aggrecan (j\lluir, 1983j. Also shown in Figurc 3-6/\ is the hctcrogcneity of the protein core that contains three globular regions: GIl the HABR located at the N-terminus that contains a small amount of KS (Poole, 1986) and a few N-Iinked oligosaccharidcs, G., located between the HABR- and the KS-rich regio·n (Hardingham et aI., 1987), and G" the corc protcin C-tenninus. A I: I stoichiomelly C:dslS between the LP and the GI binding region in c.:anilagc. iVlore recently, the o two globular regions have been extensively stu (Fosang & Hanlingham. 1996), but their functio significance has notycl been elucidated. Figure 3 is lhe accepted molecular confonnation of a PG gregate; Rosenberg Gt al. (1975) were the firs obtain an electron micrograph 01" this mole (Fig. 3-6C). In native cartilage, most aggrccans arc associ withHA to form the large PG aggregates (Fig. 3These aggregates may have up to several hund aggrecans noncovalenliy uuached to a central core via their HABR, and each sile is stabilized an LP. The filamentous HA core molecule is a n sulfated disaccharide chain that may be as long fJ.m ill Icngth. PG biochcmists have dubbed thc an "honorary" PG, as it is so intimately involve the structure of the PG aggregate in articular c lage. The stability afforded by the PG aggregates a rl"wjor functional significance. It is accepted thut PG aggregation promotes immobilization the PGs within the nne collagen meshwork, add structural stability and rigidity to thc ECM (Mo aI., 1989b; Muir, 1983; Ratcliffc et aI., 1986). thermore, two additional forms of c!crmatan sul PG have been idcntif"led in the ECM of articular tilage (Rosenberg et aI., 1985). In tendon, c1erm sulfate PGs have been shown to bind noncovale to the surfaces of collagen fibrils (Scott & Orf 1981); however, the role of derma tan sulfatc in ticular cartilage is unknown, biologically and f tionally. Although aggrecans generally have the b structure as described abovc, they arc not st turally identical (Fosang & Hardingham, 1996). grecans vary in length. molecular weight, and c position in a variety of ways; in other words, arc polydisperse. Studies have demonstrated distinct populations of aggrecans (Buckwalter e 1985; Heincgard ct aI., 1985). The first populatio present throughout life and is rich in CS; the sec contains PGs rich in KS and is present only in a cartilage. As articular cartilage matures, other related changes in PG composition and struc occur. vVith canilage maluration, the watcr con (Armstrong & Mow, 1982; Bollet & Nancc, 1 Linn & Sokoloff, 1965; Maroudas, 1979; V 1978) and the carbohydrate/protein ratio prog sively decrease (Garg & Swann, 1981; Roughle White, 1980). This dccrcase is mirrored by a crease in the CS content. Conversely, KS. whic present only in small ,imounts at birth, incre throughollt development and aging. Thus, Chondroitin Sulfate (CS) chains C-terminal globular domain (G 3) A Proteoglycan (PG) Macromolecule 1200 nm Hyaluronan (HA) B 1DI1I'---- A. Schematic depiction of aggrecan, which is compose keratan sulfate and chondroitin sulfate chains bound c lently to a protein core molecule. The proteoglycan pr core has three globular regions as well as keratan sulfa rich and chondroitin sulfate-rich regions. B. Schematic resentation of a proteoglycan macromolecule. In the m trix, aggrecan non covalently binds to HA to form a macromolecule with a molecular weight of approxima 200:<. 106. Link protein stabilizes this interaction betw the binding region of the aggrecan and the HA core m cule. C. Dark field electron micrograph of a proteoglyc aggregate from bovine humeral articular cartilage (>:120,000). Horizontal fine at lower right represents O Reprinted with permission from Rosenberg, C, Hellmann, W Klein5chmidr. AX (975). Electron microscopic studies 0; p gfycan aggregates (rom bovine articular cartifage. J Bioi Che 250. 1877. i If II!!!.."."" ~", ~ ", ~""~"::-'~~·,.cT·","'! '.," '.,=-, ~ ----:'....,....-----~. ,IIIIIIIIII!!I,II,'IIIIlIl!!.,!,!!!,,.p. •. "!!"!'!. •••• .. "'-. CS/KS ratio, which is approximately 10: 1 at birth, is only approxirnatel:v 2: I in adult cartilage (Roughlcy & "Vhile, 1980: Sweet et al" 1979; Thonar et al" 1986). Furthermore, sulfation of the CS molecules, \vhich can occlir at either the 6 or the 4 position, also undergoes age-related changes. In utero, chondroitin-6-sulfate and chondroitin-4-sulfate are present in equal molar amounts; however, by maturity, the chondroitin-6-sulfate:chonclroitin-4-sulfatc ratio has increaseel to approximately 25: I (Roughley et a1., 1981). Other studies have also documented an age-related decrease in the hydrodynamic size of tf~e aggrecan. Many' of these early changes seen in articular cartilage may reflect cartilage maturation, possibly as a result of increased functional demand \vith increased \vcight-bcaring. Ho\vever, the functional significance of these changes, as well as those occurring later in life, is as .\'ct undetermined. WATER \Valel~ the most abundant component of articular cartilage, is n10st concentrated ncar the anicular surface (_80°;(j) and decreases in a near-linear fashion with increasing depth to a concentration of approximately 65% in the deep zone (Lipshitz et al., 1976: Maroudas, 1979). This Iluid contains many free mobile cations (e.g., Nu', K', and Cal.) that greatly! influence the mechanical and physicochemical behaviors of cartilage (Gu et aI., 1998; Lai et aI., 1991; Linn & Sokoloff, 1965; Maroudas, 1979). The fluid component of articular cartilage is also essential to the health of this avascular tissue because it permits gas, nutrient, and waste product movement back and forth between chondrocytes and the surrounding nutrient-rich synovial fluid (Bollet & Nance, 1965; Linn & Sokoloff, 1965; Mankin & Thrashel; 1975; Maroudas, 1975, 1979). A small percentage of the \vater in cartilage resides intracellularl y', and approximately 30% is strongly associated with the collagen fibrils (Maroudas et aI., 1991; Torzilli et aI., 1982). The interaction between collagen, PG, and water, via Donnan osmotic pressure, is believcd to have an important function in regulating the structural organization of the ECM and its s\velling properties (Donnan, 1924; Maroudas, 1968, 1975). Most of the \vater thus occupies the interfIbrillar space of the ECM and is free to move whcn a load or pressure gradient or other electrochemical motive forces are applied to the tissue (Gu et aI., 1998; Maroudas, 1979). When loaded by a compressive force, ap- proximately 70(;r; of the water may be moved, T interstitial fluid movement is important in cont ling cartilage mechanical behavior and joint lu cation (Ateshian et aI., 1997, 1998; Hlavacek, 1 Hou et aI., 1992; Mow et aI., 1980; Mow & Atesh 1997). STRUCTURAL AND PHYSICAL INTERACTION AMONG CARTILAGE COMPONENTS The chemical structure and physical interacti of the PG aggregates influence the properties of ECM (Ratcliffe & Mow, 1996). The closely spa (5-15 angstroms) sulfate and carboxyl cha groups on the CS and KS chains dissociate in s tion at physiological pH (Fig. 3-7), leaving a h concentration of fixed negative charges that cre strong intramolecular and intermolecular cha charge repulsive forces; the colligative sum these forces (when the tissue is immerscd i physiological saline solution) is equivalent to Donnan osmotic pressure (Buschmann & Grod sky, 1995; Donnan, 1924; Gu et aI., 1998; Lai et 1991). Structurally, these charge-charge repul forces tend to extend and stiffen the PG ma molecules into the interfibrillar space formed the surrounding collagen network, To apprcc the magnitude of this force, according to Step Hawkings (1988), this electrical repulsion is million, million, million, million, million, mill million times (42 zeros) greater than gravitatio forces. In nature, a charged body cannot persist l without discharging or attracting counter-ion maintain electroneutrality, Thus, the charged sul and carboxyl groups flxed along the PGs in artic cartilage must attract various counter-ions and ions (mainly Na', Cal., and C I") into the tissu maintain electroneutrality. The total concentra of these counter-ions and co-ions is given by well-known Donnan equilibrium ion distribu law (Donnan, 1924). Inside the tissue, the mo counter-ions and co-ions form a cloud surround the fixed sulfate and carboxyl charges, thus shield these charges from each othet: This charge shield acts to diminish the very large electrical repul forces that otherwisc would exist. The net result swelling pressure given by the Donnan osmotic p sure law (Buschmann & Grodzinsky, 1995; Don 1924; Gu et aI., 1998; Lai et aI., 1991; Schuber Hamerman, 1968). The. Donnan osmotic pres theory' has been extensivel:v used to calculate Aggregate Domain Repulsive forces due 10 fixed charge density distribution A Smaller domain Increased charge density Increased charge-charge repulsive forces B A, Schematic representation of a proteoglycan aggregate solution domain (left) and the repelling forces associated with the fixed negative charge groups on the GAGs of aggrecan (right). These repulsive forces cause the aggregate to assume a stiHly extended conformation, occupying a large solution domain, B. Applied compressive stress decreases the aggregate solution domain (left), which in turn increases the charge density and thus the intermolecular charge repulsive forces (right). swelling pressures of anicular cartilage and the intervertebral disc (Maroudas, 1979; Urban & McMullin, 1985), By Starling's law, this swclling pressure is, in llll'n. resisted and balanced by tension developed in the collagen network, confining the PCs to only 20% of their fTee Solulion domain (Maroudas, 1976; Mow & Ratcliffe. 1997; Sellow eL aI., 1995). Consequently. this swelling pressure subjects the collagen network to a "pre-stress" of sig· nificant rnagnitudL: c\'en in the absclH.':c or cx loads (Sellon cl aI., 1995, 1998). Canilage PGs arc inhomogeneollsl.\' dislri th,'oughoul the malrix, with their concenlralion erally being highest in the middle zone and low the superficial and deep zOlles (Lipshitz Cl aI., ,'Jlaroudas, 1968, 1979; Vcnn, 1978). Thc biome ical consequence or this inhomogeneous swelli havior of cartilage (caused b.v the varying PG co throughout the depth of the tissue) has recently quantitalively assessed (Sellon el al., 1998), Al sults from n:cent finite clement calculations on models incorporating an inhomogeneous P lribution show lhat it has a proround effect o interstitial counler-ion dislribulion throughou depth of the lissuc (Sun et aI., 1998). \·Vhen a compressive stress is applied lo th lilage surface. there is an inSlantaneous defo lion caused primal"i)~' by a change in the PG ecular domain, Figure 3-7 B. This cxtc.'rnal causes the inlernal prcsstll"e in the matrix ceed the swelling pressure and thus liquid w gin to flow out of" tht: tissue. As the fluid r!()\V the PG concentration increases, which in tu creases the Donnan osmotic swelling pressu the charge-charge repulsive force and bulk pressive stress unlil they are in equilibrium the external stress. In this manner, the phy chcmical properties of the PC gcl lrapped w the collagen network enable il to resisl com sion. This mechanism complements the pla~:ed by collagen that, as previollsly desc is strong in lC'nsion but wcnk in compre The abilit.'· of PCs to resist compression arises fl'om two sources: (I) the Donnan os swelling pressure associated with the t packed fixed anionic groups on the GAGS (2) the bulk eomprcssivc stillness of the coll PG solid matrix, Experimentally, the Donna motic pressure ranges from 0.05 to 0.35 (Maroudas, 1979), while the elastic modulus collagen-PG solid matrix ranges from 0.5 MPa (Armstrong & Mow. 1982; Alhanasiou 1991; Mow & Ratcliffe, 1997). Il is now apparenl that collagen and PG interact and thal these interactions m'e of functional imponance. A small portion or th have been shown to be closely associated wit lagen and may serve as a bonding agent be lhe collagen fibrils. spanning distnnccs thL too great for collagen cJ."oss·links to develop (Ba ~t al.. 1996; Mow & Ratcliffe, 1997; Muir, PGs arc also thought to play an important role in maintaining the ordered structure and mechanical properties of the collagen fibrils (Muir, 1983; Scott & Orford, 1981). Recent investigations show that in concentrated solutions, PGs interact with each other to form networks of significant strength (Mow et aI., 1989b; Zhu et aI., 1991, .i 996)..Morcover, the density and strength of the interaction sites forming the network \vere shown to depend on the presence of LP between aggre~ cans and aggregates, as well as collagen. Evidence sl.w:gests that there are fewer aggregates, and m(;~e biglycans and decorins than aggrecans, in the superficial zone of articular cartilage. Thus, there mllst be a difference in the interaction bel\VCCn these PGs and the collagen fibrils from the superficial zone than from those of the deeper zones (Poole et aI., 1986). Indeed, the inle"action bet\veen PG and collagen not only' plays a direct role in the organization of the ECr.,ll but also contributes directly' to the mechanical properties of the tissue (Kempson el aI., 1976; Schmidt el aL. 1990; Zhu el aI., 1993). The specific characteristics of the physical. chemical. and mechanical interactions between coHagen and PG have not yet been fully' deter· mi.ned. Nevertheless, as discussed above, we know that these structural macromolecules interact to form a porous·permeable, fiber-reinforced composite matrix possessing all the essential mechanical characteristics of a solid that is swollen with \vat.er and ions and that is able to resist the high stresses and strains of joint articulation (Andriacchi etal .. 1997; I-lodge el aI., 1986; Mow & Ateshian, 1997; Paul, 1976). It has been demonstraled lhat these collagen-PG interactions involve an aggrecall, an I-IA filan1ent, type II collagen, other minor collagen types, an unknown bonding agent, and possibly smaller cartilage components such as collagen type IX, recently identified glycoproteins, and/or polymeric HA (Poole ct aI., 1986). A schematic diagram depicting the structural arrangement \vithin a small volume of articular cartilage is shown in Figure 3-8. \Vhen articular cartilage is subjected to external loads, the collagen-PG solid matrix and interstitial lIuid function together in a unique way to protect against high levels of stress and strain developing in the ECM. Furthermore, changes to the biochemical composition and structur,,;l organization of lhc ECM, such as during osteoarthritis (OA), are paralleled b:y changes to t';e biomechanical proper- Hyalur n -'I,-\-----Interstltlal flU Collagen fi Schematic representation of the molecular organizatio cartilage, The structural components of cartilage, colla and proteoglycans, interact to form a porous composit fiber-reinforced organic solid matrix that is swollen wit water. Aggrecans bind covalently to HA to form large teoglycan macromolecules. ties of cartilage. In the following section, the havior of articular cartilage under loading and mechanisms of cartilage fluid flow will be discus in detail. Biol71echanical Behavior o{ Articular Cartilage The biomechanical behavior of articular carti can best be understood when the tissue is viewe a multiphasic medium. In the present context, a ular cartilage will be treated as biphasic mate consisting of two intrinsically incompressible, miscible, and distinct phases (Bachrach et aI., 1 Mow et aI., 1980): an inlerstitial fluid phase an porous-permeable solid phase (i.e., the ECM). explicit analysis 01' the contribution 01' the charges and ions, one would have to consider t distinct phases: a fluid phase, an ion phase, an charged solid phase (Gu el al .. 1998; Lai et aI., 19 For understanding how the water contributes t mechanical propcnies, in the prc.:SCnl context. arti<,:. lIlar cartilage may be considered as a Ouid-filled porous-permeable (uncharged) biphasic medium, with each constituent playing a role in the functional behavior of cartilage. During joint arlicuhllion, forces at the joint surface may vary" from alnl0S1 zero to more than ten times body weight (Alldriacchi et ai., 1997; Paul, 1976). The contact arcas also vary in a cOll1plcx manner and typically they arc only or the order of several square centimeters (Ahmed & Burke. 1983; Ateshian et al .. 1994). It is estimated that the peak contact stress may reach 20 MPa in the hip while rising from a chair and to IVIPa during stair climbing (Hodge et aI., 1986; Newberrv et aI., 1997). Thus, articular cartilage, under physiologicallonding conditions, is a highly stressed malcrial. To understand how this tissue responds LInder these high physiological loading conditions. its intl"insic mechanical properties in compression, tension, and shear must be determined. From these properties, one can un· dcrstand the load-can~ying mechanisms within the ECM. Accordingly, the following subsections \\"ill characterize tht: tisslIC' behavior under these loading Il'lOclalities. NATURE OF ARTICULAR CARTILAGE VISCOELASTICITY If a material is subjected to the action of a constant (time-independent) load or a constant deforrnation and its response varies with time, then the mechanical behavior of the material is said to be viscoelastic. (n general, the response of slich a material can be theoretically modeled as a cornbination of the response of a viscolls Ouid (dashpot) and an elastic solid (spring), hence viscoelastic. The two fundamental responses of a viscoelaslic material arc creep and stress relaxation, Creep occurs when a viscoelastic solid is subjected to the action of a constant load, Typically. a viscoelastic solid responds with a rapid initial deforrnation followed by a slow (time-dependent), progressivel)' increasing deformation known as creep until an equilibrium state is reached. Stress relaxation occurs when a viscoelastic solid is subjected to the action of a constant deformation, l~ypically. a dscoelastic solid responds with a rapid, high initial stress followed by a slow (time~dependent). progressively decreasing stress required to maintain the deformation; this phenomenon is known as stress relaxation. Creep and strc~s rL,lnxalion phenomcna ll1a~' be caused b~' differenl Illcchani:sms, For single-phase solid pol~'lTh:ric materials, these phenomena are the result of iJHcrnal friction caused b.\' the mOl ion or the long polymL'ric chains sliding over ('ach olher \\'ithin the stressed lll~lIcrial (Fling, 1981), The viscocl"stic bdmvior of tendons and ligamt::IHs is primarily eauscd bv this mechanism (Woo et aI., 1987, 1997). For bone, the long-term viscoelastic lK'hador is thought to ilL' caused h.\' a rdati\'c slip uf lamellae within the osteons along with the Ilow of the inlL'rsti tial Iluid (Lakcs & Saha, 1979). For articul,,,' cartihlgC. the compressive \'iscoelastic bdHl\'ior is prilllar~ i1~' caused h~' the Jlo\\' of tilL' interstitial lillie! and lhe frictional drag associated with this lIow (Ateshian cl aI., 1997; Mow (,t aI., 1980, 1984). In shear, as in single-phase viscoelastic pol.vmers. it is primarily c<.lusL·d b~' tilt:' motion of 10l1g pol~!J1lL'r chains such as collagen and PGs (Zhu d aI., 1993, 1996). Thc component or anicu[;.\r cartilage dscoclasticit~,caused by interstitial lluid !low is known as the hi phasic viscot.,lastic behavior (i'VlO\\' et a!., 1980), ~lI1d the COI11ponent of dscoebslicil~' caused b~' macromolecular motion is known as til(' l1ow-indepl'ndelll (Ha,\'cs Bc)dinc, 1(78) or the intrinsic visc()elastic behavior Ihe collagen-PG solid matrix. Although the deformational bdwyior has been described in terms of n lillem' elastic solid (Hirsch, 1944) or viscoclastic solid (I-laves & Mockros, 1971), lhes~ modcls fail to recognii'.e the role or water in the \'iscoelastic behavior or and thc significant contribution lilal lIuid pressurizalion pla.vs in joint load support and canilage lubrication (Ateshian L't aI., 1998; Elmol'c et aI., 1963, Mow & Ratcliffe, 1997; SokolofL 1963). Recentlv, experimental measurements ha\'c deh:nnined that interstitial fluid pressurization supports mort' than YOOk of the applied load to the canilage surface (Solt/. & Ateshian. 1998) immediately following loading. This ('lfcC can persist for more than 1,000 scconds and thus shields til(.' EC1VI and chondroc.\'tcs from the crllshing defonnations or the high stresses (20 MPa) resulting frollljoilll loading. '* or CONFINED COMPRESSION EXPLANT LOADING CONFIGURATION The loading of cartilage in vivo is cxtr('mel~" complex. To achic\'C' a better understanding of the deformational behavior or the tissue under load, an explant loading conliguration known as confined compression (lvlow et aL, 1980) has been adopted by researchers. In this configuration, a cylindrical cartilage specimen is fitted snugly il1to a cylindrical, srnZoth-walled (ideally frictionless) confining ring that prohibits motion and fluid loss in the radial direction. Under an axial loading condition via a rigid p~rol1s-permeable loading platen (Fig. 3-9;1), fluid will flo\v from the tissue into the porous-permeable platen, and, as this occurs, ~he cartilage samp~e will compress in creep_ At any tIme the amount 01 compression equals the volume of nuid loss because both the watcr and theECM arc each intrinsicall:v incompressible (Bachrach et aI., 1998). The advantage of the confined compression tcst is that it creates a uniaxial, one-dimensional flow and def()rma~ tional J1eld within the tissue, which does not depend on tissue anisotropy.' or properties in the radial diN recLion. This greatly simplir-Ies the mathematics needed to solve the problem. It should be emphasized that the stress-strain, pressure, fluid, and ion flow fields generated within the tissue during loading can only be calculated; ho\VevcI~ these calculations are of idealized models and testing conditions. There are man:\' confounding factors, such as the time~dependent nature and magnitude of loading and alterations in the natural state of prc~strcss (acting \vithin the tissue), that arise from disruption of the collagen network dur~ ing specimen harvesting. Despite limitations in de~ terrnining the natural physiological states of stress and strain within the tissue in vivo, a number of re searchers have made gains to\\'ard an understand~ ing of potential mechanosignal transduction I11ech~ anisms in cartilage through the use of explant loading studies (Bachrach et aI., 1995; Buschmann et aI., 1992; Kim et aI., 1994; Valhmu et aI., 1998) based on the biphasic constitutive law For soft hy'drated tissues (Mow et aI., 1980). N BlPHAS1C CREEP RESPONSE OF ARTICULAR CARTILAGE IN COMPRESSION The biphasic creep response of articular cartilage in a oneNdimensional confined compression experiment is depicted in Figure 3-9. In this case, a constant compressive stress (To) is applied to the tissue at tinlC to (point A in Fig. 3-98) and the tis~ Slie is allo\ved to creep to its final equilibrium strain (EX). For articular cartilage, as illustrated in the top diagrams, creep is caused by the exuN dation of the interstitial fluid. Exudation is most rapid initially', as evidenced by the early rapid r of increased deformation, and it diminishes gra ually until flow cessation occurs. During cre the load applied at the surface is balanced by compressive stress developed within the collag PG solid matrix and the frictional drag genera by the flow of the interstitial fluid during exu tion. Creep ceases when the compressive str developed within the solid matrix is sufficient balance the applied stress alone; at this point fluid flows and the equilibrium strain EX reached. Typically, for relatively thick human and bov articular cartilages, 2 to 4 mm, it takes 4 to hours to reach creep equilibrium. For rabbit c tilage, which is generally less than 1.0 111m thi it takes approximately' I hour to reach creep eq librium. Theoretically, it can be shown that time it takes to reach creep equilibrium var inversely with the square or the thickness of tissue (Mow et aI., 1980). Under relatively h loading conditions, > 1.0 J\ilPa, 50 0/() of the to fluid content may be squeezed From the tiss (Echvards, 1967). Furthermore, in vitro stud demonstrate that if the tissue is immersed physiological saline, this exuded fluid is fully coverable when the load is removed (Elmore aI., 1963; Sokoloff, 1963). Because the rate of creep is governed b.y the r of fluid exudation, it can be used to determine permeability coefficient of the tissue (Mow et 1980, 1989a). This is known as the indirect m surement for tissue permeability (k). Average val of normal hun1an, bovine, and canine pate groove articular cartilage permeability k obtained this manner arc 2.17 X 10. 15 M·'/N·s, 1.42 x 10'" M'/N and 0.9342 x 10. 15 M4/N·s, respectively (Athanas et aI., 1991). At equilibrium, no fluid flow occ and thus the equilibrium deformation can be u to measure the intrinsic compressive modulus (H 01' the collagen-PG solid matrix (Armstrong & M 1982; Mow et aI., 1980). Average values of norm human, bovine, and canine patellar groove articu cartilage compressive modulus H,.\ are 0.53, 0. and 0.55 megapascal (MPa; note 1.0 MPa = Ib/in 2 ), respectively. Because these coefficients ar measure of the intrinsic material properties of solid matrix, it is therefore meaningful to determ ho\v they vary \vith matrix composition. It was termined that k varies directly, while He' varies versely with water content and varies directly w PG content (Mow & Ratcliffe, 1997), Confining ring Impermeable platen A Unloaded (A) ~ oJ "0 w > 1 o a .~ E-' c .~ r Equilibrium (C) Creep (6) E a 0; B o U_A_·___•__• :e> 0- w • •__• Time B A. A schematic of the confined compression loading configuration. A cylindrical tissue specimen is positioned tightly into an impermeable confining ring that does not permit deformation (or fluid flow) in the radial direction. Under loading, fluid exudation occurs through the porous platen in the vertical direction. B. A constant stress (T" applied to a sample of articular cartilage (bottom left) and creep response of the sample under the constant applied stress (bottom right). The o No exudation C - - =-.........._----...- i 6 Copious fluid exudation Time Equilibrium deformation 1 drawings of a block of tissue above the curves illustra creep is accompanied by copious exudation of fluid fr sample and that the rate of exudation decreases over from points A to B to C. At equilibrium (EO'-'), fluid flow and the load is borne entirely by the solid matrix (poi Adapted from MOH~ VC, Kuei, S,C, Lai, Wl'vl., f:.'( af. (1980 Biphasic creep and stress relaxation of articular cartilage in pression: Theory and experiments. J Siomech Eng, 102, 73 STRESS-RELAXATION RESPONSE CARTILAGE IN COMPRESSION viscoelustic stress-relaxation response F,,,,,.t,r,"a, cartilage in a I D compression experidepi,ek'ci in Figure 3-10. In this case, a con,·,,,,,,rtression rate (line t,,-A-B or lower left figto the tissue until lin is reached; b¢ypncl point B, the deformation Uo is maintained. cartilage, the t.ypical stress response this imposed deformation is sho\\'11 in the flgurc (Holmes et aI., t 985; Mow ct aL, 1984). During the compression phase, the str rises continuously until (To is reached, CO sponding to lI o, while during the stress-relaxat phase, the stress continuously decays along the cu B-C-D-E until the equilibrium stress (U X ) is reach The mechanisms responsible for the stress and stress relaxation are depicted in the lower p tion of Figure 3-10. As illustrated in the top d grams, the stress rise in the compression phase is sociated with fluid exudation, while stress relaxat is associated with fluid redistribution within porous solid matrix. During the compressive pha Copious fluid exudation Fluid redistribution (no exudation) t 1I t A B I I III c D Equilibrium deformation E I---"'L-o-a-d-in-g---ll jf------"S-tr-e-ss-'e-Ia-x-a-'i-o-n------ cw B E w ~ a. 6 uo- - - - - - - B C D E ;-~----....- ...- - I o Controlled ramp displacement curve imposed on a cartilage specimen commencing at to (bottom left) and the stressresponse curve of the cartilage in this uniaxial confinedcompression experiment (bottom right). The sample is compressed to point B and maintained over time (points B to E). The history of the stress and response shows a characteristic stress that rises during the compressive phase (points to to B) and then decreases during the relaxation phase (points B D) until an equilibrium is reached (point E). Above these curves, schematics illustrate interstitial fluid flow (represented by arrows) and solid matrix deformation during th compressive process. Fluid e~udation gives rise to the pea stress (point B), and fluid redistribution gives rise to the stress-relaxation phenomena. the hi!:!h stress is generated bv forced exudation of the in~crstilial fluid and the c~mpaclion of the solid matrix 11('<\1- the surface. Stress relaxation is in turn causcd bv the relicf 0" rebound of the high COIllpaction r~gi(}n ncar the surface or the solid matrix. This stress-relaxation process will cease when the compressive stress developed within the solid ll1alrix reaches the stress generated by the intrinsic COIllpressive modulus of the solid matrix corresponding to U o (Holmcs ct 'II., 1985; Mow ct aL, 1980, 1984), Analysis of this stress-relaxation process leads to lhe conc"!usion that under physiological loading cOl1clilions. excessive stress levels arc difficult (() maintain because stress relaxation will rapidly a(lenuat~ the stress developed within the tissue; this must ncces~ sarily lead to the rapid spreading of the contact area in tht:: joint during articulation (Atcshian el aI., 1995, 1998; tvlow & Atcshian, 1997), Rcccntlv, much focus has been on the inholllogencitv o(I-IA with carLilagc depth (Schinagl ct aI., 1996, '1997), Bascd on thi~ data, from an analysis of the stress-relaxation experiment it was round that an inhornogencous tissue would relax at a faster rate than would the unirorlll tissue (\Vang & 1\110\V, 1998). iVloreovcr, the stress, strain, pressure, and Ouid flow fields within the tissue were significanllv altered as well. Thus it seems thal the variation~' in biochemical and structural composition in the layel"s of cartilage pl"ovide anothet- challenge to understanding the environn1ent of chondroc.vtes in situ. PERMEABILITY OF ARTICULAR CARTILAGE Fluid-filled porous materials ma~' or may not be permcablc, Thc ratio of nuid volumc (VI) to thc total volume (\IT) of the porous material is known as the porosity (f3 :; ;: ViI VT); thus, porosity is a geOJ1lclric concept. Articular cartilage is therefore a material of high porosity (approximately 80%), If the pores are int~rconnected, th.; porous material is permeable. PermeabilHv is a measure of the ease with which fluid can f1(~W through a porous material. and it is inversely proportional to the frictional drag ex~ ened by the fluid flowing through thc porous, permeable material. Thus, permeability is a physical concept; il is a meaSure of the resistive force that is required to cause the fluid to Ilow at a given speed through lhe porous-permeable material. Thb fric~ tional resistive force is ~enerated bv rhe inleraction of the intcrstitial fluid-and the p;'rc walls of the porous-permeable material. The permeability coefficient k is related to the frictional drng coefficient K by the relationship k ~ W/K (Lai & Mo\\', 1980), A {j'cular cartilage has a \'(.'1:-' low !kTlllcabilily a thus hi~h frictional l"I.... sistin= I"oret's arc g('n~rat when n~,id is caused tn flo\\" through the poro solid matrix. In the previous sections on c;:\rtilage \·iscoclast ity \\·e discw;st.'d the process of lluid !low throu articular cartihH.!e induced bv solid matrix compre sion and how ll~is process i;lfluences lI1L' viscoda tic behavior or the tissuC'. This process ;:l1so provid an indirect method to dctcrmine the permeability the tissuc. In this section, we discuss the cxpc mCJ1t~11 method uscd to direct I.\' measure the perm ability codTicknt. Such an L'xperim(:nt is depict in Fi~urL' 3-1 I.·\. Here, a specimen or the tissue held fixed in ~\ chamber subjected to the action o pressure gradicl1l: tilL' imposed upstrt.'al11 pl\:ssu PI is greater than the dOwJ1slrt.:;:\Il1 pressure P2' T thick~css 01" tlte specimen is denoted by han the cross-sectional an..-a 01" pt.'nneatiol1 is ddin b\· A. Darcy's law, L1sed to determine the perme~\b it'\" k fron~ this simple t.'xpcrimental St.'ltlp, ~·id k' = Qh/A(Pi-P2). where Q is the \'olumelric d chargL' per unit time through the specimen who area 01" permcm ion is A ([\llo\\' & R;:ltcl irr(" 1997). U ing low pressures, approxiJ11atel~' 0_1 ivlPa. t mClhod was f1rslused to determine the permeabil of articular cartilage (Edwards, 1967; lVlnroud 1975). The value ~I" k obtained in this mann ram!ccl from 1.1 x 10. 1 < m·l/N·s 10 7.6 ;:.< 10. 1 ' m·:/N In ~~ddition, using a uniform straight tube mod the ;:\\IL'rage "pore diameter" has been estimatt.:d a 11111 (Maroudas. 1979), Thus. the "pores" within tinllar cartilage are of mokcular size.:', The pL'nnt.'abilit~'or Hnicular cartibgc under co pressive strain and at high physiological pressu (3 MPa) was first obtained b~' j\Aansour and lV'1 (1976) and later analvzed by Lai and ,Vlo\\' (198 The high pressure and compressh"e strain conditio examined in these studies more c1osd~' resem those conditions found in dianhrodial joint loadin In these l.'xperimcnts, k was measured as a functi two variables: the pressure gradient across specimen and the axial compressive strain applied the sample. Tht.· results from these experiments shown in Figure 3-11 B. Permeabilit~· decreased exp nentially ~lS a function of both increasing compr sive slr~lin and increasing applied Iluid pressure, was hlter shown, however, that. the dependence o Ion the applied fluid pressurt.: derives from co paction of the solid matdx that. in lurn, results fr the frictional drag ("used by rhe permeating nu (Lai & MO\\', 1980), From the point of view of p or Fluid pressure P Rigid porous blocks ,, { Fluid pressure P2 '- Fluid flow "~ "e I i ~ C><I;:, .. ,;.;.;+:.:.:. :.:.:-: .;.:.:.... '- J - I , , , , - I, 1Fluid1flow I I • • ! ! ~ I -'-- 0 h L ~ x •. :c Articular '" §'" cartilage -UIlIU ..I . ,1 ~ A. Experimental configuration used in measuring the perme· ability of articular cartilage, involving the application of a pressure gradient {P,-Pl)/h across a sample of the tissue (h =:: tissue thickness). Became the fluid pressure (PI) above the sample is greater than that beneath it (PJ, fluid will flow through the tissue. The permeability coefficient k in this experiment is given by the expression Qh/A(PI-P/), where Q is the volumetric discharge per unit time and A is the area of permeation. Adapted from Torzilli. PA, & MoJ.v, VC (976). On rhe fundamental fluid cranspon mechafllsms rhrough normaJ and a.'" Applied pressure difference 16 14 (P,-P,) r 10 ... 0.342 MPa r 00.689 MPa ~ 8 1.034 MPa 01.723 MPa 6 4 2 r o[ 0 B .0.069 MPa .0.172 MPa 12 8 16 24 32 40 Applied Compressive Slrain {%} parhologic caail,1ge (Juring function. I. The formulation. J Bio- meeh. 9(8), 5:11-552, a, Experimental curves for articular cartilage permeability show its strong dependence on compressive strain and applied pressure. Measurements were taken at applied pressure differential (p\-p)) and applied strains. The permeability decreased in an exponential manner as a function of both increasing applied compressive strain and increasing applied pressure. Adapted from L,1i, 1,1I/,'\11.. & MoV'-/, II C. (/980). Drag-induced compre5sion of ,1rtiwlar (,milage during a pe-rmeaiion experimenr. J Btorheology, 17. J 11 . ~--------------------------------- ';S,n~IC[lIre, compaction or the solid matdx 9j~ porosity' and hence the average "pore decreases diameter" " \\Iithin the solid matrix; lhus, solid matrix compaction increases fTictional resistance (MOWCl aI., 1984). ... >",:1'he nonlinear permeability of articular cartilage 'X de.monstrated in Figure 3-11 B suggests that the lis<~/!7:9~ has a mechanical feedback sYSlem that may .J;s,e~'ve important purposes under physiological con:'o'ditions. When subjecled 10 high lo"els through the \mechanism of incrcnsed frictional drag against in:)~rslitial l1uiel 110w. the tissue will "ppear sLirfer "nel :';L'h will be more difficult to cause fluid exudation. Re· (cent analyses of articular cartilage compressive /,',/ ':,s,trcss-relaxation behavior have validated this con;C:"'c'ept and its importance in the capacity of the interg, stili,,1 Ouid La support load (Ateshian eL al.. 1998: ;'::;',":,~oltz & Ateshian, 1998). Moreover, this mechanism ;< /al~o is important in joint lubrication, /9BEHAVIOR OF ARTICULAR CARTILAGE §;&,UNDER UNIAXIAL TENSION ~~?''.::~:;rhe mechanical behavior of articular canilage in 3~~;'{enSion is highly complex. In tension, the liss~le is ,:~.~; A strongly anisotropic (being stiffer and stronger for ;'::;;--" '.C:; 'i"; ".'., specimens h"rvested in the direction parallel to the split line pattern than those harvested perpendicular to the split line pattern) and strongly inhomogeneous (for mature anirnals. being stiffer and stronger fOl' specimens harvested from the superficial regions (han those harvested deeper in the tissue) (Kempson. 1979: Roth & Mow, 1980), Intercstingly, articular cartilage rrom immature bovine knee joints does not exhibit these layered inhomogeneous variations; however, the superficial zones o both ITlature and immature bovine cartilage appear to h"ve the ,,,me tcnsile stiffness (ROLh & Mow 1980). These anisotropic and inhomogeneous characteristics in mature joints arc believed to be caused by the vaI)'ing collagen and PG stnlctural organization of the joinl surface and the layering structural arrangements found within the tissue_ Thus, the collagerH'ich superficial zone appears to provide the joint cartilage with a tough wear-resistant protective skin (Sellon et "I.. 1993) (Fig. 3-3;\), Articular cartilage also exhibits viscoelastic behavior in tension (\-Voo ct "I.. 1987). This viscoelastic behavior is allributable to both the internal friction associated with polymeric motion and the flow of the interstitial fluid. To examine the intrinsic mechanica -i· response of the collagen-PC solid matrix in tension, it is necessary to negate the biphasic fluid now effects. To do this, one mllst perform slow, low strainrate experiments (/\kizuki et aI., 1986; Roth & Mow, 1980; Woo et aI., 1987) or perform an incremental strain experiment in which stress relaxation is al- ;.: ;;. lowed to progress IowaI'd equilibration at each incremenl of strain (Akiwki et aI., 1986). Tvpically, in a low strain-rate (or near-equilibrium tensile) experiment. a displacement rate of 0.5 em/minute is L1sed and the specimens usually arc pulled to failure. Unfortunately. lIsing these procedures to negate the ef~ feet of interstitial Ollid now also negates the Illanires~ tation of the intrinsic viscoelastic behavior of the solid matrix. Thus, only the equilibrium intrinsic mechanical propenies of the solid matrix ma~r be determined from these tensile tests. The intrinsic viscoelastic properties of the solid matrix Illust be determined from a pure shear study. The "equilibrium" stress-strain curve ror a specimen of articular cartilage tested under a constant low strain-rate condition is shown in Figure 3-IL Like other f,brous biological tissues (Iendons and ligaments), articular cartilage tends to stiffen with increasing strain when the strain becomes large. Thus, over the entire range of strain (up to 60 0m) in tension, articular cartilage cannot be described b~: a single Young's modulus. Rather, a tangent modulus, defined by the tangent to the stress-strain CUI'\'C. must be used to descl'ibe the tensile stiffness of the tissue_ This fundamental result has given rise to the wide range of Young's modulus. 3 to 100 ,\lIPa, reponed for anicular cartilage in tension (Akizuki et aI., 1986; Kempson, 1979; Roth & Mow, 1980; Woo et aI., 1987). At physiological strain levels, however, less than 15% (Armslrong et aI., 1979) of the linear Young's modulus of anicular cartilage ranges between 5 and 10 MPa (Akiwki et aI., 1986). Morphologically, the cause for the shape of the tensile stress-strain curve for large strains is depicted in the diagrams on the right or Figure 3-12. The initial toe region is caused by collagen fiber pull-out and realignn'lCnt during the initial pan ion of the tensile experimel1l, and the final linear region is caused by the stretching of the straightcnedaligned collagen fibers. Failure occurs when all the collagen fibers contained within the specimen are ruptured. Figure 3- I 3/\ depicts an ullslretched articular cartilage specimen, while Figure 3-138 depicts a stretched specimen. Figure 3-14, A & B shows' scanning eleclron micrographs or carlilage blocks under 0 and 30% sU'elch (right) and the COlTesponding histograms of collagen fiber orientation +i TenSIle modulus. E I I '=~ Ci!t - ~~~ Failure c .; 1 .!---- ~ 2 iii i ~--\ ~~~ Linear regio I I .------- L ,=------------~ ~ ~~~ Toe region ~_S-'-'''-;n-.-'_(_~--U-L-O-)------ Typical tensile stress·strain curve for articuliH cilrtilage dril\Nings on the right of the ClJrve shm.", the configu of the collagen fibrils al vcUiOllS stages loading. In th region. collagen fibril pull·out occurs c15 the fibrils ali themselves in the direction of the tensile load. In the eM regiol1, the aligned collagen fibers are stretched failure occurs .--i ------------ delermined from tilL' ~cnnnil1g dcctron microg pkltln..: s (h.'h). Ch.'arl.\· it can he seen lhal ll1l.' gcn lletwork within c~lrlil~\gL.' l"L'sponds to le sll\.. s~ and strain (\Vada 1..'\:. Akizuki, 19~7l. If the mok'ctll~lr st!1H.:(urc or L'(Jllag('~Il, Ilk' orga lion of thL' collagen fibers within the collagclloll \\"01''', or tilL' collngl'1l IiIK'!' lToss-lillking is al (such ~lS tin\! OcclIlTing in mild fibrillation or OA tensile prolKTlies of the I1L'I\\'ork \\'ill ch~lIlge. Sch l't al. (1990) ha\'c shown a definitive' t\:.-btionsh 1\\"C'I.'11 collag.. . n hydroxypyridinillm cn)ss-linkin lensik' stilflk.'SS ~II)(I stn.. ,lgtll of nOI'mal h(wirH.~ hlg(·. Aki/.uki et al. (1086) showcd lhat pr()grc dcgr~\dnlion or lUI man kllL"C joint cartilag:(.', from Of\. ~'iclds n progn..:'ssin· (k'kriorati propl".'1'liL'S IJI" Ihe collagcn-PG matrix. Similar rl.'sults 11<.1\'(' bccll obsen'\..-d n.. .c\..' 'lI1ill1"lll1odcls of O!\ (Guil", l't ,Ii.. 1004; SCt!(Hl It)t).t). l()g\..,tllcr, lhese obSCI·V~llioJls sllpporl tll'..' that disruptioll of tilL' colbgcn IlL'lwork is ~\ kL'~' ill thL' initial CH'IHS iL'adillg 10 11K' dC\'L'lopmcllt o Also. loosening of til\.' collagen network is gCI1\..'ral librillation 10 Ilk' illl rinsic Icnsik' licved to h(' responsible for tlte incrcased swc hCllci..' W~Il\..T COI1H.'nl. or ostL'oartlll'ilic car (Mankin &: Thrasher, 1973: iVlaroudas. 1979). \,Vc ;:d]"L'ad~' discussed htl\\" iJ1cl"L'asl'd walcr cOlllClll Unloaded ..AIf--+ Collagen +--f-water Proleoglycan i ' E.R., Lal: V:l.M.• & Mow V.C (1984). A continuum rheory and a cartirage. J Bio experiment for the ion-induced swelling behavior mech Eng, 106(2), 15/-158 to decreased corripresstve stiffness and increased perrneability of arti~ular cartilage. BEHAVIOR OF ARTICULAR CARTILAGE IN PURE SHEAR In tension and cOIllpression. only the equilibrium intrinsic properties of the collagen-PG solid malri~ can be determined. This is because a volumetric change always occurs within a material when it is subjected to uniaxial tension or compression. T volumetric change causes interstitia) nuid now a induces biphasic viscoelastic effects within the sue. H, however, articular cartilage is tested in pu shear under infinitesinlal strain conditions. pressure gradients or volumetric changes will 'produced within the material; hence, no interstit fluid flow will occur (Hayes & Bodine, 1978; Z et aJ. 1993) (Fig. 3-15). Thus, a stead v dynam pure shear expct'iment C£ln be used [0 assess Tension 0% • 501 40 1 C n=203 x = 52.0' " 23.0' ; ~ 30 1 20-i £ 1 ~ lmmJlLill.IILillilllllLffi,. o 45 90 Degrees A Tension 30% • 50 i n = 145 J:::~! '8.9~ 'km= :: lOi0-' o I 17.6':> In,m_r!L~ __,. 45 90 Degrees B Direction of Load Collagen fibril alignment is dearly demonstrated by the scanning electron micrographs (X10.000) (right) of cartilage blocks under 0% stretch (A) and 30% stretch (8). The histograms (left), calculated from the micrographs. represent the percent of collagen fibers oriented in the direction of the applied tension. At 0% stretch the fibers have a random orientation; however. at 30% they are aligned in the direction of the applied tension. Reprinred with permission (rom Wada, T.. & Akizuki, S. (1987). An ulcrasuucllIral sllIdy of solid matrix in articular cartilage under uniaxial tensile stress. J Jpn Orthop Assn. 61. intrinsic viscoelastic propcrLics or the collagen-PG solid matrix. In a steady dynamic shear experiment. the viscoelastic properties of the collagcn-PG solid matrix are detenl1ined by subjecting a thin circular wafer of tissue to a steady sinusoidal torsional shear, shown in Figure 3-16. 1n an experiment of this type, the tissue specimen is held by a precise amount of compression between two rough porous platens. The lower platen is attached to a sensitive torque transducer and the upper platen is attached to a precision mechanical spectrometer with a senrocontrolled de mot01: A sinusoidal excitation signal may be provided by the motor in a frequency of excitation range of 0.01 to 20 hertz (Hz). For shear strain magnitudes ranging from 0.2 to 2.0°10, the viscoelastic propcnies arc equival~nl1y defined the elastic storage modulus G'; the viscous modulus G" of the collagen-PG solid matrix ma determined as a runction of frequency (Fling, 1 Zhu et aI., 1993). Sometimes it is morc convenient to determ the magnitude of the dynamic shear modulus given by: IG*I' = (G')' + (G")' and the phase shifl angle given by: l) = tan' (G"/G') The magnitude or lh~ d~'namic shear modul a measure of the tala I resistance offered bv the coelastic material. The value of 0, lhe angl~ betw [ I I Unloaded Pure shear I Collagen +----fWater Proteoglycan B Schematic depiction of unloaded cartilage (A), and cartilage subjected to pure shear (8). When cartilage is tested in pure shear under infinitesimal strain conditions. no volumetric changes or pressure gradients are produced; hence, no interstitial fluid flow occurs. This figure also demonstrates the functional rote of collagen fibrils in resisting shear deformation. ','the steady applied sinusoidal strain and the steady ~ipusoidallorqllcresponse, -.0," is a measure of the total ':";Edctional energy dissipation within the ll1aterial. EoI' a pure elastic material with no internal fric,:'ti'onal dissipation, the phase shift angle 8 is zero; for :iijJurc viscous fluid, the phase shih angle 8 is 90t '. )/;·The magnitude of the dynamic shear modulus for 'i':t1()rmal bovine articular cartilage has been measw\:d to range from I to 3 ~llPa, while the phase 'r shift angle has been measured to range from 9 to 20' (Hayes & Bodine, 1978; Zhu Cl aL. 1993), Thc inirinsk transient shear stress-relaxation behavior of ::':tI1t~ collaucn-PG solid matrix along with the steadv '\Iynamic . . .s hear properties also ha; been llleasUl'e~1 ;',:. (Zhu el aI., 1986), Wilh bOlh lhc sleady dynamic and '::'thc. transient results, the laller investi'gators showed . ~at the quasilinear viscoelasticity th~ory proposed '?;:'.py Fung (198l) for biological materials provides an <"accurale dcscription of the flow-independent visi; coelaSlic behavior of lhe colla!!cn-PG solid malrix, '.:':/,}<'jgure 3-17 depicts a comparis--on of the theoretical 'prediction of the transient stress-relaxation phe;[lomenon in shear with the results from Fung's t 981 'quasilinear viscoelasticity theory. /:.. From lhese shear sluciies. it is possible to obrain SOme insight as to how the collagen-PG soHd matrix functions. First, we note that measurements of PG >:. SOlutions at concentrations similar to those found in ':>a. nicular cartila(Tc eo in Silu .'vidd a ma~miLudc of shear 0:;" modulus 10 bc of thc order of 10 Pa and phase shifl angle ranging up to 70" (Mow et aI... 1989b; Zhu ct aI., 1991, 1996), Therefore, it appcars that the magnitude of the shear modulus of concentrated PG so- SIN (f n I 'd I .. 111 I I ~I Time Steady sinusoidal torsional shear imposed on a specimen in pure shear. The fluctuating strain in the form of a sine wave with a strain amplitude filand frequency f. Iht: ullla~I..·1l 111..'1\\1 Irk. 1')I,.·rJllib till..' PC ~I. to rl..':--.i:--t c{JlI1l)rl..·.... ~ion ( 192-l: Marolld.. I:--'. 1~79: ,'vl0\\·I..\: Rall.:lilll', 1997 cOlllll for :--.tH.'h Fix('d Char~l..· DI..'I1:--.ily (FeD, 1 cartila1!l'. a lriplla~i(.· Ilk,dlancl-l..·kdnl\,:llelTlica 1..·!L·I...·ln,I.\·k· Ihcor,\' \\":I~ d ...·\·t.. !opl..·d [kit rJl(ldd~ "l~ a mi'\tlIrl...· Ihn..·\,· misdhk' pklsl..·s: a cklr~l pkhl..· rl..'j)rL·SI..·llIill~ IhL' \.:(jIl"I~I..·n.PC IlL'I\Uli phasL' rl..'l.,rL·~I..'llling th\,o inll..'rstiti .. d \\·<llI..'r, <lll ph"lse cOlllpri~illg tlk' llltlnO\·i.lk'llt L'alillil ' a ion CI a~ \\"1..,11 as Dllk'l' 1l111llh·.. dl..·1l1 Slk'l..:iL's C"I' (ClI L'I <II., llJlJ:{: l.'li 1..'1 al..llJ01 ). I Il I hi~ I h IOlal slrl..·s~ i:- ~in'll h.\· IIIL' slim 1\\'0 1I..'rllls CT,,,h.1 .~. Ulh"'i. \\"hel·I..' (T,,·"·I .llId (T,:',,·I "Irl' Ihl' ~ ill cc.,II"I~l:1I Ill't\\"ork 1.0 11 i5' C ·uc0 '= 0.. 0004 12"" 36.2 c = 0.13 0.8 ~ (lr U. c .Q 1ii x ~ ! a; c: 0.' j ! j ! ! '0 ~ U ~ ., '0 or a: 0.2 0.0 0 10 20 30 Time (sec) Typical stress-relaxation curve after a step change in shear strain. expressed in terms of the mean of ten cycles of stress relaxation normalized by the initial stress. The solid line represents the theoretical prediction of the quasi lin· ear viscoelasticity theory, Adapted from lilu, WB .. Lai. WA1., & Mav/, Vc. (/986). Intrinsic quasi.. /inear viscoelasric beflMior of the extracellular matrix of cartilage. Trans Orthop Res SoC, 11,407. ); IUlion is one hundred thousand times less and the phase angle is six to seven times grcmcr than lhat of anicular cartilage solid matrix. This suggests lhat PGs do nOl function in situ to provide shear stiffness for articular carlilage. The shear stiffness of arLicLIlar cmotilage Illust lherefore derive (Tom its collagen content, or from the collagen-PG interaction (Mow & Ralcliffe. 1997). From Ihis interpretation, an increase in collagen, which is a much more clastic element than PC and the predominant load-carrying element of the tissue in shear. would decrease the frictional dissipation and hence the observed phase angle. SWELLING BEHAVIOR OF ARTICULAR CARTILAGE The Donnan osmotic swelling pressure, associated with the densely packed fixed anionic groups (SO,. and COO·) on the GAG chains as well as the bulk compressive stillness of the PG aggregates entangled tri.'\ sln:ss <Inti inli:rsliliailluid pr\'·~:--.un... r",,:-I..·quilibl·illill. (T'l,,,,, is gi\·I..·!l h~' I hI..' [)OIlIl ..Hl prcssur..... IT hcl' disCllS~i(J1l Ik·lu\\·l. J)erin'd till..' fulltl ..t1n.... nl .. t1 laws (II' IllI..'C!J"lIlics <.llld d.\"ll~tlllics rather Ill"lll through IIh..· ad 11m: c tion of L'xisting :--pi..·i.:i ..l1i·I.I..'d thl·{Iril...·:-- k.g .. Grod/.insk.\", [YS/a,hL this 11'ipkl.... ic Ilk'or,\" ~\ ",\..-'1 I!lLTln(ldYIl~IJllicalh' Ik·rllli ...... ihk C li\'\..· law.. . to dl''''lTibl..' lit\..-' till'lL'-dL'[k'lllk'llI l·lll..'lllic~d, Ill\..-'Cll~lnil·;d, :IIHI L'k'l'lriL"al pI'OpL cll<:II'gL'd-!lH!ralL'd so!'t tiSSlll ;\'11. 1I'L'O\·\..-'1', Ill\.. ."ic 11lulti"ek'ctrol\'!L' tlwol',\' h<l h l 'L'1'l shown tirl·l.\' l'on ... istl'11{ \\'ith thl' ",pl'ci<tlih·d clas nWlic PI'I..'SSUl'L' thL'(lly fut' ch'll'~l'(.1 POIYnll.' ti011S, plh.'llOllll·llologic;,,1 {1'~lnSpol't II1L'ori tilL' hiphask theor.\· (DOIlIl .. IIl, 192-l: I(;'ltch Curran. 1973: .\'Io\\' 1..'1 .. 11 .. 19~O: On... a~L'r. 193 \\'hkh In\\''''' bL'I...·1l frL'lI11L'ntly llsL·d 10 slud.\· f.. lcl..'ls (II' aniCllbl' clrlilagL·. Th...' Iriphasic IhL'or~' has bl...·I...·!1 used sliccl... describe lll .. tny clf till..' 1l1L'l'll"111(I-l'kcII'Oclll.. ha\·ior:-. or articubr cartilage. Tl1l..'s...· include diclion frl..·e S\\·L·lIing lIIHk'r c1l1..'lllicalload L'a!' dCIK·IH.k·IlCL· or h.nlr"llllic pl·rnll..·i.lhilil.\· \\" llo!llirH.·;11' depcn(k'llCC stl'L'allling pOIL'llt FeD: cLll'ling or Glnil ..lg,l· l"l.\·LT:-: pn.'-qrl'ss: ..Inc! Ilcgatin' osmotic flows: s\\'clling ..Ind l responses or cells 10 11SIlH>lic SIH)(.:k IO<lding; inlllll'lln:.' or inhomog,('IlL'OUS fi.'\l·d dl<:lr~c (CUd al.. 1993. 1997. 1<.)%; Lai ,'1 "i.. 1<,)<,)1 al.. J 998; Sellon L·t al... 1Y9t': SUIl ('I aI., 199 \'idillg Illorl..' \·I..'rsalilit.\·. Ih...• triphasic Ih...·ol:· gcneralizcd to indlldl' lllulti-dcCII'OI.\·ICS in I (Gil L'I al.. 1Y9~). From analysis llsin~ !l1l' triphasic theor. cOllles clear thai thl' swelling hdw\'lor or lh (:an hI..' I'L'sponsibiL- for ..I sigllific<IlH fractio compn,:ssi\'c IOi.ld-bL.'~lriJlg Ci.lp~\i.:it~' artic ..-\1 (lr ur or or or tilagc at equilibriulll (Mow & Ratcliffe, 1997). For example, the triphasic theory' predicts for confined· compression at equilibrium that the total stress ((Jl"I'd) acting on the cartilage specimen is the SlIlll of the stress in the solid matrix (U"'lid) and the Don· nan osmotic pressure (Ulh'id = n). The Donnan osmotic pressure is the s\\.'clling pressure caused by.' the ions in association \vith the FeD and represents the ph:vsicochcmical motive force for cartilage ~\Vcl1ing (Fig. 3-18). From the classical theOl}' for osmotic pressure, the Donnan osmotic pressure caused by the excess of ion pnrticlcs inside the tissue is given by:: IT = RT[<[,(2c+c')-2<!,*c*j + p= where c is the interstitial ion concentration, c'" is the external ion concentration, c" is the FeD, R is the universal gas constant, T is the absolute temperature, (t) and 4)'" are osmotic coefficients. and peo is the osmotic pressure caused by the concentration of PG particles in the tissue, usually assumed to be negligible (La! et aI., 1991). For a lightly loaded tisSll~, ~he s\vclling pressure ma:v co~~tribule significantly to the load support. But for highly loaded tissues, such as those found under physiological conditions and certainly.' for dynamically loaded tissues, the interstitial fluid pressurization (Jtlllid) \vould dominate; the contribution of this swelling pressure to load support would be less than 5(jf; (Soltz & Ateshian, 1998). As with the biphasic theory, the triphasic n1echano~ electrochemical theory can be used to elucidate potential mechanosignal transduction mechanisms in cartilage. For example. because of their potential effects on chondrocyte function, it is important to describe and predict electrokinetic phenomena such as streaming potentials and stream~ ing currents (Gu et aI., 1993, 1998; Katchalsky & Curran, 1975; Kim et al" 1994) that arise from ion movement caused by the convection or interstitial fluid flow past the FCD of the solid matrix. As a second example, the pressure produced in the interstitial fluid by polyethylene glycol-induced osmotic loading of cartilage explants (Schneiderman et aI., 1986) was recently shown to be theoretically nonequivalent to the pressure produced in any other common Iv used mechanicallv loaded explant experime~t or by hydrostatic -loading (Lai et aI., ] 998). In light of this finding, earlier interpretations of biological data from studies making such an assumption of equivalency should be reVisited. 0.4 Swelling pressure rr 0.3 0.2 - 0.1 o 0.15 0.5 1.0 Bathing Solution Concentration c' (M) Swelling pressure of articular cartilage versus bathing s tion concentration (c*). At equilibrium, the interstitial f pressure is equal to the swelling pressure, which is defin by the tissue Donnan osmotic pressure (IT). Lubrication or Articular Cartilage As alread.y discussed, synovial joints are subjec to an enormous range of loading conditions, under normal circumstances the cartilage surf sustains little weac The minimal wear of norm cartilage associated with such varied loads indica that sophisticated lubrication processes are at w within the joint and within and on the surface of tissue. These processes have been attributed to a bricating fluid-film forming between the articu cartilage surface and to an adsorbed boundary bricant on the surface during motion and load The variety or joint demands also suggests tha number of mechanisms are responsible for arthrodial joint lubrication, To understand diarth dial joint lubrication, one should use basic e neering lubrication concepts, From an engineering perspective, there are fundamental types of lubrication. One is bound lubrication, which involves a single monolayer o bricant molecules adsorbed on each bearing surfa Subchondral bone d;J <:>\.kV (Q:> ~ Load Articular cartilage Synovial fluid gap Articular cartilage A Subchondral bone Hydrodynamic d;J <:>\.kV yn n (Q:> d;J <:>\.kV n Load and Motion Load and yMo/ion ~ Load and yMotion • p +f-o • B Squeeze-Film c Boosted A, In hydrodynamic lubrication, viscous fluid is dragged into a convergent channel. causing a pressure field to be generated in the lubricant. Fluid viscosity. gap geometry, and relative sliding speed determine the load-bearing capacity. B. As the bearing surfaces are squeezed together, the viscous fluid is forced from the gap into the transverse direction. This squeeze action generates a hydrodynamic pressure in the fluid for load support. The load-bearing (a- The other is fluid-film lubrication, in which a thin fluid-filn1 provides greater sllrrace-to~sllrraceseparation (Bowden & Tabor; 1967). Both lubrication types appear to occur in articular canilage under vaJying circulllstances. Intact synovial joints have an extremely low cocrf!cicnl or friclion, approximately o Weeping pacity depends on the size of the surfaces. velocity of a proach, and fluid viscosity. C. The direction of fluid flow under squeeze-film lubrication in the boosted mode for joint lubrication. D. Depicts the Weeping lubrication hypothesis for the uniform exudation of interstitial fluid f the cartilage. The driving mechanism is a self-pressuriza of the interstitial fluid when the tissue is compressed. 0.02 (Dowson. 196611967; Linn. 1968; McCutch 1962; Mow & Atcshian. 1997). Boundary-lubrica sllIfaces typically have a coefficient of friction one two orders of magnitude. higher Lhan surfaces lu cated by a fluid-111m, suggesting LhaL synovial joi are lubricated. al Icast in part. by the fluid-f ?,\/,.. :.': :t'ni~~Aanism. It is quite possible that synovial joints ;;{',.~. ::.'\i~.{tihe:;mechanism that will most effectively provide ;ft\:. fU$rica~ion at a given loading condition. Unresolved, .;~:;' ~23:Jnc}Y:fllFjS the manner by which synovial joints gcn- ';%.;0;;:~~;>tethe fluid lubricant film. ;/:; ;:';..c;. ;,~Z ""''',.'. ~~r ' }-iN,., Fti.JIP-fILM LUBRICATION ;;;.;;iir$Jdi~i!ri]m lubrication utilizes a thin film of lubl'i"H:~'~'.> c~nt that causes a bearing surface separation. The '6.·Jld.~a on the bearing is then supported by the pres";j~\.tt·~that is developed in this fluid-film. The nuid.'";:' ,..~';dihn';thickness associated with engineering bearings C/. 'isitsually less than 20 f.l.m. Fluid-film lubrication re' ... q"jres a minimum nuid-fllm thickness (as predicted ,i':i>Y/a"speciflc lubrication theOIy) to exceed three ~f~;tioles' the combined statistical surface roughness of cahUage (e.g., 4 to 25 f.l.m; Clarke, 1971; Walker et ,al:;)970). If fluid-film lubrication is unachievable , ',;':,b~C~lllse of heavy.. and prolonged loading, incongru'e'nL'gap geomclI)', slow reciprocating-grinding 1110tion, or low synovial fluid viscosity, boundary lubrication must exiSl (Mo\\' & Ateshian, 1997). The two classical modes of fluid-film lub"ication " defined in engineering are hydrodynamic and squeeze-film lubrication (Fig. 3-19, A & B). These modes apply to rigid bearings composed of relatively undeFormable material such as stainless steel. l-lydrodynamic lubrication occurs when nonparallel rigid bearing surfaces lubricated by a nuid-film move tangentially with respect to each other (i.e., slide on each other), forming a converging wedge of Ouie!. A lifting pressure is generated in Ihis wedge by the fluid viscosity as the bearing motion drags the fluid into the gap betwcen lhe surfaccs, as shown in Figure 3~ 19A. In contrast, squeezc-f-Ilm lubrication Occurs when the bearing surfaces 1110\'e perpendiculad.''''- toward each other. A pressure is generated in the fluid-film as a result of the viscous resistance of the fluid that acts to impede its escape from the gap (Fig. 3-19B). The squeeze-film mechanism is suffJcientto cany high loads for short durations. Eventually, however, the fluid-film becomes so thin that contact between the asperities (peaks) on the two bearing surfaces occurs. Calculations of the relative thickness of the fluidfilm layer and the surface roughness are valuable in establishing when hy'drodynarnic lubrication may exist. In hydrodynamic and squeeze-film lubrication, the thickness and extent of the nuid-fllm, as well as its load-carrying capacity, are characteristics independent of the (rigid) bearing surface material properties. Thcse lubrication characteristics are in- stead determined by the lubricant's properties, such as its rheological properties, viscosity and elasticity the film geometry, the shape of the gap between th two bearing surfaces, and the speed of the relativ surFace motion. Cartilage is unlike any man-made material with respect to its near Frictionless properties. Classica theories developed to explain lubrication of rigid and impermeable bearings (e.g., steel) cannot fully explain Ihe mechanisms responsible for lubrica tion of the natural diarthrodial joint. A variation of the hydrodynamic and squeeze-film modes o fluid-film lubrication. for example, occurs \vhen the bearing material is not rigid but instead rela tively soft, slich as with the articular cartilage cov ering the joint surface. This type of lubrication termed elastohydrodynarnic, operates when th relatively soft bearing surfaccs undergo either sliding (hydrodynamic) or squceze-film action and the pressure generated in the fluid-film substan tiallv deforms the surfaces (Fig. 3-19, A & B) These deformations tcnd to increase the surfac area and congnlcl1cy. thus beneficially alterin film geomCli)'. By increasing the bearing contac area, the lubricant is less able to escape from be tween the bearing surfaces. [t longer-lasting lubri cant film is generated. and the stress of articula tion is lower and 1110re sustainable. Elastohydro dynamic lubrication enables bearings to greatl increase their load-carrying capacity (Dowson 196611967,1990). Note that several studies have shown tha hyaluronidase treatment of synovial fluid, which de creases its viscosity (to that or saline) b~; causing de polymerization of HA, has lillie effect on lubricatio (Linn, 1968; Linn & Radin, 1968). Because nuid-fllm lubrication is highly dependent on lubricant viscos ity, thcse results strongly suggest that an alternativ 1110de of lubrication is the primary mechanism re sponsible for the low frictional coefficient of joints. BOUNDARY LUBRICATION During diarthrodial joint function. relative motio of the articulaling surfaces occurs. In boundary lu brication. lhe surfaces are protected by an ad sorbed layer of boundary lubricant, which prevent direct. surface-to-surface contact and eliminate most of the surface wear. Boundarv lubrication i .' essentially independent of the phy~ical propertie of either the lubricant (e.g .. its viscosit~,) or th bearing material (c.g., its stiffness). instead de pending almost entirely on the chemical propertie of the lubricant (Dowson, 1966/67). In synovial joints. a specific glycoprotein. "Iubricin," appears to be the synovial fluid constituent responsible for boundary lubrication (Swann ct aI., 1979, 1985). Lubricin (25 X 10·: In\\') is adsorbed as a macromolecular monolayer to each articulating surface (Fig. 3·20). These two layers. ranging in combined thickness frol11 I La 100 11m, arc able to carry loads and appear to be effective in reducing friction (Swann et aI., 1979). More recently, Hills (1989) suggested that the boundary lubricant found in synovial fluid was more likely to be a phospholipid named dipalmiloyl phosphatidylcholine. Allhough experiments demonstrate that a boundary lubricant can aCCOlint for a reduction of the friction coefficient by a factor of tbreefold to sixfold (Swann et aI., 1985; Williams et aI., 1993), this reduction is quilc modest corn pared wiLh the much greaLM cr range (e.g., up to 60 fold) reported earlier (McCutchen, 1962). Even so, these results do suggest that boundal)1 lubrication exists as a complementary mode of lubrication. M -_._-- MIXED LUBRICATION There are two joinl lubrication scenarios thal can be considered a combination of fluid·fiIm and boundary lubrication or simpl~' mixed lubricalion (Dowson, 1966). The first case refers to the temporal coe\istence of fluid-film and boundary lubrication at spatially distinct locations, whereas the second case, termed "boosled lubrication," is chanlclerized b~" a Scanning electron miuograph of the surface of huma articular cartilage from a normal young adult showing the typical irregularities characteristic of this tissue (": 3,000). ;.ldap;ed from Armsrrong. C G.. 8 t.J1o'.·.~ V C fl980r Fncrlon. fu!mCM!ol) dfld \~/ear of Synovkll Jomrs. In' O'."If.'fJ. ) Goodfe!lo~·'1. <1.'1(1 P B(Jl!ouqh {Eds.i. Scientl1,( Fou Traumatology fpp 223-232j Lo don Wilham Neinermcl!lft "ons of Or;hopa~dlcs cwd .----------------Articular surface ~n~1\\l~lf2t±- Lubricating glycoprotein Articular surface Boundary lubrication of articular cartilage. The load is carried by a monolayer of the lubricating glycoprotein (lGP). which is adsorbed onto the articular surfaces. The monolayer effectively serves to reduce friction and helps to prevent cartilaginous wear. Adapred from Armsrrong. C G., & Mo'."/, v: C (1980). Friction, rubrication and ~'/ear of synovial joints. In: R. Owen, J. Goodfellow, and P Bullough (Eds.). Scientific Foundations of Onhopaedks and Traumatology (pp. 223-232). Londofl: Wilfiilm Heinermann. • shirl or 1111id-film lo bOl1ndar~-" luhrication with l 0\'(.'1' the same loc'-llion (\Valkcr I.'l aI., \070). The articular cartilagl...' :-;urf<'KL~. like all surface not pL'rfCCtl.v smooth; as(k'ritics project out from surface (Clarke, 1971; Gardner oS: McGillivray. 1 Rc(IIl..~r& Zimn~', 1970) (Fi~s" 3-3/3 and 3·21). In s ovialjoints. situations may occur in which lhe ll film thickness is or the san'll.' order as the 111('..\11 ticular surface aspt..'rity nV..tlkcr ('t .. II", 1970), Du such insl . Hlct..'S, b()undar~' lubrication betweell asperities may COllle into pla~·. If this occurs mixed mode of lubric..ltinn is npcr"lting, with joint surface load sustailk'd b~' both the lIuiclpressure in "l1"eas nOIlCOlllact .. lIlel b~' the bCH ..Iry lubricant lubricin in the .. m:.'as of i.lspcrit.\' c tact (showll in Figure 3-22)" In this mode of mi luhl"icntion. it is probable lit..\( Illost 01" the fric or (which is still extrl...·l11d~-" low) is gellL'ratcd in Adsorbed boundary lubricant Pressutized lIuid ~ -o~m'x;\;~,'~; Boundary lubricated asperity contact I Articular surface Schematic depiction of mixed lubrication operating in ar· ticular cartilage. Boundary lubrication occurs when the thickness of the fluid~film is on the same order as the roughness of the bearing surfaces_ Fluid-film lubrication takes place in areas with more widely separated surfaces. Adapred from Armstrong, e.G" & Mo~-v. Ve. (1980), Friccion. Ivbr;(Miofl clod wear of synovial joints. In: R. Owen, 1. GoocH(!lIow, ilnd P. Bullot/gll (Eds). Scientific Foundations of Orthopaedics (lnd TraumMology (pp_ 223-232), London: William Heinemwnn. • I I i boundary lubricated areas while mOSl of the load is carricd by thc fluid,filrn (Dowson, 1966/1967, 1990), The second mode of mixcd lubrication (boostcd lubrication) proposcd by Walkcr ct aL (1968, 1970) and !'vlaroudas (1966/1967) is based on the movc, mCI1l of fluid from the gap between the approaching arlicular surfaces into the articular cartilage (Fig. 3-19C). Spccifically, in bOOSICd lubrication, articular surfaces arc believed to be prolected during joint loading by the uhrallltration of the synovial fluid through the collagcn-PG matrix. This ultrafiltration pcrmits the solvent component of the synovial Iluid (water and small electrolytes) lo pass into the articular t.::artilage during squeeze-film action, yielding a conc~ntrated gel of HA protein complex that coats and lubricates lhe bearing surfaces (Lai & fvlow, 1978). According to this theory, it becomes progressively more difficult, as the two articular surfaces approach cnch othec for the HA macromolecules in the synovial nuid to escape from the gap bet\veen the bearing surfaces because they are physically loa large (0.22-0.65 fJ.m), as shown in Figure 3-23. The Water and sll'lall solute molecules can still escape into the articular cartilage through the cartilage surface and/ol' laterally into the joint space al the pcI'iphc.)' of thc joint, Thcol'clicall'csults by I-Iou ct aL (1992) pl'cdict lhat lluid cnt.)' into the cal'tilagc, bearing surl"ace is possible, leading them to suggest that boosted lubrication mav OCCUI: The role 01" this HA gel in joint lubrication·' remains unclear, ho\\'- ever, particularly in \'iew of the findings by Li (1968), which dcmonstratcd thai purificd I-IA acts a poor lubricanl. To summarize, in any bearing, the effective mo of lubrication depcnds on thc applicd loads and the relative velocit~' (speed and direction of motio of the bearing surfaces. Adsorption of the synov fluid glycoprotein, lubricin, to articular surfac seems to be most important under severe loadi conditions, that is, contact surfaces with high load low relative speeds, and long duration. Under the conditions, as the surfaces arc pressed together, t boundary lubricant monolayers interact to preve direct contact between the articular Stll·races. Co versely, Ouid-mm lubrication operates under less vere conditions. when loads arc low and/or oscill in magnitude and when the conlacting surfaces a moving at high relative speeds. In light of the vari demands on diarthrodial joints during normal fun tion, it is unlikely that only a single mode of lub calion exists. As )/ct. it is ,impossible lO slate de nitclv under which conditions a particu lubrication mechanism may operale. Neverthele using the human hip as an example, some gene statements are possible. L Elastohyclroclynamic lluid-films of both the sliding (hydrodynamic) and the squeeze typ probably play an important role in lubricati the joint. During the swing phase of walking when loads on the joint are minimaL a subslantiallayer of synovial lluid-film is probab maintained. After the first peak force. at hee strike, a supply of nuid lubl"icanl is generate by articular cartilage. Howcvcl: this nuid-filn thickness will begin to decrease under the high load of stance phase; as a result. squeeze-film action occurs. The second peak force during the walking cycle, just before t toe leaves the ground, occurs \vhcn the joint is swinging in the opposite direction. Thus, is possiblc that a fresh supply of lluicl,film could be generated at toc-orr. thereby provid ing the lubricant during the next swing phas 2. With high loads and low spccds of "dativc motion, such as during standing. the fluidfilm will decrease in lhickness as lhe fluid is squeezed olll from between thc surfaces (flu film), Under these conditions, the fluid exud from the compressed articular cartilage cou become the main cO!1tributor to the lubrica ing film. 3. Under extreme loading conditions, such as during an extended period of standing follow :~ ~'.''.'._--~~"~,-,.,,"'''',!,,,_,_,_----~,.....= ,.,.,,"'"<" "- " ', ,",'!'='-' ' r-'- - - - -~"=,., , ,.7' ' '"' ' ' ,) :~ , "'"r Solute and small particle flow I T-=-0.02-1 urn ~ -=. ,., - Hi Articular surface 1-=-- "- -- -- ' -- -- .. - -11- - -1'1- --11-- - - H- - -1/- - -\/771J/ 77<ttl 777:JJ/ 7 7'ty 7r:Jj/ 7:JJ? 7 7 I 7 7 7 SUrf;:iCC ilmlEL rI ;'cW) :i:r' ri,;;Cfor-r!O:C:CL:ic::-', "" Ultrafiltration of the synovial fluid into a highly viscous gel. As the articular surfaces come together. the small solute molecules escape into the articular cartilage and into the lateral joint space, leaving the large HA macromolecules that, be- edU"\:, of their' "ile', afe \H1i]bir:, tu C'~(ilP(" TlH.'se Hi, nhH:rc" rnoluui(,s forrn d ((mu:,n\r,',ted ~;e! ic'"" th"Hl i l,r11 thick ti' lubricate's the "nicul.Jr' surfi'l(OS, This hypothc,,,iled IL;bric,:;t moclc' is lC'ln:(~d "i)()o51cci lubriCdtion " ~ ing impact, the fluid-film Ina)' be clilninatcd, allo\ving sllrface~to-sllrracecontact, The surfaces, however, will probabl.\' still be protected, either b.\' a thin la)'Cl" of ultrafiltratccl s.. . ·novial fluid gel (boosted lubrication) or by" the adsorbed lubricin monolayer (boundmy lubrication). :\III](lll~~lt i\J:li.,! i:-- p~lrlili(IIl!..:d h,:l\\l'l'll llll' "o :llH,lllllid pll~l:-'l'" o! ~\ hipll~l"il' Ill~\llTi~d (\10\\ l'l ~ jl.lSOJ, .'\ll',,!JI~lll 1]l)l)7) lkri\l'd ~llll'\Pl\'""i(l!l lor t l'lkl'li\\: ~or 1\1l'~\"\1I'l'd) l'(ll'llll'il'lll \\~l" lkpl'lldl'lll :-.okh (II' friction th lo h\ thl' :--olid In~ltl'i\ (c',~ .. lhl' dilll'I\:IlC hl'I\\l'l'll tOI~tl 1(l~\l1 ;\11(.1 tll;\l "lIpj1ol'll'd h\ h.\dl'o:-: OIl thl' proportioll or 1Ill' :-'lIpplll'1l'd Ii . . ' j11\',":-;[II\' ill llll' l]lIidl. Till' ilnplil';llion 01 "lIch ROLE OF INTERSTITIAL FLUID PRESSURIZATION IN JOINT LUBRICATION During joint articulation, loads transmitted across a jointma.\' be supported by the opposing joint surfaces via solid-to-solid contact, through a fluidfilm la,\'er, or b.\' a mixturc of both, Although rIuidfilm lubrication is achievable, its contribution to joint lubrication is transient, a C()J1sequcnce of the rapid dissipation of the fluid-filrn thickness by joint loads, Witb tbis caveat, Atesbian (1997), adopting the theoretical framc\vork of the biphasic theory (Mow et aL, 1980), proposed a mathematical formulation of a boundary friction model of articular cartilage to describe the underlying mechanism behind diarthrodial joint lubrication, in particular, the time-dependcnce of the friction coefficient for cartilage reported during creep and stress-relaxation experiments (1\'lalcolm, 1976; McCutchen, 1962), l'\j1I'l'",,,i(lll i" lilal Ihl' !I'il'li(l]l;tl !1),Opl'l'til'" (d c\n L\~:'-l' \';1]'\ \\ilh lillll' dlll'ill~: ;\ppli,'d 1(1;ldill~, ~l 1\'l 1101l ollhl' IllllT"llli~d l]lIid ;llld \.'(dl;l~:l'll-P(; I1I~l illll'I'~\l'll(lIl" 111;\1 ~i\\' ri"l' to thl' 1]0\\ "lkpl'lldl'111 \ <..'(ll,Lt:--lic Pl'(lPlTlil'" 01 111l' li","lIl' lk:-,crilwd \..';\r ~\Ild "ll(l\\ll in l:i~',Ul\'" ,;_l) ;llld 3" 1O. To \'alidall' Ili:-, l1lodl'!. ;\ll':-,lli~\ll dl'\'l'lopl'd ~\ IHl IU;ldill,~ l'\j1o,.'l'illll'lll lkll "lllkTilllpU:-'l'd ~l I'ri....-tl(j lorqlll' 10~ld 011;\ ""';ll'til;\~:'-l' l'\pLllli 111H,hT~~oill~:'- l'J'l 1();HJin~, ill ;1 Ulillill,'d l'()lllp)'l'S"ioll l.'ollligl1r~\ (Fig, .~-~-L\l ("\tl',,hi~lll l'l ;11, ll)l)Sl. \lnrl' Spl'l L'all.\, ;1 \.'\lilldril';J1 hip!J;l."il' l·;ll'lil~l!.-',l' plug \\'~l:-' l'O pl'l':-""l'd ill ;\ l'()lJ1illing rillg k,~2.., prohihiling r~ld lllolion ;\lll! nuil! l'\ud~\tion) 1l1ll!lT ~\ l'onstalll a plil'd IO;ld ~_',l'lllT~lll'd h\ ~lfl illlPl'ITlll'~\hk' rigid pLl llwl \\~l" r()lalill~,:'- ;11 ;\ prl':-:;\.'I'illl'l! ~lll!--,-lllar :-'jk'l'd, T :-'lIrr~\l>l' or thl' plllg oppO:-,itl' till' pl~\ll'll \\~l~ prl's ;lg;\ilbl ;\ li\l'd ri~2.id PO]'OU:-' IIIll'\' \\lllTl'h.\ tIll' int di::,-it~\ti(lll or tll~,' Ihe l';\)'liL\!2.l' \\'lll1 lhl' rOIl!..'.ll :-'lll'!;\i.X POI'lI\!:-' liltL'!' prl:\'l:l~tcd it I'n)lll n)t:ltill~!:. III l , w Confining chamber - - / - - 7+H---+-- Impermeable rigid platen Cartilage Porous filter --/-----11-- --l---+':~:',,-:~.:.; :::'::l~::..\":.~:~~~.;: I• i A ~ Fluid exudaricn 0.15 I 0.13 0.10 ~ 0.08 0.05 0.03 BOO B 1200 1600 2000 2400 Time(s) Experimental configuration superimposing a frictional torque with creep loading of an articular cartilage explant in confined compression (Ateshian et aI., 1998). A. Note that fluid exudation occurs on the opposite face of the tissue exposed to the frictional load, indicating that the frictional properties of cartilage are not dependent on the weeping of interstitial fluid to the lubricating boundary. B, Note that effective friction coefficient <.... r!') varies with increasing proportion of load on the solid matrix, as can be seen from the theoretical curve for V-rll as a function of time during the experiment. Adapred from Mow, VC & Areshian, G.A (1997). Lubrication and wear of diarthrodia! joints. In 11.e. l'.;low g. We. Hayes (Eds.), Basic Biomect1anics (2nd ed., pp. 275-315). Philadelphia: Lippincott-Raven manner, a frictional torque was developed in the lis~ Sue. Because the application or a torque load that yields pure shear, under inflnitc~imal deformations, induces no volume change in the tissue or associated nuid exudation, the load generated by the frictional torque is independent of the biphasic creep behaviOl" of the tissue. Theoretical predictions, which closely match expel"imcntal results, show that during initial loading, when interstitial pressurization is high, the friction coefficient' can be very low (Fig. 3-2413). As creep equilibrium is reached and the load is transferred to the solid matrix, the friction cocrficient becomes high (e.g. 0.15). The time constant for [his transient response is in excellent agreement with observed experimental results (Malcolm, 1976; J\l1cCutchcn, 1962), Another imponant resull or this work is that nuid pressurization can function in joint lubrication without concomitant fluid exudation to the lubricating boundary as is proposed for weeping lubrication (McCutchen, 1962) (Fig. 3-190). Equally significant, this lubrication theory is capable of explaining the obscl\'cd decrease of the effective friction coefficient with increasing rolling and sliding joint velocities and with increasing joint load (Linn, 1968). Recently. the inlcl'Stilial fiuid pressurization within cm·tilage during uniaxial creep and stress relaxation experiments was sliccessfully measured (Soltz & Aleshian. 1998). As predicted by the biphasic theOly, they found that interstitial Ouiet pressurization supported more that 90'/0 of the load for several hundred seconds following loading in confined compression (Ateshian & Wang. 1995). The close agreement their measurements with biphasic theoretical predictions represents n rnajor advancement in the understanding of diarthrodial joint lubrication and provides compelling evidence for the role of interstitial Ollid pressurization as a fundamental mechanism underlying the load-bearing capacity in cartilage. It is emphasized that while the collagen-PG matrix is subjected to hydrostatic pres~ sure in the surrounding interstitial Ouid, it does not expose the solid matrix (nor encased chondrocytcs) to deformation, presumably causing no mechanical damage. or Iv,:l '1111~h'\"II""lld;k't" \11111;11,:1 ])vl\\l'l'll til", til"~ (d I Ill' 1\\1) l'~II'lil;l~l' '111'1;ll'I.." of Articular Cartilage 'Ncar is the unwanted removal of material from solid surfaces by mechanical action. There are two components of wear: interfacial wear resulting from Lhe interaction of bearing surfaces and fatigue wear resulLing from bearing deformation under load. lnterfacial wear occurs when bearing surfaces come into direct conLact with no lubricanL film (boundary 01' Iluid) separating them. This type or wear can Lake place in either of two ways: adhesion or abrasion. Adhesive wear arises when, as the bearings come into contact, surface fragments adhere to each other and are torn off from the surface during sliding. Abrasive wear, conversely, occurs when a soft material is scraped by a harder one; the harder material can be either an opposing bearing or loose particles bc(ween the bearings. The low rates or interracial wear observed in articular cartilage tested in vitro (Lipshitz & Glimcher, 1979) suggest that di- \\ fli dkt,.·lih· l rukd (Jill. Tltl..· lI11tllipk· 11111<,11.·,. . . 1Il!ll \\l!llill!~ ill ,,'(I1II..'l·rl ;ll'l' Illl' 1111'I..,ll:llli'JII' t \'"1.::11' pi :lrlit"IILII' ,::llliLI.!-'-\." IlllJik ... ·k iIlIVr!;ll·i:d <.·nl1<.'k- ........... ;ldh<.",j\l· ;lIld :d)r~l..... i\<.· \\l·:lr 111:,\ \:11"1 ill :111 illqxtirl·d Ill" . . k·~! . . ·It",·r:ll<."d '.\llll\i:d jllill\ [)h.' l·;II'liLI.!-'-l· :-'lll'I;ll'l' 'll,t:lill' 11111';I'll'll\.:lllJ':tl :llHI'll!" dl'I..Tl·:I'V' ill 111:1". il hl·;,,:Ollh."' :-'Idi :11. !lJSh: :\nn"'l \1t,I.I.. 19~~: Sl·tloll 1:1 :d. I(jl..)-ll. Thu,.... lItJid Irc 111ur<.' rk·nlJl·;lhk· L\ki/ilki 1,:1 Illhl'il';llll !dlll 'l'p:lr;llil'I~: dll' hl';lrill~~ ,111'I:ll' Il,__ ;d~ Tlli~ !\\ .... ·I..'ll ' hd\\l'l'l1 11 lil .... i.:anil:lgc 11lhri...:;Llill~ Illiid ll'IHII 10:-.... 01 f;li.x, pl"! IhrClll~1i :1\\;1\ I11UI\.' L·:l ... ih I hI..· prllhahilil\' III dil\·t,.·1 i.-"Clll inlTl·;l'\.'" 11lL' ;1:-'PlTilil" ;llld V\;ll'I'J"h;lk, tIll' ;lh 't,.·l'~:-'. Ll\i~\II..' \\\,';\1" hl·;ll·;ll.;! '1lI'Lt<.·l·~ I"l· ... IIII,... Ill ld ,"'UI"I~ll"... :-\l'·:-'llrl~iI.,."L" ,.:lllll:tci !lUI Ii-tllli 1111..· ;H,:L" ti('ll (d l1'li,Tl""'i..'iJpL,: d;ll)I;I.:,~L· \\ililin tIll: hV;II'il kri:d until·' l"l·p,:lili\L· ... (rL· ......... ill~. Bl·;lrill;;: , L.illlrl' ,ll;t-' 11 ... '\.'111· \\illl Illl" rl'pl":t1nl ,lpplil':l lli~1i l('~ld ... O\er;l !"c!;tli\l"h :-hl.n lhTi(ld (II' \\t 1\'PI..'1 j I illil lh(lll~h 1IJ\\l..'l' I d' It lilt· j( );Id ... (1\l'1" ;111 l'\ I \.'lldl"d j1l"l'il 1\\ 11l~1[.!llillllk· til' 1!J;11l !lJl" lll:lll"l"i;d":-. thll ... I..' 1(,;((../ ... l1la\ Ik· lilt i1ll;11l" .... 11\·11,;2111. l lii~tll· \\l':11". r...:' .... l!ltill~ il"ll!lll..·\l·lil':llh I"l·Pl';\II.·d Ill:lli(1I1 nldll· hl':lrill~ 111;1ll'1'i;t! ..... L';ll1 1:1 hi..' pl:tl ill \\cll·lllbril..":lIl'd lk·;ll'ill~:-'. III :-\lll,\'j~d joinl Wear I~ill..·h lll..'l..lll' "j\t,.. \\l·;lr ill dh,,"V ","\pl..·rillh:l1h, 111'\\l'\<.·I". Ill~ld jllilH ..... IhL· ..:\I..'lil.·~d \:Iri:llillll i dlll"ill~" l)h\,i(ll()~il·:t1 11111,t :ll ...':111.-.1..':- I"l'pl'lili\l' ;lrti ...: ll!:tr \:;:It"lil;l~~l.' ~lr""~~il1~,~ 1ll:1IiullJ. III :u..ldi!ltlll. dUrill~~ rU!:llioll ;tllt.! ,lid ~pl.·\.·ilil..' I"l"giflll :\11ll (llil" 01 01 IItL· ;ll"ll,,:u!:tr :-'Url~H:l· Ill\,· !lJ;ld ... ,d l'olll;\i,:I ;\I·l':I. ·'1111 !\'pl" ~tl"l·:-...;ill~ 11l;ll aniL'u!:tr 1\'gilliL l,(';ld~ iIIlPl): :lrliL'uL,lr L"'lrlil~I~I..· a 1'1..' ~11PPIll"!I.."d 1ll;lIl"i,\ ~1lh.1 h.\· In' 1111..' l'ullag rL·",i~I;:1l1(l' gt,.·lllT~llcd IIll· Ill()\·l·IIII..'nl t1tl"(Jll~lllllll [Ili..' Ill;ll!"i\. '1'1111" joint Illtl\ ...'IlIL"111 ;IIHI ll';ldill~ \\ill 1..__ ~\Il"'L' SI1,,-',,:-.ing til" I hI..' ...;11Ii<l Ill:!l!"i\ 1) I\·p 1\' :IIH.II'\..'IW:I1l"d ""\lI and irllhihilioll (JI 1111..' lis:-'llL'·:-' illl\,"l".... lili;d llllid l\: ;\tl'shiall, Il)')/l. TIII.., ...;\" P1'l)l'l'~:"l'S ~'.i\'"," Ii...;\,.' ptl:-.:-.ihk· 1l11,(!l'llli:--llIs h.\ \dti\:h f:Jli~ll"'· d;:lln:I~ ,It,.·L·lllllll!:lll' ill ;lrlil..'ld;ll· 1..";ll'lil;l~l·: disnlpliot'! C(JlI'l~l·II-PG ~(llid 11l:llri\ :lIld PC '\\';\~h Uti!.·· First. !\"Pl'tili\\' ",'(luld d I:-ru pi ti'L· l·tdl:t~~l·ll-PC; Illalri\ \."(JII;I~L·1l Ii IlLT:-.. I hl" :-'t PC 1ll:ll l'l'lIk:-.. :11)(1'01' Iltl' illll'd;:Ii..·I..' hl'I\\n"l1 Ii\l·~l' 1\\1 PUllL·"t",.:\ PU\llt!:ll' Il-,polhl",is i.. . 111.,\ li~lIL' i ..... Ihl' l"L· . ..;ltll ld';l It.'lbiJ...· I':lihll\' lii~",'l" 111.,:1\\0,"1, (FrL'I..·lll;lll. t,.·;lI"ti\;, PI' I hI.' l·' 1075)" ;\Isu. :IS di:-l ~bove pronounced changes in the articular cani- (-~e-':l;G population have been obscn'cd with age :ric\i_lisease (Buckwalter et aI., 1985; iVluil~ 1983; \R<Juahlev et at. 1980; Sweet et aI., 1979). These PG ';. chal~ges "could be considered as part of the aCCUl11l1idrcd,tissue damage. These molcculaJ- stnlclural :-'~ha'nges would result in lower PG-PG irHcraction ;~ites and thus lower network strength (Mow et aI., 1989b; Zhu et aI., 1991, 1996). Second, repetitive :and massive exudation and imbibition of the intcr<~iitii:d fluid may cause the degradcd PGs to "wash , ollt~~;::from the ECM, with a resultant decrease in stff0e~s; and increase in permeability of the tissue tha('jp::lurn defeats the stress-shielding mechanism oUnierstitial fluid-load support and establishes a vidous/cycle of cartilage degeneration. A third mechanism of damage and resultant articular wear is associated with s~'novial joint impact loading-that is, the rapid application of a high l8"ad. \,vith normal physiological loading, articular cartilage undergoes surface compaction during the compression with the lubricating fluid being exuded lhi'ough t.his compacted region, as shown in Figure 3..:10::';\5 described above, however, fluid redistl'ibutiol}.',\vithin the articular cartilage occurs over time, which relieves the stress in this compacted region. Thi~'process of stress relaxation takes place quickly; the stress may decrease by 630/0 within 2 to 5 seconds (Ateshian et aI., 1998; ""lo\\' et aI., 1980). If, howevel~ loads arc supplied so quickly that there is insufficient time for internal Iluid redistribution to reHeve the compacted region, the high stresses prodliccd in the collagen-PG matrix may induce damage (Newberry et aI., 1997, Thompson et aI., 1991). This' phenomenon could well explain why Radin and Paul (1971) found dramatic articular cartilag damage with repeated impact loads. These mechanisms of wear and damage may b the cause of the commonly obser'ved large range structural defects observed in anicular canilag (BlIliollgh & Goodfellow, 1968; Meachim & Fergi 1975) (Fig. 3-25, A-C). One such defect is the spli ting of the cartilage surface. Ven ieal sections of ca tilage e:\hibiting these lesions, known as fibrillatio show that they eventually cxtend through thc fu depth of the articular cartilage. (n other specimcn the cartilage layer appears to be eroded rather tha split. This erosion is known as smooth-surfaced d structive thinning. Considering the variety of defects noted in arti ular cartilage, it is unlikely that a single wear mec anism is responsible for all of them. At any give site, the stress histor:v may be such that fatigue [he initiating failure mechanism. At another, the l brication conditions may be so unfavorable that i terfacial wear dominates the progression of cart lage failure. As yet. there is little experiment information on the type of defect produced by an given wear mechanism. Once the collagen-PG matrix of cartilage is di rupted, damage resulling froll1 any of the three we mechanisms mentioned becomes possible: (I) fu ther disruption of the collagen-PG matrix as a resu of repetitive matrix stressing; (2) an increase "washing out" of the PGs as a result of violent nu movement and thus impairment of articular car lage's interstitial Ollid load SUppOrl capacity; an (3) gross alteration of the normal load carria mechanism in cartilage, thus increasing friction shear loading on the articular surface. I I I Photomicrographs of vertical sections through the surface of articular cartilage showing a normal intact surface (A), an eroded articular surface (8), and a vertical split or fibrillation of the articular surface that will eventually extend through the full depth of the cartilage (C). Phocomicrographs provide chrough che courcesy of Dr. S. Akizuki. Nagano. Japan All these processes will accelerate the rale of interfacial and fatigue wear of the already disrupted cartilage microstructure. Hypotheses on the Biomechanics of Cartilage Degeneration ROLE OF BIOMECHANICAL FACTORS Articular cartilage has only a limited capacity for repair and regeneration, and if subjected lO an abnormal range of slress~s can quickly undergo total failure (Fig. 3-26). It has been hypothesized that failure progression relates to the following: (1) the magnitude of the imposed stresses; (2) the total numbe,· of sustained stress peaks; (3) 1I1C changes in the intrinsic molecular and microscopic structure of the collagen-PG Illalri.'\; and (4) the changes in the intrin- Cartilage Structure t h. Biochemical Composition t CaliagenlPG ECM Mechanical....- Joint Propenies Loading + Spatial and Temporal Slress-Slrain, Pressure and Fluid Fields 1 ChanreYleS ·41---- Cell Slimuli Synthetic Activities ~ Cartilage Function Physical Activities Flow diagram of the events mediating the structure and function of articular cartilage, Physical activities result in joint loads that are transmitted to the chondrocyte via the extracellular matrix (ECM). The chondrocyte varies its cellular activities in response to the mechano-electrochemical stimuli generated by loading of its environment. The etiology of osteoarthritis is unclear but may be traced to intrinsic changes to the chondrocyte or to an altered ECM (e.g" resulting from injury or gradual wear) that leads to abnormal chondrocyte stimuli and cell activities. ~i<: IlK'chanical prop<..Tt~· 01" the tissue. Tht: nH)SI im ponam I"ailure-inilialillg rac((1l' appears I(J he I "loost..'ning:" of tile c()lla~cn nctwork tllal al!c,ws a normal PC c.\pansiclIl and thus tissue :-,weJ[i (:\ilaroudas, 1976; McDL'vilt &: Muir, 1976). Asso alL'd willl Ihis Chi.U1gc is a dL'Cn:i.ISL' in canibg.1.: stil Ill:SS and an incrl.:ase in canilage pi..'rI1lL'abil (Allman el aI., 19S-l; Armstrong &. Mow, 198 Guilak L"l aI., 1994. SL"lton el aI., 1994). h,,'h which aller canilagL' fUllction ill ~\ diunhrodialjoi during joilll mOlion, as showll in Figure 3-27 (!vlo '" Atcshian, 1997). ThL' magnitude of the stress sustained b,\' lhe a ticular cartilage is dctennined b,\' both the toti.lllo on the joint and how tlwt kli.\d is distributed ()\ the i.\1·ticul~ir surface conl~lCt arca (Allmed 6.: Burk 19i'3; Annslrong L'I aI., 1979; Paul. 1976). :\n.\· h,:nsl..' stress COllcL'nlration in Ilk' l'ont~\Ct i.lrea w pla,\' i.1 pl·imi.llY roIL, in tissu<..' (k'gL'Ilt.'ration, A lar Ilumb<..'r or well-knowll condilions GillSi...' L'.\cL'ss stress conccnlrations in anicular canilage alld sult in caniJage failure. Most of Illese stress L'once II':Hiolls are C~ltISI,:d h.\· joint :,urfacL' irH..' ongruit.\', sulting in i.lll i.\hnornwtly small conti.lCt are E.xi.lmplcs or conditions causing stich joint inco gruilil.:s include 0:\ subscqllL'fl( to congeniwl a L"!i.lbular dysplasia, a slipped cnpilal femoral <.:pi .,"sis. and intra-articular fractures. Two fUrih eXi.\mpk-s arL' knet: joint l1lcnisct.:uoll1~·, \\'hich eli inak's the load·dislributing function of the lll~n ellS ([vlow Cl aI., 1992), and ligi.llllCnl I"llpllll"<"', whi allows cxccssi\·<..' mo\"t:'1lll'1lI and the gen<..'r;:llion abnormal mechanical ~trLSSl:S ill Ihe afk'clI:d jo (Allman el aI., 1984; Guilak Cl ai, 1994; ivkDe\"ilt Mui,·. 1976; Sellon el al.. 1994). In all the ,11)O cases, abnorllwl joint ~\rtj(-ulation increases t slJ'i..'SS acting 011 tilL' joint slIrf~lce, which aPlk'ars predispose Ihe caniiagL' to railllr<..'. iVIHcroscopiG\II~', stress IOGlJiz<.ltion and cOllce tration at the joint surfaces li~I\'L' i.\ I"url!ler dlt.'C High COntact pressures bctW('L'n Ihe anicular su faces decrease Ill<..' pr()babilit~· of fluid-fIlm lubric tion (Mow &. Aleshian, 1997). Subscqllcnl <.lelL surl'''lI.:e-lo-surracc contacl i.\Slx.'rities will cau microscopic stress concentrations that arc rt.'SI)(> ble for further tissue damage (Atcshian et aI., 199 1998; !\leshian '" Wan!'. 1995) (Case Study 3-1). The high incidencL" or speciiic joint dcgerlt'rati in indiddllals with certain occupmions. such football pl<.l~·L'rs' knccs and ballet danccrs' i.lnkl can bL' <,,'xplainL'd b\ tilL' incn:i.lsL' in high i.lI1d i.lbn Illi.llioad rn:.·qllcnt:~· and magnitude sllslainc:d b~' t joinls Ihese: individuals. II has becn suggL'slt or or •I Deformation , fluid QA Progression Normal Cartilage ..--. I exudal.on ~ PG Loss ColJagen Damage I Load . i.~yll. : ~ -::t_~+f (e.g., inlerleukin-I) (Ralcliffe el aI., 1986) a growth Factors (e.g .. transforming growth fact bela I) also appear lO pia\' an importanl role in O Another contributing factor to the etiology of ma~' be age-related changes to the chondroc (Case Study 3-2). I I I ,--'-- I I I I I ., _.1 _'--_ _--' IMPLICATIONS ON CHONDROCYTE FUNCTION The ECM modulates the transmission or joint lo to the chondroc,yte, acting as a transducer that c 1Fixed Charge Density - - ! Swelling Pressure ! Frictional Drag - t Hydraulic Permeability verts mechanical loading to a plethora or envir mental cues that mediate chondrocyte function. healthy articular cartilage. loads from norm joint function motion result in the generation Increased Matrix Deformarion Increased Fluid Flow Acts 10 diminish cartilage load-bearing properties. Knee Meniscectomy 0rty-year-Old man who had a meniscecwmy 10 years figure illustrating how osteoarthritic changes to the col'aqen·PG network can compromise the ability of articular ca'til'lOe to maintain interstitial fluid pressurization, which ""'~"'''o< the tissue's load-bearing and joint lubrication ca· Loss of PG and damage to the collagen fibers result an increased hydraulic permeability (decreased resis- !'c!"""e to fluid flow) and supra-normal loads and strains on solid matrix (and chondrocyte). F ago in his right knee. Currently, he is sufiering pain as- sociated with movement, swelling, and lirnilaiions of knee motion (Fig. 3-1-1), The history of knee meniscectomy not only implies an alteration in joint surface congruence but also the elimi· nation of the load-distribution function of the meniscus. The effect is an abnormal joint. characterized by an increase in the stress acting on the joint surface that results in cartilage failure. Most of these stress concentralions are ;~ ~.;,f' thiLin some cases, OA mav be caused by deficicn'<eie'~3~' the mechanisms th~t act to Il1ini~llizc peak '~'ff6rc,~s~on the joints. Examples of these mechanisms ~jh~~l~(le the ~clive processes of joint nexion and _~:~,q~e, lengthening and the passive absol"ption of ':;k,,,by the subehondral bone (Radin, 1976) and ciscus (Mow el al" 1992), ~generativc changes to the structure and com~gion of articular cartilage could lead to abnorkti:ssue swelling and funZtionally inferior biome~ ;;,iji8<11 properties. In this weakened state, the '~-,:J~ge ultrastruclure will then be gradually de.-g.t~d by stresses of normal joinl articulalion (Fig. ~m)j7;OAmay also arise sec~ndarily from insult ~o :' ':- ~~i~tIinsic molecular and microscopic structure of /;f.~tfi~2so;qagen-PG matrix. Manv conditions may pro:\-.Iil<?!~:~,l!ch a brea~down in .~latrix .integrity;· the~e -;111~e degeneratIon assoclatcd wllh rheumalOId --·.1isJ joint space hemorrhage associated with 'p.pilia. various collagen metabolism disorders. }~S§.tl.e degradation by proteolytic enzymes. The _,'l1C: c ': of soluble mediators such as cytokincs caused byjoint surface incongruity, resulting in an abnormally sma!tconraet area. This small contact area will suffer hign.c9_ntact pressure. decreasing the probability of fluid-film lubrication, and thus the actual surface~lo~ surface contact will cause microscopic stress concentrations that lead to damage. Case Study Figure 3-1-1, ~------_ .._-----~ Osteoarthritis eventy.year-OJd woman, overweight, with OA of the S right hip joint with associated symptoms of pain, limi- tation of motion. joint deformity, and abnormal gait (Fig. 3-2-1). OA is characterized by erosive cartilage lesions, carti- lage loss and destruction, subchondral bone sclerosis and cysts, and large osteophyte formation at the margins of tho joint (Mow & Ratcliff., t997). In this case. roentgenograms of the right hip of the patient show a decrease in the interarticular space and changes in bone surfaces as sclerolic and osteophyte formations. nle most severe alterations are found at the point of maximum pressure against the opposing cartil(1ge surface, in this case at the superior aspect of the femoral head. ... t1I........ 10\\ j1l..'l'Illt.,<thilit.\·, thl..· n(lnll~tI em il"UIlIIll..·IH ell illl..· ChllJJdrlil:.\·!I..· i df)lIlill~lil..'d h.\· h.nlrli:-'I~ltic 1'1'1.......:-'111"1..' ill the illtL·r tili~d lluid. \'~tl·i(jth phl..'lI(lllIL·II~ origil1;llitl2:-' IrOll1 inIL'I"slili~t111uid Ilo\\' l',\ist ~IS \\l·11. llJ)plil'~\kd ill I..'nh;lIll..·ill~ Illitril...'llt dilfliSioll, illlL'!" :-.{iti~d lIuid 11<1\\- (i.I.,:.. unhoulld \\'~tlL'rJ ~i\'L'~ ri"L' 10 l'1..'J111I~\1' ....lillHlii Ill' ~lll 1..·ll..'cll'k,d 1l~lllll"L', n~lnwl.\' ... 1l"l'~Hllillg p(ItL"nti~d~ and (,.·UITL"JllS {Fr~lllk Gr()t!l.insk\-. 11.)~7: Gil L't ~J1. 190.). 1t)t)t'). III ,Iddi~ lioll, illk'rslili~d lluid 110\\' tlJr()lI~h tilL' "mi.lll p0l"es a.... :-.(J(..·i~·lk·d with tilt.' ....olid Il1;'HI·j\ (-?iO IlIll) or Ilol"lna \:al'lila~l..', \\hkh olll'r 1..·(Ilbi<.IL-rahk rl..·si:-'l~lIKl· ti Jluid 111)\\· (:\lanlud;'I:--, I ':lit): ,\kCtltchL'Il, 1<)6]: \tluw 1...[ ~d .. ll..)~..j.). \:~1I1 2:-'i\'L" l"iSL' [0 ~\ Jlll'l'h~\llk;l1 phl'Jl()1I1 1,.'lla tl..'l"lllL'd Iluid-indlll'l'd 1\1~\Il"i.\ compadion (Lli I.\. \-10\\·, 1'J~OI. ThL' frktiollal illk'l"~lcli(Jn hL"l\\-t.·\,.'1l inkl"slili,d Illiid and sldid ~\rl' <l l"L·....ull drag l"L'"ist~lllL·L· to r(JI·I..·I..·d i1U\\ Ihrullgh Ihl.." pllroll . . ·lkTlIll..·ahk ",::ll"lila~l' 1I1~\1I"i,\ ~\Ild ~l \'iSCllll." sl'IL'~Il" sll'l'ss L'.\l·l·ll'd hy IhL' intl..'r... tiliailluid. Gi\'(,.·ll tlll" nominal !lo\\" l'aks 01 lhl.' ilHt.'l":-.tilial i111id llh..·IUiulll.."d I..·adil..'r ~1IH.1 Ih. 10\\· pl·rrl1l..'~\hilil\· "I lIll..' l';ll"lil;qll' Irl'lll'i\. cholldruC.\ II..' pl'l'ct.·ptilll1 of thi .... frictional illh-'ra(,.·,itlll f(Jl"i..·l· is likL'h to Ik' dUlninalL'd \1\' thl..' dl"~\g rL'si:'oI~IIlCL' !lO\\ Ihl"()lJ~h [III..' IlJ~\lri\ l"allll'I' [ll~\Jl h.\ diJ'l·\.."[ \,i:,c()\.IS ~hl..'al· :'trL'S~ l,Jl I hI..' cL·II. This friL"tjonal dr~lg forc(. 1..:<111 pr<'(\UL''''' ....u!id 11\;,\[,.;\ I..krOnll~t(i(1I1 011 1111..' urde (II' 15 10 .'(V-" Frum tilL' disctls:-.ioll ahO\'L', i..'l'lolldI'Ol·.\-IL' (1...'1'01" m~I[i()Jl L>~\1l hI..' UlJJ~i<.lI,'I"L·d 10 hI..' gO\·lTI1L·d hy [hrL'e cuupkd loadjn~ Illl..'l'hallislll~: dirl..'ct I-:C.\;, <.IL-for tl1~llioll: Ho\\··;ndllL·l·d L'(ll1lp~\ctiotl: and fluid prl'S* sllri/'llilln. III OJ\. I hI..' iIlI..T(,.·asl..·d lis~lll..· pt.·nlll·'lhil~ il.\" diminis!k· ... carlibgL"~ !lormal flu;d pr..."~~lIre Itl~I<.I·SllPP()I"[ IllL'L·l1ani."lll. Thus, [i'lL'!"...' i." a shirt or !O,ld support Ollto thl..' solid matrix. I..·i.lusing sllprallol"nwl ,slrl..·S::>L·:' and s!l·ailb 10 hI.." illlposl.'d 011 1111..' chondniL·.\·!I..'S (Fig. 3-271. Th..:·si.' ahnormally high Stl'L'SS ~lnd . . tl'~lill (('\-C Is , and utili..'! rn(",,,,'k\llo·l'iL'I.:II'Ol"Jh'llli(,.·~\ll·h~lll!:!I...'S [Iw[ ~\rL" nwnik~l('d \\'itll 0;\, C=ill triggl..·1' an irl1hal~IIKI..' of <.:IHIll <',,"01..'.\'1\.' an;:lholic and I..·alaholic al'li\'jtk"s, further l,.'Ollll"ibuling In ~l \'icious c.n·1t.- Ill' progr...'s~i\"i.: Glr!ilagl" dl.."gl.."nl..·r~lli(lll. Itllk...' d, changes to tht.' bin L'hL'lnkal l:Olllpositioll ~llld s[l"lIclUI"l,: oj' cal'til~\g Ctlll Iwu..· a profound inlp~lcl Oil tiSSlll..· and (,.'!lOIl d n)cy{t.' fUlll'1 ion. \\'i t h IntI! t iLl i:,dpli nary collaho* r~ltioll:' =md ~1Il appropri::lli..' thl·Ol"I..'lil"=-d rl'~IIllC* \\,(H-k, such as th ...' biphas:ic tl1 ....<I1·.\·, iJlsight~ illW 'Ii'lL' factors that g()\·1..'1'1l c!londnlC\[<.' f"llIKlion L'arlil<lgL' s[ruclllr...' <lnd l'UI'H,:lioll, ~\IH.l lhc diology or 0:-\ I..·an hI..' obtain..:d, or « or or Case Study Figure 3·2·1. mcchano-e1ectrochemical stimuli (e.g,. hydrostatic pressure, SlI-ess and strain fields, streaming potenti~t1s) that promote normal caI·lilage maintenance (by the chondroeytes) and normal tissue function (Fig, 3-26). However, when the integrity of the collagen-PC network (the transducer) of articular cm'tilage is compromised. such as from trauma or disease. normal joinl articulation leads to abnonllal Illcchano-clcctrochemical stimuli, with ensuing abnormal ECM remodeling by the chondrocytes and debilitated tissue function. In the absence of joint loading, the normal environment of the chondrocyte is characterized by the pre-stress established by the balance between tension in the collagen flbers and the Donnan osmotic pressure. During joint loading. by vit·tue or the lis- , ~.,t,--. ....'.~ ., .. " REFERENCES SumrlzalY The function of anicular cartilage in clianhrojoints is to increase the area of load distribution ;'.(thereby reducing the stress) and provide a smooth, (\,vear-resistant bearing surface. i;~(2) Biomechanically, articular cartilage should !J~~viewed as a multiphasic material. In tcrrns or a }biphasic material. articular cartilage is comprised ,i' porous-permeable collagen-PG solid matrix (approximately 25% by wei weight) filled by the fl'e'~ly, movable interstitial fluid (approximately i5r~',:bY wet weight). In addition to solid and Fluid :tIlel,~e', (~xists an additional ion phase when CO!1si~f~r'ing articular cartilage as a triphasic medium. /;:The ion phase is nccessa'"y to describe the swcl>~cling and other electromechanical behaviors of the .of :'~~/tissue. 'lH~:-'{~' lmporlanl biomcchanical properties of articular ;:',c~rlilage are the intrinsic material properties of the soli~1 matri:\ and the frictional "csistance to the flow 9f/':,~nterstitial fluid through the porous-permeable :spHcl,';',JTlatrix (n parameter inversely proportional to ,~h8i+issue permeability). Together, -these parameters ,;~;~,~efine the level of interstitial fluid pressurization, a :"'-};£6ajor detenninant of the load-bearing and lubrica/iZ.J!ibn capacity of the tissue, which can be generated in )s:cartilage. '~,ti!i-~t;Damage to articular cartilage, from whatcver /~%g~use. can disrupt the normal interstitial fluid loadf;"9~~ring capacity of the tissue and thus the normal "IH§n,fation process operating within the joint. ~n~tTf()re, lubrication insufficiency may be a pd:m~~;y':(aclor in the etiology of OA. /i?~@;,- :.' ;"G~hr~J'Nhen describing articular cartilage in the conframework such as the J~.Jjlphasic. triphasic. or multiphasic theories. it is ':ltnqssible 10 accurately predict Ihe biomechanical be?:~}i~viors of anicular cartilage under loading and to W~1,~!:;lJ~idate the underlying mechanisms that govern ';r1M,01oad-bearing and lubrication function. Further'~?,r~',:')nsights into the temporal and spatial nature '?Rr.\p·?'phy;ical stimuli that may affect chondrocyte ·;flinction in situ can be gained. ,:D'Jexl ()f a rigorous theOl-etical ,. (~j'r!;: - 4?£1.CKNOWLEDGMENTS (~~'i::;';"-' -f~~O;:his'wark \vas sponsored bv the atianal [nstitutes }iitiR+lealth grants AR41913 ,{nd AR42850. ,:~:<ii~BX~(/ ~. ;;,::/';:;<,-, !\hllh:d, i\"\1.. (\: Burkt.:, D.L. (1983). In vilro mcnSliremen st:llil: pressure distribution in synovial joints-P,lrl 1: i:1I surface of the knee. J Biolllech Ell.!!,. 105, 216. Akizuki. 5., \lo\\'. "-,lulieI'. F.. (.'t a1. (1986). Tensih.: p "nies of knel." joint cartilage: 1. Influence of ionic: co tion. wcighl bearing. and fibrillalion on the tCllsill." mo IllS. J Orr/lOp U/.:.\', 4. 379. Ahm~ln, R.D., Tt:nl'baum, J., Lilla, L. l'l "I. (1984). Bio Ch~1I1ic.:l1 <lnd bioc:ht:mical propenh:s of dog cartilage in peri mentally induced oSlcoarthritis. All/I Rhl!/lI11 Dis. 43, :\ndri~u.:chi, T.P.. Natar~lj'tn, R.N., ~ Hurwitz. D.E. (19 Musculoskclc·t~d dyn,llllics, locomolion, anti dink'll ~l l.:;llioll. In V.C. Mo\\' & \V.e. H<'lyl'S (Eds.). Bw;;c I!wpw:dic BiViIlt.'dlllllics (2nd l."d.) (pp. 31-68). Phila phia: Lippincott-RHvcn. Armstrong, CG .. Balmllli, A.S., ~\: Bardner. 0.1.. (1979) \'itl'o measurement of articular c;lrlilage deformation lh~ intacl human hip joinl und"r load. J B(m" loi/a S 61,1.744. :\rmslrong. CG., & Mow, V_e. (J980). Friclion, lubricil1ion w,,'ar or syno\'i,al join Is, III Owen, R.. Goodf...!low, J. (\: lough, P. (Eds.). Sci~'IlIiric FOlll1dt/tions or OrtJ/(JI'l/~'tlic:s 7iwf/lllUO/Ogy (pp 213-232). London: William 1-ldncnn<'H ArmSIr<)IH!. e.G., & l\.fo\\'. V.e. (1982). Varialions in dll." Irinsic-mcchank<.tl propt:r1ies of hum'lll anil:ulrtr ci.lni Wilh agL', degencr;ltion. and w:lI..:'r COrllenl. } 8(111~' J v.e., Sltff::, 64..1, 88. Ateshian, G.A. (1997'). Theoretical formulation for bound rrktion in articular cartilage. J !3iOll/cc), ElIg, 119, Sl. :\LL'shi~tn, G.A., K\\'~tk, S.D.. Soslowsk.\', LJ., et al. (199n..:'w sll:n:ophotogr:lInmctry lll..:'lhod for dctermining situ conlact are;\s in dianhrodinl joinls: A compar sludy. J Biol1/~'dlallics, 2i, III. Atc:shian. G.A., Lai, W.M., Zllli. \V.B., (:t a!' (1995). An :lsy totic solution for the contact of twO biphasic c~nibgc \.'1':;. J Biol1lccJulIIics. 27, 1301 i. Ateshi:lll, G.A., & W'\Ilg. H. (1995). ;\ thl'oH:tical solution til(.' friclionless rolling contact of cylindrical biphasic ticular cartibge layers. J BiolIlCC/HlIlics, 28,1341. Att:'shian, G.A., \Vallg, H., & Lai. W,M. (1998). Thl..' roll.' o tcr::;titi,d fluid in pressurization ilnd surface porositie the boundary friction of articular cartilage. 1 hi};ol 120.241. .-\Ieshian. G.:\., Warden. W.H., Kim. J.J .. d al. (1997). Fi ("-,formation biplwsic malerial properties of bovine ,Ir !;Ir cartilage from confined compr..:ssion experimellt BiulI/t.'c/UlIlics. 3D, 1157. Athan<lsioll, K.:\ .• Ros('llwasser, i\·I.P., Buckwallt.'r. J.:\., e (1991). Inlerspecies comparison of in sill! Illcdmn propl."rlies or diswl rcrnor,ll cartil:tgc. J art/lOp Ik.., 9, Bachrach, N.M., VlIlhmu, We., Stazzonc, E.J., et 'II. (19 Chang~s in protcoglycan synthesis rutes of chondroc in articular cartilage arc associ~lted with the time de dCllt changes in lilt: lllC'chanie<d environment. 1 13i ch(lIlic~, 28, 1561. Bachr'Kh. j\!'~C Mow. \i.C, & Guilak. F. (1998). Incompr ibility or the solid matrix of articular cartilage under hydrostatic pressure. J BioJ11£'cluwics, 3/, 4015. Bateman, J.E, Lamanue, S.R., & R,II11SIw,w, J.A.~'1. (1996). lagen sllpcrbmily. In W.D~ Comper (Ed.). Extracellular Irix (Vol. 2. p. 226i). :\lIlsll'rdam: Harwood Academil: P Bollet. A.1., &: Nance. J.L. (1963). Bio<:hcmical findillg~ in normal and oSh:oarthritic <:tnit:ubr canibgc, II: Chon· droitin suJf~\IC concl.'lltralion and ch'lin length. and W;:lt('l' ,.\[1<1 ash content. J Clill l'II't'St, oJ). 11 iO. Bowden, F.P.. & T~lbor. D. (1967). rfiction /llId Lllbrictlli()J/. London: \lC1hucn Pubs. Broom, N. D" &. Sil~·n-R()bens. H. (1990). Collagl'n·collagcn \'crsus collagen.prolCoglyc:w interaclions in the de!l:rmin;:ltion of c.lt·til~lgc stn..'ngth . .·Irthrilis 101i.'1II11, 33, 13[2. Buckwa!l\:'r, J.,\., KucHner, K,E., &: Thonar. E.J.-M.:\. (1985), ,\g('-rl"l<lted changl's in articuhlr c:lnilage proleoglycans: Eleclron Illicroscopit: sllIdics. J Orr/lOt' Re:., 3, 251. Bullough, P.G .. &: Goodf,.'!low, J. (1968). The significance of th{' fillc structures of ~lrticllbr cartilage, J BOllI.' Join/ Surg 50B, 852. Bullough, P.G .. &. Jag,-Illnalh, A. (1983). The lI\orphoJog~' of the calcific'ltion front in articul:lr l.:anilagc . .I Bouc Joiut SlIrg, 65B. 72. G:trdlll·r. 5.1.. ...\: .\-kGjllin;ty. D.C. (1971). Li\"ing :lrticul:I l';lrtilagl' i:- nOI SlIllll__llh. Till..' :-trth,:tl1l"l' (,t 1ll:lIll!ll;di;lll an a\"i;lll joint :-Ilrbl.:l':- dCllHJIl:·dr:lll..'d in \'i\"o b~ inlflli:r~i(l indd('llt light miu(Js("op,\". ..1111/ Rhl:/1l1l Ois, 30, 3. Garg.II.G., & Swann. n.:\. (1981). :\g,:-rdatct! ch:tng,·s lllthl chl..'lllicd l:omposition (Jf h(I\'inl' :Inil'ld~\r ,;;anilagt Hiot.:ht'lll J. 193, ·!.';9. Gll, \\'.Y., Lai. \\'.\1. l..\: .\11)\\", v.e. !,1993J. Tr,-lll:-pon of fILl Buschm'llln. l\I.D .. Gluzband. Y.,\., Grodzinsky, A.1., L't :\1. (1992). Chondrocylt'S in agnrosl' culture s~'nthcsize a Illechanic,lIly functional i.'xtracdltll;'lr m:ltrix . .I Or/hop Res. 10. 745. Buschmann, :\-1.0.. &: Grodzinsk...... '\.J. (1995). A lIlolc~lllar modd of protL"()gl . . . can-nssoci~\{cd ckctrostatk forcl.'s ill l,;arti!:l~(' mechanics. } BhlllH'ch Ell!!" 117, I iO. Clark, J.~,i.·(1985). The organization ~f coll:lgl'll in cr.. . ofr;lctun:d rabbit ~niclllar cartilage: A s'::lnning l'll'.:troll microscopy stud...... J Or/hop lk~, 3. li_ Clarke. I.e. (19il). Articldar cartilage: A re\'ie\\, and scanning l']cctron microscope study-I. The inll'r1crritorial fibrillar '.l!"chitccltIl'c. J BClIIl' Joim Surf;. 538. 732. Donohue. J.M .. Buss, D.. Oegcma, T.R .. (.'t al. (1983). The dfects of indirect blunt trauma 011 adult canine articular Gujbk. F.. R:lIcliffL' . .-\ .. L:tlll·. :\ .....·1 ;t!. (1994)..\kch:1I1ic and bindlcmil';d l·hallgn ill lhl' Slllh·rlil.:ial ZUlll..' of ;-\rtil..· 1;11' canil:lgl' in a ~'~lnil1(' Illo(kl or nSll..'o:lrthrilis. J Orr/w Rt's. 12, .ti-L flardillgh:llll. T. .\1.. l..\: .\Iull', H. (197-+1. IlYllluroni( :H.'id ill ca lilagl' and prokogl~'C:1l1 :lggn..'~~l\ion. Hiu(·/tt·//t.l. 1.39. 56S 1·lardilJ.~dlall1. T.E .. !.h:ardrn(Jrl'·G;lrt~. \1., ~ Dunklill. D.G (llJ8iL Pn,Il'in d(lll"lain Stl"lll"llll"e of Ih ...· aggrq~~l!illg pl'O teogl.\T;11l frolllc\rti1:lgl·. Tnm., On/t{J!, Rt'." ,)0('.11,61. Ibsc;dl. V.C. (l<Jill. 11I11..·r~lC(i(JllS (If l'anilag...· pro(('oglycitl with h,\';duronil' acid . .1 S/lpl"iul!l!i SII'I!ClIll"t·. 7. 101. Ib~·l·':-. W.c.. & Bodilh', :\.1. (1978). Flo\\'-indi..'peIHknt \'i; cOl:l:!stir propl·nil..'s of articul;l], <.:ani!:lgl' m;\lri.". J Bitt/IIt c/ulIIics. II. 407, 1[;1.\·('S. W.C .. l..\: \Iockros, I...F. ([971 Oi.Viscoi..'!astic propl'rlii. of human anit:ubr canil;lgl'.) :\PI'I Phy.~i(Jj, 31.562_ 1I~I\\'killgs. S.\\'. (19SSI.:\ Brid l1ililor.,· of Tillie.' Frolll tilt: Bi Ba/l~ 10 l3!ack 110/n. Xl ·\\· York: B:IIlI<lm Book::.. Hcinl'gard. 0 .. \\'it:slall(kr. J,. Shl'dlan. J., l'l ;d. (1983). S(' :tration ;'tlld . . haral·ll.'riz,llinn (If 1\\'0 poplll;ltions of ;1g.gr g:lllng proll..'ogl~"l';ln:- from i.':lnib~l'. Bif,d/"III J. 215. 95, Hdminl'n. H.1,. Ki\'ir:llll;l, J.. T;lI11l1li . .\1.. L'l :.1. (Eds.) (19~i ./oil// l.orldill,!.:: Biology (lild 11(.'111111 (>1" ..\I"/iclilal' Sll"/Iel((1"" Bristal, V.K: Wright &: Sons. Pubs. lIill:-. B.A. (1989), Oligolallll'llar luhri\:alion of join\:- hy su fan' ~H:li\'c phnsphlJlipid. J Rlll'lIlJI. I-b. S2-91. Hirsch. C, (19-+4). TIll' p;lIhngL'!ll'sis of l'llOndronwlacia of 11 p~lli..'lb . .-\0(1 Chi)" Sl.'{/!ul. 83 lSuppl), I. 111:1\·a<:l,k. :\-1. (1995). Till' rnll' of s~--n('\'ial fluid fillr;llioll h l·~lI"lilag ..., in luhril':llion of s~-Il(J\"i:1I join Is: IV. Squl.'l.'1....· .fil c;\l·tibge. J BOlle JoiJl/ SlIrg. 65A. 948. Donnan, F.G. (1924). The theory of membrane equilibria. Ch.:mical RCl';C\\', J, 73. Dowson, D. (1966/1967). Modes of lubricalion in human joints. Proc IlIs{ Mt!ch ElIg, IBIJ. 45. Dowson, D, (1990). Bio·tribology of natural and rcp!accIlH:nt joints. In V,C. :"'Io\\', A. Ralcliffe, S-L.Y. Woo (Eds.l Biollll.'('''(/Ilks f)ial'thmdia! .foill/S (pp. 305-345). New York: or Spri ng~r- Verlag. Edwards. J. (196i), PhysiGII charach:risti(:s of <lnicubr canilage. Proi' IIl.H :\ft:dl Ellg, ISIl. 16. Elmorc, S.~·I., Sokoloff, L., ~orris, G.. et al. (1963L Nature of "imperfect" clasticity of anicLll'lr CarLil~lgc. J :\pplid Physiol, 18. 393. Eyr('. D.R. (1980). Collagl'll: Molecular di\'ersity in lhe body's prol('in scaffold. Scicl/n:, 207, 131 S. Fos~Hlg, :\.1., & Hardingh:lln. T.E. (1996). ;\\;.Hrix protcoglycans, In W.O. Comper (Ed.) Ex/racdlulur .\lll/rix (Vol. 2. pp. 200-229). Amsl(:rd~II11: Harwood ;\c;ldcmic Pubs. Frank. E.H .. &: Grodzinsky. A.J. (198ia). Cartilage dcclromechanics-I. Ekclrokint:tic transduction nnd dfects of pH .1lHI ionic strength. J Biol1lec!ulIlics, 30. 615_ Frank, E.H., & Grodzinsky, A.J. (1987b). Cartilnge electrollll'chnnics-ii. " continuulll model of l·;:u·tilagc declrokinetics and corn:l<:ltion with L'xperiments. J BiOlltt:chol1ics. 20, 629. Freeman. M.A.R. (19i5). The fatiguc c;lrtilage in the pathogenesis of ostcoanhro$is. :\cw OrtllOfJ Scal/d. 46, 313. Fung, Y.C. (1981). Ouasi-Iinear \'iscch:lasticity oj" sol'ttisslles. In Biolllcdlllllics: JleclulI1;cal Properlics of Lil'illg Tissues (p, 226). New York: Springl'r·Veriag. or and ion:' lhrough ,\ POr<HIS-Pl'l"llll..·;lbll' ch;lrgl·d·h.nlrall ti:-SliC. and slrl'~llnilig p(lll'llti;1! d'lla on nnrllwl h()\"inc ;1 licltlar Glnilagl..·. .Il3iol//<,clllll/;("S, 1r;. iU9. Cll, "·,Y.. La;' W..\1. . .:i.: .'1 ow. V.c. (1997). ,\ triph:l"ic allaly:-. or l1l'g,lli\'l: Olillloti<.· flo\\':" lhrough chMgL'd hydr,lll'd so li:-SllCS. J BiOlIl,·c!Wllic.....W. i I. CII. \\'.Y., Lti. \\'..\I.,..\: .\Io\\". \'.c. {199S).:\ llli\tllrL' thl·ory { dl'trgl..'d h.nll';llCd sott lisSl!l..·s (ont;lil\ill~ lllultl-eIN trolYll'S: PaSSI\'l' tr:lllspon and s\\'dling lll'h,,\·iors. J J3i I1ICc!t EII~, 102. 169. lubrkatlo11 for :t.\ial ;,.nnllll·lr'.\ Linder high !o;,ding cond tiOIl~,.1 BioIlH·C!z(/uic.', 28. 1199. Hodge. \\'...\ .. Fij:lll. R.S .. Clrlson,. 1\. .. 1.'1 al. (1%61. Conl.l prl':-iSllrl' in till' human hip joinl Illl·~I:-illrl'd in yi\·o. Pro Na/f :lclIe! Sci. US.--\. 83. 28i9. Holmcs, ~1.H .. L.li, W.M .. l\: !\lo\\". V.C, (1985). Singular p~' tllrb:ition ~Inalysis (Ill the nonlinear. f10\\".dqll'lldenl, COI prL'ssi\'c slrl'ss-rl'laxalion bdl;wior of ~1I'lkul;lr (~lnihlgL'. Biollll'ch I.:I/~, Wi. 206. HOll. J.S., .\'Io\\". V.C .. L:ti. \\".\1 .. ('I al. (1992). :\n :In;d.\'sis till' SqIH.'l'i'.L'-fillll lubrication llll'ch~llllsm for :1I·tleul:H· c~l tilagl·. J Riollll'dul1Iic.... 15, 147. Huhkr;llllZ. W. (I S9S). L'dwr dil' Spaltrichlllngl'll lkr G lellkknorp..::J. \'~~l"lliIlldl/lll~ell dl'l' :llIrI/olI/iscln'lI C:;csdlsc!za 12. 1-+8. ~ I;~;::~~~{:;IA~. :"i &ill Curran. P.F. (1975). NOI1t.'qllilibriIUJI ThaBiophyjics (4th l.'d.L Camhridgl': Han'anl Prl.'ss. G.£.. Tuke, M_A .• Dingle. J.T.. ct al. (1976). The dof proteolytic enZYllll's on thl.' Illl"chanical propt:.'l"ties hunHlfl 'll·ticuIOlr canilagl.'. Biocltl!/IJ 13iopll,\'... :\cra, I. G.E. (1979). Mechanil.:i.ll propl.'rtics of <Inicular In ~'I.A.R. Frl.'cman (E,L), .4t/lIlt :\aicillar C(lai· cd .. pr. 333--t.l-1). Tunbridge Wells. U.K.: Pitman v.J., Sah. R.I..., Grodzinsky. A.J .. ct al. (1994). :Vh:chani1'C!;ulalion of cartil;lge biosyllthetic b\:h~l\'ior: Ph~'sical Arch Biocln'/It Biophy,,,', 311, 1. Gu, W,Y.. & Mo\\', \!.e. (1998). On the conditional <Qui"""",,·" of chemic:11 loading and ml'ch;lllical loading articular C;lrt ilagc. 1 Biol/1('c:/wllics 3 I (111. 1181-11 $5. W.:'\'I., & Mow, \I.e. (1978). Ultrafiltr,ltiol1 of synodal '';'?:'"'''' by canilagc_ J Ellg .tlt:clt Dil· ..I8CI":.:. /0./. 79. & Mo\\', r.e. (1980). Dnlg-induccd t'ompression of ,,''''ulal' cartil;\gt: during a permeation c.xpl,'rimenl. J Bio17. III. W.M .. I-lOll, J.S .. &. Mo\\', V.C. (1991). A triphasic theory the s\\'dling .lIld deformation beh~l\'iors of anicular canil'a.". 1 Bicn/It'cll EII~. 1/3. 2-t.5. & Saha. S. (1979). Cemenl line motion in bone. Sci- 204. 50 I. ,\'t., & \-Vt~iss, C. (1975). Rl'\'iew of nnicubr t·;\rtil;lgc resl.'arch. Anhrilis Rhett/II, 18. 553. (1968). Lubrication of animal joints: l. The mechJ BilJllli'Challics. ,. 193. EC.. is: Radin. E.L. (1968). Lubric<llioll of :-lnimal joints; III. The eHcc! of certain chtmic~\1 altnations of thl' cart iIag~ rind lubricant. .. \f/hi/is Rlh·UII/. J 1. 674. F.C.. & Sokoloff. L. (1965). MO\·l.'IlH:nt and composition of inlt:'rslitial fluid of cartilage. Arthritis Rhc.'ll!ll, S, 481. H.. Ethcredgl', R.. &. Glimcher, M.J. (1975). In vitro of articular c~ll"Iilage. I: I-Iydl'o'\~<prolinc, h(,'.'\and amino acid composition or bO\'inl: i.\rticular c<l.nibge as It function of depth from thl.' surface: hydroxyproline cont"nl of the lubric;II11 and the weur debris as a 1ll~;lsure of wear. 1 B01l1! 10i1l1 Slfrg. 5i..t. 527. Lipshiti'., H., Etheredge, R.. is: Glimcher, ~1.J. (1976). Changes ill the he.'\os:lJlllne contt:nt and swelling ratio of articular (:'Il·tibge as r11l1~lioll:; of depth from Ih" Sur!";,lC\..'. J BOlle loilll Surg. 58..\. 11-t.9. Lipshitl.. 1-1., &. Glimchcr. \1.J. (1979). In \'itro stud it'S of the wenr of articular cartila!.!e. 1I'cm: 52, 297. Malcolm. L.l. (1976) . ..t/l ~t'xpt:rill1ollal illl't':,tigatioll of 1111: {i'jet/vllal al/{l d,,{or1lltlliOlUl! I'espOIIScs o( or/iclllar cartilage illwrt,u:es 10 .wl/ic (lI/(/ dYl/(lIllic loadillg. Doctoral thesis. Universitv of California. San Diego. :\lankin. 1-1.:\., & Thrasher, A.Z. (1975). Water conlent and bindin~ in normal and osteonrthritic human cartil:-lgc. J BOlle J~i/II Sun.:, 5i:\, i6. Mansour. J.,\'I..& ~to\\", v.e. (19i6). The permeability or articu!;\r canilag..: ulU!l'r comprt'ssin:' str,lin lind at high pressures. J BOlle lohn SlIl',!;, 58;\, 509. :\laroudas, A. (196611967). Hyaluronl( acid films. Proc IllS! Jkch Ell!;. 1.011£1011,1811. 122:. Maroudas, A. (1968). Physicochl.'llIical properties of c~\l·tilage in Ii£ht of ion-exchangL' theory. Biophy,'; J. S. 575. Maroud'IS. A. (1975.1. Biophysical chcmislr~' of cartilagin ti~Slh:S with sllI.:cial refen::l1cC' 10 solute ~ll\d fluid transp Biorhco!ogy. 12,233. "taroud~ls. ..\. lI976). 8;1blll.:e bC(\\'l.'l.'1l swelling pn:ssun: coll~lg"n tension in n(lrl1l~a1 alld ,k·gelll.'ratc c3nilage. IIItc.', 260. 80$. i\larou,!:Ls, A. (1979). Ph~'sicochcmkal propenies of anicu cHrtilnge.ln M.A.R. Frl.'C'man (Ed.). Ildlill ,.\riiCl1ltu· Ca la;:c: (2nd cd .. pro 215-290). Tunhridge Wdls. EnghllHI: mall '\h:dical. :\broud:\s, A.. \bchteL L, Grushkn, G.. \..'t al. (1991). Th\.. fecI of osmotil.: ~llld Ilh:ch;wical preSSllrl:S Oil Wi.ller pa tioning in ~lJ'lit-lIl<lr (,:Hrtibgl·. Biocltt'l1I Biophys .. 10i3. 285. \lcCutchcn, C.\\'. (19621. The frictional propi..·nies of :Ini joints. it-'eM.5, 1. ,\kDevitt, C.A .. Xluir. H. (1976). Biochemical chall~e th,' cartilage of ill(' knce in eXpl.'rilll('nt~d <llld lIatul;d h:o'lI·(hritis in Iht: do~. j BOIlt! 10ill/ S/lr~. 588. 9.J. . .\-kachilll. G., & Ft:rgie. I~..\. (1975). Morph~logic~d pattern articular cartilage fibrillation. J Pm/lal, 115, 231. \'\(1\\', V.C .. AlllO~zky, S.P., &. Jadson, D.\V. {1992L K Jlt'lIis('/ls: Basic ami Clillical F(J/(II(huiol/.~. N(,'\\, 'Y Ra\'en Press. ,\10\\', V.e.. & Ateshian, G...\. (1997). Lnbrit~llion and \\"(,';1 dianhrodial joints. [n V.e. Mo\\' &. W.e. Ha~'t's (Eds.), sic BiOUlt'c/wuics (2nd l'd .. pp. 175-315). Phibddph Lippincot t-R~l\"l'n. \lo\\', v.e" Atcshi'Hl, G.A .. Lai, W.M .. i..'l a1. (1998). Effl.'ct fixed charges on the sll'ess-r(,'laXlltion bdl:l\"ior of h~'dra soft lissues in a confint'd compression pl'Obkm. /I Solid... & SIT/tc!/lres. 35, ~9.J.5-4961. :\low, V.C., Gibbs. ,\I.e., LaL \\'.\1.. l.'t Ill. (1989a). Biphnsic dentation of ~Il·ticular Glrlibge-Part II. A llurlh:ric:11 a rithm ;\Ild an l'xpt.:riml.'U1;'ll stlld~'. 1 Bi()UI~'(:I/(l/1ic..;. 22, 8 '\'lo\\', V.C., Holmes, ;\1.1-1 .. &. LaL \\'.\'L (198"+). Fhlid tl'i.ll1s and mechanical propl.'nies of articular c,lrtilage: A rev J Hiol/lt'cha /lies. I?, 377. \lo\\". \I.e.. Kuei, S.c.. LaL W.'\-1.. ('t al. (1980). Biphash: cr and stress rdaxOIlion of anicul;ll' L·'lrtilage ill comprcss Thl'ory and cxpl.'rimellts. J Biol1lt'dl EJlg. 102. 73. Mow. V.c.. LaL W.;\I. . ..\: Rt.'dlcr. I. {197-H. Some surface c aCIl.'ristics of articular cartilag~s. A SC;-\lliling electron croscopy study and a th~'orl'\ic~d model ror the d.vna inli,:ractioll of synovial fluid and articular cilrtilage. J IIleclumics. 7. -t.-I9. Mow, V.e. . ..\: R:llcliffi..'. :\. (1997). Structure :Hld funclion articular c::Irtili1gc and menisclls. In v.e. ;"low l\! W Hayi..'s (Eds.), B(I.'ii~· Ori"o!'(/I:dic: Biollll.·chal/ics (2nd pp. lI3-11i). Philadelphia: LippincOIl-Ra\"l.'ll. \!o\\', V.e., Zhll, W.B., Lai, W.M., ct al. (1989b). The in ence of link protein stabilization on 1h..: viscoelastic p ertici' of protcoglYC<1I1 aggregates. Bioclt':lll Biophys .992, 201. :'I·luil". H. (1983). Proteogl~'cans as org:lnizl.'rs of the extra lular matrix. Bim.:ht'JIl Soc TraJJs, I 1.613. Myers. E.R .. Lai, \V.M .. &. ~·Iow. V.C. (1984). A continuulll ory ;Illd an expcriment for the iOIl-inducl'd swelling beh ior canila£c. 1 Biol1wdl r.II!.'" 106(21, 151-15S. N\..'\\·b..:rry, Zukosky, D.K:, & Hau!. R.C. (1997). Subf tllrc insult to a knt:.·c joint causes alter~l(ions in the b :-lnd in the rUflctional stiffness of o\'erlying cartilage. 1 Ihop Rc:s, 15. -t.50. '* W.N.. Onsagcr. L. (1931J. Reciproc:l1 relations in irn.'vcrsible processL's. I. Phys ReI' 37, "+05. Pau!. J.P. (1976). Force actions transmilted b~' joints in the human bod\'. Proc Ror Soc Loud, /92B. 163. Poole. A.R. (1986). Pr~)[coglycans in health and diseasc: StrucLure and function. Hinchem J. 236. 1. Radin. E.L.. and Paul, 1.1.. (1971), RespollSc of joints to impact loading. l. In vitro wear. Arthritis Rhellill. 14, 356. Radin. E.L. (1976). Aetiology of osteoarthrosis. CIiIl RhculJI Dis. 2, 509. Ratcliffe. A.. &. ~'Iow, \f.e. (1996), Articular cartilage. In \V.D. Compel' (Ed.), Extrace/llIlar Alatrix (Vol. 1, pp. 234~302). AmsLcrdam: Harwood Academic Pubs. Ratcliffc. A.. 'I\ler. J.. &: I-Iardingham. TE. (1986). Articuhlr cartilage culture with inlerlcukin I: lncreascd releasc of link pr~)tein. hyaluronate-binding region and other proteoglycan fragments. BiochclII J, 238, 571. Redler, I.. &. Zimny. M.L. (1970). Scanning electron microscopy of normal and abnormal artlcuhll' curtilagc and synovium. J BOl/c Joint SlIrg. 52/\. 1395. Rcdlcr. L, Zimny. :'.'1.L., \Ianscl!, J .. et al. (1975). Thc ultrastructure and hiol11cchanical significancc of the tidemark of articular canilagc. Clill Or/hop Rei Res, / /2.357. Rosenberg. L.. Choi. H.U .. Tang. 1..-H .. et a!. (1985), Isolation of dennatan sulf.ltc proteoglycans from mature bovine artindar cartilage. J Bio! Cht'lJl, 260, 6304. Roscnberg, L.. Hellmann, \V., &. Kleinschmidt, A.K. (1975). Electron microscopic studiL's or proteoglycan aggregatL'S from bo\'ine articular cartilage. ./ Bio! Chew. 250, 1877. Roth, V.. &: Mow. V.c. (1980). The intrinsic tensile bchavior of the nl<\trix of bovine articular canilagc and its variation with age . ./ BOHi.' ./oilu SlIrg, 62A, 1102. Roughlev. P.J.. &. Whitc, R.J. (1980). Age-related changes in t';e structure of thc proteoglycan sub~lnits from hum7'1ll articular cartilage. J BioI Choll. 255, 2!i. Roughley. P.l .. \Vhite. R.l., & Santcr. V. (198 I). Comparison of proteoglycans extracted from high- and low-weight bearing human articular cartilage, with particular referencc to sialic acid content. J BioI Che/1/, 256, 12699. Schinagl. R.M .. Gurskis. D., Chen, A.C., et a1. (!997). Depthdependent confined compression modulus of fullthickness bovine articular cartilage. J Ort/lOp Res. /5,499. Schinagl. R.M., Ting. M.K .. Pricc. J.H .. el a1. (1996). Video microscopy to quantitate the inhomogeneolls equilibrium strain within articular cartilage during confined compression. illIll Biolllt'd Eng. 24. 500. Schmidt, :\-;1.B., Mow, V.C .. ChUll. L.E., et al. (1990). Effects of proteoglycan extraction on the tcnsile behavior of artiClIJar canilage . ./ Or/hop Rcs, 8. 353. Schneiderman. R., Keret, D., &. Maroudas, A. (1986). Effects of mechanical and osmotic pressure on thc rate of glycosaminoglycan synthesis in adult femoral head cartilage: An in vitro study. J Orthop Res. 4, 393. Schubert, i\'1.. &. Hamerman, D. (1968). A Primer 011 COIIlICCtive Tisslle Biochelllistry. Philadelphia: Lea &: Fcbiger. Scott . .I.E.. & Orford, C.R. (1981). Dennatan sulphate-rich protcoglycan associates with rat tail-tendon collagen at the d band in the gap region. BiochelJl J, /97,113. Selton, L.A., Gu, \V.Y., Lai, \V.M .. et a1. (1995), Predictions of the swelling induced pre-stress in articular cartilage. In A.P.S. Selvadurai (Ed.), .lleclulllics of Porous Atcdia (pp. 299-311). Kluwer Academic Pubs. Dordrecht. the Netherlands. SCtlUIL L.A .. \leJ\\', V.e. \Iuller, F..J .. L'l :d. (199..+1. :\lcch<H pnq)l:l'liL's of c:lninL' :lnictllar c;lrlil:tgL' :11"L' signil'kanLly Il'rL'd following tr:lflSacliofl or IbL' ;lI'llL'l'ior l"l"tlCiak l nh:nl. 1 Of/!zop Res. /2.451, SL'llon. L.A., Toll.\':lIl1a. fl .. l\: \10\\'. V.e. (1998). S\\'L'lling curling bl'ha\'ior oi aniL'ular canibge. J BiuJllt'c1/ r /10.355. Setton. L.A .. Zllli. \\'.13.. & :..,,10\\-. V.e. (19931. Thl' biph poro\'iscol'l:lstic lk'h:l\'ior 01' anlcular cartib1:!c in C prL'ssion: Role or thl' surracc zone. J Biollleclwl/ics. 581. Sokoloff. 1.. (1063). ElasliciL,V of articular cartilagL': Ei'fcc iOllS :lIld \'iscOlls solutions. Sci(l/ce. /4/. 1055. Soltz, \1.A .. & A(l'shi:1I1. G.A. (1998). ExpnilllL'nlal \'nif tion and Ihl'()]"L'lic,d prediction of canil:lgL' inkrsli r1uid pressurizalion al an impl'rmL'able cont:lct intcr in conl'inl'd corTlpn:ssion . .I Bio!!/cchanics. OCt 31 ( 927-934 StoL:kWL'lI, R.S. (1979). Biolo.!!.y COJ"/i!o!.!.c Cd/so Cambrid C:\lllbridgl' Uni\'crsit.\' Prcss. Sun, D.N .. Gu. \\'.1'.. Guo. X.E .. L·t al. (199S1. ThL' infhll'llcL inhomogelleous fixed charge dL'l1sitv on canil,q!l' nlccha L'leetrochclllical belw\'iors. '!i'mls On/lUp Res Soc. 23. 4 Swann. I).A .. Radin, E.L .. & Hcndn:lL R.B. (1979). Thl·lu c:ltion of :lrliclIbr cartilage b~' syn(wial I'luid gl,\'co tL'ins. Arthriris Rhein. 12, 665. SW:\l1Il. fL\ .. Sih'L'r, F.H., Sla~'lL'r. H.S .. ct :d. {1%5L ThL' m t'cubr structure ~lnd lubricating :lcti\'it,\' of lubricin f bo\"inc and human s~'nO\'ial fluids. BiucltoJl J. 225. 19 S\\'cL'l. \l.B.E .. Thonar. E../.-\1.A., & \1:lrsh . .I. (1979). A rL'lated ch,lngcs in protcoglYC:1I1 structurL'. Arch Hioch or Biopltys. / 98. 439~448. Thompson, R.C .. Ocgcma. TR .. Lcwis . .1.1" L'l al. (1991 L tcoanhrotic changcs al'tcr anHe Iransarticular load. animal model. J BOHL' lohll Slirfi. 73:\. 990. Thonar. E..I.-\L·\ .. Bjornsson, S., & KueHner. K.E. (19 Age.related changcs in canil:lge proteoglycans. In Ku,-'ltner. R.S .. SclJle~'crbach. &. V. C. Hascall (Eds.). A 1111/1' Clirrilllg.c Biochcl!/islJY (pp. 273,-287). Nl'\\' Y Ra\'en Prcss. Torzilli. P.A .. &: \Iow. V.e. (19761. On the t'lInd:lI11elltal f Iransport lllechanisms through normal and palhologic tilage during function. 1. The formulation . ./ Biolllcch, 9 541-552. Torzilli. P.A .. Rose. D.!:., & Dethl'mcrs, S.A. (I9821. Equ rium watn partition in articular cartilage. Biorhw/ogy, 519. Urban . .l.P.G .. & \-lc\luliin . .I.F. (1985). Swclling prcssur thc intervertcbral disc: Influcnce of collagcn and pro glycan content. Biorhc%gy, 22. 145. Valhlllll. W.B .. StazmDc. E.J .. Bachrach. N.;vl . L'l a!. (19 LO<ld-controlkd compression of articular cartilagL' duces a transicllt stimulation of aggn:can gene cxprcss Arch Biocltc/!/ Biophys 353, 29. Venn, \'I.F. (1978). Variation of chemical compOSition age in hllman remora I head cartilagc. Ann RheulII Dis, 168. \Vada. T. & Akiwki. S. (1987). An ultrastructural stud solid matrix in articular cartilagc under uniaxial ten stress. J .Ifill On//tJfJ Assn. 61. \Valker. P.S .. Dowson. D.. Longfeild. \I.D .. et al. (19 "Boosted lubrication" in syno\'ial joints b~' rluid ent lllent and enrichment. :Inn !VIC/IIII Dis, 27, 511. Walka. P.S .. Unswonh, :\ .. Dowson, D.. I,:t a!. (1970). ~-lod(.' of ag!!rcgation of hyaluronic acid prntdn cOll1pk:x on tilL' sll~facc of articular cartilage. .. \/11I Wlt'lflll Dis. 29. 591. \V;JOg. C.B., &. Mow, V.C. (199S). Inhomogendty of aggreg,HL' n;odulus affects c~rtil:tgL' cornpn.:ssi\·I..' strcss-n:la:Xi1tion behavior. 1i'alls Or/hop Uc:s Soc, 13( I). 481. Weiss. Roscnbl.:rg. L., & J-1e1fL'1. A.J. (1968). Alluhraslrllctural study of !lumlal young adllli hum;11l anicular cani· lage. J BO//(: 10illl SlIrg, 50A, 663. \Villiams, P.F., j1owdl. G.l.., &: Laberge, M. (1993). Sliding friction analysis of phosphatidylcholinl: as a boundary lubricant for anicul<"lr cartilage. Prot Ills 1 l/ech EIl.!jI'S, c.. l07, 59. Woo. S.L.-Y., Mow, V.c., &: Lai, W.~,t. (1987). l3iorncchanical propcnics of ;Irticubr cartilage. In /-IllIu/book of Bioengineering (pp. 4.1-.....·H. New York: McGrmv-Hill. Woo, S.L.Y.. LL'wsay G.A., Runco, 1.1., ct:ll. (199;). Struc aml function of (t:ndolls and ligalllt:nts. In V.C. ~·1o W.e. Hayes (Eds.), Basic Orthopaedic BiollleclUlllics ( cd., pp. 109-2.5 I). Phibddphia: Lippincolt-Raven. Zhu, \V.B .. Iatridis, J.e.. Hlibczjk. V.. Cl al. (1996). DCle n;ltioll of collagen-prot('oglycan interactions in vitr BioJl1cclwnics. 29, 7i3. Zhu, \-V.B., Lli, \V.~-I., &: J'I·low. v.e. (1986). Intrinsic qu linl..'ar viscoebstic bL'havior of thc c.\tracelllllut' m;:Hrl cartilagL'. 7i'tIIIS OrthofJ Res Soc, 1/, 407. Zllu, \V.B., Lai. W.i\1., &. I\\OW, V.e. (1991). TIl(: dl.'nsity sll"L'ngth of proteoglycan-proli..'oglycan interaction site C(IIlCenlrakd solutions. J Biolllt'C!liIllics, ],1, 1007. Zhll, W.B., ,\In\\', V.C .. Koob. '1'.1 .. ct al. (1993). Viscoela shl.."u· propi..'rtil..'s of articular cartilagL' and the <:ffcc!:. glycosidase treatments. 1 Ortlwp Res, I J. 771. AIRTI'CULARCARTILAGE ,streSses sustlllned by cont::ict joint surfaces ." '.' '" ".",.'.::~::~><,,:,>~' To provldoa smoothwear·reSistantbearing s'urfac~ecrease b1ction~' '~:,', '.•···HI.,hlvSpeclalizedTlssu'e ~ ;' I 'I EXTRACELLULAR COMPONENTS CELlULAR COMPONENTS Extracellular M3trix-ECM- 1 \~ ,-'/,.,'-' " POROUS PERMEABLE SOUD PHASE .! -Qrganlt Component--: i .A ;~TERS~IT.I~]f3~!·[)rHASE ,:;;; ~!ceir'~~~~!%!~i;'1f'~~~&!,i~'> 0, :":1 ~ j :1 Biph;uic Str'uc(urc II :,:,'1 BIOMECHANICAL PROPERTIES \! Animtropy,Viscocl;micity. Swelling Bchavior-ColllprCHivc lo~d bearing c~pJcity- .~, FLOW, CHART 3·1 Articular cartilage structure and biomechanicaJ properties. ~ '~\j "Thh How ChMl is d~si9ncd fOI das>room or group disCLl~~io". Flow than i~ not mCi'lr'lt \0 be cxh<lustive. Results in Disruption of collagen-PG solid matrix PG "wash out" Gross alteration of the normal load cartilage mechanisms ~fLOW CHART 3'2", Articular cartilage wear mechanisms. ~ "This flow chart is designed for c1assroorn or group discussion. Flow ch.Jr1 is not meant to be exhaustive. ~ ;;; E ~ -"! .2 -.; v ~x ~ ~' \J '"C' ~ v >- '"0 0 ~ Ci ei 0 > 't;; . 0 0 0 0 .0 8 ~ E 5 E 0 0 .;0 ~ .- ~ .,; c '" ~ ." ~ '" ~ .;0 ~ v £ .~ ." ~ ~ ·v ~ ~ ~ E V ." ~ e:;; ." 0 , 0 ~ 's1;' <;-S :E :; , !1. g .~ ~ '0 ~ ~ "'6 E 0 e • 0 v ~ " ~ iii § 0 ." g ~ 0 C .~ 0 ."- ,J.' t , 0 .,; ~ i= • ! .';".'i~'- .;. !j '.: ./ ! Biomechanics of Tendons and ligaments Margareta Nordin, Tobias Lorenz, Marco Campello Introduction Composition and Structure of Tendons and Ligaments Collagen Elastin Ground Substance Vascularization Outer Structure and Insertion Into Bone Mechanical Behavior of Tendons and Ligaments Biomechanical Properties Physiological Loading of Tendons and Ligaments Viscoelastic Behavior (Rate Dependency) in Tendons and ligaments Ligament Failure and Tendon Injury Mechanisms Factors That Affect the Biomechanical Properties of Tendons and Ligaments Maturation and Aging Pregnancy and the Postpartum Period iVlobilization and Immobilization Diabetes lvIellltus Steroids Nonsteroidal Anti-Inflammatory Drugs Hemodialysis Grafts Summary References Flow Charts • Introduction The three principal structures that c10selv surround, connect and stabilize the joints of the ~keletal svst~Il1are tendons, ligaments, and joint capsules. ~\l­ tIlollgh these structures arc passive (i.e., thev do not acth'ely produce motion as do the l11usck:s), each pl.iJ's an essential role in joint motion. The role of the ligaments and joint capsules, \vI11c.h connect bone with bone, is to auament the I-IICchanical stability of the joints, to gl1i(iL~joint motion, and to prevent excessive motion. Ligaments anel joint capsules act as static restraints. Tile functicm of the tendons is to attach muscle to bone and to transmit tensile loads from muscle to bone, thereby.' producing joint motion, or to maintain the hpcly posture. The tendons and the muscles form the muscle-tendon unit, which acts as a dvnamic n> straint. The tendon also enables the mtls~le bell\' to p¢at an optimal distance from the joint on whi~h it acts without requiring an extended length of muscle ~ between origin and insertion. Tendon and ligament injuries and derangements a~'e common. Proper management of these disor(leI'S requires an understanding of the mechanical properties and function of tendons and ligaments ZtIld their capacity' for selF-repair. This chapter discusses the following: Composition and structure of tendons and ligaments Biomechanical properties and behavior of normal tendon and ligament tissue Biomechanical properties and behavior of injured tendon and ligament tissue tutes a large portion of the organic matrix of b and ~artilage and has a unique mechanical supp ive function in other connective tissues such blood vessels, heart, ureters, kidneys, skin, and li The great mechanical stability of collagen gives tendons and ligaments their characteristic stren and flexibility. Like other connective tissues, tendons and l ments consist of relatively few cells (fibroblasts) an abundant extracellular matrix. In general, cellular material occupies approximatelv 2000 of total tissue volume, while the extraccliular ma accounts for the remaining 80°/(;. Approxima 70t;0 of the matrix consists of water, and appro mately 30% is solids. These solids are collag ground substance, and a small amount of c1a; The collagen content is generallv over 75 % and somewhat greater in tendons than in ligame (Kassel', 1996); in extremitv tendons, the solid m rial may consist almost e~1tirely of collagen (up 99% of the dry weight) (Table 4- I). The structure and chemical composition or l ments and tendons are identical in humans and many animal species such as rats, rabbits, dogs, monkey's. Hence, extrapolations regarding~ th structures in humans can be made from the res of studies on these animal species. COLLAGEN The collagen molecule is synthesized hv the fl blast \vithin the cell as a la;'ger precurso·r (proco gen), which is then secreted and cleaved extrace larl y' to. become collagen (Fitton-Jackson, 19 (Fig. 4-1). Tendons and ligaments, like bone, composed of the most common collagen molec Several factors that affect the bioI11echanical u t <:> function of tendons and liuaments are auin co "" co' ')fe'(1nancy, mobilization and immobilization, diabetes, nonsteroidal anti-intlammatory drug (NSAID) use, and the effects of hemodialvsis. Biomechanical considerations regarding graft~ are also given. Tendon I Cellular Material: Composition and Structure of Tendons and Ligaments Tendons and ligaments are dense connective tissues knc)\vn as parallel-flbered collagenous tissues. These sl:arsely vascularized tissues are c0l11posed largelv of- collagen, a fibrous protein constituting approximately one third of the total protein in the bodv (White, Handlel~ & Smith, 1964). Collagen consti- i I EX~~:~~I~'ar Matnx I 20% 20% Fibroblast 80% 80% 60-80% 60-80% Solids: 20-40% 20-AO% Collagen: 70-80% slightly hig Type 1 90% 9S-99% Type 3 10% 1-5% Ground substance 20-30% slightly En(l0!0f!On f:1l1nd!e ProcoHagen Fibroblasl Collagen molecules - - - - - - - - - - - - - _ . _..._._---- - _ _...... ----._.--.~--- ._--._--------- Schematic representation of wllagen fibrils. fibers. and bun- f'xtrac('lIuiar lll<1!rl)( ill cl PiHt111e-! ilflangelll('llt {o form m;· dles in tendons and collagenous ligaments (not drawn to crofibrils clnd then fibrils, The staggered array of the mole (tIles, in which each overlaps the other, gives .1 banded ap- scale). Collagen molecules. triple helices of coiled polypep- tide chains. are synthesized and secreted by the fibroblasts. These molecules (depicted with "heads" and "tails" to repre- sent positive and negative polar charges) aggregate in the type I collagen. This Il101eculc consists of three polypeptide chains (ex chains), each coiled in a lefthanded helix wilh approximately 100 amino acids, which give it a total molecular weight or approxi. mately 340,000 daltons (Rich & Crick, 1961) (Fig. 4-2). Two or the peptide chains (called ,,·1 chains) are identical. and one differs slightly (the 0:-2 chain). The three a-chains are c0l11binccl in a right-handed triple helix. which gives the collagen molecule a rodlike shape. The length or the molecule is approxi. mately 280 nanOllleters (nIT}), and its diameter is approximately 1.5 nm. Almost two thirds of the collagen molecule consists of three amino acids: glycine (33%), proline (15%), and hydroxyproline (15°/(;) (Ramachandran, 1963). Every third amino acid in each 0: chain is glycine, and this repetitive sequence is essential for the proper rormation of the triple helix. The small size 01' this amino acid allows the tight helical pack· ing of the collagen molecule. C'vlorco\'cr. glycine enhances the stability of the molecule by forming hydrogen bonds among the three chains of the superhelix. Hydroxyproline and proline form hydrogen bonds, or hydrogen-bonded wal~r bridges. within each chain, The intra- and inlcrchain bonding. or cross-linking. between specific groups on the chains is essential (0 the stability of the molecule, Cross-links are also formed between collagen molecules and are essential to aggregation at the ,0 the Coll<1gen iibri!s under the electron pcarancE' llIicro,:>wpe. Tilt:' hb!!l ... <)~J9rcgate further into hbers, ~"... I1l<:h (omc together into dcn<,cly piicked bundles. lihril k'\'l'l. IL i...; till.' I,.'I'{):-;:,,-!illkl,d dl~II·~h.:[l'I' ur Llll'l' 1~lgl'll lihrib ikll I~h'l'~ ,"'II'l'll~lh [0 thl' li:--:-il.l('.'" IIl C( lmpOSl' and <:t1I< I\\':-- I hl.':--l' I b:"Ul'S I( I l"tl1"ll.·1 il III und 1ll1.'("II<tllic~d :--lrl':":--. \\'illlill 111(' lihril:-., !Ill.' mulecul <:tl"t' ;Jppar('Il!I~' tTUss-lillked h.y "hl'<:\(.I·I(l-i~lil" illte <:ll:tiulb (Fi~_ -l~ I L hUI inlcrlihrilbr IT(I:-..... -linkill~ o rnOl"L' complL'\ n~llllr(' <:d~(l rn~l.'" on·lll'. III 111..'\\-1\- fcwlllt:d ,,:(lll~l~l'l1. thl' CI'(l:-.:--·lillk:-- ~\1\':' r :lli\"\,.,h- k\\' :lnd ~lrt' rl.~dlll.'ihl\:: 1111.' l.'(llb~l.'l1 b :,>olll ill rJt'lIlral sail :"ohlliol"l:' <:!llll ill ~Icid :"(IIUliull:--, a 1Ilt' i.:ro:--~-Ii111..:-; a I"l' hI i rI., t'a~i I.,' dL'Il~llll rt'd hy hl.'at. t.:olla~t'n a~L':-', lilt' 1111<:11 rlullIhtT nl rl'lilldhk' tTOS link:-. dl'LTl'a:-'L':-' lei ~l tllinilllurn ~b ~l 1~lr~l' lllllllhl'r :--(~lhk', nonrl,dul'ihk· lTCls:"-lird,:,, ~lrl' fOI"t'lll'd. \bIUI l'<"lI~l~t'n i~ nul sU[lIhk, ill Ilt'lllral :"i.dl ~(lIt1ti(Jll:" 01' ~ldd solulio1\:", and il :-illlyi\'l'~:\ hi~llL'l' dl'I1;,lllll'~l tl'l1IjWl':lllll'l', (1"-"0[' ~\ I\~\'il'\\' or lTo:"s-link:\1!l' ill ~'() 12l'll, Sl.'l' Viidik, D~lllil'lsl'll, t~ Chlulld, 10~2,l .:\ lihril is !"(Irllll,d h," till' a~~I·q.!.ali(l1l Sl'\'t'l l,'olla~l'1l lllolL'ctlk's ill ;1 qU<lltTn~lIY Slrtlc!url.'. Th strllclllrl.', in \\·Ilh.:h t"ach Illnkndt' O\'t'!·bp:-. I otht'r, is I\'spollsihk for 1I1l' J'L'lk'alillt! IXIII((:-; o :':'t'J"Yl,d Oil thl' lihril", lllH..kT till' t,k-l'II'OIl lllilTOS(O or (Fi~, --l-2: ~l'l' :ds(l Fig, "'''''). Till' qll,llt'l'll:II-," Slrll...: lu or l.T)lIa~cn rdalt'S 10 till.' or~~1I1i:t<:lli(l1l III l.·ollagl.: lllolt'ClIll's il110 :l :"l~lbk. 10\\' l.'lll'r~i,:lic bi(llo!!ic unit ha:--t'd 011 :J rq:!.ul<:\1· a:':"(ll"ial iOll (~r ~\(,.l.iaCt'lll ';11 t'''.'lIks' h~ISil' <:llId ~lcidil' :llIlino ~ll·ids. B., :In:ll1gi ~\d.iacl'lll col1at!.I.'l't 1I1pk'l:uks in a qLl~lnl'l'-st~lg1:-~t'r. o om 64 Fibril 000 , 0 or;: Overlap zone 010 r ;:r: Packing of 'molecules I I Hole zone Micrafibrils ~~~~=~'~'~'~~~~~~~~~ I I ................. ..-",/ ~~/ / ... -' ........... rCollagen r'.'='='=====:2~80~nm~======):J~ i' molecule !... I ........ .J . ,. , - , ! .................. " ....................... .......... i a.2~~~t 1.5 nm ~, 1 Schematic drawing of collagen microstructure. The collagen molecule consists of three alpha chains in a triple helix (bonom). Several collagen molecules are aggregated into a staggered parallel array. This staggering, which creates hole zones and overlap zones, causes the cross-striation " (banding pattern) visible in the collagen fibril under the electron microscope. Adapled from Prockop, 0.1.. & Guzman, ;\! A. (1977). Col/agen dise,)ses and the biosynthesis of collagen Hasp Pract. Dec. 61-68. positely charged amino acids are aligned. This stabl.c structure will require a great amount of energy ,':~ri,d force to separate its molecules, thus contribut:.ipg to the strength of the structurc. In this way, orgU)lizcd collagen molecules (five) form units of microfibrils, subfib"ils, and fibrils (Fig. 4·3) (Simon, 1994). The fibrils aggregate f'urther to form collagen fibers, which are visible under the light microscope. ;;./(~hese fibers, which range from I to 20 J.L111 in cliam''.;:,'',¢ter, do not branch and l11av be many centimeters "j<lI1g. They reflect a 64-nm periodicity of the fibrils .:!: and have a characteristic undulated form. The fibers aggregate further into bundles. Fibroblasts are aligned in rows between these bundles and arc elongated along an axis in the direction of ligament or ;'i':. ~.enclon function (Fig, 4-4). ';/: The arrangement of the collagen j-ibers differs somewhat in the tendons and ligarncnls and is Suited to the function of each stru~ture. The fibers composing the tcndons have an orderl.y, parallel arrangcn)cnt, which cquips the tendons to handl lhe high unidirectional (uni4lxial) lensile loads t which they are subjected during activity (Fig. 4-4A) The Iigamenls gencrally sLisrain tensile loads in on predominant direction but may also bear smalle (cnsile loads in other directions; their fibers may no be completely parallel but arc closely interiaccd with one another (Fig. 4-48). The specific orienta tion of the fiber bundles varies to some cxten among the ligaments and is dcpendcnl on the func tion of the ligament (Arnie! et al .. 1984), The metabolic tUI'nover of collagen may be stud ied by tritium labeling of h)-!droxyprolinc or glycin and by autoradiographic methods, Studics in ani mals have shown that the half-life of collagen in ma lure animals is very long: the same collagen mole cules may exist throughout the animal's adult lire however, in young animals and in physically altered (e.g., injured or immobilized) tissue, the turnover i acceleratcd. Rabbit studies have shown melaboli activity to be somewhat greater in ligaments than in tendons, probably because of different stress pat terns (Amie! et al .. 1984). ELASTIN The mechanical properties of tendons and liga ments are dependent not only on the architectur and properties of the collagen fibers but also on th proportion of elastin that these su'uctures contain The protein elastin is scarcely present in tendon and extrcrnily ligaments, but in elastic ligamcnt such as the Iigamenlum Oavum, the proportion o elastic fibers is substantial. Nachemson and Evan (1968) found a 2 to I ratio of clastic to collage fibers in the ligaments flava. These ligamcnls, which connect the laminae of adjacent vertebrae, appea to have a specialized role, which is to protect th spinal nerve roots from mechanical impingement to pre·stress (preload) the motion segment (th functional unit of the spine), and to provide som intrinsic stability to the spine. GROUND SUBSTANCE The ground substance in ligaments and tendons con sists of proteoglycans (PGs) (ur> to approximatel 20% of the solids) along with stnlctlll,JI glycoprotcins plasma proteins, and a variety of srnall n1olcculcs The PG units, macromolecules composed of variou sulfated polysaccharidc chains (glycos<lminoglycans bonded to a corc protein, -bind to a long hyaluroni acid (HA) chain La form an cxtremely high molecula Fibroblasts Crimp Fascicular membrane Schematic representation of the microarchitecture of a tendon . • • Nearly parallel bundles of collagen fibers Parallel bundles of collagen fibers I~= ~ Fibroblasts ligament Tendon A B Schematic diagram of the structural orientation of the fibers of tendon (A) and ligament (8); insets show longitu- dinal sections. In both structures the fibroblasts are elongated along an axis in the direction of function. Adapted from Snell, R.5. (984), Cfinical and F~ln(tional Histology for Medical Stuejents. Boston: Liule. Brown weight PC aggregate like Lhm found in Lhc ground substance of anicular cartilage (set.:' Fig. 3-6). The PC aggregates bind most of the extracellular water of the ligament and tendon. making the matrix a highl~' structured gel-like material rather than an amorphous solution. Furthermore. by acting as a ccment-likc substance betwcen the coHagen microflbrils. they ma),' help stabilize the collagenous skeleton of tendons and ligaments and contribute to the o\'crall strength of these composite structures. Only' a small number of these molecules exist in tendons. however, and their imporlance for its biornechanical properties has been questioned, VASCULARIZATION Tendons and ligaments have a Iimitcd vascularization. which alTeets dircctl~' their healing process [lnd metabolic activit~;. Tendons receive their blood supply dircctly from \'cssels in the perim.vsiulll. the periosteal insertion. and the surrounding tissuc via vessels in the para tenon or mesotenon. Tendons surrounded b~) paratenon have been referred to as vascular tendons. and those sllrround~d by a tendon sheath as avascular tendons, In tendons surrounded by' a paratcnon, vessels enter rrom many points on the pcripher.v and anastomose with a longitudinal svstem or capilbries (Fig, 4-5), India ink-injected (Spalteholz technique) into the calcaneal tendon of a rabbit, illustrating the vasculature of a paratenon-covered tendon. Vessels enter from many points on the periphery and anastomose with a longitudinal system of capillaries. Reprinted wirh permission from Woo, S.L. Y, An, K.N., Amoczky, D.V.M., et at. (1994). And!Om}~ biology. and biomechanics of rhe iendon, ligament, and meniscus. In S.R. Simon (Ed) Orthopaedic Basic Science (p. 52). Rosemont, Ii: MOS. ··f " • The vascular pattern for tendons surrounded by tendon sheath is different. Here the meso tenons a reduced to vincula (Fig. 4-6). This avascuJar regio led a variety of researchers to propose a dual pat way for tendon nutrition: a vascular pathway, an for the avascular regions, a synovial (diffusio pathway. The concept of diffusional nutrition is primary clinical significance in that it implies th tendon healing and repair can occur in the absen of adhesions (i.e" a blood supply). Conversely, lig ments in comparison with surrounding tissue a pear to be hypovascular. However, histological stu ies reveal that throughollt the ligament substan there is a uniform multivascularit)', which orig nates from the insertion sites of the ligament. D spite the small size and limited blood flow of th vascular s)'stem, it is of prima(y importance in t maintenance of the ligament. Specifically, by pr viding nutrition for the cellular population. this va cular system maintains the continued process matrix synthesis and repair. In its absence, damag fTom normal activities accumulates (fatigue) an the ligament is at risk for rupture (Woo et aI., 1994 Ligaments and tendons have been shown in bo human and animal studies to have a variety of sp cialized nervc endings and mcchanorcceptors, Th play an important role in proprioception and noc ception, which are directly related to the functio ality of joints. the tendon substance. ReprinlGd wirh permission from Woo, S.L. Y, An, K.N., Amoczk}~ D. VM., et al. (1994). Anatomy, biolog and biomechanics of (he rendon, ligament, and meniscus, In S.R Simon (Ed.). Orthopaedic Basic Science (p. 52). Rosemont. IL: MOS. OUTER STRUCTURE AND INSERTION INTO BONE Certain similarities arc found in the outer structure of tendons and ligaments, but there are also important differences related to function. Both tendons and ligaments arc surrounded by' a loose areolar connective tissue. In ligaments, this tissue has no specific name, but in tendons it is referred to as the para tenon. More structured than the connective tissue surrounding the ligaments, the paratcnon forms a sheath that protects the tendon and enhances gliding. In some tendons, such as the flexor tendons of the digits, the sheath runs the length of the tendons, and in others the sheath is found only! at the point where the tendon bends in concert \vith a joint. In locations where the tendons are subjected to particularly high friction forces (e.g., in the palm, in the digits, and at the level of the wrist joint), a parietal synovial layer is found just beneath the paratenon; this synoviun1-like membrane, called the epitenon, surrounds several flber bundles. The synovial fluid produced by the synovial cells of the epitenon facilitates gliding of the tendon. In locations where tendons are subjected to lower friction forces, they are surrounded by the panltenon only'. Each fiber bundle is bound together by the endotenon (Fig. 4-1), \vhich continues at the musculotendinous junction into the perimysium. At the tendo-osseous junction, the collagen fibers of the endotenon continue into the bone as Sharpey's perfOl·ating fibers and become continuous with the periosteum (Woo et aI., 1988). The structure of the insertions into bone is similar in ligaments and tendons and consists of four zones; Figure 4-7 illustrates these zones in a tendon. At the end of the tendon (zone 1), the collagen fibers intermesh with fibrocartilage (zone 2). This fibrocartilage gradually becon1es Inineralized fibrocartilage (zone 3) and then merges into cortical bone (zone 4). The change from more tendinous to more bony n1aterial produces a gradual alteration in the mechanical properties of the tissue (i.e., increased stiffness), which results in a decreased stress concentration effect at the insertion of the tendon into the stiffer bone (Cooper & Misol, 1970). Merhanical Behavior of Tendons and Ligaments Tendons and ligaments arc viscoelastic structures \vith unique mechanical properties. Tendons are Electron micrograph of a patellar tendon insertion from dog, showing four zones (:<25,000). Zone 1, parallel coll gen fibers; zone 2, unmineralized fibrocartilage; zone 3, mineralized fibrocartilage; zone 4, cortical bone. The lig ment-bone junction (not pictured) has a similar appearance. Reprinted with permission from Cooper, R.R. & Misol, S (1970). Tendon and ligament insertion. A light and electron microscopic study J Bone Joint Surg, 52A, I. • strong enough to sustain the high tensile forces th result from muscle contraction during joint moti yet are sufficiently' flexible to angulate around bo surfaces and to deflect beneath retinacula to chan the flnal direction of muscle pull. The ligaments pliant and Oexible, allowing natural movements the bones to which they attach, but are strong a inextensible so as to ofTel' suitable resistance to plied forces. Analysis of the mechanical behavior of tendo and ligaments provides important information the understanding of injury mechanisms. Bo structures sustain chiefly tensile loads during n mal and excessive loading. When loading leads injury, the degree of damage is affected by' the r of impact as well as the amount of load. VIVI~,-."MI."."'L PROPERTIES means of analyzing the biomechanical properof tendons and ligaments is to subject specimens tensile deformation lIsing a constant rate of elonThe tissue is elongated until it ruptures, and resulting force, or load (P), is plolted. The resultload-elongation Clll-ve has several regions that the behavior of the tissue (Fig. 4-8). The first region of the load-elongation cuniC is the "toe" region. The elongation reflected in region is believed to be the result of a change the wavy pattern of the relaxed collagen fibers. In region, the tissue stretches easily, without much and the collagen fibers become straight and their \vavy appearance as the loading proa",:sSl:s(Hirscb, 1974; Woo et a!., 1994) (Fig. 4-9, A B). Some data suggest, however, that this elongamay' be caused mainly by' interfibrillar sliding 4 3 2 I 1mmLI I I I Elongation (%») _ Load-elongation curve for rabbit tendon tested to failure in tension. The numbers indicate the four characteristic regions of the curve. (1) Primary or "toe" region, in which the tissue elongated with a small increase in load as the wavy collagen fibers straightened out. (2) Secondary or "linear" region, in which the fibers straightened out and the stiffness of the specimen increased rapidly. Deformation of the tissue began and had a more or less linear relationship with load. (3) End of secondary region. The load value at this point is designated as PI;,,' Progressive failure of the collagen fibers took place after PI'" was reached, and small force reductions (dips) occurred in the curve. (4) Maximum load (Pm.,,) reflecting the ultimate tensile strength of the tissue. Complete failure occurred rapidly, and the specimen lost its ability to support loads. Adapted from Carls/edt, c.A. (7987). Mechanical and chemical factors in tendon healing: Effects of indomethacin ancl surgery in the rabbit. Acta Orthop Scand Suppl, 224. ~~--------- Scanning electron micrographs of unloaded (relaxed) a loaded collagen fibers of human knee ligaments (x 10,000). A, The unloaded collagen fibers have a wavy configuration. S, The collagen fibers have straightened under load. Reprinted with permission from Kennedy, J. C, Havjkins, R.I. Willis, R.B., et al. (7976). Tension stuclies of hu man knee ligaments, Yield point, ultimate failure, and disrup of the erueiare and tibial collateral ligaments. J Bone Joint Su S8A.350. and shear of the interfibrillar gel (ground substan (for review, see Viidik, Danielson, & Oklund, 198 As loading continues, the stiffness of the tis increases and progressively greater force is requi to produce equivalent amounts of elongation. T elongation is often expressed as strain (e), whic the deformation of the tissue calculated as a p centage of the original length of the specimen strains arc increased (strain values of between 1.5 and 4% [Viidik, 1973]), a linear region will follow the toe region. This sudden increase in slope represents the second region in the diagram and con"csponds to the response of Ihe tissue to further elongation (Diamant et al.. 1972). Following the linear region. at large strains the stress-strain ClIlve can end abruptly or curve downward as a result of irreversible changes (failure) (Woo et aI., 1994). Where the curve levels off toward the strain axis, the load value is designated as Plin , The point at which this value is reached is the yield point [or the tissue. The energy uptake to PHn is represented by the area under the ClIl·ve lip [0 the end of the linear region. \Vhen the linear region is surpassed, major failure of fiber bundles occurs in an unpredictable l11anne1'. \Vith the attainment of maximum load that reflects the uhimatc tensile strength of the specimen, complete Failure occurs rapidly, and the loadsupporting ability of the tendon or ligament is substantially reduced. The modulus of elasticity for tendons and liga~ ments has been determined in several investigations (Fung, 1967, 1972; Viidik, 1968). This parameter is based all a linear relationship between load and deformation (elongation), or stress and strain; that is, the stress (force per unit area) is proportional to the strain: elastic deformation that they endure under tensi strain and the storage and loss of energy. Durin loading and unloading of a ligament between tw limits of elongation, the elastic fibers allow the m tcrial to return to its original shape and size aft being deformed. Meanwhile, part of the enelspent is stOt"cd; what is left will represent lhe energ loss during Ihe cycle and is called hysteresis. Th area enclosed by the loop represents the energy lo (Fig. 4-1 I). PHYSIOLOGICAL LOADING OF TENDONS AND LIGAMENTS The ultimate tensile strength (P,,,:..,,) of ligamen and tendons is of limited interest from a function standpoint because under normal physiologic conditions in vivo these structures are subjected a stress magnitude that is only approxirnately on third of this value. The upper limit for physiologic strain in tendons and ligaments (when running an jumping, for example) is from 2 to 5% (Fung, 1981 E = rIlE where E = modulus of elasticity (T = E = stress strain In the toe portion of the load-elongation curve (or the modulus of elasticity is not constant but increases gradually. The modulus stabilizes in the fairly linear secondary region of the curve. The load-elongation CUPie depicted in Figure 4-8 generally applies to tendons and extremity ligaments. The curve for the ligamentum flavulll, with its high proportion of elastic fibers, is entirely different (Fig. 4-10). In tensile testing or a human ligamentum flavlIIll, elongation of the specimen reached 50% before the stiffness increased appreciably. Beyond this point. the stiffness increased greatly with additional loading and the ligament failed abruptlv (!'eached P"",,), with little further deformation (Nachemson & Evans, 1968). The proportion of elastic proteins in ligaments and capsules is extremely important for lhe small stress~slrain curve), ~1.'· o -70 Elongation (~o) Load·elongation curve for a human ligamentum flavum (60 to 70% elastic fibers) tested in tension to failure. At 70% elongation the ligament exhibited a great increase stiffness with additional loading and failed abruptly with out further deformation. Adapted from Nachemson, A.L., & Evans, J.H. (1968). Some mechanical properties of tile third hu man lumbar imerlaminar ligamenr (ligamentum flavum). ) B;omech, 1, 21/-220. • ,--? . 1 cycle ~Peakload ~ Elongation Typical loading (top) and unloading curves (bottom) from tensile testing of knee ligaments. The two nonlinear curves form a hysteresis loop. The area between the curves, called the area of hysteresis, represents the energy losses within the tissue. '0 ,Few studies or loading of tendons or Ii!!aments in \:ivo have been perrorm~d. Kear and Smith (1975), using the strain gauge method, measured the rnaxinUll strain in the latcral digital extensor tendons of sheep. The strain reached 2_6t}~ while the shl:cp were trolling rapidly and decreased when the trolting speed decreased. This maximal strain occurred for only 0.1 second during each stride. The maximal load imposed on the entire tendon was approximately 45 newtons (N). These results suggest that during normal activity, a tendon in vivo is subjected LO less than one fourth of its ultimate stress. VISCOELASTIC BEHAVIOR (RATE DEPENDENCY) IN TENDONS AND LIGAMENTS Ligaments and tendons exhibit viscoelastic, or ratedependent (time-dependent), behavior under loading; their mechanical propelties change with different rates of 1041ding. v"hen ligament and tendon specimens are subjected to increased strain rates (loading rates), the linear portion of the stress-strain curve becomes steeper, indicating greater stiffness of the tissue at higher strain rates. v\lith higher strain rates, ligaments and tendons in isolation c',," store more energy, require marc force to I1.lptu and undergo greater elongation (Kenncd.", Hawki Willis, & Danylchuk, 1976). Dw-ing cyclic testing of ligaments and tendo where loads are applied and released al specific tenmls, the stress-strain curve is displaced to riglll along the deformation (strain) axis with ea loading cycle, revealing the presence of a nonel tic (plastic) cOlllponcnt; thc amount 01" perm nent (nonrecoverable) deformation is progressiv greater with every loading cycle. As cyclic load progresses, the specimcn also shows an increase clastic stiffness as a result of plastic dcfonnati (molecular displacement). IVlicrofailure can oc within the physiological range if frequcnt loading imposed on an already damaged struclUrc wh lhe stiffness has decreased. Two standard tests that reveal the viscoelastic or ligaments and tendons are the stress-relaxat test and the creep test (Fig, 4~12). During a stre relaxation test, loading is hailed safely below linear region of the stress-strain curve and the str is kept constant over an extended period. The str decreases rapidly at first and then more sl()\v \Vhen the stress-relaxation test is repeated cy cally, the decrease in stress gradually becomes l pronou need. During a creep lcst,loading is halted saFely bel the linear region of the stress-strain curve and stress is kept constanl over an extended period. T strain increases relatively quickly at first and th more and more slowl.':. \Vhen this tcst is perform cyclically. the increase in strain gradually becom less pronounced, The clinical application or a constant low load the soft tissues over a prolonged period, which ta advantage of the creep response. is a llseful tre ment for several types of deformities. One exam is the manipulation of a child's clubfoot by subje ing it to constant loads by means of a plaster ca Another example is the treatment of idiopathic s liosis with a brace, whereby constant loads are plied to the spinal area to elongate the soft tiss surrounding the abnormally curved spine, More complex viscoelastic beh41vior is observed the entire bone-ligament-bone complex_ Antel cnlciate ligaments (ACLs) in knee specimens tak from 30 primates were tested in tension to failure a slow and a fast loading rale (Noyes el aI., 1976). , the slow loading rate (60 seconds), much slower th that of an injury mechanism in vivo, the bony ins tion of the ligament was 'the weakest component the bone-ligament-bone complex.. and a tibial sp bone-ligament-bone complex \vas nearly' the sam These results suggest that as the loading rate is i creased, bone shows a greater increase in streng than does ligament. "'"o --' Load relaxation (length held constant) Ligal1zel1t Failure and Tendon Injury Mechanisms Injury mechanisms arc similar for ligaments an Time A tendons, therefore the following description of lig ment injury and failure is gencrall~1 applicable tendons. \Vhen a ligament in vivo is subjected Creep phenomenon (load held constant) c o .~ E loading that exceeds the physiological range. micr failure takes place even before the yield point (P" is reached. \"'hen Ph" is exceeded, the ligament b gins to undergo gross failure and simultaneolls the joint begins to displace abnormally. This d placement can also result in damage to the su " <ii o Time B The viscoelasticity (rate dependency, or time dependency) of ligaments and tendons can be demonstrated by two Microfaiture standard tests: the load-relaxation test and the creep test. A, Load relaxation is demonstrated when the loading of a specimen is halted safely below the linear region of the load-deformation curve and the specimen is maintained at a constant length over an extended period (i.e., the amount of elongation is constant). The load decreases rapidly at first (Le., during the first 6 to 8 hours of loading) and then gradually more slOWly, but the phenomenon may continue at a low rate for months. B, The creep response takes place when loading of a specimen is halted safely be· low the linear region of the load· deformation curve and the amount of load remains constant over an extended period. The deformation increases relatively quickly at first (within the first 6 to 8 hours of loading) but then progres~ sively more slowly. continuing at a low rate for months. avulsion was produced. At the fast loading rate (0.6 seconds), which simulated an injury mechanism in vivo, the ligament was the weakest component in t\Vo thirds of the specimens tested. At the slower rate, the load to failure decreased by 20%, and 30% less energy was stored to failure, but the stillness of the o 2 3 4 5 6 Elongation (mm) Progressive failure of the anterior cruciate ligament from cadaver knee tested in tension to failure at a physiologic strain rate. The joint was displaced 7 mm before the liga ment failed completely. The force-elongation curve gene ated during this experiment is correlated with various de grees of joint displacement recorded photographically; photos correspond to similarly numbered points on the curve. Reprinted wirh permission from Noyes, F.R., and Graae E.S. (976). The strengrh of the anterior cruciare ligament in h mans and Rhesus monkeys. Age-relared and species~relared changes. J Bone Joint Surg, 58A. 1074-1082. 350 spectivcly to (I) the load placed on the ACL duri tests of knee joint stability performed clinical (2) the load placed on this ligament during phys logical activity, and (3) that imposed on the lig ment during injury from the beginning of micr failure to complete rupture. Microfailurc begi even before the physiological loading range is e ceeded and can occur throughout the physiologic range in any given ligament. In fact, under expe mental testing, the ultimate tensile loaels-or t load at failure for human ACL-is between 340 a 390 N (Case Study 4-1), Complete failure (3) ~ -0 "0 200 -' o o 2 4 3 5 6 7 8 --- Joint Displacement (mm) ACL Failure: Failure of the ACL Associated With High Strain and Stress The curve produced during tensile testing of a human an- ~,. terior cruciate ligament in vitro (20) (see Fig. 4-13) has been converted to a load-displacement curve and divided into three regions correlating with clinical findings: (1) the load imposed on the anterior crudale ligament during the anterior drawer test; (2) that placed on the ligament during physiological activity; and (3) that imposed on the liga w 25-year·old male occasional soccer player injured . his ACl as a result of an abnormal torque in rotation A of the knee. The player locked his foot on the ground and pivoted on his lower limb to produce a high rota· , tional torque on the knee, which increased tensile loads the ACL, men! from partial injury to complete rupture. It should be on noted that the divisions shown here represent a general- The first region of the load-displacement curve shows a normal physiological loading response. In the microfailure region, the increase in strain deformation leads to high internal stress and finally a complete rupture. Experimental testing in vitro in human ACl yielded a point of failure between 340 and 390 N. The knee with the ACl injury will increase intra-articu- ization. Microfailure is shown to begin toward the end of the physiological loading region, but it may take place well before this point in any given ligament. rounding structures, such as the joint capsule, the adjacent ligaments, and the blood vessels that supply [h~sc structures. Noyes (1977) demonstrated the progressive failure of the ACL and displacement of lhe tibiofemoral joint by applying a clinical test, the antel-iOl- drawer '." tc;-st, to a cadaver knee up to the point of ACL failure 'AFig, 4-13), At maximum load, the joint had displaced several millimeters. The ligament \vas still in '~o,ntinuit.y even though it had undergone extensive ~jacro- and microfailure and extensive elongation, 'In Figure 4-13, the force-elongation curve generated during the experiment, indicaling where Illicrofail.,:. ure of the ligament began, is compared with val"ious stages of joint displacement t-ecordcd photographically, Correlation of the results of this test in vitro with clinical findincrs sheds Jjcrht on the microevents that take place in the ACL during normal daily activitv and during injuries of variou~ degrees or s~verity. I~ Figure 4-14, the curve for the experimental study On cadaver knees that was presented in Figure 4-J3 been converted into a load-displacement curve divided into three regions, corresponding re~ lar joint motion, producing abnormally high stresses on other joint structures such as cartilage, which can lead to osteoarthritis. A deficiency in joint stability that results from ACl impairments will increase the likelihood of ex· periencing the "giving way" sensation or functional instability, thus affecting activities of daily living such as gait, jogging, and squatting (Case Study fig 4-1- 1). ~ o 2 3 4 5 6 Joint Displacement (mm) Case Study Figure 4-1-1. 7 8 Ligament injuries arc categorized clinically in three ways according to degree of severity. Injuries in the first category produce negligible clinical symptoms. Some pain is felt, but no joint instability can be detected clinically, CYCI1 though microfaillire of the collagen fibers may have occurred. Injuries in the second category produce severe pain and some joint instability can be detected clinically. Progressive failure of the collagen fibers has taken place. resulting in parlial ligament nlplure. The strength and stiffness of the ligament may have decreased by 50% or more, mainly because the amount of undamaged tissue has been reduced. The joint instability produced by the partial rupture or a ligament is oflen masked by muscle activity, and thus the clinical test for joint stability is usually performed with the patient under anesthesia. Injuries in the third categOlY produce severe pain during the course of trauma with less pain after injury Clinically, the joint is founelto be completely unstable. Most collagen fibers have ruptured, but a few Illay still be intact, giving the ligament the appearance of continuity even though it is unable to SUppOrl any loads. Loading of a joint that is unstable as a result of ligament or joint capsule rupture produces abnormally high stresses on the articular cartilage. This abnormal loading of the articular cartilage in the knee has been correlated with carl~,' osteoarthritis in humans and in animals. Although injury mechanisms are generally comparable in ligaments and tendons, two additional factors become important in tendons because of their attachment to muscles: the amount of force produced by contraction of the muscle to which the tendon is attached and the cross-sectional area of the tendon in relation to lhat of its muscle. A tendon is subjected to increasing stress as its muscle contracls (see Fig, 6-10), When the Illuscle is maximally contracted, the tensile stress on the tendon reaches high levels. This slress can be increased fl.lrlher if rapid eccentric contraction of the muscle takes place; for example, rapid dorsiflexion of the ankle, which does not allow for reflex relaxation of the gastrocnemius and soleus muscles, increases the tension on lhe Achilles tendon, The load imposed on the tendon under these circumstances may exceed the yield point. causing Achilles tendon rupture (Case Study 4-21The strength of a muscle depends on ilS physiological cross-sectional area. The larger the crosssectional area of the muscle, the higher the magni;:' lude of the foree produced by the contraction and thus the greater the tensile loads transmitlcc! through the tendon. Similarly, rhe larger the crosssectional area of the tendon, the greater the loads it can bear. Although the maximal stress to failure for a muscle has been difficult LO compute accurately, such measurements ha\'c shown that the tensile strength or a healthy tendon 111<1)-.' be mon: than t\Vice that or its Illuscle (Elliot, 1967), This finding is supponed clinically by the fact that muscle fUplures are more common than are ruptures through a tendon. Large muscles usually' have tendons with larg~ cross-sectional areas. Examples arc the quadriceps - - - --------<0 -rend on Injuries: Achilles Tendon Injuries in Runners. Which Result From a High Strain Rate .·.. A middle-aged male marathoner engaged in a strenuous running activity expenenced pain and a popping sensa~ tion in his posterior calf. An overuse injury is diagnosed. The first region of the load-deformation curve ShQ\.V5 a normal physiological toe-loading response. In the linear region, high load is producing a higher deformation within the tendon structure. When the Achilles tendon is subjected to higher strain rates during frequent loading ;_ ~ycles and insufficient time is allowed for the healing process, the result is an overuse injury. Histological studies ot',these injuries reveal a pathological pattern described as ~'angiofibrotjc hyperplasia," which suggests a degenera" tive process. This failure in tendon remodeling frequently ,. occurs before the abrupt ruplure of the tendon. Relative aydscu!arity, inflammatory disease, and other local factors also contribute to midsubstance ruptures (Case Study 4·2-1), 250,00 OVERUSE 200,00 ~ Toe region Linear region ~ RUPTURE yield and failure region 150,00 u .3 100.00 50,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 Def9rmation (mm) Case Study Figure 4·2·1. muscle with its patellar lendon and the triceps Slifac muscle with its Achilles tendon. Some small muscles have tendons with large cross-sectional areas; such as the plantaris. which is a liny muscle with a large tendon. did nOl differ signillcanLi~1 (Kaspcrczyk et aI., 1991). This may be due 10 the fact lhal only lhe ACL was taken from donors in \I,,1hom no vascular or cardiopulmonary disease and no ostcoanhritis of the knee was found on autopsy. Factors That Affect the Biomechal1ical Properties of Tendons and Ligaments PREGNANCY AND THE POSTPARTUM PERIOD Numerous factors alTecl I he biomcchanical properties of tendons and ligaments. The most common are aging. pregnancy, mobilization and immobilizalion. diabeles. steroids. NSAID use. and hemodialysis. The biomcchanical properties of grafts are also d,iscllssed because reconstruction, particularly of the anterior and posterior knee ligaments, is common. A common clinical observation is thc incrcasecllaxity of the tendons and ligaments in the pubic arca during later stages of pregnancy and the postpartum period. This obsen"~llion has been confirmed in animal studies, Rundgren (1974) found lhmlhe tensile strength of the lendons and the pubic symphysis in rats decreased at the end of pregnancy and during the postpartum period. Stiffness of these stnlctures decreased in the early postpartum pCI'iod but was later restored. MOBILIZATION AND IMMOBILIZATION MATURATION AND AGING , / i': The physical properlies of collagen and of lhe lissues it composes are closely associaled with the number and quality of the crossMlinks within and between the collagen molecules. During maturation (up to 20 years of age), the number and quality of cross-links increases. resulting in increased tensile strength of the tendon and ligament (Viidik, Danielsen, & Oxluncl, 1982). An increase in collagen fibril diameter is also observed (Parry et aL. 1978) wilh high variabilily in size (range 20-180 nm) (Strocchi. el aL. 1996) noted in lhe ~'oung «20 years), The diameler in adults (20-60 veal's) and in lhe elderly (>60 years) decreases remarkably (120 and 110 nm) but \vith a marc even distribution. Slrocchi et al. (1996) investigated age-related changes in human ACL collagen fibrils and reports an increase of libril concenlration from 68librils/mu' in the young to 140 fibrils/mu' in the elderly. Howevel; Amiel (1991) reports Ihal lhe waleI' content and the collagen concentration decreases significantly in the medial cruciate ligament of 2 M , 12-, and 39-month-old rabbits, After maturation, as aging progresses, collagen reaches a plateau with respect to its mechanical properties, afler which the tensile strength and stiffness of the tissue begin to decrease. This may be the result of an increase in small collagen fibrils. Conversely, when the ACL of younger donors (average age 30 years) was compared wilh the ACL of older donors (average age 64.7 years), the material properties (strain. elastic modules, and maximum stress) oF"'-';- , Living tissues are dynamic and change their mechanical properties in response to stress, which leads to functional adaptation and optimal operalion of the tissue. Like bonc, ligament and tcndon appear to remodel in response to the mechanical demands placed on them; they become stronger and stiffer when subjected to increased stress and weaker and less stiff when the stress is reduced (Noyes el aL. I 977a), Physical training has been found to increase the tensile strength of tendons and of the Iigament M bone interface (Woo et aL. 1981), Tipton and coworkers (1970) compared the slrenglh and sliffness of medial collateral ligaments from dogs that were e~cr­ ciscd strenuously for 6 weeks with the values for ligaments from a control group of animals. The ligamcnts of the exercised dogs were stronger and stiffer than lhose of lhe control dogs. and the collagen fiber bundles had largel- dianlelers. (mmobilization has been found to decrease the tensile strength of ligaments (Newton et al., 1995; Walsh et aL. 1993), Noyes (1977a) demonstrated a reduction in the mechanical properties of the bone-ligament-bone complex in knees of primates immobilized in body casts for 8 weeks. When lested in tension to failure, the ACLs from these animals showed a 391'10 decrease in maximum .load to failure and a 32% decrease in energy stored to failul-e compared with ligaments fl'om a control group of animals (Fig. 4-15A), The immobilized ligaments also displayed more elongation and were significanlly less stitT than the control spcc..:imens (Fig, 4-158), Amid and coworkers (1982) showed a similar de- crease in the strength ancl stiffness or lateral collateral ligaments in rabbits immobilized for 9 weeks. As the cross-sectional area of the specimens did 110t change Significantly, the degeneration of mechanical properties was altributed to changes in the ligament substance itself. The tissue metabolism was noted to increase, leading to proportionally more immature collagen with a decrease in the amount and quality of the cross-links between collagen molecules. Newton et al. (1995) reported that the crosssectional area of ligaments in immobilized n1bbit knees waS 740/0 of the control valuc. In Noves' (I977a) experiment, assessment of the effects of a reconditioning program initiated directly aftcr the 8-\\'cek immobilization pel"iod demonstraLcd that considerable time was needed 1'01' the immobilized ligaments to regain their former strength and stiffncss. After 5 months, the reconditioned ligaments still showed considerably less stiffness and 20% less strength than did ligaments from control animals. At 12 months, the reconditioned ligaments had strength and stiffness values comparable to those of control group liga- ments (Fig. 4-1 SA). Woo et "I. (1987) found that the stress-strain characteristics after rcmobilization return to normal but that the energy-absorbing capabilities or the bone-ligament complex improved but did not return to normal. DIABETES MELLITUS The term diabetes rcfcrs to disorders characterized by excessive urine excretion. Diabetes mellitus is a metabolic disorder in which the ability to oxidize carbohydrates is more or less completely lost. This is usually caused by pancreas insufficiency and a disturbance of the normal insulin mechanism, rcsulting in hyperglycemia, glycosuria, and polyuria. Diabetes mellitus is known to cause musculoskeletal disorders. Diabetics compared with nondiabetics show higher rates of tendon contracture (29 \IS. 9%), tenosynovitis (59 vs. 7%), joint stillness (40 vs. 9%), and capsulitis (16 \'s. 1%). Diabetes also causes osteoporosis (Calvallo et aI., 1991; Lancaster et aI., 1994). Duquette (1996) examined the effects of diabetes on the properties of the collateral knee ligament in rats. The tissue elastic properties did not differ between the diabetic and the control group. The vis- 100 60 -" 0,2 ~~ l'l E c ~ "~·x E "'" 60 40 Q..E 20 o Maximum load 10 failure Energy slored 10 failure -- ,.- " " "" " A' .------- Immobilized --~8 weeks) I \ \ \ Elongation (mm) B A A, Maximal load to failure and energy stored to failure for primate anterior cruciate ligaments tested in tension to faHure. Values are shown as a percentage of control values for three groups of experimental animals: (1) those immobilized in body casts for 8 weeks; (2) those immobilized for 8 weeks and given a reconditioning program for 5 months; and (3) those immobilized for 8 weeks and given a reconditioning I program for 12 months. 8. Compared with controls. ligaments immobilized for 8 weeks were significantly less stiff (as indicated by the slope of the curve) and underwent greater elongation. Adapred from Noyes. FR. 0977(1). Functional proper~ ties of knee ligaments and alterations induced by immobilization. (lin Onhop, 123,2/0-242. component of the tissue response, however, \vas increased in the hy'perglycemic group. Insulin therapy seems to lessen such alterations. Lancaster, ct al. (1994) examined the changes in the mechani~ cal properties of the patellar tendon in diabetic dogs. The results showed the stiffness of the canine patellar tendon-tibia complex in a phy'sioiogical range of loading was 13% greater than in the control group. There was no difference in the strength of the tendon between the groups, but the mode of failure was dif~ ferent. In the control group, failure was caused by substance and avulsion failure, whereas failure of the tendon in the diabetic group was caused by tensile fractures of the patella (Lancaster et al" 1994). COllS STEROIDS Corticosteroids, when applied immediately after injury, may' cause significant impairment of the bio~ mechanical and histological properties in liga~ ments. Corticosteroids also are known to inhibit collagen synthesis in vitro (\'Valsh et aL, 1995). \Viggins et a!. (1994) described these results in rabbits and implied that an acute injured ligament treated with corticosteroid injections may not \vithstand the mechanical loads of an earl)!, vigorous rehabilitation. Noyes et a!. (1977b) reported decreased ligament stiffness, failure load, and energy absorption in monkey ligaments after injection of long-acting corticosteroids. These findings were tin1e- and dosage-dependent. After application of a dosage that was approximately 10 times an equivalent human dose. only! minimal changes were found after 6 \veeks, but after IS weeks the maximum failure load (20%), energy absorption prior to failure (11%), and stiffness (11 %) decreased significantly. After application of a dosage equivalent to the human dose, the maximum failure load (9%) and the enabsorption (8%) decreased significantly. Campbell ct al. (1996), however, showed that a single injection of long-acting corticosteroid does not cause histological differences in rats with acute injured ligaments as compared \vith rats with acute ligament injury and no injection of corticosteroids. lv1cchanical testing showed no significant difference in ultimate load or ultimate stress in the two groups. Oxlund et a!. (1980) reports that local injections of corticosteroids every 3 days for 24 days increase the tensile strength and maxin1um load stiff~ ness of muscle tendons but decrease the strength of the bone attachments of ligaments. Laborator).' investigations established the presence of estrogen receptors in hun1an ACLs. Liu et al. (1997) reports that physiological levels of estrogen reduce the collagen production by 40(;1(; and at phar macological levels of estrogen, collagen production is decreased by' more than 50%, Estrogen fluctua tions may alter ligament metabolism and ma change the composition of ligament, rendering i more susceptible to injury. NONSTEROIDAL ANTI-INFLAMMATORY DRUGS NSAIDs (which include aspirin, acetaminophen, and indomethacin) are frequently used in the treatmen of various painful conditions of the musculoskeleta system. NSAIDs are also widely used in the treat ment of soft tissue injuries such as inflammatOl)' dis orders and partial ruptures of tendons and liga ments. Vogel (1977) found that treatment wit indomethacin resulted in increased tensile strengt in rat tail tendons. An increase in the proportion o insoluble collagen and in the total collagen conten also was observed. Ohkawa (1982) found increase tensile strength in the periodontium of rats after in domethacin treatment. Carlstedt and associate (1986a, 1986b) found that indomethacin treatmen increased the tensile strength in developing an healing plantaris longus tendons in the rabbit an noted that the mechanism for this increase wa probably an increased cross-linkage of collagen mol ecules. These animal studies suggest that short~tern administration of NSAlDs would not be deleteriou for tendon healing but instead \vould increase th rate of biomechanical restoration of the tissue. HEMODIALYSIS Tendinous failure resulting from chronic renal failur does occur, \vith tendon rupture reaching 36% amon individuals receiving hemodialysis. Hyperlaxity o tendons and ligaments \vas found in 74%, patella tendon elongation in 49%, and articular hypermobil ity in 51 % of individuals receiving long-term he modialysis (Rillo et a!., 1991). Dialysis-related amy loidosis may cause the deposition of amyloid in th synovium of tendons. The major constituent of th amyloid fibrils is the beta 2-microglobulin (Morita e a!., 1995; Honda et a!., 1990; Bardin et a!., 1985). GRAFTS Reconstruction of torn ligaments, especially or th anterior and posterior cruciate ligament, is now frequent procedure. The need for reconstruction related to age, activity level, and associated injurie Grafts derived from different individuals of the same species ar~ called allografts; grafts derived from the same individual are called aUlograhs. Allografr tissue preservation is done through freezcdrying and low·dose irrndiation to reduce rales of rejection ancl infection and to limit effects on the structural properties. Bone-patellar. tendon-bone, and Achilles tendon are llsually L1sed as allograft tis- sue, whereas the central tissue of the patellar tendon is commonly llsed as autograft tisslie. Shino ct al. (1995) L1sed fresh-frozen allogenic Achilles, tibialis anterior or posterior. and peroneus longus or brevis tendons for ACL reconstruction in humans. Specimens were procured during secondlook anhroscop~'. Several years after reconstruction. the allografts had collagen fibril profiles that did not resemble normal tendon grafts or normal ACL. Strocchi et al. (1992) L1sed patellar tendons that had been autograftcd to reconstn1C[ [onl ACLs. Follow-up biopsies were performed 6, 12. and 24 months after surgel}'. During this time, the aUlOgraft underwent considerable changes. and after 24 months the autograft had the appearance of normal ligament tissue. Strocchi suggested that the patellar tendon autograft is a valid functional ACL substitution for patients who desire to perform normal mechanical activity. Corselli el al. (1996) reports thal replacement tis- sue undergoes extensive biological remodeling and incorporation. However. even a fully incorporated graft will never duplicate the native ACL but works instead as a check reign that increases the knee function_ Tohayama et al. (1996) stated that the graft elongation at the time of implantation influences the long-term outcome of ACL rcconstruc· tions. at least in the canine model. They compared those cases where the graft elongation behavior was within the 95% confidence limit of normal ACL (group 1) with those cases where the graft elongation behavior was more than lhe 95(1'0 confidence limit of the normal ACL (group 2). Group 2 had sig- nificantly less inner stiffness of the graft than did group 1. Group 2 showed a significantly increased anteroposterior laxity. but there was no difference in ultimate failure load and absorbed energy. substantial proportion of elastin. which lends these structures their great elasticity. 2 The arrangement of the collagen fibers is nearly parallel in tendons. equipping them to with- stand high unidirectional loads_ The less parallel arrangement of the collagen fibers in ligaments al~ lows these structures to sustain predominant tensile stresses in one direction and smaller stresses in other directions. ,;:~~)': At the insertion of ligament and tendon into stiffer bone. the gradual change from a morc fibrous [0 a morc bony rnaterial results in a dccreased stress concentration effect. ~ ~ Tendons and ligaments undergo deformation before failure. ,,""hen the ultimate tensile strength of these structures is surpassed. complete failure DC· curs rapidl~!. and their load~bearing ability is sub~ stantially decreased. 5. Studies suggest that during nonllal activity. a tendon in vivo is subjected [0 less than one fourth of its ultimate stress. '/'~:(: Injul"\! mechanisms in a tendon are influenced by"{he ~lm~unt of force produced by the contraction the muscle to which the tendon is attached and the cross-sectional area or the tendon in relation to that of ilS muscle. or 7 i The biomechanical behavior of ligaments and tendons is viscoelastic. or rate·dependent. so thal these structures display an increase in strength and stiffness with an increased loading rate. S An additional effect of rate dependency is the slow deformation, or creep. that occurs when tendons and ligamcnts are subjected to a constant low load over an extended period; stress relaxation takes place when these structures sustain a constant elongation over time. /{gi,::': Anina results in a decline in the mechanical ,/),,, :;;. b properties of tendons and ligaments, th~1t is, their strength, stiffness, and abilily to withsland defor- mation. 10 Pregnancy, immobilization. diabetes. steroids. NSAlD Lise, and hemodialysis affect the biomechan- Summary f~1-~Tendons and extremity ligaments are composed largely of collagen, whose mechanical stability gives lhese structures their characteristic strength and flexibility. The ligaments flava of the spine have a ical properties of ligaments and tendons. 11 Allografts and autografts are useful in ligament reconstruction but material propenies do not re'turn completel~" to normal levels. 12) Ligaments and tendons remodel in response to the mechanical demands placed on them_ REFERENCES :\micl. D.. Kuipl'l'. S.D.. \\'all"C('. C.D .. Harwood. E V'lndl'berg. J.S. (1991). :\gt:.rdall'd properties of llledi.d (;olbt· l.:ral Iig~lIncnt and anterior cruciall' ligament: ..\ morpho· logic :lnd coll'lgen maturation study in thl' r'lbbit. J Gerem/ol. 46(.J). BI56-BI65. Amid. D•• Fr,lIlk, c.. I-Iar\\'ood. E. l.'1 OIL (198-1). Tendou~ and ligaments: A morphological and biochl'lllical comp~l'ison. lOt/hop nt's. 1.157. :\mit'1. D., Woo. S.L.Y.. Harwood. F.L.. l't a!. (1982). Thl' dfl.'ct of immobilization on colbgen lUl'IlO\"l'J" in t:onlll,,'cli\'c tis· suc. i\ biochl..'rnical-biomechanical C01Tl'btion. Ac/o Or. !hop Scalld, 53, 323. Bardin. T.. Kuntz, S., Zingr.d"f, J.. Voisin. ~'I'I.C., ZelmaI', A. (1985). Syno\'ial amyloidosis in patients undergoing long:. terlll hl..'lllodialysis . .·\J'/l1l"it;., Rheum, 28(9}. 1052-1058. Campbell, R.B .. Wiggins. \-1. E.. Cannislra. L.;\L F"dalc. P.P.. Akclm.m, E. (1996). Influencc of stcroid injc(:tion in li~:lment hc~t1ing in thl..' r.lI. Cli" OnllOp. 332. i.J1-253. Carlstcdl, C.:\. (1987). Mt:chanical and cht.'f1li(~11 f:u;tors in [('ndon healing: Effects of indolTlcthacin and sun.!cn· in th..: 1'00hbil. Aew On/WI' Scawl SIIPI,I, IN. ~ . , Carlstcdt, C.:\ .. Madson. K.. &. Wn:dmark. T. (l9Soa). The illflllencc of indolTlethacin on collagen synthi.'sis durin!! lell' ([cm healing in lhi.' rnbhit. Pros/ag!(/I/{!i,ls. 32. 353. C~Il'lstcdl, C.A .. ~bdsOTl. K" & W!'cdrnark. 1". (1986b). Till' influent:l.: of indtlllll.'thacin on lendon hl'aling. A biolJlcchall' itnl nnd biochcmical sllJd~'. ..lFell OrtllOl~1 7i'allll/{/ So!'!.;. 105. 332. ' (:11'\,(\110. A.. Ordonez. M.A .. Garcia. H.. T;tpia. J.C .. Bcbv:\c, 1.. V:tlC:.'llCi;l. J., r..lorl'ira. ~','l. (1991). Adult llon·insll1in-dt:pcndl'lll diahetic: Limitation of <lJ'licular Illobi1itv ~\lld soft tissue involvcment. Rev lIed Chil,119(9). 10Io-1()2l. Cooper. R.R .. ~ Misol. S. (1970). Tendon .Ind lig:lIllCIH inserlion. ;\ light and electron microscopic sllLdv. J B{J!/~' Joiul SU,,!;. 52;\, I. . Corsl.'tti. J.R .. J'lckson. D.W. (1996). Fnilul'c of :lIl1Crior nud· ~IlC Iig<lll1l'llt rl'conslruction: The biologic basis. Clill Or. fhop. 325. .J2--t9. Diamant. J .. Kdlcr. r\ .• Bael', ..\., Lill, M.. :\lTid~l'. R.G. (1972). Colhlgen: Ultraslructure and its relations ~ 10 mechanic,,1 propl..'ni<.'s as a function of aging. Prot· R Soc Loud iBio/}, /80, 293. DuquclIc. J.J .. Grigg, P.. &. Hoffman. A.H. (1996). The effect (If diabetes on the viscoelastic propcrtil's of r~\[ kncl' li!'!H· ments. J Biolllt.!ch Ellg. 118(·0. 557-56.J. Elliot!. O.H. (1967). Tht: biomcchanical propl'rties of ll~lldoll in relation 10 muscular strcn"lh . .-\111/ PIn's ,\led 9 1 Fillon·Jne.:kson, S. (1965), Antc~cdcnt pha~cs'in ;lla'tJ"i~ formalion. In 5tl'/lt"lll1"1..' alld PI/llelioll OJ"COIIIIt:ctiw {jllll Skelewi Tis.~lIes (p. 277). London: ButtCl'worth. Fun~, .Y.C.~. (1981). Biomeclumics: ;\kcllflllical Propr:f,i.:s o{ Llvl1Ig T,ssue.'s (p. 222). Nc\\' York: Springer· Veri:.!!. Fung..r.C.B. (967). Elasticity of soft lissucs in sim-pk don· g'lllon. :Im J Phvsiol, 213. 1532. Fung. Y.C.B. (1971). Strl.'ss-str.lin-hislo/,v rdations of sof! lis· sues in simplc elongation. In Y.C_ Fu·ng. N. PelTOnl', & t\'1. l~nlikcr (Eds.). Biolllechallics: Its FOlll/dations awl Ob;ec. IlvcS (pp. 181-20$). Englewood CliiTs. NJ: Prentice-Hall. HIrsch, C. (1974). Tensile properties during tendon healinl!. ..lCla Or/hop Scellld. SIIPl'l 153. - Honda. K.. I-lara. M., Ogura, Y.. Nilll.:i. 1-1 •• i\lirnura, N. (I990 ~~t<l-~-microgl()hulin ~llnyloidosis in hemodialysis pa t~l·ntS: An autopsy study of intcrH·rtt:hr.t1 disks and POSlC nor hg<lOll·lltS . ..\eUl Patlwl lpII . .JO( II). S20-826. Kaspl'I"czyk. W. J.. Rosoch:l. S .. BO$t:h. U .. Ocsh~rn. J.1.. Tsch crnl·. I-I. (1991). Ag(' .\CtiviIY .Hld stl'<.'ngth of knee.: li!:.! ments. U,,{alkhimrr,. 94(7).372-375. Kassel". J. (Ed.) (19961. Onlwp/li.'dic I\lI00dccly,i.' Update f/?IIll! S{lfdy Syllal}f/,~, P'.ll'k Ridgl'. IL: American Acadl'Ill ot Onhopatdic Slll"~c.:ons. Kcar, i\l.. & Smith. R.N. (1975), A flll'thnd for l'l'corJin u W don stnlin in sheep during locomotioll, Aew On/lOp S:flll -16. 896. Kl'llIll'dy, J.C., l-bwkins. RJ .. Willis. R.B .. d :.1. (1976). Ten sion Sllu.!i(·S of humall knec ligaments. Yield point, ult matc failure, and disruption of thl.: cl'tlciatc and tibbl co "Heral ligaments. 1 BOlle.' loillT S/f/'~. 58:\. 350. Lancaster. R.L., Haut. R.C.. DeCamp. C.E. (199.J). Chan2es thl' lllech:luic,,1 propenil's of p.ltdl.ll· tl'lldOll pr..:par.~tio of spolHancously dii:lbl'lic do,gs tIIH!l'r 10nll·tC'I"1ll insuli th~rapy. J Biomf!cil. 27(8). 110.1-1108. Lill. S.H .. AI-Shaikh. R...\., Panossi~lll. \1.. Fil1{·rman. G.A Lane, J.M. (1997). Estrogen ;tlTccts l!Ie cellular Illctabo lisll.l of .lh(.' .aolt'rior t:l'l1ciall' ligaml'lll. ,\ potential expl n:tllnn lor lemalL' ~llhlt.'lic injury..'\111 1 Sports ,I/i.'d. 25(5 71J4-709. :\-lorita. H.. Shinzaalo. T.. Cai. Z.. Dadd. G.. i\lizulani A Habuchi, H.. Ito, :\1.. Asai, J. 1.'1 ;d. (1995). Basic fibrol;la growth faClor-hepatin sulphate complex in the huma dia1:o'sis-relalcd amyloidosis, Vir"hol\'s Arch. ';27(4 395-400. Nachclllson. A.L .. & E\'ans, J.I-I. (1968), Some mechanic propertit:s of tht: third hUl1wn lumb;\[' intcrlaminar lig men! (ligalll(.'ntllm lhvlllll). J BicHIIIN,:". I. 211-220. Nl·wton. P.O., \Voo. S.L.. ~ladenrw, D.A .. Akeson. \V.H (1995). Immohilization of tli..: knee:.: joint alters Ihl' m dwnic~d and uilrastrucllll',ll propenil's of the r.lhbit anle rior crucialC Iigaml'llt. J Orthop UI.:.... 13(2). 191-200. Noyl's. F.R .. t:l al. (l977a). Funt:lional propl'nics of knee li :1I11,;:-Uls and alterations induced b\' immobilization. Cl~ On/lOp. 123, 210-1-12. . Noyes. F.R .. Grood. E.S.. Nussbaum N.S .. Cooper. S.M (1977b). Effect of intra<lrlicuhlr corticosleroids on li!:.! mCIH properties: A biollll'chiHlic;1! and hislOlo!'!ical stl~ in Rhesus knees. Clill art/lOp. 123. 197-209. ~ No)'l's. F.R .. and Grood, E.S. (1976). Tht: sm:ngth of the antl riot' cruciate ligament in hUIlIi.111S and Rhl'SllS monkl'\' r\gc-rdaled and specil's.rl'lated chan!!,:;.. J BOlli: loint SI/·, 58..\, 1074-1082. ~ Ohkawa, S. (1982). £ffl'ns of orthodolltic forces and anti· i n<lmn~a(or~' drugs on the !llct:hallical sll'l:~ngth of the per odoll II II III 111 Ihl' ral mandibular first molar. ..\/1/ J artho 8/, 498-502. Oxlund, H. (1980). The influence of a lot:~ll injt:t:lion of cort solon the Illcchanic,,1 prolll'nies of tendons ;lnd lit!amelH and thl.' indirect effeci 011 skin . ..\eta Or//top St:tllld, 51(2 231-138_ P.llTY, D.:\.D., Barnes. G.R.G ..•lIlel Cr;Ii!!, :\.5. (1978).:\ com p"rison of the size distribution of 1.'~lI;l!.!cn fibrils in co nective tissucs as a function of age anl possible rdatio between fibril size ;\Ild llll'chanicO'1 propcrlil's. Proc R So Lrmd. 203. 305-321. Prockop. D.J .. Ii: Guz.man, N...\. (1977). Collagl.:1l dist.'asi,.'s ;\ltd the bios,\'lllhcsis of collagen. /-Iosp Prw.:t, Dec, 61-68. Rarnachalldr<\ll. G.N. (1963). Molecular strllC!lln: of collagen. lilt Rev COllllcet TisslU' Uc.~·, /, 127. Rich, A., Crick, F.H.C. (1961), The molecuhll' structure of col· lagen. J .Hol Bioi, 3. 483. Rillo. D.L. Babini. S.M., B<lsn;lk. A., Wrlincr. E., BaHwclwl1. E.. Cocco. J.:\. (1991). Tendinous .md Iig:Ulll'lltoliS hyperlaxity in patients reechoing long-Icrm hemodialysis. J Rht'lIl1ullof. /8(8), 1227-1231. Runc.lgrcn, A. (1974). Physical properties of conncctin: Lissuc as influenced by single and n::pc<I!L.'d pr~gn;\ncics ill thl.' r;IL .·kltl Pilysiol Scam!, Sl/ppl.. 417. Shino, K., Oakt-s, B.W.• I-Ioribe. S.. Nnkaw, K.. Nakamurn. N. (1995). Collagen fibril popuhltions in human ;:Ulterior cru· ciate lig;unent allografts. Electron micfoscopic analysis. A/1/ } Sport.~· Mal, 23(2), 203-208. Simon, S.R. (1994). Orthopedic Basic SciclI"'~. Rosemont, IL: MDS, Snell. R.S. (198-1). Clillicld al/{! FI/Ilctiomd Histology {or ,\It,d· icaf SIUt/CIIIS. Boslon: Lillk" Brown. Stfocchi, R., Dl' P<lsquale. v., Facchini. ,\ .. Rnsp,ulli, M .. Zaffagnini, 5 .. ~'larcacci, ~1. t 1996). Agc:·rcbted changes in human anterior cruciatt: ligament (ACL) collagen fibrils. Ira! ..lWI/ E",bryui. /01(-1).213-120. Slrocchi, R.. DePasqu:lIL', V.. Guizzardi, S., ~I:lrcacci, M.. Ruggeri. A. (1992). Ultr.\slructural modificalions of patdlar tendon fibres lIscd as 'Ulterior cruci.\te ligament (ACI..) repbcclllcnL/uJ! } Allat EII/bryo!, 97(4). 221-228. Tipton, C. =\<1. , James, S.L., Mcrgner, W., Cl al. (1970). Influ· encc of e-xcn::ise on slrl.:ngth of medial (ollatcralliganH:nls of dogs. Alii) Physio!, 2/8. 894. Tohayama. 1-1 .• Beynnon. B. D.. Johnson, R.J., Renstrom, P.A .. Arms S.W. (1996). The effect of <lnlerior cruciate Jig.lInent graft c1ong'llion at the lime of implantation on the biolllechanical behavior of the graft and knee. ,.\/1/ j Sports .\lcd, 24(5),608-614, Viidik, A. (1968). EJ<lslicil)' <IIul lensile strenglh of Ihe antl.'· rior cruci<llc ligament in rabbils :lS influcnced by Iraining. Acta Ph\'sio! Scallll, 7~, 372. Viidik, A. '(1973). Functional propertics collagenous tissues. IlIr Rei' COllllcct Tissue Res, 6. 127. or Viidik, A., D'lllidscn, C.e.. Oxlund, H. (1982). Fourlh tnlem,ltional Congress of Biorhl'oJogy Symposiulll on Mcc.:hHnical Propcnics of Living Tisslles: On fundamental and plll:llOnlcnological models, SlrllC(llrc and mcch:lJlical properties of collagen, elastic and glycosaminoglyc'1I1 complexes. Biorll/:o!o,~y. 19,437. Vogel. 1·1.e. (I 97i). ~'kcll<:lnh.:al and chl'mic:d properties of nil'· iOlls conneclive lissue org.tns in rats as iniluenced by nonsleroidaJ anlirheum:ltic drugs. COI11/l!ct Tisslle! Res. 5, 91. Walsh, W.R., Wiggins. r-,·1.E .. Fadak P.O., Ehrlich, !\·I.G. (1995). Errc:cts of dd;tyed steroid injection on ligamenl healing using a rabbit medial colialen,1 ligament model. BilJllwt..:,-;a!s. 16( I 1), 905-910. Walsh,S .. Fr'lIlk. c.. Shrive. N., Han, D. (1993). Knee immobiliz;Hion inhibits biolllcchanic<llmatllnHioll of the r:\bbit medinl collateral lig:ll11l:111. Clill Orrhop. 297, 253-261. \Vhite, A., H;:lndler, P., & Smith, E.L. (1964). Prillciples o{Bio. cI/!:/IIistry. Ncw York: McGraw-Hill. \Viggir\,o;, [VI. E., Fadale, P.O., B<lITach, H., Ehrilich, ,,"I.G., Walsh. W.R. (1994). Healing ch<lr;tClcrislics of a type I col· Iag('1l01lS SlnlClUre treJh:d wilh corticosteroids. .-\1/1 j SPOI"/S ,\kd, 22(2), 279-188. Woo, S.L.Y.....\n, K.N., Arnoczky, D.\I.\1.. Fithian, D., & ~ly('rs B. (1994). Anatomy, biology, and biomechanics of the lendon, ligaml'nl, ;Iud m..:niscus. In S.R. Simon (Ed.). Or· rJ/Of/at'die Basic SCit'IICt' (p. 51). Ros\,'monl. Il: .-\AOS. Woo, S.L.Y., Gomez, \-1.,.\.. Siles, T.J., Cl :11. (1987). Thc bin· mechanical .llld morphological clwngcs in the Illcdi:d (01latcr,tl ligamelll of the rabbit after imlT1obilization (lnd rc[T\obiliz;llion. ) !3lme )0;1// SlIrg, 69..1(8),1200-1211. \Von, S.L.Y., Gomez, M.A., Amid, D., Ritl('r, ~·l.A., Gdbt.'rman, R.I,!.. Akcson, W.H. (1981). The effects of e.\crcise on the biomcchanical and biochemical propcnies of swine digiwl flc:.\or leudons. } B;ol1l~!c.:h Ellg, 103, 51. Woo. S.L.Y. (1988). Lig;\mellt, tendon. :HH.l join! capsule insertions tu bone. In S.L.Y. Woo, & J. Buckwalter(Eds.), Illjllly (II/(/ Repair of th~ ,HlISCU!oskduli So{t Ti...... uc:s (pp. 133-166). Park Ridge. II. :\meric.w Academy of Or· Ihop<:ledi(..· Surgeons. ,,\', " ST~UC,;,UR~ ;;' ,~ ',,'- Nutrition. mct.'lbolism CEL(:FIBROBJ-ST't;\, ! .~, ~'\.\, \},o' Synthesize Ex(r!:lcclJuI3r Mmix 81 l "" EXTRACEllltLAR ." -\ STRONG AND FlEXIBlES, -{ The -Collagen type J: suPPOrt tensile loads mcc~anjcal response is dependent on the rate at which thl! loads arc applied -Ground $ubsL.,ncc Gel.like matcrialfSt.abilizc collagenous skclC!ton Resistance of high tensile forces with limited elongation 'El:min pS;.;~"NEUAAl Nm:ORK' M~chjlnoreccp[ors , :.~\:' ,- ·Proprioception ·Sensorimotor Mechanism '~\, '- ',;:'-c:,,;,,' FLOW CHA~T 4-1 Common structure and mechanical properties of tendons and ligaments. * *This flow chart is designed for classroom Of group di~cussion. Flow chart is not meant to be exhaustive. .. ..~.~."" ,~,."""".~~ ..•""",,,,,,,,,,,,,,,,,,'",,"_~,_,,~,,,,.,,-"",",,,,,,,,,,,~,_,,,,,!.'<=J ••""......" ~. ;~ ;.' ,( ci ,~ y, "• x0 0 ~ 9 " 0 •~ .~ to w "- Ci -5 ;;; u ·c m .c u w E c "0 m ~ 0 10 ~ ~ 0 "0 c ~ '\:z:: ):Sf <5c .to 0 , 0 E 0 • ~. C 0 .~ ~ 'C c. 0 0 '<;" E 0 ~ ::: u .2 •m ~ C ." ,,:,;iO{:'; i,;vi 0 -"~/,< ~ ';:>'i',·o- ;:{;y;, , / .- to 0 -5 • 0 ;0 i'o ~ s;;:e 6 ,~:<t ;;i ~;; ';~ I' 'r." l: The Cell andfor eXlracel1ubr matrix components , 1? I t' '{ Mechanical Caused InjurJes Injuries in association with local strain type and/or level of stress ;, i{ ~: :{I '" ~I j t ),I! Stre5s +- Str:lin J l Microtr.lumJS that exceed the reparative procc5S ... )l t _._- Stress -Thermal injuries Tendon 'Gun injurie5 'Inherit injuries 'Paratcnonilis ·L1cer3.tions 'Corticosleroids ·Tcndinosis 'Bone fr.lCIUre5 'Endocrinologic and metabolic discases: diabetes -Tendinitis 'Colli5iol1s Tendon's Jtuchment -Rheumatoid arthritis 'ApophY5it is -Congenital diseases -EnthcsopJtics 'Inf~ctions \) i RUPTURE , "1 :l); TOlal/PJrti;a1 ,if FLOW CHART 4·3 Tendon injuries. Clinical examples. * "This flow chart is designed fOf cla5sroom Of group discussion. Flow chart is not meant to be exhaustive. '/01 . .".._. -'--~---'-'-=-- J Function Structure ViSCOElASTIC. MECHANICAC PROPERTY FLOW CHART 4·4 ~This Ligaments structure and mechanical properties.* (PG, Proteoglycan) flow chart is designed for classroom or group discussion. Flow chart is not meant to be exhaustive. .:; The Cell and/or cxtracellular matrix componcnts Injurics in a~sociation Injuries in association with local strain type and/or Icvel of stress with diseases *1 Dccre.lsc in mechanical properties ), ~ i: +- ~i! ;J Stress Stress ·Endocrinologic ;lOd metJbolic diseases: Diabetes Mellitus 'Collisions -l~cerations Tot~I/PJrtial I t -Gun injuries RUPTURE '1 unit area 'Bone fnlCtureS ''.t.,."I 'FLOW CHART 4·5 t Amount of load per 'Iatrogenic Injuries: Corticosteroids I I -+ Strain t Sinin r.ltC Microtraumas that exceed the reparative process Ligament failure. Clinical examples.'" -This flow chart is designed for classroom or group discussion. Flow chart is not meant to be exhaustive. Biomechanics of Peripheral Nerves and Spinal Nerve Roots Bjorn Rydevik, Goran Lundborg, Kje/l O/marker. Robert R Myers Introduction Anatomy and Physiology of Peripheral Nerves The Nerve Fibers: Structure and Function Intraneural Connective Tissue oi Peripheral Nerves The Microvascular System of Peripheral Nerves Anatomy and Physiology of Spinal Nerve Roots Microscopic Anatomy of Spinal Nerve Roots Membranous Coverings of Spinal Nerve ROOlS The Microvascular System of Spinal Nerve Roots Biomechanical Behavior of Peripheral Nerves Stretching (Tensile) Injuries of Peripheral Nerves Compression Injuries of Peripheral Nerves Critical Pressure levels Mode of Pressure Application Mechanical Aspects of Nerve Compression Duration of Pressure Versus Pressure level Biomechanical Behavior of Spinal Nerve Roots Experimental Compression of Spinal Nerve Roots Onset Rate of Compression Multiple levels 01 Spinal Nerve Root Compression Chronic Nerve Root Compression in Experimental Models Summary References Flow Charts • ._~ .. _-... _-~ Introduction The nervouS s)'stcm scn'cs as the bod.:,/s control center and communications nct\vork. As slIch, it has three broad roles: it senses changes in the bod)! and in the external environment, it interprets these changes, and it responds to this interpretation b)· initiating action in the form of muscle contraction or gland secretion. For descriptive purposes, the nen'OllS s)'stem can be divided into two parts: the central nervous system, consisting of the brain and spinal cord, and the peripheral nervous system, composed of the various processes that extend from the brain and spinal corel. These peripheral nerve processes proinput to the central nervous system from senreceptors in skin, joints, muscles, tendons, viscera, and sense organs and provide output from it to effectors (muscles and glands). The peripheral nervous system includes 12 pairs of cranial nen'es and branches and 31 pairs of spinal nerves and their branches (Fig. 5- lil). These branches arc called peripheral nerveS. Each spinal nerve is connected to the spinal cord a posterior (dorsal) root and an anterior root, which unite to form the spinal nerve intervertebral foramen (Fig. 5-1, B-D). The noste,ric)r roots contain fibers of sensory' neurons conducting sensory information from recepin the skin, muscles, tendons, and joints to the celHr.al nervous system) and the anterior roots con~ mainly fibers of motor neurons (those that conimpulses from the central nen'ous system to targets such as muscle fibers). Sh,ortlv after the spinal nerves leave their interforamina, they divide into two main the dorsal rami, which innervate the musand skin of the head, neck, and back. and the generally.' larger and more important ventral rami, innervate the ventral and lateral parts of structures as well as the upper and lower ex~ ,,.,'m;t;,>< Except in the thoracic region, the ventral do not run directly to the structures that they' innervate but first form interlacing networks, or with adjacent nen·'es (Fig. 5-1A). This chapter focuses on both the peripheral nerves spinal nerve roots, which cOlltain not only nerve fibers but also connective tissue elements and vascular structures that encompass the nerve fibers. The possess some special anatomical properties that may sen/e to protect the nente from mechanical damage, for instance, stretching (tension) and com- pression. In this chapter, the basic microanatomy the peripheral nentes and the spinal nCIYC roots revicwed \vith special reference to thesc buil mechanisms of protection. The mechanical behav of peripheral nen'es that are subjected to tension compression is also described in some detail. Anatomy and Physiology Peripheral Nerves of The peripheral nervcs are complcx composite str tures consisting of nen'e fibers, connective tiss and blood vessels. Because the three tissue cleme that make up these nen'es react to trauma in dif ent ways and may each pia)' distinct roles in functional deterioration of the nerve after inju each element is described separatel)'. THE NERVE FIBERS: STRUCTURE AND FUNCTION The term nerve fiber refers to the elongated pn)c (axon) extending from the nerve cell bod)' along w its myelin sheath and Schwann cells (Figs. 5~2 5-3). The nerve fibers of sensory neurons cond impulses frolll the skin, skeletal I11uscles, and jo to the central nervous svstem. The nerve fibers of motor neurons conve)' impulses from the cen nen'ous system to the skeletal muscles, causing m cle contraction. (A detailed description of the chanics of muscle contraction is given in Chapter The nen'e fibers not only transmit impulses also sen'e as an anatomical connection between nerve cell body' and its end organs. This connect is maintained by axonal transport systems, thro which various substances synthesized within cell body (e.g., proteins) are transported from cell body to the periphel)' and in the opposite di tion. The axonal transport takes place at speeds t Vat)' from approxilnatcly 1 to approximatel)" mm per day. Most aXO!lS of the peripheral nervous system surrounded by multilayered, segmented coveri known as myelin sheaths (Fig. 5-3). Fibers with covering are said to be myelinated, whereas th without it (mainly small sensol)' fibers conduct impulses for pain from the skin) are unmyelina The myelin sheath of the axons of the periph nerves is produced b).' flattened cells called Schw cells arranged along the axon (Fig. 5-3). A sheat formed as the Schwann cell encircles the axon Vertebral Intervertebral foramen body Ventral Cl-::>"_--., C2 C3 C4 Cervical nerves B Vertebral arch Dorsal Ventral Thoracic nerves Lumbar plexus Spinal nerve Cauda equina ~~ ~~ MOlor _ -->"",,--,-,'00' I---~'-~~ Dorsal root ganglion Coccygeal nerve A A, Schematic drawing of the spinal cord and the spinal nerves (posterior view). The spinal nerves emerge from the spinal canal through the intervertebral foramina. There are 8 pairs of cervi<al nerves, 12 pairs of thoracic nerves,S pairs of lumbar nerves, 5 pairs of sacral nerves, and 1 pair of coccygeal nerves. Except in the region of the 2nd to the 11 th thoraci< vertebrae (T2-T11), the nerves form complex networks called plexuses after exiting the intervertebral forClmina. Only the main branch of each nerve, the ventral ramus, Sensory rOOI o is depicted. Adapted from rorrord, GJ & Anllgnostako5. N.fl. (J 984). Principles oj Anatomy and PhySiology (.1lh ed.J. New York: Harper & Row. B, Cross-section of the cervi<al spine showing the spinal cord in the spinal canal and the nerve roots exiting through the intervertebral foramina. C, Cross-section of the lumbar spine shOWing the nerve roots of the cauda equina in the spinal canal. 0, Each exiting nerve root complex in the intervertebral foramen consists of a motor root, a sensory root, and a dorsal root ganglion. • \vinds around it many times, pushing its cytoplasm and nucleus to the outside layer. Unmyelinated gaps called nodes of Ranvier lie between the segments of the myelin sheath at approximately I to 2 I11Ill apart. The myelin sheath increases the speed of the conduction of nerve impulses, and insulates and maintains the aXOIl. Impulses arc propagated along the unmyelinated nerve fibers in a slow, continuous way, whereas in the myelinaled nerve fibers the impulses "jump" al a higher speed from one node of Ranvicr to the next in a process called saltatOl)" conduction. The conduction velocity of a myelinated nerve is directly proportional to the diameter of the fiber, which lIsually ranges [Tom 2 La 20 J-l111. MOlar f-ibers -. f: innervate skeletal muscle have large diameters, do sensor~y' fibers that relay impulses associated touch, pressure, heat, cold, and kinesthetic such as skeletal muscle tension and joint poSensory fibers that conduct impulses for dull, diffuse pain (as opposed to sharp, immediate pain) the smallest diameters. Nerve fibers are packed closely in fascicles, \vhieh are further arranged into bUIl1(II·es that make lip the nen'c itself. The fascicles arc the functional subunits of the nCI-ve. the same nerve. \!Vhere the nerves lie close to b INTRANEURAL CONNECTIVE TISSUE OF PERIPHERAL NERVES Successive layers of connective tissue surround the nerve fibers-called the endoneurium, perineuriUl11, ,'mel epineurium-and protect the fibers' continuity (Fig. SA). The protective function of these connec- Sensory nerve root Sensory cell body in dorsal rool ganglion ! Spinal nerve \ \ Motor cell body in anterior horn of spinal cord Dorsal \ Dorsal ./ ramus / Ventral Peripheral nerve ... '- Motor nerve rool I nerve fiber 1 ~'----------- Schematic representation of the arrangement of a typical spinal nerve as it emerges from its dorsal and ventral nerve roots. The peripheral nerve begins after the dorsal ramus branches off. (For the sake of simplicity, the nerve is not shown entering a plexus.) Spinal nerves and most peripheral nerves are mixed nerves: they contain both sensory (afferent) and motor (efferent) nerve fibers. The cell body and its nerve fibers make up the neuron. The cell bodies of the motor neurons are located in the anterior horn of the spinal cord, and those of the sensory neurons are found in the dorsal root ganglia. Here, a motor nerve fiber is shown innervating muscle and a sensory nerve fiber is depicted innervating skin. Adapted from Rydevik, B., Brown, MD., & Lundborg, G. (/98·4). Pathoanatomy and pathophysiology of nerve root compression. ? tive tissue la)rers is essential becausc nerve flbcrs extremely susceptible to stretching and comp sion. The outcnTlOst layer, the epineurium, is loc between the fascicles and superficially in the ne This rather loose connective tissue layer serves cushion during movements of the nel\'e, protec the fascicles from external trauma and maintain the oxygen supply sy'stem via the epineural bl vessels. The amount of epineural connective ti varies among nerves and at different levels wi Spine, 9, 7. .... ,':;.- .. or pass joints, the epineurium is often more ab c1ant than e1scwhere, as the need for protection be greater in these locations. The spinal nerve r are devoid of both epineuriurn and perincuri and the ncrvc fibcrs in the nel\'e root may therc bc morc susceptible to trauma (Rydevik et aI., 19 The perineurium is a lamellar sheath that compasses each fascicle. This sheath has great chanical strength as well as a specific biochem barrier. Its strength is demonstrated by the fact the fascicles can be inflated by fluid to a pressur approximately! 1000 mm of mercury (I-Ig) beFore perineuI'ium ruptuI·es. The barrier function or the perineurium che cally isolates the l1el\'e fibers From their surrou ings, thus preserving an ionic environment of th terior of the fascicles, a special milieu interieur. endoneurium, the connective tissuc inside the f cles, is composed principally of fibroblasts and lagen. The interstitial tissue pressure in the fasci the endoneurial fluid pressure, is normally slig elevated (+ 1.5 ± 0.7 mm Hg [Myers & Pow 1981]) comparecl \\'ith the pressure in surround tissues sLIch as subcutaneous tissue (-4.7 ± 0.8 Hg) and muscle tissue (~2 ± 2 mm Hg). The elev endoneurial fluid pressure is illustrated by the nomenon whereby' incision of the perineurium sults in herniation of nerve fibers. The endoneu fluid pressure may increase further as a resu trauma to the nerve, with subsequent edema. S a pressure increase ma.y affect the microcircula and the function of the nerve. THE MICROVASCULAR SYSTEM OF PERIPHERAL NERVES The peripheral nerve is a well-vascularized struc containing vascular networks in the epineurium perineurium, and the endoneurium. Because inlpulse propagation and axonal transport dep Node of Ranvier Schwann cell Myelin sheath \ \ Axon Node of Ranvier Schematic drawings of the structural features of a myelinated nerve fiber. Ad(1fJted from Sunderland 5. 0978j. Nerves and Nerve Injuries (2nd ed.). Edinburgh: Churchill Livings/one. • on a local oxygen supply, it is natural that the microvascular system has a large reserve capacity. The blood supply to the peripheral nerve as a whole is provided by large vessels that approach the nCl\'C segmentally along its course. \,Vhen these local nutrient vessels reach the nerve, they divide into ascending and descending branches. These vessels run longitudinally and frequently anastomose with the vessels in the perineurium and endoncuriul11. Within the epineurium, large arterioles and venules. 50 to 100 fl-l11 in dial11etel~ constitute a longitudinal vascular system (Fig. SA). \Vithin each fascicle lies a longitudinally oriented capillary plexus with loop formations at various levcis. The capillary system is Fed by arterioles 25 to 150 fLrn in diameter that penetrate the perineurial membrane. These vessels run an oblique course through the perineurium, and it is believed that because of this structural peculiarity, they arc easily closed like valves in the event that tissuc prcssurc inside the fascicles increases (Lundborg, 1975; Myers et aI., 1986). This phenomenon may explain why even a limited increase in endoncurial fluid pressure is associated with a reduction in intrafascicular blood flow. The built-in safety system of longitudinal anastomoses provides a wide margin of safety if the regional segmental vessels arc transectcd. In an ex- Epineurium Perineurium Endoneurium Schematic drawing of a segment of a peripheral nerve. The individual nerve fibers are located within the en· doneurium. They are closely packed in fascicles, each of which is surrounded by a strong sheath, the perineurium. A bundle of fascicles is embedded in a loose connective tissue, the epineurium. Blood vessels are present in all layer of the nerve. A, arterioles (shaded); V. venules (unshaded). The arrows indicated the direction of blood flow. Adapted from Dahlin, L.B., Rydevik, B., & Lundborg, G. (1986). The parhophysiology of nerve entrapments and nerve compression injuries. In A.R. Hargens (Ed.). Effects of lv1echanical Stress on Tissue Viability. New York: Springer-Verlag. • peri mental animal in vivo model, it is extremely difficuh to induce complete ischemia to a nerve by 10surgical procedures. For example, if the whole sciatic··til."'1i nerve complex of a rabbit (15 em long) is surgically Sepal"aled from its surrounding structures and the region"11 nutrient vessels arc ClIt. there is no detectable reduction in the intrafascicular blood flow as studied by intravital microscopic tcch· Y"iGucs. Even if such a mobilized nerve is ellt distally or proximall~'. the intraneural longiwdinal vascular systems cnn maintain the microcirculation at least 7 l~ 8 ern rTom the clil end. If a non mobilized nerve is 0"'.,'" cut, there is still perfect microcirculation even at the vcr")' tip of the nerve; this phenomenon demonstrates the sufficiency of the intraneural vascular collalcrals. Howevcr~ othcr sllldies in rats indicate .that stripping the epineural circulation rronl nerve bundles causes demyelination of subpel'ineural nerve fibers. "swdling" of the most caudal part of the respective dorsal neryc root, called the dorsal rOOl ganglion. The dorsal root ganglia arc locatcd in or close to the intervertebral Foramen. Unlike the nCn!c roots, the dorsal root ganglia arc not enclosed by cerc brospinal fluid and the meninges. Instead. they are enclosed by both a multilayc:rcd connective tissue sheath, similar lo the perineurium of the peripheral w 4 2 Anatomy and Physiology of Spinal Nerve Roots ~ t I I f ! 3 In the earl~' embryological developmental stages, the spinal cord has the same length as the spinal column. However, in the full~: grown individual, the spinal cord ends as the COIlUS mcdullaris, approximately at the level of the fit'st lumbar vertebra. A nerve root that leaves the spinal canal through an intervertebral foramen in lhe lumbar or sacral spine therefore has to pass from the point where it leaves the spinal corel. which is in the lower thoracic spine. lO the point of exit li'om the spine (Fig. 5-5). Because the spinal cord is not present be.low the first lumbar vertebra, dte nervous content of the spinal canal is only comprised of the lumbosacral nerve roots. This "bundle" of nerve roots within the IUI11bar and sacral pan of the spinal canal has been suggested to resemble the tail of a horse and is therefore often called the cauda equina, that is, lail of horse. Two different lypes of nerve roots are ("ound within the lumbosacral spine, ventral/motor roots and dorsal/sensory roots. The cell bodies of (he motor axons are located in the anterior horns of the gray malleI' in the spinal cord, and because these nerve roots leave the spinal cord from the ventral aspect, they are also called ventral roots. The other type of nelVC root is the SenSOl)', or dorsal root: As the name suggests, these nerve roots mainly COll1 prise sensory (Le., afferent) axons and reach lhe spinal cord at the dorsal region of the spinal cord. The cell bodies of the sensory axons are located in a 2 .. 3 7 5 8 6 The intraspinal nervous structures as seen from behind. The vertebral arches are removed by cutting the pedicles (1). A ventral (2) and a dorsal (3) nerve root leave the spinal cord as small rootlets (4). Before leaving the spinal canal, the dorsal root forms a swelling called the dorsal root ganglion (5), which contains the sensory cell bodies. before forming the spinal nerve (6) together with the ventral nerve root. The nerve roots are covered by a cen· tral dural sac (7) or with extensions of this sac called nerve root sleeves (B). ReprodtlCec/ with permission ;,om O/marker. K. (J 99/ J. Spinal nerve rOOI compression. ACt/Ie compression of the cauda equina sludied in pigs. Acta Orthop w Scand. 62. Suppf 242. • nerve, and a loose connective tissue layer called epineurium. ""Vhen the ncnlC root approaches the intervertebral foramen. the root sleeve gradually encloses the nerve tissue more lighLly. The subarachnoid space and the amount of cerebrospinal fluid surrounding each nen'c root pair will thus become gradually reduced in the caudal direction. Compression injury of a nerve root may induce an increase in the permeability of the endoneurial capillaries. resulting in edema formation (Olmarker et aI., 1989b; Rydevik & Lundborg, 1977), This can lead to an increase of the imraneural fluid and subsequent impairment of the nutritional transport to the nerve (Myers & Powell. 1981; Myers, 1998), Such a mechanism might be particularly important at locations where the nerve roots are tightly enclosed by connective tissue. Thus there is a more pronounced risk for an "entrapment syndrornc" within the nerve roots at the intcrvenebral foramen than more central in the cauda equina (Rydevik et aI., 1984), The dorsal root ganglion, with its content of sensory nCt've cell bodies, tighlly enclosed by meninges, might be panicularly susceptible to edema formation. MICROSCOPIC ANATOMY OF SPINAL NERVE ROOTS There are two microscopically different rc.:::gions of the ncnlC roots. Closest to the spinal cord is a central glial segment comprised of glial cells and therefore resembles the microscopic organization of central nervous structures at the spinal cord or the brain. This glial segment is transferred to a nonglial segment in a "dome-shaped" junction a few millimeters from the spinal cord. This nonglial segment is organized in the same manner as the endoneurium of the peripheral nerves, that is, with Schwann cells instead of glia cells. However, some small islets of glia cells also are found in this otherwise "peripherally" organized endoncurium. MEMBRANOUS COVERINGS OF SPINAL NERVE ROOTS The axons in the endoneurium are separated from the cerebrospinal fluid by a thin layer of connective tissue called the root sheath, This root sheath is the structural analogue to the pia mater that covers the spinal cord. There are usually 2 to 5 cellular layers in the root sheath, but as many as 12 layers have been identified, The cells of the proximal part of the outer layers of the root sheath arc similar to the pia cells of the spinal cord, and the cells in the distal part are more similar to the arachnoid cells of the spinal dura. The inner layers of the rOOl sheath are comprised of cells that show similarities to the cells of the perineurium of peripheral nerves. An inlerrupted basement membrane encloses these cells separately. The inner layers of the root sheath constitute a diffusion barrier bet\veen the cl1do~ neurium of the nerve roots and the cerebrospinal fluid, This barrier is considered to be relatively weak and may only prevent the passage or macromolecules. The spinal dura encloses the nerve roots and the cerebrospinal fluid, When the two layers of the cranial dura enter the spinal canal. the outcr layer blends with the periosteum of the pan of the laminae of the cervical vertebrae facing the spinal canal. The inner layers join the arachnoid and become the spinal dura. [n contrast to the rooL sheath, the spinal dura is an effective difTusion barrier: The barrier properties are located in a connective tissue sheath between the dura and the arachnoid called the neurotheliunl. Sirnilar to the inner layer or the root sheath, this ncurothclium resembles the perineurium of the pel'iphcral nen'es. It is suggested thal these t\VO layers in facl form the perincuriulll when the nerve rooL is transformed to a peripheral nerve upon leaving the spinal. THE MICROVASCULAR SYSTEM OF SPINAL NERVE ROOTS Information about the vascular anatomy of the nerve roots has mainly been derived from studies on the vascularization of the spinal corel. Therefore. the nomenclature of the various vessels has been somewhat confusing. A SLlllllTIat·y of the existing knowledge on nerve root vasculature will be presenteel below, The segnlental arteries generally divide into three branches when approaching the intervertebral foramen: (1) an anterior branch that supplies the posterior abdominal wall and lumbar plexus. (2) a posterior branch that supplies the paraspinal muscles and facet joints, and (3) an intermediate branch that supplies the contents of the spinal canal. A branch of the intermediate branch joins the nCI"ve root at the level of the dorsal root ganglion, There are usually three branches from this vessel: one La the ventcal root, one to the dorsal root, and one to the vasa corona of the spinal corel. :7:. !I The branches to the vasa corona of lhe spinal corel, called medullal)' arteries, arc inconsistcnt. 7 to 8 remain or the 128 from the cmbryologiperiod of life, and each supplies more than one segment or the spinal core!. The main medullary in the thoracic region of the spine was discovered by Adamkiewicz in 1881 and still bears his name. The medullary arteries run parallel to the ~~-nt:rve roots (Fig. 5-6). In humans, there are no connections between these vessels and the vascular net work of the nerve roOlS. Because the medullmy feeder artcrics only occasionally supply lhe nerve roots with blood, they have been referred to as the extrinsic vascular system of the cauda equina. The vasculature of thc nerve roots is formed by branches from the intermediate branch of the segmental arlclY distally and by branches from lhe vasa corona of the spinal cord proximally. As opposed to the medullat)' arteries, this vascular network has been named the intrinsic vascular system of the cauda equina. The dislal branch to the dorsal root first forms lhc ganglionic plexus within the dorsal root ganglion. The vessels run within the outcr layers of the root sheath, called cpi-pial tissuc. As there are vesscls coming from both distal and proximal directions, the ncr've roots are supplied by two separate vascular systems. The two syslems anastomose at approximalely two thirds of the nerve rool lenglh from the spinal corel. This location demonstrates i:\ region of a less-developed vascular nClwork and has been suggested lo be a panicularly vulnerable site of lhe nerve rools. The arterics of the intrinsic system send branches down to the decper parts of thc nerve tissue in a Tlike manner. To compensate for elongation of the nerve roots, the m·teries are coiled both longitudinally and in thc steep running branches bClwccn the different fascicles (Fig. 5-6). Unlike peripheral nerves, the vcnules do not course togethcr with the arteries in the nerve roots but instead usually have a spiraling course in the deeper pans of the nerve. There is a barrier of the endoneurial capillaries in periphcral nerves called the blooel-nerve barrier, which is similar to the blood-brain barrier of lhe central nervous system (Lundborg, 1975; Rydevik & Lundborg. 1977). The presence of a corresponding barrier in netlle roots has been queslioned. If presem, a blood-nerve barrier in nerve rOOlS does not seem to be as well developed as in endoneuria I capillaries of peripheral nerve, which implies that edema may be formed more easily in nen.,'c roots lhan in peripheral nelves (Rydevik el aI., 1984). M Schematic presentation of some anatomical features of t intrinsic arteries of the spinal nerve roots. The arterioles within the cauda equina may be referred to either the e trinsic (1) or the intrinsic (2) vascular system. From the su perficial intrinsic arterioles are branches that continue al most at right angles down between the fascicles. These vessels often run in a spiraling course, thus forming vasc lar "coils" (3). When reaching a specific fascicle they branch in a T-like manner, with one branch running cranially and one caudally, forming interfascicular arterioles (2b). From these interfascicular arterioles are small branches that enter the fascicles, where they supply the endoneurial capillary networks (2c). Arterioles of the extrinsic vascular system run outside the spinal dura (4) and have no connections with the intrinsic system by local va cular branches. The superiicial intrinsic arterioles (2a) are located within the root sheath (5). Reproduced with permi sion from Olmarker. K. (1991 J. Spinal nerve rooe compression. Acuee compression of the cauda equina seudied in pigs. Acta Qnhop Scand, 52. Supp1242. Biomechanical Behavior Peripheral Nerves of External trauma to the extremities and nerve e trapment may produce mechanical deformation the peripheral nerves that resulls in the deterior tion of nerve Function. If the mechanical trauma e ceeds a certain degree, the nerves' buill-in mech nisms of prolcction may not be sufficient, resulli in changes in nerve sLructure and function. Co . mon modes of nerve injlll)' are stretching and co pression. which may be. inflicted, respectively, rapid eXlension and crushing. STRETCHING (TENSILE) INJURIES OF PERIPHERAL NERVES ties, and the nerve behaves 1110re !ike a plastic m terial (I.e., its response to the release of loads is i complete recovery). Although variations exist in the tensile streng of various nerves, the maximal elongation at t elastic limit is approximately 20(~fi, and comple structural failure seen1S to occur at a maximu elongation of approximately 25 to 30(k;. These v ues are for normal nerves; injury to a nell/e may duce changes in its mechanical properties, name increased stiffness and decreased elasticit\'. Stretching, or tensile, injuries of peripher nerves are usually' associated with severe acciden such as when high-energ.\' tension is applied to t brachial plexus in association with a birth~relat injur.\'. as a result of high-speed vehicular co!lisio or after a fall from a height. Such plexus injuri may result in partial or total functional loss of som or all of the nerves in the upper extremity, and t consequent functional deficits represent a conside able disability in terms of sensory and motor lo The outcorne depends on which tissue componen of the nerves are damaged as well as on the exte of the tissue injury'. Of clinical importance is the o servation that there can be considerable structur damage (perineurial sheath injuries) induced stretching with no visible injury' on the surface the nerve (Case Study 5-1). Nerves arc strong structures with considerable ten- sile strength. The maximal load that tained b~/ call be SlIS- the median and ulnar nerves is in the rangc of 70 to 220 newtons (N) and 60 to 150 N, rcspectively. These f-igurcs afC of academic interest only because severe intraneural tissue damage is produced b:v tension long before a nerve breaks. A discussion of the elasticity and biomcchanical properties of nerves is complicated by the fact that nerves are not homogeneous isotropic I11aterials but, instead, composite structures, \vith each tissue component having its O\VI1 biomechanical properties. The connective tissues of the epineurium and perineurium arc primarily longitudinal structures. \,Vhen tension is applied to a nen'e, initial elongation of the nerve under a ver.v small load is followed by! an interval in \vhich stress and elongation show a linear relationship characteristic of an elastic material (Fig. 5-7). As the limit of the linear re~ gion is approached, the nerve fibers start to rupture inside the endoneurial tubes and inside the intact perineurium. The perineurial sheaths rupture at approximately 25 to 30i?-c elongation (ultimate strain) above in vivo length (Rydevik et aI., 1990). Aftcr this point, there is a disintegration of the elastic proper- " (L ~ ~ ~ ~ iii 12 11 10 9 8 In-situ strain: 11.0 :!:. 1.5% In-situ stress: 0.05 :c: 0.3 MPa 7 6 5 4 3 2 1 0 Ultimate stress: 11.7 :.:: 7 MPa Ultimate strain: 38.5 :::: 2.0'% 0 4 8 12 16 20 24 28 32 36 40 Strain ('%) The stress-strain behavior of a rabbit tibial nerve. The nerve exhibits a low stiffness toe region of approximately 15 % and begins to retain significant tension as the strain increases beyond 20%. Reproduced with permission from Rydevik. 8.L., K\·van, M.K., Myers, R.R., et al. (1990). An in. vitro mechanical an histological study of acute stretching on rabbit tibial nerve. J Orthop Res, 8, 694-707. "-- ,-"" ----0 I . B~ichial Plexus Palsy "ri.~.i-ipg the birth process, a newborn suffered a trac- .~fit9~. injury in his left brachial plexus. A few months .-0'.> .1,: '~t~~:·.h~'presents with the upper left arm in a static posi'~i?r-.()fadduction, internal rotation of the shoulder, exi~h?i{)pof the elbow, pronation of the forearm, and flexi'9r'l9fthe wrist. He does not respond to sensory stimulus 'iri;~i~>shoulder and presents biceps and brachioradialis ~',-"~fje.x.ia. A sudden deformation and high tensile stress in· . j.. '~'. 'f',""'·. rr~~.in the (5-(6 nerve roots affected the mixed (motor :'.'~·ilnd·.s.en.sory) neural functions, mainly the muscles respon·",-slbh~,'for the scapulohumeral rhythm (see Chapter 12). .1' ;:\)'AHE.rb's palsy is diagnosed. The sudden elongation J7;r~uHer~~d during the traction can lead to structural dam~;.~ge an9 reduction in the transverse fascicular cross-sec{ftio.f1~(area, producing impairment of the intraneural :~",,:vascu.lar flow and impulse transmission. 1', ·.In less severe cases, functional restoration may occur t.- within·weeks or months. In more severe cases, healing may "i·i:~~X~k(place during the first 2 to 3 years, but if the structural :~r~erveJnjury is severe, considerable long-term functional :s<'disability can result, If structural derangement of the nerve :i·::i.trunk has taken place, nerve grafting may be required. ~. ):i:' , " J High-energy plexus injuries represent an extreme type of stretching lesion caused by sudden violent trauma. A differcm stretching situation of considerable clinical interest is the sutllling of the two ends of a CuL nerve under moderate tension. This situation occurs when a substantial gap exists in the continuity of a nerve trunk and the restoration of the continuity requires the application of tension to bring the neryc ends back togethel: The moderate, gradual tension applied to the nerve in these cases may stretch and angulate local feeding vessels. It ma.y also be sufficient to reduce the transverse fascicular cross-seclional area and impair the intraneural nutritive capillary flow (Fig. 5-8). As the sutured nelve is stretched, the perineurium l'ightens; as a result, the endoneurial fluid pressure is increased and the intrafascicular capillaries may' be obliterated. Also, the flow is impaired in lhe scgmcn· lal, feeding, and draining vessels, as it is in larger vessels in the epineudulll, and at a certain stage lhe intraneural microcirculation ceases. Intravital observations of intraneural blood now in rabbit tibial nel'ves (Lundborg & Rydevik, 1973) showed that an elongation of 80M induced impaired venular no\\' and that even greater tension produced continuous impairment of capillary and arteriolar now until. at 15% elongation, all intraneural microcirculation ceased completely. For the same nerve, an elongation (strain) of 60/0 induced a reduction of nerve action potential amplitude by 709'0 at I hour with recover)' to normal values during I-hour restitution. At 12 elongation, conduction was completely blocked by hour and showed minimal recovery (Wall et a 1992). Such data have clinical implication in ner rcpail~ limb trauma, and limb lengthening. A situation of even more gradual stretching, a plied over a long time, is the growth or intraneu tumors such as schwannomas. In this situation, t nerve fibers are forced into a circumferential Cou around the gradually expanding tumor. Function changes in cases of sllch very gradual stretching a oftcn minimal or nonexistent. COMPRESSION INJURIES OF PERIPHERAL NERVES It hns long been known that compression of a ner can induce symploms stich as numbness, pain, a muscle weakness. Thc biological basis for the fun tional changes has been investigatcd eXlensiv (Rvdevik & Lundborg, 1977; Rydevik et aL, 1981), these investigations (Fig. 5-9), even mild compre sion was observed lo induce structural and fun lion'll changes, and the significance of mechanic Elongation Schematic representation of a peripheral nerve and its blood supply at three stages during stretching. Stage I: T segmental blood vessels (S) are normally coiled to allow for the physiological movements of the nerve. Stage II: U der gradually increasing elongation, these regional vess become stretched and the blood flow in them is impaire Stage III: The cross-sectional area of the nerve (represen within the circle) is reduced during stretching and the in traneural blood flow is further impaired, Complete cessa tion of all blood flow in the nerve usually occurs at approximately 15% elongation. Adapted from Lundborg. G. Rydevik. 8. (1973). ENeas 0; strecching the tibial nerve of the rabbit: A preliminary study 0; (he intraneural circulation and barrier (uncrion of the perineurium. J Bone Joint Surg, 558. 3 Recording electrodes and amplifier r1./,...,/ Stimulating electrodes Compression chamber I Compressed ajr inlets Schematic drawing of an experimental setup for studying deterioration of nerve function during compression. Adapted from Dahlin, L.B.• Rydevik, 8.• & Lundborg. G. (1986). The pathophysiology of nerve entrapments and nerve compres' sian injuries. In A.R. Hargens (Ed.). Effeers of Mechanical Stress on Tissue Viability. New York: Springer- Verlag. sure levels (approximately 30 to 80 mm Hg) may in ducc intrancural edema. \vhich in turn may becom organized into a fibrotic scar in the ncn'e (Ryclevi & Lundborg, 1977), Compression at approxirnately 30 111m Hg als brings about changes in the axonal transport sys tems, and long-standing compression may thus lea to depletion of axonally transported proteins dista to the compression site. Such blockage of axona transport induced by local compression (pinching may cause the axons to be marc susceptible to ad ditional compression distally. the so-called double crush syndrome. Slightly higher pressure (80 mm Hg, for example causes complete cessation of intraneural blood flow the nerve in the locally compressed segment be comes completely ischemic. Yel, even after 2 hour or more or compression, blood flo\\' is rapidly· re stored whcn the pressure is released (Rydevik ct aI 1981), hen higher levels of pressure (200 to 40 mm Hg, for example) applied directly to a nen'c ca induce structural nerve fiber damage and rapid de terioration of nerve function, with incomplete rc covel)· after even shortel· periods of compression Hence, the magnitude of the applied pressure an the severity of the induced compression lesion ap pear to be correlated. factors such as pressure level and mode of compression became apparent. Mode of Pressure Application Critical Pressure Levels Experimental and clinical observations have re· vealed some data on the critical pressure levels at which disturbances occur in intraneural blood flow, axonal transport, and nerve function. Certain pressure levels seem to be well defined with respect to structural and functional changes induced in the nerve, The duration of the compression also influences the development of these changes, At 30 mm Hg of local compression, functional changes may occur in the nerve, and its viability may be jeopardized during prolonged compression (4 to 6 hours) at this pressure level (Lundborg et aI., 1982), Such changes appear to be caused by impairment of the blood flow in the compressed part of the nerve (Rydevik et aI., 1981), Corresponding pressure levels (approximately 32 111111 Hg) were recorded close to the median nen.re in the carpal tunnel in patients with carpal tunnel syndrome, while in a group of control subjects the pressure in the carpal tunnel averaged onl~,.. 2 mm Hg. Longstanding or intermittent compression at lo\\' pres- The pressure level is not the only faclOr that influ ences the severity of nerve injlll)' brought about b compression, Experimental and clinical evidenc indicates that the mode of pressure application also of major signiflcance. "Its importance is illu trated by the fact that direct compression of a nerv at 400 mm Hg by means of a small inflatable cu around the nCI've induces a more severe nerve injur than does indirect compression of the nerve at 100 mrn Hg via a tourniquet applied around the extrem ity. Even though the hydrostatic pressure acting o the nerve in the former situation is less than ha that in the latter, the nerve lesion is more sever probably because direct compression causes a mor pronounced deformation of the nerve (especially a its edges) than does indirect compression, in whic the tissue layers between the compression devic and the nerve "bolster" the nerve, One nlay also con clude that the nerve injllly caused by compression not directly related to the high hycfrostatic pl'essu in the center of the conipressed nerve segment bu instead is more dependent on the specific mechan cal deformation induced by the applied pressure, Mechanical Aspects of Nerve Compression ;!: I Electron microscopic analysis of the deformation of the nerve fibers in the peroneal nerve or the baboon hind .limb induced b~,' tourniquet compression demonstrated the so-called edge effect; that is. a specific lesion was induced in the nerve fibers at both edges of the compressed nerve segment: the nodes of Ranvicr were displaced toward the noncompressed parts of the nct·ve. The nerve fibers in the center of the compressed segment, where the hydrostatic pressllt·c is highest. generally were nOl affected acutely. Tht.: large-diameter nerve fibers were usually affected, but the thinner fibers were spared. This finding confirms theoretical calculations that indicate larger nerve fibers undergo a rdatively greater deformation than do thinner fibers at a given pressure. It is also known clinically that a compression lesion of a nCI-vc IIrst affects the large fibers (e.g., those that can)' molOr function). while the thin fibers (e.g.. those that mediate pain sensalion) arc often preserved. The intraneural blood vessels have also been shown to be injured at the cdges of the compressed segment (Rydevik & Lundborg, 1977), Basically, the lesions of nerve libers and blood vessels seem to be consequences of the pressllre gradient. which is maximal just al the edges of the compressed segment. In considering the mechanical effects on nCI-ve compression, keep in mind that the effect of a given pressure depends on the way in which it is applied, ils magnitude and duration. Although pressure may be applied with a variety of spatial distributions. two basic types of pressure applications arc generally encountered in experimental settings and in pathological conditions. One type is uniform pressure applied around the entire circumference of a longitudinal segment or a nerve or exu'emity. This is Ihe kind of purely radial pressure thai is applied by the common pneumatic tourniquet. It has also been used in miniature apparatlls to produce controlled compression of individual nerves (Rydevik & Lundborg, 1977) (Fig, 5-9). Clinically, Ihis Iype of loading on a nerve probably occurs when the pressure on the rnedian nenre is elevated in the carpal tunnel. producing a characteristic syndrome. Another type mechanical action takes place when the nerve is compressed laterally_ This is the kind of defol-mation that occurs if a nelve or extremity is placed between two parallel nat rigid sur faces that arc then moved toward each othel~ squeezing the nerve or extremity. This type of deformation occurs if a sudden blow by a rigid object squeczes a nervc against the surface of an undcrly~ ing bone_ h may also occur when a spinal nerve compressed by a herniated elisc (Case Study 5·2). The details of the deformation or a nerve may quite difTerent in these two cases of loading. In un form circumferential compression like that appli Sciatic Pain 35-year-old male construction worker has chronic low back pain radiating below the- Ie-f! knee that is more severe with lifting activities and prolonged posi, tions. After a careful examination. certain neurological signs were found. Positive straight leg raising and L5 motor and sensory functions were affected. A MRI shows a herniated disc at level L4-L5 with posterolateral protrusion. which laterally compresses (he left L5 nerve root. Compression of !he nerve deforms it toward a more elliptical shape. increasing strain and stress loads_ The effects of the pressure and mechanical deformation resultant from the load affects the nerve tissue, its nutrition, and the transmission function. in· flammation of the nerve root, induced by the nucleus pu!posus, may sensitize the nerve rOOt so that mechanical ne-rve root deformation causes SCiatic pain (Case A Study Fig 5·2·1), or R .:.-- ~I _.' Case Study Figure S-2·1. y factor at both high and low pressures, but ischemia pl[t~·s a dominant role in longer-dunHion compression. This phenomcnon is illustraled b\' the fact that direct ncr'vc compression at 30 111111 Fig for ) to 4 hours produces reversible changes, whereas pro· longed compression a(}()\'c this time period at this pressure level ma~' cause irreversible damage [0 the nCI"e (Lundborg CI aI., 1982; Rydevik el aI., 1981), Corn pression at 400 I11Ill Hg causes a much more severe nervc injur~1 arter 2 hours than after 15 minules. Such information indicalcs that Cn:n hiuh pressure has to "acl" for a certain period of lime for injur.v to occur. These data also give some informalion about the viscoelastic (time-dependent) properties of" peripheral nCI"\'t: tissue. Sufficient time OJ! f!51 elapse for permanent deformation to develop. i , , I I . .-: ,'.' --I------1I-- X .... -,' ";.:,, • I' :' .. ,-", "/': ", " '; ,';"" '/,:,". :," '.: 'f;-i::;: Pressure dislribution ..--, ---c___ A - , . ---..." " ' .. " :',:' B' C' E E' F i,' G . . '",.,. I F' G' x ,, B "iG, 5-11 A, Theoretical displacement field under lateral compression as a result of uniform clamping pressure. B, The origi, nal and deformed cross-sections are shown for maximum elongation in the x direction of 10. 30, and 50%. The vectors shown from A to fJ:, B to a', and so forth, indicate the paths followed by the particular points A, B, and so forth during the deformation. I • or that this kind deformation can trigger r-iring of nerves, resulting in a sensation or pain when the ne"'e Gbers are laterallv compressed. The details or such deformation of nerves and their functional consequences have not been studied extensively and require further research for their elucidation. Duration of Pressure Versus Pressure Level Knowledge is limited regarding the relalive importance o[ pressure and time, rcspectively, in the production of nerve compression lesions. Mechanical faclors seenl to be rehllively more imporlant al higher than at lo\ver pressures. Time is a significant Biol17echal1ical Behavior (Jf Spinal Nerve Roots The nerve roots in the thee;:d sac lack epinei"fl-tl:lw and perin~uril1m, but under tensile loading they exhibit both elasticity and tensile strength. The ultil11,tle load rnr ventral spinal nelye roots from the thecal sac is belween 2 and 22 ! , and 1'01' dorsal nerve roots from the thecal sac the load is between 5 and 33 N. The length of the nerve roots I'rorn the spinal corel to the foramina \'aries From appro:dI11ntely 60 mm at the L I level to approxilmllely 170 mm at the S I level. The mechanical propertics of human spinal nerve roots arc c1ifferelll for an~1 given nerve roOl at its location in the cenlral spinal canal ancl in the lateral intel'vertebral foramina. The ultinwtc load for the inrr4lthecal portion of human SI nel'VC rOOIS at the S I level is appro:"\imatcl~; 13 N, and thal 1'01' the roraminal portion is approximately 73 N. For human nCI'vc roots at the L5 level. the corresponding values arc 16 Nand 71 N, respectively (Fig. 5-12). Thus, the "allles ror ultimate load are approxil11«lel~" five times higher ror the foraminal segment of the spinal ner'vc roots than for the intrathecal pan ion of the same nClVC roots under tensile loading. However, the cross-sectional area of the nerve root in the intervertebral fOI'(llllcn is sie:nificantlv larger than that of the same nerve root in lhe . thecal sac; thus, the ultimate tensile stress was more comparable for the two locations. The ultimate strain LInder tensile loading is 13 La 19% for the human ne,Ye root at the L5-S I level (Fig. 5-13), The nerve roots in the spine arc nol static slruclures: they 1110\'C relati\'e to the surrounding tissues - - Ultimate load: Human sinal nerve roots u '" S 2 rn'" § 5 o'--_-'-_-';:-=----'-'-_-'-_-';-;:"'-_-'-_..J Sl LS N"4 N~4 Diagram illustrating values for ultimate load obtained for human spinal nerve roots under tensile loading. INR, intrathecal nerve roots; FNR, foraminal nerve root. Note the marked difference in ultimate load for the intrathecal and the foraminal portions of the nerve roots. Error bars indi- cate standard deviation. Reproduced with permission from Weinstein, IN., Latviotte, R.. Rydevik, B., er al. (1989), Nerve. In J, \IV, Frymoyer & S.L. Gordon (Eds.). NevI Perspectives on Low Back Pain (Chapter 4, pp. 35-130). Park Ridge. IL: MOS. {Based on a workshop arranged by the National Institutes roots normally' adhere lo the surrounding tiss above and below the intervertebral disc they n:rsc, compression may give rise to intraneural sion, Spencer and associates (1984) measured contact force between a simulated disc herniat and a dclcwmed nerve 1'001 in cadavers. Taking area of contact into aeCOUlll, they assumed a con pressure of approximately 400 mm Hg. \<\'ith duced disc height. the contact force and pressure tween the experimental disc herniation and nerve root was reduced, They suggested that th findings may explain in pan why sciatic pain is lieved after chcmol1ucleolysis, and as disc degene tion progresses over time and the disc height ther decreases, [n central spinal stenosis, the mechanics or ne root compression arc completely diFferent. Un these conditions, the pressure is applied circum cntially around the nerve roots in the cauda cqu at a slow, gradual rate, These different deformat factors, together with the fact that the nerve ro centrally within the cauda equina differ comple from the nerve roots located more laterally, clos the discs, ma.y explain some of the dillcrcnt syrm toms found in spinal stenosis and disc herniatio of Hea/rh Ultimate strain: Human spinal nerve roots (NIH) in Airlie, Virginia, USA, May 1988.} _ 1S with every spinal motion. To allow for such motion the nerve roots in the intervertebral foramina, for example. must have the capacity to glide. Chronic ilTitation with subsequent fibrosis around the nerve roots. in association with conditions such as disc herniation and/or Foraminal stenosis. can thus impair the gliding capacity of the nel"VC roots. This produces repeated "microstretching" injuries of the nerve roots even during normal spinal 1110Vcments, which might be speculated to induce yet FUrl her tissue iITitation in the nerve root components, The normal range of movements of nerve root.s in the human lumbar spine has been measured in cadaver experiments. It was found that straight leg raising moved the nerve roots at the level of the intervertebral foramina approximately 2 to 5 mm. Certain biomechanical Factors are obviollsly involved in the pathogenesis of val'ious symptoms induced by nen'e root deformation in association with disc herniation and spinal stenosis and resulting in radiating pain. In disc herniation, only one nerve root is usually compressed, Because individual nerve l c .~ en 10 r"n § 5 S o'----'---';:-J'-----''----'-------',-:-'---'-S1 LS N.=4 N=4 Ultimate strain for human spinal nerve roots under ten loading. INR, intrathecal nerve root; FNR, foraminal ne root. Reproduced with permission from Weinstein, IN., L,1M otte, R" Rydevik, 8., er at. (1989). Nerve. In I \tV Fryrnoyer & Gordon (Eds.). New Perspectives on Low Bc1Ck Pain (Chapter pp. 35-130). Park RieJge, Ii: MOS. IBased on cl workshop arranged by the NatioTlallnstitut.es of Health (NIH) in Air/ie, ginia, USA, May /988.J COMPRESSION OF SPINAL has been moderate interest in the past to nerve root compression in experimental modEarly sludies in the 1950s and 1970s found that roots seemed to be more susceptible La COI11-';l;'pl'essie,n than did peripheral nerves. During recent howevcl~ the interest in nerve rool patho;ggph,ysiology has increased considerably and a oumof studies have been performed thal arc rc.vic,\Vc:ct below. Some years ago. a model was presented to evalthe effects of compression of the cauda in pigs, \\!hich for the first time allowed experimental, graded compression or cauda nCl"Ve rOOlS at known pressure levels (01marker. 1991) (Fig. 5-14). In this Illodel, ,he calida equina was compressed by an inflatable balloon that was fixed to the spine. The cauda cquina "could also be observed through the translucent balloon. This mndel made it possible to stud~' the flow in the intrinsic nerve roOl blood vessels at various pressure levels (Olmarkcr el aI., 1989a). The experiment was designed in a way that the pressure in the compression balloon \Vas increased by 5 mm Hg evcry 20 seconds. Blood flow and vessel diameters of the intrinsic vessels could simultaneously be obsClycd through the balloon using a vilal microscope. The average occlusion pressure for the arterioles was !"ound to be slightly below and directly related to the systolic blood pressure. and the blood Ilow in the capillary networks was intimately depcndent on the blood flow of the adjacent venulcs. This corroborates the assumption that venular stasis may induce capillary stasis and thus changes in the microcirculation of the nerve tissue and is in accordance with previous studies in which such a mechanism has been suggested as involved in carpal tunnel syndrome. The mean occlusion pressures for the venules demonstrated large val"iations. However, a pressure of 5 to 10 mOl Hg was found to be sufficient for inducing venular occlusion. Because of retrograde stasis, it is not unlikely to assume that the capillary blood flow will bc affected as well in slIch situations. In the same experimental SCt-lIp. the effects of gradual decompression, after initial acute compression was maintained for only a short while. were studied. The average pressure for starting the blood lIow was slightly lower at decompression than at comprcssion For artcrioles. capillaries. and vcnulcs. _ . . . .'--->-->--- Schematic drawing of an experimental model. The cau equina (A) is compressed by an inflatable balloon (8) t is fixed to the spine by two l-shaped pins (C) and a ple glass plate (0). Reproduced with permission from O/marke Rydevik. 8., & Holm. S. (1989a). Edema (ormation in spina/ nerve roars inducecl by experimental. graded compression. experimental srudy on the pig cauda eq'Jina ~'.Jirh special re 1 ence co differences in effects between rapid and slow onsec compression. Spine. ld. 579. >------- Howevcc with this protocol a full restoration or blood flow did not occur until thc compression lowcred from 5 to 0 mill Hg. This observation ther supports the previous hypothesis that vasc impairment is present even at low pressure leve A cornpression-induced impairmenl of the culature may thus be one mechanism for nerve dysfunction because the nutrition of the nerve will be affected. However, the nerve rools will derive a considerable nutritional supply via d sion from the cerebrospinal nuicl. To assess compression-induced eFfects on the tolal conlr tion to the nerve rools. an experiment was desig in which jH-labelcd melhyl-glucose was allowe be transported to the nen/c tissue in the c pressed segment via both the blood vessels and cerebrospinal fluid diffusion aflcI- systemic in tion. The results showed that no compensa mechanism from cerebrospinal fluid diffusion could be expeclr:d at the low pressure levels. On the COil Lntl)·, to 111m Hg compression was sufficient to induce a 20 to 300/0 reduction of the transport or mClhyl·glucosc to the nerve r001S, as compared with the control. \·Ve know 1"1'0111 experimental studies on peri phend nen!cs that compression Illay also induce an increase in the vascular permeability, leading to an intraneural edema formation. Such edema may' increase the cndoncuriul Fluid pressure, which in turn may impair the endoneurial capillary blood flow and jeopardize the nutrition of the ner'vc rOOls. Because the edema Llsually persists for some lime after the rClllo\'al of a compressive agenl. edema may negativel~r affect the nerve root for n longer period than the compression itself. The presence of intraneural edema is also related to the subsequent formation of intraneural fibrosis and may therefore contribute to the slow recaVCI)" seen in some patients with nellie compression disorders. To assess if' intraneural edema also may form in nerve roots as the result of cornprcssion, the distribution of Evan's blue-labeled albumin in the nerve tissue \vas analyzed after compression at various pressures and at various durations (Olmarker el aI., 1989b). The study showed thal edema was formed even at low pressure levels. The predomi· nant location was at the edges of the compression zone. The function of the nelYC roots has been studied by direct electrical stimulation and recordings either on the nenie itself or in the corresponding muscular segments. During a 2- hour compression period, a critical pressure level for inducing a reduction uf MAP-amplitude seems to be located between 50 and 75 111m Hg. Higher pressure levels (100-200 mm Hg) may induce a tOlal conduction block \vith var.ying degrees of recovery after compression release. To stud)! the effect's of compression on sensory nerve fibers, electrodes in the sacrum were used to record a compound nerve action potential after stimulating the senso(v nerves in the lail, that is, distal to the compression zone. The resuhs showed thaI the sensory fibers are slightly more susceptible to compression than are the motor fibers. Also, the nerve roots are more susceptible to compression injury if the blood pressure is lowered pharmacologically. This further indicates the importance of the blood supply to maintain the functional properties of the nerve roots. ONSET RATE OF COMPRESSION One faclor that has not been fully recognized compr-ession trauma of nerve tissue is the onset r of the compression. The onset rate, that is. the ti from start to full compression. may vnry clinica rrom fractions of seconds in lraUI11atic condition months or years in association with degenerat processes. Even in the clinically' rapid-onset ra there ma:v be a wide variation or onset rales. \V the presented model. it was possible to vary the set time of the applied compression, Two onset ra have been investigated, Either the pressure is p sent and compression is started by flipping switch of the compressed-air system used to inl the balloon or the compression pressure leve slowl~1 increased during 20 seconds. The first on rate was measured at 0.05 to 0.1 seconds. thus p viding a rapid inflation of the balloon and a ra compression onset. Such a rapid-onset rate has been found to ind more pronounced effects on eden1a formati mcth~ll-glucose transport. and impulse propagat lhan the slow-onser rale (Olmarker, 199 I). Rega ing methyl-glucose transport, the results show t the levels within the compression zone arc m pronounced at the rapid than at the slow onSd r at corresponding pressure levels. There was als striking difference between the two onset ra when considering the segments ourside the co pression zones. 1n the slow-onset series, the le approached baseline values closer to the compr sion zone than in the rapid-onsel series. This m indicate the presence of a more pronounced ed zone edema in the rapid-onset sel"ics, with a sub quent reduction of the nutritional transport in nerve tissue adjacent to the compression zone. For the rapid-onset compression, which is lik lo be more closely related to spine trauma or d herniation than to spinal stenosis, a pressure of mill Hg maintained for only 1 second is surflcien induce a gradual impairment of nc,'ve conduct during the 2 hours studied aftcr the cornpress was ended. Overall, the mechanisms for these p nounced differences between the different on rates arc not clear but may be related to dilTeren in the displacement rates of the compressed ne tissue toward the uncompressed parts, as a resul the viscoelastic propcnics of the nellie tissue, S phenomena may ICfld nOl only ro structural dam to the nerve fibers but also to stn.lcturfll change the blood vcssels with subsequcnt edema formati The gradual formation or intraneural edema may also be closely" related to observations of a graduall.\' increasing dilTcrcncc in nerve conduction impair- ment between the two onset rates (Olmar-ker et al.. 1989b). MULTIPLE LEVELS OF SPINAL NERVE ROOT COMPRESSION .' . ••1. !'atients with double or multiple levcls of spinal stenosis seem to have morc pronounced symptoms than do patients with stenosis only at one level. The presented model was modified to address this interesting clinical question. Using two balloons at two adjacent disc levels. which resulted in a 10-mm uncompressed nerve segment bctwcen the balloons, induced a much more pronounced impairment of nerve impulse conduction than previollsl~' had been round at corresponding pressure levels (Olmarker & Rydevik. 1992). For instancc. a pressurc of 10 mm I-1g in two balloons induced a 60% reduction of nerve impulse amplitude during 2 hours of COIllpression, whereas 50 111m I-Ig in one balloon showed no reduction. The mechanism ror the diFrerence between single and double compression may not simply be based on the fact thal the nerve impulses have lo pass more than one compression zone at double-level compression. There may also be a mechanism based on the local vascular anatomy of the nerve roots. Unlike for peripheral nerves, there are no re· gional nutritive arteries from the surrounding structures to the intraneural vascular system in spinal nerve roots. Compression at two levels might therefore induce a nutritionally impaired region betwcen the two compression sites. In this way. the :.~: segment nffected by the compression would be "', widened from one b,~lIoon diarneter (10 mm) to two balloon diameters including the inLCrjacent nerve segment (30 mm). This hypothesis was partly confirmed in an experiment on continuous analyses of the total blood flow in the ul1compresscd nerve scg· ment located bet ween two compression balloons (Takahashi et al.. 1993). The results showed that a 64% reduction of total blood now in the uncoll1pressed segment was induced when both balloons were inflated to 10 mm Hg. At a pressure close to the systemic blood pressure lhere was complete ischemia in the nerve segment. Thus, experimental evidence shows that the blood supply to the nerve segment located between two compression sites in nerve roots is severely impaired although this Tlel'VC segment itself is ullcomprcssed. Regarding n conduction, the effects were much enhanced i distance between the compression balloons wa creased frorn one \'cn~bral segment La two v bral segments (Olmarker & Rydevik. 1992). Thi dicates that the functional impairment may directl:v related to the distance between the compression sites. CHRONIC NERVE ROOT COMPRESSION IN EXPERIMENTAL MODELS Th~ discLlssion of compression-induced effect nerve roots has deall primarily wiLh acute c pression, that is, compression that lasts for s hours and with no survival of [he animal. To be mimic various clinical situations, compres must be applied over longer periods of time. T arc probably man." changes in the nerve Li such as adaptation ofaxons and vasculature, will occur in patients but cannot be studied in perimental models using only I to 6 hours of c pression, Another important factor in this con is the onset rale that was discllssed previousl) clinical syndromes with nerve root compress lhe onset time may in many cases be quite s For instance, a gradual remodeling of the v brae to induce a spinal stenosis probably lead an onset lime of many years. Jt will of cours dirficult to mimic such a situation in an exp mental modcl. It will also be impossible to control over the pressure acting on the nerve r in chronic models because or the remodeling adaptation or the nerve tissue to the applied p sure. However, knowledge of the exact pressur probably of less importance in chronic tha acute compression situations. Instead, chr models should induce a conLrolled compres with a slow onset time that is easily reproduc Such models may be well suited for studies pathophysiological events as well as inlcrven b~' surgery or drugs. Some attempts have b made to induce such compression. Dclamarlcr and collaborators (1990) present model on dog cauda equina in which they appli constricting plastic band. The band was tighte nrollnd the thecal sac to induce a 25, 50, or 750/ duction of the cross-sectional area. The band left in place for various limes. Analyses were ~forl1led and showed both stnlctural.and functi changes that were proportional to the degree of striction. Experimental study to analyze the effects on nerve conduc· from Cornefjord, 1''11., 5ato, K., O/marker, K., et 011. (1997), tion velocity of nucleus pulposus (1), the combination of nucleus pulposus and compression (2), and compression only (3). The nucleus pulposus and the constrictor were applied to the first sacral nerve root in pigs. The contralateral nerve root served as a control. Reproduced with permission morJe! for chronic nerve root compression studies. Present To induce a s!c)\ver onset and more controlled compression, Cornefjord and collaborators (1997) used a constrictor to compress the nerve roots in the pig (Fig. 5-15). The constrictor was initially! intended for inducing vascular occlusion in experimental ischemic conditions in dogs. The constrictor consists of an outer metal' shell that on the inside is covered with a material called amaroid that expands when in contact with fluids. Because of the metal shell, the amaroid expands inwards with a maximum expansion after 2 weeks, resulting in compression of a nerve root placed in the central opening of the constrictor. Compression of the first sacral nerve root in the pig resulted in a significant reduction of nerve conduction velocity and axonal injuries using a constrictor with a defined original diameter. An increase in substance P of a porcine mode! (or controlled slow-onset compression analyses of anatomic aspects, compression onser rate, and phologic and neurophysiologic effects. Spine, 22. 946-957 in the nen'(' root and the dorsal root ganglio lowing such compression also has been fo Substance P is a neurotransmitter that is re to pain transmission. The study may thus pr experimental evidence that compression of roots produces pain. The constrictor model has also been use stud.\' blood flow changes in the nerve root v lature. It could then be observed that the b Flow is not reduced just outside the compre zone but significantly reduced in parts o nerve roots located inside the constrictor. In context, note that in case of disc herniation nerve root may become sensitized by substa from the disc tissue (nucleus pulposus) so mechanical root deformation can induce nounced sciatic pain. ',:",f;1>Thc peripheral nerves arc composed of nerve }t;,,:rs, layers or connective tissue, and blood vessels. ;~{2 The nCIYC f'ibers arc extremely susceptible 10 ruuma but because they arc surrounded by sliccessive layers of connective tissue (the epineuriulll and 1?~rinc·urillm)J they arc 111cchanically protected. ;:.'$§';}Strctching induces changes in intraneural bl~~d noll' m~d nerve Iiber S[;'llctllre before the trunk ruptures. Comrression of a nerve can cause ItlJury to nerve fibers and blood vessels in the nerve, 'mail,lv at the edges of the compressed nerve segV>:ment. but also by ischemic mechanisms. j"."'j 5- Pressure level, duration of compression. and mode of preSSllre application arc signillcant variables in the development of nen'c injury, 6.. Spinal nerve roots arc anatomically different from peripheral nerves and therefore react differently to mechanical defonnalion, T: Spinal nerve roots arc morc susceptible than peripheral ncrves to mechanical dcformation, mainly because of the lack of protective connective tissue layers in nerve roots. Myc:rs. R.R .. ~Itlrabmi, H., .& "(}WI:I!. H.C. (1986), Red uel"\'l' hlood flow in l.'delllatous nCllropathics-A bi chanica I lIle(,:h.lOism . .11;LTol'tlscu!al' Ues. 32. 145-151 Myi.'rs, R.R" 8.: Powell, H.e. (1981). Endoncuri,d fluid sure in peripheral neuropathies. In A.R. Hargens Ti..; SIIC flilid PreSSl/re tlml C01lllJOsithm (p. 193). Balti Willi.II11S &. Wilkins. Olillarkl"', K., Rydc\·ik. B., &: Holm. S. (1989a). Edema r( lion in spin£llnl'l'\'l' roots induced by experimelltal. g compression, An experirnenud Sllld~' on lhe: pig c equina with special reference 10 dilTerences in effec 1\\,(.'(.'11 rapid .. Iud slow onset of compression. Spiw:, I..J Olmarkcr. K., Rydc\'ik. B.. Holm. S .. 1..'1 ill. (l9&9b). EffcC CXpl'l'illlental graded compression on blood flow in s nl,.·l'vc roms. r\ vilal microscopic study 011 Ihe po caudn equinll. J Orthop Res, 7,8Ii. Olmarkcl'. K. {1991l. Spinal nerve rOOl compression. compression of dle cauda equina studied in pigs . .. IeU f/mp Scallll, 62. Suppl 2·n. Ollllarker. K. 6.: Rydc\·ik. B. (1992). Single: \'('rstls doubk· COllllll·l,.·ssion. '\n experimental slud.... un till..' porcine c l'qUill,1 with 'lll.t1ySl'S of nen·c impulse condu<:lion pr lies. Cfill On/lOp. 279, 3539. Olmarker. K. & Hasuc, :\1. (1995). Classificalion and p physiology of spinal pain syndromes. III J.1'$. \\\:inSl B. Rydl'\'ik (Eds.), Essc/llial... Ihe Spillt'. RJ\'l'n !\l,.·w York. NY. R~·(!l'\'ik. B.L., Kw~n, \I.K., ;-"'Iyel's, R.R .. l'l al. (1990). vitro mechanical and histological study of <ll..:lHe stret on r;,hbit tibial nerve . .I Onhop Ucs. B, 694-701. Ryd(·\·ik. B. & Lundborg. G. (1977). Pl:rIllt.'ahilily of ncur:ll micl'U\'l'ssels :lnd perineurium following : gradl,.·d cxpl'l'illlenlal ncrn: cOlllprcs~ion. SC(/Iu} J Na:> or COI1Slr Sl/".~. REFERENCES CorndjonJ. lVi., Saw. K.. Olmarkcr. K.. ci :d. (1997). A nHldcl ror chronic n('IT":" roOI compression studies. PrCSi..'IlI~llion of a porcine model for contrnllt·d slow-onset comprl'ssion \\"ith an.tlyscs of anatomic ::lSPCCIS. comprl,.·ssion onSel r~lle, and morphologic and neurophysiologic crfeels. SpilIC, 22. 946-957, [hillin. L.B .. Ry(kvik, B.. &: Lundborg. G. (1986). Thc p.IIIHI' physiology of nelTe eHtrapmellis and nervi..' cOlllpr":"ssion injuries. In A.R. Hargens (Ed.). E!l;'cIS .\kclwllic:al Stress 011 Tisslle Viability. New York: Springer-Verlag. Delamartcr, R.B" Bohlman. I-U-I., Dodge, L.D" el a!. (1990). Experimental lumbar spinal SlCllosis. Analysis of the cortic~d evoked potentials, microvasculature and histop<1thCII· ogy. J BOIlt: JOillf SI/r:.:, 72:\. 110-120. Lundbor!!, G., &: Rvdcvik, B. (1973), Effecls of strctchin!! Ihe libial ~\cl"\"e of lill,.· rabbic: :\ prdimin.ary slUdy of tilt: i;llraneural cin:ul~lliol1 and Ihe b.uTicr funclion of Ihe per· inelll'iul11. J BOllI: JOill1 SlIrg, 558, 390. Lundborg. G. (1975). Structure :\Jld flllll:tioll of the inlra· ncural micrO\'essl,:ls as related lO Iraum:l. edcma formation and nerve fUllclion.) !JOIII.' )oim Surg. 57..1. 938. Lundborg, G., e{ .11, (1982). Median nel'\'e compression ill the C,I1'p,ll tunnel: The fUllclion:l1 respons(.· 10 cxpcrillH.'nlally induced controlll·d pressure. .f f-/mu} SlIrg, 7, 152. Myers, R.R. (1998). :\Iorphology or the peripheral nervous System nnd its relntiollship to neuropathic pain. In T.L ';aksh. C. Lynch III. W.M. Zapol.:-"-1. ~'1azc, J.E Bicbuyck . .& L.J. Saidman (Eds."J. Aw:...tht'S;u: Bio}o~ic FouI/datiolls (pp. 483-514). Phil'lCldphia: LippiIKOll·R~I\,l,.'n. or II. 179. Rydc\'ik, B., Lundborg. G., &. Bagge, U. (1981). Eff(,., gra<k·d compr('ssion nn inlr,1I1cural blood flow. An in sllld.... 011 rabbil tihi.tl ner\'e. J flawl Sure;. 6. 3. Ryd":"\'ik, B.. Bro\\'ll, :\I.D .. &: Lundborg, G. (1 P<llhoanatolll~' and pathoph ....siolog.\· of nel'\'e root pression. Sp;Ilt.'. 9. 7. Ryd(.·\'ik. B.L.. Kw<tn, M.K .. Myers R.R., et :11. (1990). vitro mechanical :\n(\ histological study of ,H:llle strel on r .. hbil tibi:ll nen·l'. J Orrhop U,·.... 8. 694. Spencer. D.L.. ~'lillcr. J.I\., .& BCriolini. J.E. (198·.J). Th fects of intervertebral disc space narrO\\"ing nil the co force bl·twcen the Ill'rn' rOOt and a simlll:llcd disc p sion. Spille 9, 422. SUl1ckrbnd. S. (1978). Nen'c.>: alld Nave III;tlri(~s (2nd Edinburgh: Churchill Livingslone. Takah;lshi, K.. Olm:-.rkc.·r. K., Holm, S .. cl £II. (1993). Do I('vel cauda ('quina cqmprcssion. An experimenlal wilh conlinuous monitoring of intraneural blood fl Onhop Res, II, 104, TOrlora, GJ., & Anagnoslakos. N.r. (1984). Prillci"l AlUlIOI1lY (/ltd Physio}ogl' (41h cd.). N(.·w York: Harper & Wall. E.J .. M:ISSic, J.B.. Kwan. M.K .. CI 011. (1992). E lllcnta! stretch neurop'lthy. Changcs in nerve condu under tension,) BOllt' .foitlf Stl,.~, N-B. 126, Weinstein. J.N" L.li\lotll'. R.. Rydcdk, B. el 011., (1989). N In J.W. Frymoyer 8.: S.L. Gordon (Eds.)_ Nell' 1'('I"spccti 1.0\1' Back Paill (Chapll'r 4, pp, 3S-130). Park Ridg( '\AOS. [Based 011 a workshop ,IIT,lngcu by Ihe N:llion stiwtes of Health (NIH) in AirJil'_ Virginia, USA. May Nerve structure Structural composition and function affected Protects from external trauma -Biochemical barrier properties 'Preserves ionic environment 'Axon's protector -Nutrition Delivery of blood .. Local oxygen supply D Illness or injury _ Peripheral nerve's structure and alteration. Clinical examples.* "This flow (hart is designed for classroom or group discussion. Flow chart is not meant to be exhaustive. Returned blood flow Tl<lnsmission of impulses; intl<lcellular tl<lnspon of material Of importJ.nce (or impulse transmission Myelin production and synthesis of trophic factors Biomechanics o Skeletal Muscle Tobias Lorenz, Marco Campel/o adapred fr Mark I. Pieman, Lars Peters Introduction Composition and Structure of Skeletal Muscle Structure and Organization of Muscle Molecular Basis of Muscle Contraction The MOtor Unit The Musculotendinous Unit Mechanics of Muscle Contraction Summation and Tetanic Contraction Types of Muscle Contraction Force Production in Muscle Length-Tension Relationship Load-Velocity Relationship Force·Time ReJ(ltionship Effect of Skeletal lvIuscie Architecture Effect of Prestretching Effect of Temperature Effect of Fatigue Muscle Fiber Differentiation Muscle Injuries Muscle Remodeling Effects of Disuse and Immobilization Effects of Physical Training Summary References Flow Charts IJ~S'.r' Introduction The muscular sYSlCll1 consists of three III usc Ie.:: l.ypes: the cardiac muscle, which composes the hc[trt; the Sl1100th (nonstriated or involuntary) muscle, which lines the hollow internal organs; and the skeletal (striated or voluntary) muscle, which auaches to the skeleton via the tendons. The focus of this chapter is the role and function of skeletal muscle. Skeletal muscle is the most abundant tissue in the human body, accounting for 40 to 45% of the tolal body weighL. The human body has morc than 430 skeletal !TIuscles, found in pairs on the right and left sides of the body. The most vigorolls movements are produced by fewer than 80 pairs. The muscles provide strength and protection to the skeleton by distributing loads and absorbing shock; they enable Ihe bones to move at the joints and provide the maintenance or body posture against force. Such abilities usuall~' represent the action of muscle groups, not of individurll mllscles. The skeletal muscles perform both dynamic and static work. Dynamic work permits locomotion and the positioning of the body segments in space. Static work maintains body posture or position. In this chapter we describe the composition and structure of skeletal muscle. the mechanics of muscle contraction. force production in muscle. Illuscle fiber differentiation, and muscle remade-ling. Composition and Structure o( Skeletal Muscle An understanding of the biorncchanics of Illuscle function requires knowledge of the gross anatomical structure and function of the musculotendinous unit and the basic microscopic structure and chemical composition of the muscle fiber. STRUCTURE AND ORGANIZATION OF MUSClE The slructuralunit of skeletal muscle is thc muscle fiber, a long cylindrical cell with many hundreds of nuclei. Muscle fibers range in thickness from approximately 10 to 100 fim and in Icngth from approximately 1 to 30 cm. A muscle fiber consists of many myofibrils, which are invested by a delicate plasma membl·ane called the sarcolemma. The sarcolemma is connected via vinculin- and dystrophinrich costameres with the sarcon1ctric Z lines, which represent a pan of the cxtramyofibrillar cytoskcle- ton. The myofibril is made lip of several sarcome that contain thin (actin), thick (myosin), clas (titin), and inelastic (ncbulin) filaments. Actin a myosin are the contractile part of the myofibr whereas titin and nebulin nrc part of the intramy fibrillar cytoskelcton (Stromer et aI., 1998). T myof-ibrils are the basic unit of contraction. Each fiber is encompassed b:y a loose connect tissue called the endomysium and the fibers arc ganized into various-sized bundles, or Fascicles (F 6·1, A & B), which arc in (urn encased in a den connective tissue sheath known as the perimysiu The muscle is composed of several fascicles s rounded by a fascia of fibrous connective tis called the epimysium. In general, each end of a muscle is attached bone by tendons, which have no active contract properties. The muscles form the contractile co ponent and the tendons the series clastic comp nent. The collagen fibers in the perimysium a epimysium are continuous with those in the t dons: together these fibers act as a stnlctural fr,\I work for the attachment of bones and muscle fibe The perim)'sium, endomysium, epimysium, and s colemma act as parallel elastic components. T forces produced by the contracting muscles transmitted to bone through these connective sucs and tendons (Kassel'. 1996). Each muscle fiber is composed of a lal'ge nu ber of delicate strands, thc mvol'ibrils. These the contractile clements of muscle. Their structu and function have been studied cxhausti\'c1y Iighl and eleclron microscopy, and [heir his chemistr~' and biochclllistry have been explain elsewhel'c (Alvidson et aI., 1984; Guvton, 198 Approximately I p.m in c1iamctcl: the myofibrils parallel to each other within the cytoplasm (s coplasm) of the muscle fiber and extend throug out its length. They vary in number from a few several thousand depending on the diameter of ll1uscle fibel~ which depcnds in turn on thc type ll1usclc fiber. The transverse banding pattern in striated m cles repeats itself along (he length of (he mus fiber. each repeat being known as a sal'comerc (F 6-IC). These striations are caused by thc individ myofibrils, which are aligned continuously throu Ollt the muscle fiber. The sarcomere is the fu tional unit of the contractile syslem in muscle. a the events that take place in one sarcomere arc plicated in the others. Various sarcomere build myofibril, various myof"ibrils build the muscle fi and various muscle fibers build the muscle. 1 Schematic drawings of the structural organization of muscle. A, A fibrous connective tissue fascia. the epimysium. surrounds the muscle, which is composed of many bun- dles. or fascicles. The fascicles are encased in a dense con· / nective tissue sheath, the perimysium. B, The fascicles are composed of muscle fibers, which are long, cylindrical, multinucleated cells. Between the individual muscle fibers are capillary blood vessels. Each muscle fiber is sur· rounded by a loose connective tissue called the endomysium. Just beneath the endomysium lies the sarcolemma. Epimysium a thin elastic sheath with infoldings that invaginate the fiber interior. Each muscle fiber is composed of numerous delicate strands-myofibrils, the contractile elements of muscle. C. Myofibrils consist of smaller filaments that form a repeating banding pattern along the length of the myofibril. One unit of this serially repeating pattern is called a sarcomere. The sarcomere is the functional unit of the contractile system of muscle. D, The banding pattern of the sarcomere is formed by the organization of thick and thin filaments, composed of the proteins myosin and actin, respectively. The actin filaments are attached at one end but are free along their length to interdigitate with the myosin filaments. The thick filaments are arranged in a hexagonal fashion. A cross-section through the area of overlap shows the thick filaments surrounded by six equally spaced thin filaments. E, The lollipop-shaped mol· ecules of each myosin filament are arranged so that the long tails form a sheaf with the heads, or cross-bridges, projecting from it. The cross· bridges point in one direction along half of the filament and in the other direction along the other half. Only a portion of one half of a filament is shown here. The cross-bridges are an essential element in the mechanism of muscle contraction, extending outward to interdigitate with receptor sites on the actin filaments. Each actin filament is a double helix, appearing as two strands of beads spiraling around each other. Two additional proteins, tropomyosin and troponin, are associated with the actin helix and play an important role in regulating the interdigitation of the actin and myosin filaments. Tropomyosin is a long polypeptide chain that lies in the grooves between the helices of actin. Troponin is a globular molecule attached at regular intervals to the tropomysin. Adapted from Williams, P & War".vick. R, (1980). Gray's Anatomy (36ch ed., pp. 506-515). Edinburgh: Cht/rdlill A S',glc [ muscle Ilber (cetl) B c Thin Cross section a level of a band IllamentlJl (actin) .,,''''':'' , ThICK o filament (mYOSin) .;.... : " ,,':'.' ", , ',:: . :.:: : ' ..... Sa(comer,:, organilalion Myosin { Mamenl E \ Troponin Livingstone. Each sarcomere is composed of the Following: I, The thin filaments (approximately 5 nm in di- ameter) composed of the protein actin 2, The thick filaments (approximately 15 nm in diameter) conlposed of the protein myosin (Fig, 6-1, D & E) 3, The elastic filaments composed of the protei titin (Fig, 6-2) A, The inelastic filaments composedof the proteins nebulin and tilin Aetin, the chief component or the thin filamen has the shape of a double helix and appears as tw strands of beads spiraling around t.:ach othcl: Two additional proteins, troponin and tropomyosin, are important constituents of the actin helix because thev appear to regulate the making and breaking of cOI~taclS between the actin and myosin filamel1ls during contraction. Tropomyosin is a long polypeptide chain that lies in the grooves between the heiices of actin. Troponin is a globular molecule atlaehed at regular intcr'vals to the tropomyosin (Fig. 6-1, D & E). The thick filaments are located in the central region of the sarcomere, where their ordcrl.y, parallcl arrangemcl1l gives rise to dark bands known as A bands because they arc strongly anisotropic. Thc thin filaments arc atlachcd at either end of the sarcomere to a structurL' known as the Z line, which consists of shon elements that link the thin filaments of adjacent sarcomcres. defining the limits of each sarcomere. The thin filaments cxtcnd From the Z line toward the center of the sarcomere, where they overlap with the thick fibmcnts. Recently it was shown thal there is a third set or n1YOfibril filaments in the vertebrate striated Illuscles. This connecting filament, named titin, links the thick Filaments with the Z line (elastic I band region of titin) and is part of the thick filaments (A band region o[ titin). This filament maintains the central position of the A band throughout contraction and relaxation and might act as a template during myosin ~,sselllbly. Myosin, the thicker filament, is composed of individual molecules, each of which resembles a lollipop with a globular "head" projecting from a long sha or "tail." Several hundred slIch molecules are pack tail to tail in a sheaf with their heads pointed in on direction along half of the filament and in the opp site dircction along the other half, leaving a hca free region (the H zone) in between. The globul heads spiral about the myosin filament in the regio where actin and myosin ovcrlap (the A band) and e tcnd as cross-bridges to interdigitate with sites o the actin filaments, thus forming the structural an I'unctionallink between the two filament types. The intramyofibrilh.,r cvtoskeleton includes i elastic nebulin filaments, which span from the line to the actin filaments. Ncbulin might also act a template for the thin filament assembly. Titin is I j.1m long. It is the largest polypepti and spans from the Z line to the IVI line. Titin is a clastic filament. The part between the Z linc an myosin has a string-likL' appearance. Titin has be suggested to contdbutc greatly to the passh'c for development of muscle during stretch (Fig. 6·2). also might act as a template for the thick filame assembly (Linke et aI., 1998; Squire et al., 199 Stromer et aI., 1998), The I bane! is bisected by the Z lines, which conta the portion of the thin filaments that does not overl with the thick filaments and the clastic part of tit in. the center of the A ballet. in the gap between the en or the thin filaments, is the H zone, a light band co taining only thick filaments and that part of tilin th is integrated in the thick filaments. A narrow, da area in the center of the H zone is the M line, 1'1 M-line A-band part 01 titin Extent of one !!lin molecule The arrangement of titin molecules within the sarcomere. Adapted from Craig. R.(J994). The stftJ(ture of the contract filaments. In A.G. Engel & Fraflzini·Armscrong (eds.). !vlyology (2nd ed.• p. 150). New York: McGr<)w-Hill, Inc. •..-,r; •.. __. allel to the myofibrils and lend to enlarge and f at the level of the junctions between the A an bands, forming transverse sacs, or the terminal ternae, that surround the individual myofi completely, The terminal cisternae enclose a smaller tub tha1 is scpanllcd from them by its own membra The smaller tubule and the terminal cisternae ab and below it are known as a triad. The enclo duced by transversely and longitudinally oriented proteins that link adjacent thick filaments, maintaining their parallel Hrrangclllel1L The various areas of the banding patlern arc apparent in the photomicrograph of human skeletal muscle shown in Figure 6-3. Closely correlated with the rcpealing pattern or the sarcomcrcs is an organized network of tubules and sacs known as the sarcoplasmic reticulum. The tubules of the sarcoplasmic reticulum lie par- A ----"'.... Sarcomere r _--~"" _~ } Sarcoplasmic reticulum Mitochondrion { I~M --- Line t z Line H Zone B A. Single muscle fiber with three protruding myofibrils. B. Electron photomicrograph of a cross·section of human skeletal m c1e. The sarcomeres are apparent along the myofibrils. Characteristic regions of the sarcomere ar~ indicated. Myo,libril Mitochondrion Sarcolemma 'l Band ~ ill E 0 u m A Band If) I Band I z· Triad I ! \j Sarcotubule Terminal cisternae DD-------- 1 Diagram of a portion of a skeletal muscle fiber illustrating the sarcoplasmic reticulum that surrounds each myofibril. The various regions of the sarcomere are indicated on the left myofibril to show the correlation of these regions with the sarcoplasmic reticulum, shown surrounding the middle and right myofibrils. The transverse tubules represent an infolding of the sarcolemma, the plasma membrane that encompasses the entire muscle fiber. Two transverse tubules supply each sarcomere at the level of the junctions of the A band and I bands. Terminal cisternae are located on each side of the transverse tubule, and together these structures constitute a triad. The terminal cisternae connect with a longitudinal network of sarcotubules spanning the region of the A band. Adapted from Ham, A.W & Cormack, D.H, (7979). Histology (8th ed.J. PiJifadefphia: IS, Lippincott tubule is part or the transverse tubule system, or T system, which are invaginatio!1s of the surface membrane of the fiber. This membrane, the sarcolemma, is a plasma membrane that invests every striated muscle (Fig. 6-4). Molecular Basis Contraction of Muscle The most widely..' held theory of muscle contraction is the sliding filament theory, proposed simultaneously by A.F. Huxley and H.E. Huxley in 1964 and subsequently refined (Huxley, 1974). According to this theory, active shortening of the sarcomere, and hence of the muscle, results from the relative m ment of the actin and myosin filaments past on other while each retains its original length. force of contraction is developed by the m heads, or cross-bridges, in the region of overla tween actin and myosin (the A band), These c bridges swivel in an arc around their fixed pos on the surface of the m,vosin filament, much the oars of a boat. This movement of the c bridges in contact with the actin filaments duces the sliding of the actin filaments tc)\var center of the sarcomere, A muscle fiber con \vhen all sarcomere shorten simultaneously all-or-nothing fashion, which is called a twitch Because a single movement of a cross-bridge duces only' a small displacement of the actin ment relative to the myosin filament, each in ual cross-bridge detaches itself from one rec site on the actin filament and reattaches itself other site further along, repeating the process f six tirnes, "with an action similar to a man p on a rope hand over hand" (Wilkie, 1968). cross-bridges do not act in a s)-'nchronized ma each acts independently, Thus, at any given mo (mly approximately half of the cross-bridges ac generate force and displacement, and when detach, others take up the task so that shorten maintained, The shortening is reflected in th comere as a decrease in the I band and a decre the H zone as the Z lines move closer togethe width of the A band remains constant. A key..' to the sliding mechanism is the calciu (Ca~'), which turns the contractile activity on off. Muscle contraction is iniliated \vhen calci made available to the contractile elements ceases when calcium is removed, The mecha that regulate the availability of calcium ions contractile machinery are coupled to electric e occurring in the muscle membrane (sarcolem An action potential in the sarcolemma provid electric signal for the initiation of contractile it,v. The mechanism by which the electric signa gel'S the chemical events of contraction is kno excitationNcontraction coupling. \Vhen the motor neuron stimulates the m at the neuromuscular junction (Fig, 6-5A) an propagated action potential depolarizes the cle cell membrane (sarcolemma), there is a \vard spread of the action potential along system. (Details of this process arc given in F 6-5, A-C and in Box 6-1, which summarize events during the excitation, contraction, an laxation of muscle, Figure 6-5D shows the Molor end plate A B (1 ) Synaptic vesicles (3) o c Schematic representation of the innervation of muscle fibers. A, An axon of a motor neuron (originating from the cell body in the anterior horn of the spinal cord) branches near its end to innervate several skeletal muscle fibers, forming a neuromuscular junction with each fiber. The region of the muscle membrane (sarcolemma) lying directly under the terminal branches of the axon has special properties and is known as the motor end plate, or motor end plate membrane. The rectangular area is shown in detail in B. B, The fine terminal branches of the nerve (axon terminals), devoid of myelin sheaths, lie in grooves on the sarcolemma. The rectangular area in this section is shown in detail in C. C. Ultrastructu the junction of an axon terminal and the sarcolemma. Th vagination of the sarcolemma forms the synaptic trough which the axon terminal protrudes. The invaginated sarcolemma has many folds, or subneural clefts, which grea crease its surface area. Acetylcholine is stored in synaptic cles in the axon terminal. Band C. adapted from Brobeck, I (Ed.) (/979). Best and Taylor's Physiological Basis of Medical Pra (10(h ed., pp. 59-1/3). Baltimore: Williams & Wilkins. D. Cros bridge cycle of muscle contraction. Events During Excitation, Contraction, and Relaxation of Muscle Fiber 1. An action potential is initiated and propagated in a motor axon. 2. This action potential causes the release of acetylcholine irom the axon terminals at the neuromuscular junction. 3. Acetylcholine is bound to receptor sites on the motor end plate membrane. 4. Acetylcholine increases the permeability of the motor end plate to sodium and potassium ions. producing an 1 1. Actin activates the myosin ATPase found on the myosin cfOss-bridge, enabling ATP to be split (hydrolyzed.) This process releaSES energy used to produce movement 01 the myosin cross~br;dges: A, I'll ' ATP --; A ' Iv1 + ADP + P, 12. Oar-like movements of the cross bridges produce rela4 tive sliding of the thick and thin filaments past each other. end-plate potential. S. The end-plate potential depolarizes the muscle mem- 13. Fresh AT? binds to the myosin cross-bridge, breaking brane (sarcolemma), generating a muscle action pOlen- the actin-myosin bond and allowing the cross-bridge to tiallhat is propagated over the membrane suriaee. dissociate from aGin: 6. Acetylcholine is rapidly destroyed by acetylcholinesterclse A . rv1 .;-. ATP -) A + tvl . ATP on the end plate membrclne. 7. The muscle aGion potential depolarize') the transverse tubules. 8. Depolarization of the transverse tubules leads to the re- 14. The ATPase hydrolyzes the myosin ATP complex to rhe M . AlP complex, which represents rhe relaxed state of the sarcomere: lease of calcium Ions from the terminal cisternae of the M' ATP -) Iv1 ' ATP sarcoplasmic reticulum surrounding the myofibrils. These ions are released imo the sarcoplasm in the direct vicinitY of the regulatory proteins tropomyosin and troponin. 9. Calcium ions bind to troponin, allowing movement of the tropomyosin molecule away from the myosin receptor sites on the actin filament that it had been blocking and releasing the inhibition tilat had prevented actin from combining with myosin. 10. Actin (A) combines with myosin ATP (M-ATP). In this 15. Cycles of binding and unbinding of actin with the myosin cross-bridges at successive sites along the actin filament (steps 11, 12, 13, CH'\d 14) continue as long as ii)e concentration of calcium remains high enough to inhibit the action of the troponin-rropomyosin system. 16. Concentration of calcium ions falls as they are pumped into the terminal cisternae of the sarcoplasmic reticulum by an energy-requiring process that splits ATP, 17. Calcium diSSOCIates from troponin. restoring the in- state, ATP has been hydrolyzed to ADP and phosphate hibitory action of troponin-tropomyosin. The actin fila- bUt the products are still attached to myosin (receptor ment slides back and rhe muscle lengthens. In the pr~s-, sites on the myosin cross-bridges bind to receptor sites ence of ATp, actin and myosin remain in tll~ dissociated; on the actin chain): relaxed stale. A + Iv1 ' ATP --; A ' M ' ATP (,l.,!odiiied from Luciano er ,l!. (I 978j, In HUIl1"n FuncliOn <1nd Structure (Fig 5,5D/ New York: ,\,'1cGril~·/·Hi,ll: and ,jdiJpted from (I'dlg. R. 099.:1). Myology (2nd ed.. p. 162) New York: McGra:'1·;·/iIl.) lural features between actin and the cross-bridges of mvosin.) THE MOTOR UNIT The functional unit of skeletal 111uscle is the mOlor unit. which includes a single motor neuron and all of the muscle rlbers innervated by it. This unit is the smallest part of the muscle that can be made to CO!l4 tract independently. \Vhcn stimulated, all mu fibers in the mOlor unit respond as one. The fi of a Illotor unit arc said to show an all-or·none sponse to stimulation: the),..' contract either m mally or not at aIL The number of muscle fibers forming a -m unit is closely I"elated to the degree of contro quirccl of the muscle, In small muscles thal perf vcr~' fine movements, such as the extraocular I des, each motor unit may contain less than a dozen muscle fibers; in large n'lUscies that perform coarse movements. such as the gastrocnemius, the motor unil may contain 1,000 to 2,000 muscle fibers. The fibers or each motor unit arc not contiguous but dispersed throughout the muscle with fibers of other units. Thus. if a single motor unit is stimulatex!. a large ponion or the muscle appears to con· (rae£. If additional motor units of the nerve inner· vating the muscle are stimulated, the muscle corHracls with greater force. The calling in of adeli· tional motor units in response to greater stimulation of the motor nClvc is called recruitment. THE MUSCULOTENDINOUS UNIT The tendons and the connective tissues in and around the muscle belly are viscoelastic structures that help determine the mechanical characteristics of whole muscle during contraction and passive ex· tension. Hill (1970) showed that the tendons represent a spring-like elastic component located in se· rics with the contractile component (the contractile proteins or the m~yofibril, actin, and myosin), while the epimysium, perimysium, endomysium, and sm'· colcmma represent a second elastic component 10· cated in parallel with the contractile component (Fig. 6-6). When the parallel and series elastic components stretch during active contraction 01- passive extension of a muscle, tension is produced and energy is stored; when they recoil with muscle relaxation, this energy is released_ The series elastic fibers are more important in the production of tension than are lhe parallel elastic fibers (Wilkie, 1956). Several investigators have suggested that the cross-bridges of the myosin filaments have a spring-like property and also contdbutc to the elastic properties of muscle (Hill, 1968). The distensibility! and elasticity of the clastic com· ponents arc valuable to the muscle in severnl wavs: I. Thev tend to keep the muscle in readiness for contraction and assure that muscle tension is produced and transillittcd smoothly during contraction. 2. They assure that the contractile elements return to their original (resting) positions when contraction is terminated. 3. They may help prevent the passive overstretch of the contractile elements when these elements are relaxed, thereby lessening the danger of Illuscle injUl)'. PEC ,OOOOOH SEC CC I I The musculotendinous unit may be depicted as consisti of a contractile component ((C) in parallel with an elas component (PEQ and in series with another elastic com nent {SEQ_ The contractile component is represented b the contractile proteins of the myofibril, actin, and myo (The myosin cross-bridges may also exhibit some elastic The parallel elastic component comprises the connectiv tissue surrounding the muscle fibers (the epimysium, pe mysium, and endomysium) and the sarcolemma. The se elastic component is represented by the tendons. Adap hom Keefe. CA.. Neil, E., & Joels. N (1982). I.;lvsc!e and the lle!Voll$syslem. In Samson 1/</rI911(5 f.\ppli(~cJ PhySiology (i3th ed, fJP, 248~259). Oxford: Oxford Unlversiry Press 4. The viscous property of the series and para lel clastic components allows lIk'lll to abso energy proportional to the rate of force app cation and to dissipate energy in a timc~ dependent manner. (For a discussion of vis coelasticity, see Chaptcl- 4.) Thb viscoLIs property. combined with the ela properties of the musculotendinous unit. is dem slI-atcd in everyday activities. For example, whe person altempts to stretch and touch the toes, stretch is initially elastic. As the stretch is held, h ever, further elongation or the muscle results fr the viscosity of the Illuscle-tendon structure, the fingers slo\\'I~' reach closer to the noor. Mechanics of iVluscle Contraction _Electromyography' provides a mechanism for ev ating and comparing neural effects on muscle the contractile activity or the muscle itself in v and in vitro. Much has been learned by using c tromyography Lo study various aspects or the contractile process, particularly the time relationship between the onset of electrical activity in the muscle and aClltal contraction of the muscle or muscle fiber. The following sections discuss the mechanical response of a muscle to electrical (neural) stimulation and the various ways in which the muscle contracts to move a joint, control its motion, or maintain its position. time of the rnuscle so that little or no relax can occur before the next contraction is ini (Fig. 6-8). The considerable gradation of COlllraction e ited by wholc muscles is achic\·ed by the differ IIL,L,U--. SUMMATION AND TETANIC CONTRACTION f ' The mechanical response of a muscle to a single stimulus of its molar nerve is known as a twitch, which is the fundamental unit of recordable muscle activity. Following stimulation there is an interval of a few milliseconds known as the latency period before the tension in the muscle fibers begins to rise. This period represents the timc re· quit'cd for the "slack" in the elastic components to be taken up. The time from the start of tension development to peak tension is the contraction time, and the time from peak tension ulltil the tension drops to zero is the relaxation time. The contraction time and relaxation time vary an10ng muscles, depending largely on the muscle fibcr makeup (de· scribed below). Some muscle fibers contracl with a speed of only 10 mscc; others ma~; take 100 Illsec or longer. An action potential lasts only approximately I to 2 msec. This is n small fraction of the time taken for the subsequent mechanical response, or twitch, even in muscles that contract quickly; thus it is possible for a series of action potentials to be initiated before the first twitch is completed if the activity of the motor axon is maintained. \'\Then mechanical responses to successive stimuli arc added to an initial response, the result is knc)\vn as summation (Fig. 6-7). If a second stimulus occurs during the latencyl period of the first muscle twitch, it produces no additional response and the muscle is said to be completely refractory. The frequency of stimulation is variable and is modulated by individual motor units. The greater the frequency of stimulation of the muscle fibers. the greater the tension produced in the muscle as a whole. However, a maximal frequency will be reached beyond which the tension of the muscle no longer increases. \"'hen this maximallension is sustained as a result or summation, the muscle is said to contract tetanically. In this case, the rapidity of stimulation olltstrips the contraction~relaxation o t tOO 200 300 400 500 t I;+-<- - - I > s\ 82 S3 A Ilh t~ 0 100 200 300 400 I 500 I-tS3 t S, S2 B !I~ ,~ 0 t S, 100 200 300 400 I 500 ..tI.5,S3 Time (msec) C Summation of contractions in a muscle held at a cons length. A, An initial stimulus (51) is applied to the mu and the resulting twitch lasts 150 msec. The second (5 third (5 J ) stimuli are applied to the muscle after 200-m intervals when the muscle has relaxed completely, thu summation occurs. 8.5 1 is applied 60 msec after 51, w the mechanical response from 5.. is beginning to decr The resulting peak tension is greater than that of the gle twitch. C. The interval between 5.. and 5J is furthe duced to 10 msec. The resulting peak tension is even greater than in S, and the increase in tension produce smooth curve. The mechanical response evoked by 5J pears as a continuation of that evoked .by 51' Adapted Luciano, 0.5., Vander. AJ.• & Sherman. J.H. (1978). Huma tion and Structure (pp. J J3-(36). New York: McGraw,Hill. 3 c 0 'iii c 2 <l> >-- <l> > 'g 0; a: 0 is 100 ts Is is 200 300 400 is is 500 600 Jumnmrnmmnnrrmnnm ssssssssssssssssssssssssssssssssss 700 800 900 1000 Time (msec) Generation of muscle tetanus. As the frequency of stimulation (5) increases (Le., the intervals shorten from 200 to 100 msec), the muscle tension rises as a result of summation. When the frequency is increased to lOO/second, summation activity' of their motor units, in both stimulation frequency' and the number of units activated. The repetitive twitching of all recruited motor units of a muscle in an as:vnchronous manner results in brief summations or more prolonged subtetanic or tetanic contractions of the muscle as a whole and is a principal Factor responsible for the smooth movements produced by the skeletal muscles. TYPES OF MUSCLE CONTRACTION During contraction, the force exerted by a contracting muscle on the bony lever(s) to which it is attached is known as the muscle tension, and the external force exerted on the muscle is known as the resistance, or load. As the muscle exerts its force, it generates a turning erFect, or moment (torque), on the involved joint because the line of application of the muscle force usually lies at a distance from the center of motion of the joint. The moment is calculated as the product of the muscle force and the per~ pendicular distance between its point of application and the center of motion (this distance is known as the lever arm, or moment arm, of the force), Muscle contractions and the resulting muscle work can be classified according to the relationship bet\veen either the muscle tension and the resistance to be overcome or the muscle moment generated and the resistance to be overcome, as shown in Box 6-2 (Krocmcr e! 'II., 1990), becomes maximal and the muscle contracts tetanically, ing sustained peak tension. Adapled from Luciano, OS. AI, 8 Sherman, IN (978). Human Function and Structure I i 3-136). New York: McGraw-Hi//. Although no motion is accomplished and n chanical \vork is performed during an isom contraction, muscle work (physiological wo performed: energy is expended and is Illostly paled as heat, which is also called the isornetric production. All dynamic contractions involve ma.y be considered an initial static (isometric) as the muscle first develops tension equal to th it is expected to overcome. The tension in a muscle varies with the t.y contraction. Isometric contractions pro greater tension than do conccntric contrac Studies suggest that the tension developed eccentric contractIon may even exceed that oped during an isometric contraction. Thes ferences arc thought to be due in large part vat'ying amounts of supplemental tension duced in the series clastic component of the cle and to diFferences in contraction time longer contraction time of the isometric an centric contractions allows greater cross~ formation by the contractile components, thu mitting greater tension to be generated ( 1987). More time is also available for this te to be transmitted to the series elastic comp as the muscle-tendon unit is stretched. The l contraction time allo\vs the recruitment of tional nlotor units. Komi (1986) has pointed oul that concentri metric, and eccentric muscle contractions se _ Types of Muscle Work and Contraction Dynamic work: Mechanical work is performed and joint mo- 4. Iseinenial (iso, constant; inertial, resistance) contraction: tion is produced through the following forms of muscle con- This is a type of dynamic muscle work wherein the resis traction: 1. Concentric (con, together; cemrum, center) contrac- constant. If the moment (torque) produced by the musc tion: When muscles develop sufficient tension to over- tance against which the muscle must contract remains is equal to or less than the resistance to be overcome, come the resistance of the body segment, the mus- the mllscle length remains unchanged and the muscle cles shorten and cause joint movement. The net contracts isomelfically. If the moment is greater than th moment generated by the muscle is in the same di· resistance, the muscle shortens (contracts concentrically rection as the change in joint angle. An example of a and causes acceleration of the body part. Isoinertial con~ concentric contraction is the action of the quadriceps traction occurs, for example, when a constant external load is lifted. At the extremes of motion, the inertia of the load must be overcome; the involved muscles contract isometrically and muscle torque is maximal. In the midrange of the motion, with the inenia overcome, the mU'jcles contract concentrically and the torque is submaximal. 5. Isotonic (iso, constant; tonic, iorce) contraction: This ter is commonly used to define muscle contraction in which the tension is constant throughout a range of joint m()~ tion. This term does not take into aCCOLlnt the leverage'" eilects at the joint. Ho\,vever, because the muscle force moment arm changes throughout the range oi joint mo lion, the muscle tension must also change. Thus, isoton muscle contraction in the truest sense does not exist in the production of joint motion (Kroll, 1987). Static work: No mechanical work is performed and posture or joint position is maintained through the following form muscle contraction: 1. Isometric (iso, constant; metric, length) contra(ti()~,;,. :~~s. c1es are not always directly involved in the prod~qi~~:-Pr. joint movements. They may exercise either a rest'ra(ning or a holding action, such as that needed t~ ~aint~i~ th body in an upright position in opposing the force of gravity. In this case the muscle attempts to shorten (i.e., the myofibrils shorten and in doing so stretch the series elastic componenr. thereby producing tension), but it does not overcome the load and cause movement; instead. it produces a moment that supports the load in a iixed position (e.g., maintains posture) because no ," change takes place in the distance between the mlJsc!le points of attachment. in extending the knee when ascending stairs. 2. Eccentric (e.g., out of-, centrum, center) contraction: When a muscle cannot develop sufficient tension and is overcome by the external load, it progressively lengthens instead of shortening. The net muscle moment is in the opposite direclion irom the change in joint angle. One purpose of eccenlric contraction is to decelerate the motion oi a joint. For example, when one descends stairs, the quadriceps works eccentrically to decelerate flexion of the knee, thus decelerating the limb. The tension that il applies is less than the force of gravity pulling the body downward, but it is sufficient to allow controlled lowering of the body. 3. Isokinetic (iso, constant; kinetic, motion) contraction: This is a type of dynamic muscle work in which movement of the joint is kept at a constan I velocity. and hence rhe velocity of shortening or lengthening of the muscle is constant. Because velocity is held constant, muscle energy cannot be dissipated through acceleralion of the body part and is entirely converted to a resisting moment. The muscle force varies with changes in its lever arm throughout the range of joint motion (Hislop & Perrine. 1967). The muscle contracts concentrically and eccentrically with different directions of joint motion. For example, the i1exor muscles of a joint contract concentrically during flexion and eccentrically during extension, acting as decelerators during the latter. occur alone in normal human movement. Instead, one type of contraction or load is preceded by a different type. An example is the eccentric loading prior to the concentric contraction that occurs at the ankle frolll miclstance to toe-off during gait. Because muscles normally' shorten or lengthen at "at)!ing velocities and with valying amoullls of tcnsion, perfolll1ance and measurement of isokinetic work require the lise of an isokinetic dynamon1ctcl: This device provides constant velocity of joint motion and maximum cxtcIllal resistance throughout the range of 010lion of the involved joint. thereby requiJing maximal muscle torque. The use of the isokinctic dynamometer provides a method of selective training and measurement, but physiological movement is not simulated. c .2 '.0 m I I I c " f- " .~ <i ~, I I 0.5 I I 0; I a: I 1.27 I I 1.65 2.0 2.25 Sarcomere Length (JIm) ==i/i== 2.25-3.6 I'm c==:jr=z=J~\:~::\g; ::~;::~:::~\ z 20-225"m=9! M i':::": :::!::~ A Z Ir== <1.65"m==j~F= Force Production in Muscle The total force that a muscle can produce is influenced by its mechanical properties, which can be described by examining the length-tension. loadvelocity, and force-time relationships of the muscle and the skeletal muscle architecture. Other principal factors in force production arc muscle temperature, muscle fatigue, and prestrclching. LENGTH-TENSION RELATIONSHIP The force, or tension, that a muscle exerts varies with the length at which it is held when stimulated. This relationship can be observed in a single fiber contracting isometrically and tetanically, as illustrated by the length-tension curve in Figure 6-9. Maximal tension is produced when the muscle fiber is approximately at its "slack," or resting, length. ,If the fiber is held at shorter lengths. the tension falls off slowly at first and then rapidly. If the fiber is lengthened beyond the resting length, tension progressively decreases. The changes in tension when the fiber is stretched or shortened primarily are caused by structural alterations in the sarcomere. Maximal isometric tension can be exerted when the sarcomeres arc at their resting length (2.0-2.25 /J.m), because the actin and myosin fllamcnts overlap along their entire length and the number of cross-bridges is maximal. If the sareo meres arc lengthened. there are fewer junctions between the marnents and the active tension decreases. At a Sarcomere length of approximately 3.6 J.Lm, there is no overlap and hence no active tension. Sarcomcl·e shortening to less than its resting length decreases the acLive ten- Tension·length curve from part of an isolated muscle f stimulated at different lengths. The isometric tetanic t sion is dosely related to the number of cross-bridges o the myosin filament overlapped by the actin filament. tension is maximal at the slack length, or resting lengt the sarcomere (2 p.m), where overlap is greatest, and f to zero at the length where overlap no longer occurs ,....m). The tension also decreases when the sarcomere length is reduced below the resting length. falling sha at 1.6S IJ.m and reaching zero at 1.27 IJ.ffi as the exten overlap interferes with cross-bridge formation. The str tural relationship of the actin and myosin filaments at ous stages of sarcomere shortening and lengthening is portrayed below the curve. A. actin filaments; M. myo filaments; Z, Z lines. Adapted from CriJ....,;ord. CN.C & J~ N. T. (1980). The design of muscfes. In R. Owen, J. Goodfelfo P. Buffough (Eds.), Scientific Foundations of Orthopaedics an Trmilllalology (pp. 67·-74). London.' William Heinemann;')5 modified from Gordon, A.M., rluxley, A.F.I., & JlIlian, F.l. (1 Tile variation in isometric tension with SMcomere length in tebr~lre muscle fibers. J Physiol, 184, 170. sian because it allows ovedapping of the thin ments at opposite ends of the sarcomere, which functionally polarized in opposile directions. sarcomere length of less than 1.65 fJ.m. the thic aments on the Z line and the tension dimi sharply. Thc length-tension relationship illustrated in ure 6-9 is for an individual muscle fiber. If this ,lionship is measured in a whole mus.de contrac isometrically and tetanically, the tension produ by both active components and passive compon must be taken into account (Fig. 6-10). The cun/e labeled "active tension" in Figure 6-10 represents the tension developed by the contractile elements of the muscle. and it resembles the curve for the individual fiber. The curve labeled "passive tension" reflects the tension developed when the muscle surpasses its resting length and the noncontractile muscle belly is stretched. This passive tension is mainly developed in the parallel and series elastic components (Fig. 6-6). When the belly contracts, the combined active and passive tensions produce the total tension excrled. The curve demonstrates that as a muscle is progl'cs· sivdy stretched beyond its resting length, the pas· sive tension rises and the active tension decreases. Most rnuscles that cross only one joint normally are not stretched enough fOl- the passive tension 10 play an important role, but the case is different for lwo-joint muscles. in which the extremes of the Eccenlric Concentric --~ u '"o -' o .. Velocity load-velocity curve generated by plotting the veloci motion of the muscle lever arm against the external When the external load imposed on the muscle is ne ble, the muscle contracts concentrically with maxima speed. With increasing loads the muscle shortens mo slowly. When the external toad equals the maximum that the muscle can exert, the muscle fails to shorten has zero velocity) and contracts isometrically. When load is increased further, the muscle lengthens eccen cally. This lengthening is more rapid with greater loa L Length ~------ I I The active and passive tension exerted by a whole muscle contracting isometrically and tetanically is plotted against the muscle's length. The active tension is produced by the con· tractile muscle components and the passive tension by the series and parallel elastic components, which develop stress when the muscle is stretched beyond its resting length. The greater the amount of stretching, the larger the contribution of the elastic component to the total tension. The shape of the active curve is generally the same in different muscles. but the passive curve, and hence the total curve, varies depending on how much connective tissue (elastic component) the muscle contains. Adapted from Crawford. CN.C & James. N. T. (1980). Tile design of muscles. III R. Owen. 1. Goodfel/o;,t~ & p. Bulfough (£els.), Scientific Foundations of Orthopaedics and TraurnJ· tology (pp. 67-74). London: William Heinemann. I • length-tension relationship may be functi (Crawford & James, 1980). For example, the strings shorten so much when the knee is Hexed that the tension they can exert decreases siderably. Conversely, when the hip is nexed an knee extended. the Illuscles arc so stretched t is the magnitude of their passive tension tha vents flirt her elongation and causes the knee t if hip flexion is increased. LOAD-VELOCITY RELATIONSHIP The relationship between the velocity of shOl· or eccentric lengthening of a muscle and dif constant loads can be determined by plolLing t locil)' of motion of the muscle lever arm a1 va external loads. thereby generating a load-ve curve (Fig. 6-11). The velocity of shortening muscle contracting concentrically' is inverse lated to the external load applied (Guyton, The velocity' of shortening is grealest when th ~~~·_----_·_-------T Gastrocnemius Muscie Tear - -. ·. A 22·year-old male professional athlete tears his ga5___ trocnemius during a race (Fig. (56-1-1). The tensile overload that happens during strenuous eccentric and c6n~entric contractions increases the risk of injury. especiallywhen the forces involve bj-articular muscles such :asthe gastrocnemius. This indirect trauma is associated :wi!~ 111gh tensile forces during rapid contraction (high v~l_()~ity) and continued changes in muscle length. The tcrnalload is zero, but as the load increases th ele shortens morc and more slowly. \Vhcn the nul load equals the mnximal force that the can exen, the velocity of shortening becom and the muscle contracts isometrically. \,\fl load is increased still further, th~ muscle co eccel1Lrically: it elongates during contractio load~velocity relationship is reversed from the concentri-cally contracting rnuscle; the eccentrically lengthens more quickl)! with inc load (Kroll. 1987) (Case Study 6-1). status of muscle contracrion at the time of overload is uS!Jally eccentric, and failure most often occurs at or near the myotendinous junction unless the muscle has Q.eef1. previously injured (Kasser, 1996). Swelling from ·h~~{br'rh~ge occurs initially in the inflammatory phase. ,The cellu.lar response is more rapid and repair is (rlore complete'if the vascular channels are not disrupted and ·thenutritioo.of ihe tissue is not disturbed. The degree of injury from a tensile overload will dictate the poten· tia/hos< r~sponse and the time needed iar repair. FORCE-TIME RElATIONSHIP The force, or tension, generated by a muscle portional to the contraction time: the long contraction lime, the greater is the force deve lip to the point of maximurn tension. In Figur this relationship is illustrated by a force-lime for a whole muscle contl"acting isomet Slower contraction leads to greaLcr force p lion because time is allowed for the tensio duced b.y thL' cornractilc clements to be trans through the parallel clastic components to t don. AltlH)Ugh tension production in the cont component can reach a maximum in as little mscc. lip to 300 ITlSeC ma~' be needed for th sion to be transferred to the clastic compo The tension in the tendon will reach the max tension developed by the contractile element the aClive contraction process is of sufficien lion (Olloson, 1983). EFFECT OF SKElETAL MUSCLE ARCHITECTURE The Illuscles consist of the contractile comp the sarcomere, which produces active tensio arrangement of the contractile components the contractile propcnies of the muscle dr cally. The more sarcomere lie in series, the the myofibril will be; the more sarCOlllere lie lel. the larger the cross-sectional area 01" the m ril will be. These two basic architectural patte myofibrils (long or thick) affect the contractile erties of the llluscies in the following ways: C~;~e-Study Fig~re 6-1-1, ;-" -. 1. The force the muscle can produce is pro lion'll to the cross-section of the myofibr (Fig_ 6-13A). 2. The velOcity and the excursion (working range) that the muscle can produce are portional to the length of the I11vofibril (Fig. 6-138). 100 Muscles with shoner fibers and a Im'ger crosssectional area are designed to produce [orce, whereas muscles \vith long fibers are designed for excursion and velocity. The quadriceps muscle contains shorter myofibrils and appears to be specialized ror force product ion. The sartorius muscle has longer fibers and a smaller cross-sectional arca and is better suited for high excursion (Barana et aI., 1998; Lieber & Bodine-Fowler, 1993), 80 ~ ~ ~ 60 ~ 40 ~ ::; " 20 EFFEa OF PRESTRETCHING 10 It has been demonstrated in amphibians and in humans (Cuillo & Zarins, 1983) that a muscle performs more work when it shortens inlmediatcly af- 15 20 Muscle Length (cm) 100 B ter being stl'etched in the concentrically contracted state than when it shortens from a state of isometric contraction. This phenomenon is not entirely accounted for by the elastic energ~' stored in the series clastic component during stretching but must also be caused by energy stored in the contractile com· 80 ~ ~ o 60 u. ~ '0 ~ 40 ::; " 20 5 10 15 20 Muscle Velocity (cm/s) Isometric and isotonic properties of muscles with differ architecture. A. Force-length relationship. B. Force-veloc relationship. PSCA. physiological cross-sectional area. Reprinted with permission of the AmeriC<Jn Physical Therapy sociation from Lieber, RL (J 993). Skeletal muscle rnec!wnics Implications for rehabilitation. Physical Therapy, 73(2). 852. • 0F----------- pon~nl. 'It has been suggested [hat changes in the trinsic mechanical properties of l11)'oflbrils are ponant in the stretch-induced enhancement work production (Takarada et aI., 1997), Time IBm i L- _ Force-time curve for a whole muscle contracting isometri· cally. The force exerted by the muscle is greater when the contraction time is longer because time is required for the tension created by the contractile components to be transferred to the parallel elastic component and then to series elastic component as the musculotendinous unit is stretched. EFFECT OF TEMPERATURE A rise in muscle temperature causes an incre in conduction velocity across the sarcolem (Phillips & PeLrorsky, 1983), increasing the quency of stimulntion and hence tbe product of muscle force. Rising of the muscle temperat from 6 La 34 u C results in an almost linear incre or the tension/stirrness raLio (Callcr et al" 199 A rise in temperature also causes greater enzymatic activity of muscle metabolism, thus increasing the efficiency of rnusclc contraction. A further effect of a rise in temperature is the increased elasticity of the collagen in the series and parallel elastic components. which enhances the extensibility of the muscle-tendon unit. This increased prestrclch incrcases the force production of the muscle. Muscle temperature increases by rneans of two mechanisms: Fatigue in a muscle contracting isometrically. Prolong stimulation occurs at a frequency that outstrips the m cle's ability to produce suHicient ATP for contraction. result, tension production declines and eventually ce Adapted from Luciano. 0.5., Vander. A.i., &- Sherman. J.H (1978). Human Function and Structure (pp. 113-136). Ne I. Increase in blood now, which occurs when an athlete "warms up" his or her muscles 2. Production of the heal of reaction generated by metnbolism. by the release of the energy of contraction, and by friction as the contractile components slide over each other However, at low tenlperaturc (1 O('C), it has been shown that the maximum shortening velocity and the isometric tension are inhibited significantly. This is caused by decreased pH (acidosis) in Ihe muscle. The 1'1-1 plays a much less important role at temperatures close to the physiological level (Pate et aI., 1995). EFFECT OF FATIGUE The ability of a muscle to contract and relax is dependent on the availability of adenosine triphosphate (ATP) (Box 6-1). If a muscle has an adequate supply of oxygen and nutrients that can be broken down to provide ATP, it can sustain a series of lowfrequency twitch responses for a long time. The rrequency must be low enough to allow the muscle to synthesize AT? at a rate sufficient to keep up with the rate of ATP breakdown during contraction. If the fTequency of stimulation increases and outstrips the rate of replacement of ATP, the twitch responses soon grow progressively \veaker and eventually fall to zero (Fig. 6-14). This drop in tension following prolonged stimulation is muscle fatigue. If the frequency is high enough to produce tetanic contractions, fatigue occurs even sooner. [f a period of rest is allowed before stimulation is continued, the ATP concenln.l.lion rises and the muscle briefly recovers its contractile ability before again undergoing fatigue. Three sources supply ATP in muscle: creatine phosphate, oxidative phosphorylation in Ihe mitochondria. and substrate phosphor:ylation during anaerobic glycolysis. \'Vhen contraction begins, the York: McGraw-Hili. • myosin ATPase rapidly breaks down ATP. Th crease in adenosine diphosphate (ADP) and phate (Pi) concentralions resulting from this b down ultimately leads to increased rale oxidative phosphorylation and glycolysis. A short lapsc. however. lhese metabolic palhway gin lO deliver ATP at a high rate. During this vaL lhc energy for AT? formation is provided b atine phosphate, which offers the most rapid m of forming ATP in the muscle cell. At moderate rates of muscle aClivity, most o required ATP can be formed by the process o idalive phosphorylation. During intense exe when ATP is being broken down rapidly, the abilitv to replace ATP by oxidalive phosphory may be limited. primarily by inadequate deliv oxygen to the muscle b.y lhe circulatory system Even \vhcn oxygen delivery is adequate, th al which oxidative phosphorylation can pro ATP may be insufficient to sustain inlense ex because the enzymatic machinery of this pathwC relatively slow. Anncrobic glycolysis then beg contribute an increasing portion of the ATP. Th colytic pathway, although it produces much sr amounts of ATP h-om the breakdown of glucos erates at a Illuch faster rate. It can also proce the absence of oxygen, with the formation of acid as ils end product. Thus, during intense cise, anaerobic glycolysis becomes an addi source for rapidly supplying the muscle with A The glycolytic pathway has the disadvanta requiring large amounts of glucose for the pr tion of small amounts of" ATP. Thus, even th muscle stores glucose in the form of glycogen, ing glycogen supplies rnay be depicted quickly muscle aconty is intense, Finally, myosin ATPase may break down AT? faster than even glycolysis can r~p'lace it, and fatigue occurs rapidly as AT? concentrations drop. After a period of intense exercise, creatine phosphate levels have become low and much of the muscle glycogen may have been convened to lactic acid. For the muscle to be returned to its original state, creatine phosphate must be resynthesized and the u1vcofLcn stores must be replaced. Because both pl:oce~ses require energy, the muscle will continue to consumc oxygen at a rapid rate cven though it has stopped contracting. This sustained high oxy!:rell uptake is demonstrated by the fact that a person ~ontinues to breathe heavily and rapidly aher a period of strenuous exercise. v"hen the energy necessary to return glycogen and creatine phosphate to their original levels is taken into account, the ef(-icienc~. . with which muscle coovens chemical energy to work (movement) is usually no more than 20 to 250/0, the majority of the energy being dissipated as heat. Even when Illuscle is operating in its most efficient stnte, a maximum ()f only approxinwtel)! 45{>~ of the energy is used for contraction (Arvidson ct 'II., 1984; Guyton, !986). In growth biomechanics, muscle fatigue is flrst observed by the lack of coorclinalion or movement and its effect in the increasing of loads in tissue. Researchers including Bates et al. (1977) have indicated that the skill of the person in performing a given action is affected by fatigue, They studied the fatigue effect on runners and absented that runners decrease their knee extension when fatigue occurs (Bates et aI., 1977). Parnianpour (1988) studied the motion coupling of the spine at exhaustive extension Oexion. This study showed that when an individual became fatigued, the coupled motion increased and therefore the spinaltorquc increased. The most deleterious component of the ncurornuscular adaptation to the fatiguc state was the reduction in accuracy control and speed of contraction, \vhich may predispose an individual to injury if muscle fatigue occurs. Muscle Fiber Differentiation. In the preceding section, we described the major ractors that determine the total tension developed b.v the whole rnuscle when it COlllracts. Individual muscle fibers also display distinct differences in their rates of contraction, development of tension. and susceptibility to fatigue. Jvlany methods of classifying muscle fibers been devised. As early' as 1678, Lorenzini obse anatomically the gross dilTcrence bel\\,Icen red white muscle. and in 1873 Ranvier typed muscl the basis of speed of contractility and fatigab Although considerable confusion has existed cCl'ning the method and terminology for classif skeletal muscle, recenl histological and histoch cal observations have led to the identificatio three distinct typcs or muscle fibers on the bas differing contractile and metabolic prope (Brandstater & Lambert, 1969; Buchtah! & Soh burch, 1980) (Table 6-1). The fiber t~'pes are distinguished mainly by metabolic pathways by which they can gene AT? and the rate at which its energy is made a able to the conrractilc system of the sarcom which determines the speed of contraction. three fiber types are termed type I, slow-twitch idative (SO) fibers; type lIA, fast-twilCh oxida glycolvtic (FOG) fibers; and type liB, fast-twitch colytic (FG) fibers. Typc I (SO) fibers are characterized by a low tivity of my'osin ATPase in the muscle fiber thcrefore, a relatively slow contraction time. glycolytic (anaerobic) aClivity is low in this lype, but a high content of mitochondria produc high potential for oxidative (aerobic) activity. Ty fibers are difficult to fatigue because the high ra blood now to these fibers delivers oxygen and n ents at a sufficient rate to keep up with the relati slow rate of ATP breakdown by myosin ATP Thus, the fibers are well suited for prolonged, intensity work. These fibers are relatively sma diameter and so produce relatively little tens The high myoglobin content of type I fibers give muscle a red color, Type II muscle fibers arc divided into two n subgroups, IIA and IlB, on the basis of differing ceptibility to treatment with different buffers p to incubation (Brooke & Kaiser, 1970). A third group, the typc lIe fibers, are rare, undifferenti fibers, which are usually seen before the 30th w of gestalion. This fiber type is infrequent in hu muscle (Banker, 1994). Tvpe IIA and 11 B fibers characterized by a high activity of myosin ATP which results in relatively fast contraction. Type lIA (FOG) fibers are considered interm ate between type 1 and type lIB because their contt-action time is combined with a modera well-developed capacity for both aerobic (ox tive) and anaerobic (glycolytic) activity. T Properties of Three Types of Skeletal Muscle Fibers TYPE 1 Slow-Twitch Oxidative TYPE JlA Fast-Twitch OxidativeGlycolytic (SO) (fOG) Glycolytic (fG) Speed of contraction Slow Fast Fast Primary source of ATP production Oxidative Oxidative phosphorylation Anaerobic phosplloryla tion glycolysis Glycolytic enzyme activity low Intermediate High Capillaries Many Many few Myoglobin conlent High High low Glycogen content Low Intermediate High Fiber diameler Small Intermediate Large Rate of fatigue Slow Intermediate Fast fiber's also have a well-developed blood supply. They can maintain their contractile activity for rel<Hively long periods; ho\vevcr, at high rates of activity. the high rate of ATP splitting exceeds the capacity of both oxidative phosphorylation and glycolysis to supply ATP, and these fibers thus eventually fatigue. Because the myoglobin conlCIH 01" this muscle type is high, the muscle is of len cat- egorized as red III lIscl e. T:vpe liB (FG) libers rely primarily on glycolvtic (anaerobic) activity for ATP production. Few capillaries are found in the vicinity of these fibers and because they contain little myoglobin (hey are ohen referred to as white muscle. Although Lype liB fibers are able to producc ATP rapidly, they fatigue easily because their high rate of ATP splitting quickly depletes the glycogen needed for glycolysis. These fibers generally are of large diameter and are thus able to produce great tension, but only for short periods before they fatigue. It has been well demonstrated that the nerve innervating the muscle fibcr determines its t.ypc (Burke el aI., 1971); thus, the muscle fibers of each motor unit are of a single type. In humans and other species, electrical stimulation was found to change the fiber type (Munsal, McNeal, & Waters, 1976). In animal studies, transecting the nerves that innervate slow twitch and fast-twitch muscle fibers and then crossing these nen;es was noted to reverse the fiber types. After recovery from the cross-innervation, the slow-twitch fibers became fast in their con8 TYPE JIB Fast-Twitch tractile and histochemical properties and the twitch fibers became slow. The fiber composition of a given muscle pends on the function 01" that muscle. Some m cles perform predominantly one form of cont tile activity' and arc often composed mostl one muscle fiber t}'pc. An example is the so muscle in the calf, which prirnarily maint posture and is composed of a high percentag type I fibers. More commonly', however, a mu is reqllir'cd to perform endurance-type act unde)' some cirClllTlstances and high-inten strength activity under others. These mus generally contain a mixture of the three mu fiber types. (n a typical mixed muscle exerting low tens some of the small motor units, composed of ty fibers, contract. As the muscle force increa more motor units are recruited and their quency of stimulation increases. As the frequ becomes maximal. greater muscle force achieved by recruitment of larger motor u composed of type 11;\ (FOG) fibers ancl eventu type lIB (FG) fibers. As the peak muscle force creases, the larger units are the first to cease tivity (Guyton. 1986; Luciano, Vander. & Sherm 1978). It is generally, but not universally, accepted fiber types are genetically det~rmined (Costi aI., 1976; Gollnick, 1982). In the average pop tion, approximatel.'Y' 50 to 55% of muscle fi are type I, approximately 30 to 35 Cj(; are type IrA, and approximately' 15 rJr; are t.\'Pe lIB, but these percentages vary greatly among individuals. In elite athletes, the relative percentage of fiber types differs from that in the general population and appears to depend on whether the athlete's principal activit)' requires a short, explosive, maximal effort or involves submaximal endurance. Sprinters and shot putters, for example, have a high percentage of tvpe II flbers, whereas distance runners and cross~ountry skiers have a higher percentage of type I fibers. Endurance athletes may have as man.v as 80% t.ype I fibers, and those engaged in short, explosive efforts as few as 30 0/e of these fibers (Saltin et 'II., 1977). The genetically determined fiber typing may be responsible for the natural selective process by which athletes are drawn to the type of sport for which they are most suited. Because fiber types are determined by the nerve that innervates the muscle fiber, there may be some cortical control of this innervation that influences an athlete to choose the sport in which he or she is genetical I)' able to excel. Muscle Injuries iVluscle injuries comprise contusion, laceration, ruptures, ischemia, compartment syndromes, and denervation. These injuries weaken the muscles and can cause significant disability'. Blunt trauma can diminish muscle strength, limit joint motion, and flnally lead to myositis ossiflcans. lVlusclc laceration, surgical incisions, and traumatic lesion to muscle tissue and denervation weaken the ITmsclcs, sometimes significantly. Ruptures in muscles also can cause weakness. Like the other injuries, they rnay result from direct trauma, but muscle contractions against resistance also can lead to tears in muscle tissue. Acute muscle ischemia and compartment syndromes can cause extensive muscle necrosis. The many potential causes of compartment s}'ndrome all result in increased pressure \vithin a confined muscle compartment. In this case, failure to relieve the pressure rapidl.y ma}' cause complications that range from weakness and decreased motion to loss of an entire limb. Studies have shown that healthy skeletal muscle has a substantial capacity to repair itself. This repair process following a specific injury is inferred by the prior innovation pattern, vascularization, physical constraint of the surrounding tissues, t extent and condition of extracellular matrices, a the development of repair cells. Muscle injuries a important but the topic is not \vithin the scope this chapter. Injuries should be investigated ca fully if suspicion arises that a patient has mus damage. Muscle Remodeling The remodeling of muscle tissue is similar to that other skeletal tissues such as bone, articular car lage, and ligaments. Asin these other tissues, mu cle atrophies in response to disuse and immobili tion and hypertrophies when subjected to grea use than usual. EFFECTS OF DISUSE AND IMMOBILIZATION Disuse and immobilization have detrimental effe on muscle ftbers. These effects include loss of durance and strength and muscle atrophies on a m crostructural and macrostructural level, such as creased numbers and size of fiber. Biochemi changes occur and affect aerobic and anaerobic erg}' production. These effects are dependent fiber type and muscle length during immobilizati Immobilization in a lengthened position has a l deletet-ious effect (Appell, 1997; Kassel', 1996; Oh et 'II.. 1997; Sandmann, et 'II., 1998). Clinical and laboratory studies of human and imal muscle tissue suggest that a program of mediate or early motion may prevent Illuscle at phy after injury or surgery. In a study of cn injuries to rat muscle, the effect of immobilizat of the crushed limb was compared with that of mediate motion. The muscle fibers \vere found regenerate in a more parallel orientation in the m bilized animal than in the immobilized anim capillarization occurred more rapidly, and tens strength returned more quickly. Similar resu were found in a later study on the effect of imm bilization on the IT'lOrphology of rat calf musc (Kannus et aI., 1998a). It has been found clinically that atrophy of quadriceps muscle that develops while the limb immobilized in a rigid plaster cast cannot be v'ersed through the use of isometric exercises. rophy may be limited by allowing early mot such as that permitted by a partly mobile c brace. In this case, dynamic exercises can be performed. Human muscle biopsy studies have shown that it is mainly' the type I fibers that atrophy with immobilization; their cross-sectional area decreases and their potential for oxidative enzyme activity is reduced (Kannus et aI., 1998b). Earl)' motion may' prevent this atrophy. It appears that if the muscle is placed under tension when the body segment moves, afferent (sensory) impulses from the intrafusal muscle spindles will increase, leading to increased stimulation of the type I fiber. Although intermittent isometric exercise may be sufficient to maintain the metabolic capacity of the t.vpe II fibel~ the type I fiber (the postural fiber) requires a more continuous impulse. Evidence also suggests that electric stimulation may prevent the decrease in type I fiber size and the decline in its oxidative enzyme activity caused b:v immobilization (Eriksson et ai., 1981). In elite athletes, inactivity following injury, surgery, or immobilization rapidly decreases the size and aerobic capability of muscle fibers, particularly in the fiber type affected bv the chosen sport. Tn endurance athletes, type I fibers are affected, while in athletes engaged in an explosive activity such as sprinting, type II fibers are affected. stood (Gollnick, 1982; Guyton, 1986). It ap that these evcnts arc controlled or modifi both the intrafusal muscle spindles, located allel with the extrafusal fibers of the muscle and the Go!gi tendon organs, located in serie Ruptured Left Anterior eruciate Ligam A 25-year-old male, status postsurgical repair of the - ruptured left anterior cruciale ligament, had torqu measurements taken from the involved and uninvolved limb 10 weeks after the surgical procedure (Fig. (56-2-1 and repeated 6 weeks after the training began (Fig. (56 2-1 B). An increase of muscle torque is shown in the repeated isokinetic test. The initial deficit of the involved side was approximately 63~o when compared with the uninvolved side. After 6 '.Neeks of trainin~J. the deficit o the involved side compared with tho uninvolved side de creased to 43%. 187,------------163 140 E 116 <n c 93 ;:o 47 " z EFFECTS OF PHYSICAL TRAINING Physical training increases the cross-sectional area of all muscle fibers, accounting for the increase in muscle bulk and strength. Some evidence suggests that the relative percentage of fiber types composing a person's muscles rna:v also change with physical training (Arvidson, Eriksson, & Pitman, 1984). The cross-sectional area of the fibers affected by the athlete's principal activit).' increases. For example, in endurance athletes, the arca of muscle taken up by type I and type IIA fibers increases at the expense of the total area of type lIB fibers (Case Study 6-2). Stretching incrcases muscle Oexibility, maintains and augmcnts the range of joint motion, and increases the elasticity and length of the musculotendinous unit (Brobeck, 1979; Cuillo & Zarins, 1983). It also permits the musculotendinous unit to store more energy in its viscoelastic and contractile components. The events that take place during muscle stretching are complex and incompletely under- / 69 23 ----- o J-~c_-----__:-~:-''''-'---o 16 32 48 64 80 96 112 Knee motion degrees A 294 239 204 ~c 192 0 ;: 136 108 z 68 33 0 " " /. .< " "- "- / 0 16 32 48 64 80 96 , 112 Knee motion degrees B Case Study Figure 6-2-1 Isokinetic test at 180~/sec. A Measurement of the quadriceps femoris torque produc tion at 10 weeks postsurgical procedure. The dashed fi represents torque output by the involved limb. The so line represents torque output by the noninvolved limb B. Measurements of the quadriceps femoris torque pro duction at 16 weeks postsurgical procedure and 6 wee after training sessions. The dashed line represents torq output by the involved limb. The solid line represents torque output by the noninvolved limb. these fibers. The spindles respond to an increase in muscle length and the Golgi apparatus to an increase in TllUSc!C tension. The resulting spindle reflex increases muscle contraction, while the Goigi reflex inhibits contraction and enhances muscle relaxation. The intrafusal muscle spindles are or t\Vo types: primary and secondary. The primary spindles respond to changes in the rate of muscle lengthening (dynamic response) and the actual amount or lengthening. The secondary spindles respond only (0 the actual length change (static response). The static response is weak and the dynarnic response is strong: therefore, keeping the rate of stretch low may allow the dynamic response to be bypassed, essentially negating dte effect of the spindles. Conversely, the increase in rnusclc tcnsion during stretching may activate the relaxing effect of the Golgi apparatus and thus enhancc further stretching. The various 111ethods and theories of" stretching all have as a common goal inhibition of the spindle effect and cnhanccrnent or the Golgi effect to relax the muscle and promote further lcngt hen i ng. Sllmm.ary i!;< The structural unil of skeletal muscle is the fil.;·c'r, which is encompassed by the endomysium and organized into fascicles encased in the perirnyslum. The epimysium surrounds the entire musclc. 2 The fibers are composed of myofibrils. aligned so as to create a band pallern. Each repeat of this pattern is a sarcomere, the functional unit 01" the contractile system. 3 The myofibrils arc composed or thin filaments of the protein actin and thick filaments of the protein myosin, and the intramyofibrillar cytoskeleton is composed 01" the clastic filaments titin and the inelastic filarnents ncbulin. 4· According to the sliding filament theOl)" active shortening of the muscle results from the reiativL' movement of the actin and myosin filaments past one anothel: The force of contnlClion is developed by movement of the myosin heads, or cross-bridges, in contact with the aclin filaments. Troponin and tropomyosin, two proteins in the actin helix, reguIatc the making and breaking of the contacts between r-ilamenls. 5 A key (0 (he sliding mechanism is the calciu ion, which lu)-ns the contractile activity on and a 6 The mOlOr unit, a single motor neuron and musclc fibers inncrvatcd by it, is the smallesl pan the muscle thal can contral:t independently. T calling in of additional motor units in response greater stimulation of thc rnotor nerve is known recruitment. 7 The tendons and the endomysium, perim sium, sarcolemma, and epimysium reprcsent par lel and series clastic components dUll stretch w aClive contraction or passive muscle extension a recoil with muscle relaxation. 8 Summation occurs whcn mcchanical sponscs of the muscle to successivc stimuli a added to an initial response. \Vhen maximal tensi is sustained as a result of summation, the mus contracts tetanically. The muscle fiber contracts an all-or-nothing fashion. 9 Muscles may contract concentrically, eccent cally, or isometricaH)' depending on the relationsh belween the muscle lension and thc rcsistance to overcome. Concentric and eccentric contractio involve dynamic work, in which t he muscle move joint or controls its 111ovcn1cnl. ',10 Force production in muscle is influenced the length-tension, load-velocity. nne! rorce-time lationships of the muscle. The length-tension re tionship in a whole Illuscle is influenced by both tive (contractile) and passive (series and paral elastic) components. 11 Two other factors that increase force produ tion are prestretching or the muscle and a rise muscle tempenllure. 12 The energy for muscle contraction and its lease is provided by the hydrolytic splitting of AT Muscle fatigue occurs \\'!hen the ability of the mu cle to synthesize ATP is insufficient to keep with the rate of ATP breakdown during contra tion. 13 Three main fiber types have been identifie type I, slow-twitch oxidative; type IIA, fast-twit oxidative-glycolytic; and type liB. fast-twitch g colytic fibers. Most muscles contain a mixture these types. .,14 Muscle atrophies occur under disuse and i mobilization; muscle trophism can' be restor through early and active rcmobilization. REFERENCES Appell, H.J. (1997). Thl' muscle in the rch'lbilit~l!ion pro("e~s. OnJlOp{/{!t'. 26( 11).930-934. AITidson. I., Eriksson. E.. 6.:. Pitman, M. (1984). Neuromuscular basis of n:.. habilit'llioll. In E. Hunter & J. Funk (Eds.), RelllI!JiUtarioll the "ljlt/wl I\."m'c (pp. 210-234). SI. Louis: or C. V. Mosby. B;:lnkcr, B.O. (1994). Basic reaction of 1ll1lSt:!t:. In A.G. ErH!cl & C. Franzini-Al'fllstrong (Eds.), Myology (2nd cel.). N~\\' )'ork: McGraw-HilI. Inc. Baratta R.V., Solornono\\', \1., Z!lOtl, S.H. (1998), Frequency domain-based llwdds of skeletal llluscle. J ElcclroJllyogr Kilu:siol, 8(1}, 79·91. Billes. B.T.. Ostl,.'l'lling. L.R., James. S,L. (1977). Fati!!uc ef· feels in running. ) MOlor Be/ull'. 9. 203-107. - Brobeck, J. R. (Ed.) (1979). Bes/ lIml Ttl.do,.:, Physiological Bt/· sis of :\let/iclIl Practice (10th cd., pp. 59-113). B~llimore: \\'illinms &. Wilkins. Brooke. ;\'1.1-1. &. K~iscr, K.K. (1970). Thn:e lllyosin adenosine Iriphosphat~sl: systcms: The nature of t1~cir pH liability and sulfhydryl depi.'nd"llce. J l-listodl/!JIl Cy/oclli:lll. 18, 670. Blll;ht~lhJ. F & Sohl1li.\lburch, H. (1980). Motor unilS of Ill;llll11l;llian Illusdt.'. Phniol Ret'. 60. 90. Burh-, R.E .. Levinc, D.N.. Z~lj~~C. F.E. (1971). M~lnlllwlinn mo· tor unils: Physiological hislochemic~ll corrckltion in Ihrce types of motor units in cal gastrocllemius. Seicllct.', 17-1, 709. Cost ill, D.L.. Coyle, E.F., Fink. W.F., l.. eslllcs, G.R .. Wilzmann. EA. (1979). Adapt.uioll in skeletal llluscles following strengLh Lraining. J .-Ippl Pln'siol, 46.96·9, ~ Craig. R. (199-4), The structur~ of (he contract filaments. In A.~. ~ngd ,&. C. Fr<lnzini-Annstrong (Eds.), ,\(mlogy (2nd t.'cI.). New )ork: i\1cGraw·Hill, Inc Crawford. C.N.C. &. J<101es. N.T. (1980). The dl'sign of muscles .. In R. O""..'n, J. Goodfellow, ~ fl. Bullollgh (Eds.) Sci· ~1!.lilie FOtllldaliOll.' Onltopacdics (lIId Tnl1lll/tllO!ogy (pp. 0/-74). London: Willi.lm Heinemann. Cuillo, J.V. & Zarins, B. (1983). Bioml,.'chanics of the mLlSt:Ulotcndinous unit: Relation 10 athk-lic performance and in· jury. Clilt Sports ,\lcd, 2. 71. Eriksson, E., Hagglll;uk, T, Kiessling, K.H., l,.'t;:\1. (1981). EffeCI of dectrie:ll slimulation on human skelt.-Ial musde. 1m Sports :lIed. 2, IS. Galler, S.. Hilbcr. K. (1998). Tension/stirrncl-is ratio of skinned rat skelcLallllllsde fibn.' types at v:\l'iolls lCl11pCratllrt's. Ac/ Physio! Sctlml, 162(2), 119-126. Gollnick, P.O. (1982). Ri..'lationship or strcngth and cndllr<lIlce with skeletal muscle structure and llll,.·tabolic potential. lilt J Spons ,\1t:tI (SliP!'!]. 3, 26. Gordon, A.M., Huxley, A.F.!., &. Julian, FJ. (1966). The variation in isol1\l,.'tric 1i..'nsiOll with s;trC<llllt.'rt.' kngth in \'I:.'rtebratc Illuscle fibers. J Physiol, 18-1, 170. Guylon, A.C. (1986). Tex/book ,\ledieal Pltysiolog\' (7th e(l.). Philndelphia: \V.B. Saunders. . Hnm,. A.W. &. Corm:lck, D.H. (I979). Nis{olo~y (Sth cd.). Phl1adclphia: J.B. Lippincott. Hill, A. V. (1970). Firsl (//Ill L{lsi £'\"/1(:ri1l11:11/:; ;11 ,\fl/sell! ,lh'· dUlllies. Cambridgl..': C;imbridgl: Uni\'cl"sitv Press. Hill, O.K. (1968), Tl.:nsion due 10 interaction bt.~lween lhe sliding filal1lt.'nts of resting $lriatcd museit.-. Thl,.' dft.'cr of stimulatinn. Pltysio! (LOllll], /99,637. or or -' Hi!ilop, H.J. ~ Perrine, J. (1967). The isokinelic I.:Olltl.'p cn:isc. Ph."..; Ther, -Ii. 114. Huxley, A.F (1974). Muscul"r contraction. J Physi<:d, 2Huxley. A.F &. Huxll.'Y, H.E. (196-1). Organizcrs of:t disc of Ihe physical and chemical basis of l1l11st.:lllar tO tion. Pmc Sec. 8160. -433. Kannus. P., JOI.5", L.. K\·j:.;1. ,\1., Jarvinen, T" J;lninc (1998aL EHeus of immobilizaLion and sllbseqm'n and high. intensity excrcise on morphology of I'~\t tal cles. Scal/t! J ,\Jed Sci Sports. 8(3), 160-171. Kant1us. P., Jozsa, L.. Jar\'int'n, T.L., K\'ist. 1\'1.. Vie Jilrvinen, ~\'t. (1998b). Fn:e mobilization ,11ld low- to intensity exercise ill immobili/.ation-indllccd muscl phy. J ..\ppl P/I\'s;ol, 8-1(4), 1418-1424. Kassel', l.R. (1996). Celleral Knowledge. III J.R. Ka5sI.. Onltop{lcdic Kllou·l.... d~i! lIpt/tltf! 5: I-Ioml.' Sr/flh- S\' Illinois: American :\cad('my of Orthopaedic Su;·!!c(; Kedc. C.r\ .• Nl.'il. E., ..\: Jock N, (1982). \Iusch: a,;d Ih vous system. In StllJl:WI/ W,.ighl~ Applied Phrsiolo"\" cd .. pp. 248-259). Oxford: O.xford Unin.'l'siL\: Pres~~' Komi, P. V. (1986). Tbl,.' slrl,.'tch-shorti.'nin~ t:v~le and power OUlpUt. In N.L. JOllCS, N. McCan·!lt.'\·, ~ A. Con~s (Eds.) I-ItlIIltlJl "'//SciI.' POII'a (PI'. 2'7-39). paign. IL: 1-lull1an Kin .... tics Publishers. r\.I'Oellll,.'l', I\..H.E., Marr:ls. \V.M., McGlothlin, J.D., (1990). On the IlH::ISUl'l..'lllent of hllm~\ll stf'I..'Il!.nh. III elliS! [;'/",!!,OIlOlllics, 6, 199-21 O. ~ Kroll. P.G. (1987). The <fjfi:Cl oFpl"t.'l'iolls co!//rac/ioll cO 011 SU!lS(,({IIl'11I l..'cn'Jllric PO\\'('I" production iu dhol\' IIIIISC!CS. Unpllbli~lwd doctoral dissertaLioll, New University. Nt::\\· York. Lkbcr, RL. &. Bodine·Fowler, S.c. (1993). Sh'Il,.,t~d 1ll1lS ~hanics. ImpliGllions for n.'hahilit<ltion.l'h.l's Tho: n 844-856. Linke, W.A .• In:·l1lcyl,.'l', ~L '\-llInde!. P., Stockm,·icl'. Kolml,.·rL'r, B. (1998). N~lturl,.· of PEVK-titill e1asli skclet~ll muscle. PJ'(l(: NIlt! '-\1.:lId Sci liS..\. 8052-8057. Luciano, D.S., Vander, :\.1 .. .& Sherman, J.H. (1978). I I-'Il.I/clio/1 alll! Strl/('ltlrt' (pp. 113-136). New York: \k Hill. \'lul1s.al. TL. NcNl.'al. D.. & Wat(-rs. R. (1976). Eff lll,.'l"\,e stimulation on human must-Ie. :lrcll N<'lI.rol, 3 Ohir~l. Y.. Y~\slli, w., Roy, R.R., Edgerlon, \f.R. (199;). of mllscl\: ll:ngLh on 1111: r('spons\: to unloadirH2. ..le! ([Jasen. 159(2-3), 90-98. Ottoson, D. (1983). I'hysio!o.!,:y of lhi' NO"l'IllIS 8"SI<,1 78-116), New York: Oxford Univcrsitv Press. . Parnianpou[", ~...I.. Nordin, M .. Kahano\"iLz. N.. "'I al. The triaxial collpling torque generation of trunk m during isometric I..'xl,.'nioI1S and tlti.' l,.'fb.:1 of fat isoinerlial mOH'ltIents on the molor output <lnd mOH pallcrns. Spillt.', 13(9). Pate, E., Bhimani, i\l., Fnmks-Skib'l, K., Cook. R. (199 duced ('ffCCI of pH on skinned r~lhbit jlSO:lS musc chanics at high temperatures: Implications for f"l P!lysiol (Loud), -I86(Pt 3), 689-694. Phillips, C.,x. &. Pctrofsky, 1.5. (1983) . .\lrchfUlics of S alltl e(mliut.' ,\fuse/e. Springfield; Charles C. Thon1<ls S"ltin. B.• el al. (1977). Fiber types and I1K'tabolk polt of skdelal Illuscles in sedcntal"}' m;m and enduranc n(;rs. :11111 NV ..lead Sd, 301. 3. Snndm<tnn, ~I.E .. Shoemanll. J.A .. Thompson. L. V, (1998). Th..:- fib1.'r·typ ...·spccific effcct of inactivity and intcrmillellt wci2ht·henring on the !!aslroc!lcmius of 30'lllollth-old rats~ Arch PI/\'~ ,lied UCllllb;/, 79(6), 658-662. Squire. J.~1. (1997). Architecture and fUllction in thl.: muscl~ s~lI·comcrl·. e/ln Opill Stl'llCI Bio!. 7(2). 247-257. Strom~r, M.H. (1998). Thc cytoskelelon in skckt;-.l. c:lI'di:1C :lIId smooth muscl~ (dis. His/o! f/islOpa/llOl, / 3( I). 283-291. -'::' ' '-'" i.: ",' '''" T:lbrad~l, Y., (wnll\oto, 1-1., Sugi. 1-1 .. Hir~ln(), Y.. Ishi (1997). Strctch-ilH.luccd l'nhanCCIllc.'Jll of m':"chanical production in long frog single fihers and hUIll<ln mus App! Physio!, 83(5),1741-17-4$. Wilkie. D.R. (J956). The mechanical propenies of llluscl .\It'd BIIII, /2, 177. Wilkie, D.R. (1968) . .\lusc/c. london: Edward Arnold. Williams, P. &. Warwick. R. (1980). Gray:.; rlJlll/om." 06tl1 pp. 506-5(5). Edinburgh: Churchill Livingstone. z 0 ;= --:----- ~ ---------1 e- o Z .~ 0 u ~ 1< >!' z " 0 0 "9 :> ~ 0 > ~ v '" E <;; , .... 0; 0; ,.' .,; ~ '" (; c .2 :;; .~ c ,·_------·1·----------1' ',." ,. . ~ . '" <; -0 . ·· ·-~ ~ -~= 6 'g" .~ .~ "=c :, <; = <; e t ~ ~ 2 ~ ~ < § "= 0 .~ " .~ ", -5 g ;:: ........ - .. " <; '" l' C , "= ~I . ~ c. ,~:':,i$;\') E .2X ;;; v .a ;1~& 0 c ~ :s .:)~;;;: '"'" E ,.', u.... '.;'x. ~ j,t-:;:". . e2? ~ <~- ij X ~ ",:tt:./ s" .~ ." ~ v ~ ~ E -S .~ 0 0 £ 0 0 E -0c .- " .;; 0 3 0 ;;: " ." .~ ., '5 .!!i . ~ 0 ~ ~ ~ §v .ev :~ x w ~ , 0. 0 c, i; E 0 e~ 0 v 2 ." 0 C '" .~ ." :". .~ .;;" 0 3 .g .0 .a '" -,~, . "., ~. ~ Intrinsic disorders associated with muscle damage. Clinical examples. * *This flow chart is designed for classroom or group discussion. Flow chart is not meant to be exhaustive. Biomechanics of Joints Biomechanics of the Knee Margareta Nordin, Victor H. Franke Introduction Kinematics Range of Motion Surface Joint ~J1otion Tibiofemoral Joint Patellofemoral Joint Kinetics Statics of the Tibiofemoral Joint Dynamics of the Tibiofemoral Joint Stability of the Knee JOint Function of the Patella Statics and Dynamics of the Patellofemoral Joint Summary References Introduction The knee transmits loads, part III pales in motion, aids in conservation of momentum, and provides a force couple for activities involving the leg. The human knee, the largest and perhaps 1110st complex joint in the bod)!, is a two-joint structure composed 'of the tibiofemoral joint and the patellofcmoral joint (Fig. 7-1). The knee sustains high forces and moments and is situated between the body's two longest lever arms (the femur and the tibia), making it particularly' susceptible to injury. This d18ptcr utilizes the knee to introduce the basic terms, explain the methods, and demonstrate the calculations necessary for analyzing joint motion and the forces and moments acting on a joint. This information is np· plied to other joints in subsequcnt chaptcrs. The knce is particularl.. . . \vell suited for dem strating biomechanical anal.vscs of joints beca these anal.. .,scs can be simplified in the knee and yield useful data. Although knee motion occurs multancous!y" in three planes, thc motion in planc is so great that it accounts for nearly all of motion. Also, although many muscles prod forces on the knee, at any particular instant muscle group predominates, generating a force great that it accounts for most of the muscle fo acting on the knee. Thus, basic biomedwni analyscs can be limited to motion in one plane to the force produced by a single muscle group still give an understanding of kncc mc)tion and estimation of the magnitude of the principal for and moments on the knee. Advanced bioIllcchan d.\'namic anal,\'ses of thc knee joint that include Anterior cruclate ligament Medial collateral ligament Popliteus tendon Fibular collateral ligament Semimembrano Biceps femoris tendon Superficial medial collat ligament (cu Iliotibial band (cut) Interosseous membrane Transverse ligament -cr--l'- Gracillis Semitendinosus Patellar tendon B Two-joint structure of the knee. A, Lateral view of a knee joint with open growth plates. B, Anterior view without patella. Sartorius soft tissue structures are complex and still under investigation. Analysis or motion in any joint requires the lise of kinematic daw. Kinematics is the branch of mccrlanics thal deals with Illotion of a body without reference to force or mass. Anal:vsis of the forces and momenl~ acting on a joint necessitates the usc of both kinematic and kinetic data. Kinetics is the branch of mechanics that deals with the motion of a body under the action of given forces and/or moments. Kinematics Kinematics defines thl..' range of motion an scribes the sw-facc motion of a joint in three p frontal (coronal or longitudinal), sagittal, and verse (horizonlal) (Fig. 7 2, A & B). Clinical surements of joint range 01" motion defin anatomical position as a zero position for mc ment. This laxonomy will be uscd for joint m throughout this book. Other taxonomies and ence systems exist (Ant!riacchi et aI., 1979; Gr H Proximal/ dislal !j"~~;;:~i..k~~~i===::::=Fronlal ~1 plane •., I~. o k'. ~ -Sagittal plane ;,.j )1 ~ Internal/external Varus! valgus A A, Frontal (coronal or longitudinal), sagittal. and transverse (horizontal) planes in the human body performed easily for both the tibiofemoral and the patellofemoral joint. B, Depiction and nomenclature of the six degrees of freedom of knee motion: anterior posterior translation, medial/lateral transla- .,. . .<:, ";'~' , I?:A B tion and proximal distal translation, flexion-extension tion, internal-external rotation, varus-valgus rotation. Adapted from Wilson, S.A., Vigorita, V.J., & Scott, W.N (1994). Anacomy. In N. Scott (Ed.), The Knee (p. 17). Philad Mosby- Year Book. Sun tal' 1983; Kroemer ct aL. 1990; bzkaya & Nordin, 1999), but the anatomical reference system bv far is the mOSl commonlv lIsed among clinicians. the two joints composing the kne'; the tibiofemoral joint lends itself particularly well to an analysis of range orjoinl motion. Analysis of surface joint motion can be performed easily For both the tibiofemoral and the P<'11cllofcmoral joint. An~' impediment of range of motion or surface joint 'motion will disturb the normal loading pattern of' a joint a.nd bear consequences. O'r -20 !:=------;!':---7::----==---:'::-100 20 40 60 80 10 RANGE OF MOTION The range Percentage of Cycle or motion of any joint can be measured in any plane. Gross measurements can be made with a goniometer, but more specific measurements require the lise or more precise methods such as elcclrogoniomelr~·. roenlgenogJ"aphy. stcreopllologralllmetr)'. or photographic and video techniques using skeletal pins, In the tibiofcmoral jail'll. nlOtion takes place in all three planes, but the range of motion is greatest by far in the sagittal plane. Motion in this plane from full extcnsion to Fulillexion of the knee is from 0' to approximately 140°. Motion in the transverse plane. intel-nnl and external rotation. is inOuenced by the position of the joint in the sagittal plane, With the knee in Full extension. rOHHion is almost completely restl'ictcd b~f the interlocking of the femoral and tibial condyles. which occurs mainly because the medial femoral condyle is longer than the lateral condyle. The range of rotation increases as the knee is flexed, reaching a maximum at 90° of nexion; with the knee in this :- position. external rotation ranges from 0° to approximately 45° and internal rotation ranges from 0' to appmximatcly 3D'. Beyond 90' of Ilexion. the range of internal and external rotation decreases, primarily because the soft tissues restrict rOlation. Motion in the frontal plane. abduction and adduction. is similarly aFfected by the amount of joint flexion, Full extension of the knee precludes almost all motion in the frontal plane. Passive abduction and adduction increase with knee Oexion up to 30°. but each reaches a maximum of only a few degrees. With the knee flexed beyond 3D'. motion in the frontal plane again c1ecre~lses because of the limiting function of the sol't tissues. The range of tibiofernoral joint motion required For the performance of various physical activities can be determined from kinematic anal~lsis. MOlion Range of motion of the tibiofemoral joint in the sagittal plane during level walking in one gait cycle. The shaded dred indicates variation among 60 subjects (age range 20 to 65 years). Adapted {rom MtJ((a}~ M.P., Drought, A.B., Kor} R. C. (1964j. I/o/alking patterns of norma! men. J Bone Joint Su 46A,335, in this joint during walking has been measured all planes. The range of motion in the sagittal pla during level walking was measured with an e1ectr goniometer by Lamoreaux (1971) and MUlTay et (1964). Full or nearly Full extension was noted at t beginning of the stance phase (0% of cycle) at he strike, and at the end of the stance phase before LO off (around 60% of cycle) (Fig, 7-3), Maximum fle ion (approximately 60°) was observed during t middle of the swing phase (sec Chapter 18, Biom chanics of Gait. For more detailed inFormation These measurements are velocit.y-c1ependent an must be interpreted with caution. ~"lotion in the transverse plane during walking h been measured by several investigators. Using a ph lOgraphic technique involving the placement skeletal pins through the femur and tibia, Leve and associates (1948) Found that total rotation of t tibia with respect to the fernur ranged from appro imately 4 to 13' in 12 subjects (mean 8,6'). Great rotation (mean 133') was noted by Kettclkamp an coworkers (1970). who used electrogoniollletry o 22 subjects. In both studies, external rotation beg during knee extension in the stance phase an reached a peak value at the end of th~ swing pha just before heel strike. Internal rotation was note during Ilexion in the swing: phase. DDIII-~ Range of Tiblofemoral Joint Motion In the Sagittal Plane During Common Activities I Activity Range of Motion from Knee Extension to Knee Flexion (Degrees) - - - - - - - - -0-67" ------, Walking ""I Climbing stairs Descending stairs Sitting down Tying a shoe lifting an object 0-90 0-93 0-106 0-117 ';Dillil lrom Kel1clkamp el al. (1970i. fvlean for 22 subjects. A slight ddiercncc was found bel'.'.'een r19ht "nc! lei! knees (mean for right knee 68.1°, nW,ln for left knee 66,r). "These and sub:.equerH date' from lauberllhal Cl al. (1972). Mean for 30 subjects. Kcltclkamp's group (1970) also mcasurcd motion in the frontal plane during walking. In nearly all of the 22 subjects, maximal abduction of the tibia was observed during extension at heel slrike and at the beginning of the stance phase; maximal adduction occurred as the knee \Vas flexed during the swing phase. The total amount of abduction and adduction averaged 11°. Values for the rangc of motion o!" the tibiofemoral joint in the sagittal plane during several comn101l activities arc presented in Table 7-1. l'vlaximal knee flexion occurs during lifting. A range of motion I'rol11 full extcnsion to at least II r 01' flcxion appears to be required for an individual to carry out the activities of daily living in a normal manner. Any restriction of knee motion can be compensated for by increased motion in other joints. In studying the range of libiofcmoral joint motion during various activities, researchers found that an increased speed of movement requires a greater range of motion in thc tibiol'cl11oral joint (Holdcn et aI., 1997; Pcrry ct aI., 1977). As the pace accelerates from walking slowly to running. progressively more knee flexion is nccdcd during thc stancc phasc (Table 7·2), SURFACE JOINT MOTION SUrract~ joint motion. which is the Illation between the articulating surfaces of a joint. can be described 1'01' any joint in any plane with the lise of stcrcophoLogrammetric mcthods (Sclvik, 1978, 1983)~ Bc~ callsc lhese mcthods ~lre highly technical and com plex, a simpler method evolved in the nineteen ccnL'It'\, Is sLillus~d (Reuleau,\, 1876), This m~,h() called the installt cl.?nter techniqul.', allows surfa joinl 1l1Olion lO be analyzed in the sagiual an froolal planes but not in the transverse plane. T instant center l~chniqLle pl{)vides a description the rch'tive uniplanar motion of two adjacent se ments or n body and the direction of displaceme or the contact points bct\\'L~cn these segmcnts, The skeldal portion 01" a body segment is called link. As one link rotaLes ~il){H1l the other. al any i stum then... is a point that docs nOl mov(:, that is, point thal has h'ro velocil.V. This point conslitlli an instantaneous cenLer of motion, or instant ce tel: The instant center is found b.\' identirying t displacen1cnt (If two points on a link as the li moves from one position to another in relation to adjacl.?nt link, which is considered to he stalionar The points on thl.' moving link in its original po tion and in its displaced position arc designated a graph and lines an.' drawn connecling the two se of points. The pCl-pcndicular bisectors or these tw lines are then drawn. The intersection or the pe pendicular bisectors is thc instant ccntcl: Clinir::'lll~', a path\\'a~' of the instant center for joint call be determined b\' takinQ - successive roen genograms of the joint in dilTcrent positions (usual 10° apart) throughout the range of motion in o plane and applying the Rculcaux melhod for locali the inslanl center 1'01' each intcrv[ll or motion. \Vhcn the instant center palhway has been dete mined 1'01' joint motion in one plane, the surfa joint mol ion C[ln be described, For each intel'val motion, the point at which the joint surfaces ma contact is located on the roentgenograms lIsed f the instant center analysis, and a line is drawn frol ~ ~ 1Imtl!JIJI--~ I -~~-,-~----~ Amount of Knee Flexion During Stance Phase of Walking and Running Activity Range in Amount of Knee Flexio During Stance Phase (Degrees) W(llking Slow Free Fast Running Data 0-6 6-12 12-18 18-30 itom Perry ct al. (1977). Range for seven subjects. the instant center [() the contact point. A second line drawn at right angles to this line indicates the direclion of displacement of the contact points. The direction of displacement of these points throughout the range of motion describes the surface motion in the joint. In most joints, the instant centers lie at a distance from the joint surface, and the line indicating the direction of c1isplaccrncnt o!" the conwct points is tangential to the load-bearing surface, demonstrating thm one joint surFace is gliding on the other (load-bcaring) surface. In the casc in which the instant center is found on the sudace. the joint has a rolling motion and there is no gliding function. Because the instant center technique allows a description of motion in one plane only, it is not useful for describing the surface joint motion if more than I5° of motion takes place in any plane other than the one being rneasured. In the knee, surface joint motion occurs between the tibial and fcmoral condyles and between the fcmoral condylcs and the patella. In thc tibiofemoral joint, surface motion takes place in all three planes simultaneously but is considerably less in the tra verse and frontal planes. Surface motion in patellofcmoraljoint takes place in two planes sim tancoLisly, the frontal and transverse, but is greater in the frontal plane. Tibiofemoral Joint An example will illustrate the lise of the instant c ter tcchniquc to describe the surface motion of tibiofemoraljoint in the sagittal plane. To determ the pathway of the instant center of this joint dur flexion, a lateral roentgenogram is taken of the k in full extension and successive films arc Laken 10° intervals of increased nexion. Care is taken kcep the tibia parallel to the ,-ray table and to p vent rotation about the femur. \Vhcn a patient limited knee motion, the knee is Ilc,ed or e'tend only as far as the patient can tolerate. Two points on the femur that arc easily iden ned on all roentgenograms are selecled and des Locating the instant center. A, Two easily identifiable points on the femur are designated on a roentgenogram of a knee flexed 80°. B. This roentgenogram is compared with a roentgenogram of the knee flexed 90°, on which the same two points have been indi~ cated. The images of the tibiae are superimposed, and lines are drawn connecting each set of points. The perpendicular bisectors of these two lines are then drawn. The point at which these perpendicular bisectors intersect is the instant center of the tibiofemoral joint . for the motion between 80 and 90° of flexion. Courtesy of Ian Goldie, M.D. Univ~rsity of Gothenburg, Gothenburg, Sweden. natcd on each roentgenogram (Fig. 7.4..4). The films are then compared in pairs, with the images of the tibiae supcl-imposed on each othcl: Roentgenograms with marked differences in tibial alignrncnt arc not used. Lines are drawn between the points on thc felllur in the two positions, and the perpendicular bisectors of these lines arc then drawn. The point at which these perpendicular bisectors intersect is the instant center or the tibiofemoral joint for each 10° interval of motion (Fig. 7-4B). The instant center pathway throughout the entire range or knee Oexion and extension can then be ploued. In a normal knee, the instant center pathway for the tibiofemoral joint is semicircular (Fig. 7-5). After the instant center padnvay has been determined for the tibiofemoral joint. the surface motion can be described. On each set of superimposed roent-gcnograms the point of contact of the tibiorcmoral joint surfaces (the narrowest point in the joint space) is determined and a line is drawn connecting this point with the instanl centet: A second line drawn at right angles to this line inclicates the direction of displacement of the contact points. In a normal knee, lhis line is tangential to the surFace of the tibia for each interval of motion from full extension to full flexion, demonstrating that the femur is gliding on the tibial condyles (Frankel et aI., 1971) (Fig. 7-6). During normal knee motion in the sagittal plane from full extension to full flexion. the instant center pathway moves posteriody, forcing a combination of rolling and sliding to occur between the articular surface (Fig 7-6. A & B). The unique mechanism prevents the femur from rolling off the posterior aspect of the tibia plateau as the knee goes into increased flexion (Draganich et aI., 1987; Fu et aI., 1994; Kapandji, 1970). The mechanism that prevents this roll-off is the link formed between the tibial and Femoral attachment sites of the anterior and posterior cruciate ligaments and the osseous geometry of the femoral condyles (Fu el aI., 1994) (Fig. 7-6, B-D). Frankel and associates (1971) determined the instant center pathway and analyzed the surface motion of the tibiofemoral joint from 90· of Ilexion to Full extension in 25 normal knees; tangential gliding was noted in all cases. They also determined the instant center pathway for the tibioFemoral joint in 30 knees with internal derangement and found that, in all cases, the instant center was displaced from the normal position during some portion of the motion examined. The abnormal instant cen- Semicircular instant center pathway for the tibiofemor joint in a 19-year-old man with a normal knee. tel' pathway for one subject, a 35-~"ear-old m with a bucket-handle derangement. is shown Figure 7-7. If the knee is extended and Hexed aboul an normal instant center pathway, the libiofem joint surfaces do not glide tangentially throu out the range of motion but become cither tracted or compressed (Fig. 7-8). Such a kne annlogous to a door with a bent hinge that longer fits into the door jarnb. If the knee is c tinually forced to move about a displaced ins center, a gradual adjustment to the situation be reflected either b~: stretching of the ligam and other supporting soft tissues or by the im sition of abnormall~' high pressure on the art lar surfaces. Internal derangements of the libiofcmoralj may interfere with the so-called screw-h mechanism, which is external rotation during tension of the tibia (Fig. 7-9). The tibiofem ~ioint is not a simple hinge joint; il has a spiral helicoid, motion. The spiral motion or the t about the femur during flexion and extension sults from the anatomical configuration of medial femoral condyle; in a normal knee, th condyle is approximately 1.7 ern longer than th lateral condyle. As the tibia moves on the femu from the fully flexed to the fully extended pos tion, it descends and then ascends the curves the n1edial femoral condyle and simultaneous rotates externally!. This motion is reversed as th tibia moves back into the fully flexed positio This screw-home mechanism (rotation around th longitudinal axis of the tibia) provides more st bility' to the knee in any position than would simple hinge configuration of the tibiofemor joint. Matsumoto et al. (2000) investigated the ax of tibia axial rotation and its change with kne flexion angle in 24 fresh-frozen normal knee ca daver specimens ranging in age From 22 to 6 years. The magnitude and location of the long tudinal axis of tibia rotation were measured 15° increments between 0 and 90° of knee fle ion. The magnitude of tibia rotation was 8° at Gliding A 8 c I D 1mIlI'------I _ A. In a normal knee, a line drawn from the instant center of the tibiofemoral joint to the tibiofemoral contact point (line A) forms a right angle with a line tangential to the tibial surface (line B). The arrow indicates the direction of displacement of the contact points. Line B is tangential to the tibial surface, indicating that the femur glides on the tibial condyles during the measured interval of motion. B. Pure sliding of the femur on the tibia with knee extension. Note that the contact point of the tibia does not change as the femur slides over it. Eventually impingement would occur if all surface motion was restricted to sliding. Round points delineate contact points at the femur and triangles delineate contact points at the tibia. C. Pure rolling of the femur on the tibia with knee flexion. Note that both the tibia and the femoral contact points change as the femur rolls on the tibia. Also note that with moderate flexion, the femur will be~ gin to roll off the tibia if surface motion was restricted to rolling. D, Actual knee motion including both sliding and rolling. Abnormal instant center pathway for a 35-year-old man with a bucket-handle derangement. The instant center jumps at full extension of the knee. Adapted from Frankel, VH., Burstein, A.H., &- Brooks, D.B. (1971). Biomechanics of in ternal derangement of the knee, Parhomechanics as determin by analysis of the instant centers of motion. J Bone Joint Surg, 53A, 945. of knee flexion. The tibial rotation increased rapidly as the knee flexion angle increased and reached a ma:dmunl of 31 0 at 30 0 of knee flexion. IL then decreased again with additional flexion (Fig 7-10). The location of the longitudinal rotational axis was close to the insenion of the anterior cruciate ligament (ACL) at 0° of flexion. At continuous flexion up to 60°, the I'otational axis moved toward the insertion of the posterior cruciale ligament. Belween 60 and 90° of flexion, the rotational axis moved anteriorly again (Fig 7-11). This study showcd thaI lhc rolational axis remains approximately in the area between the two cruciate ligaments. Any change of direction and tension of the cruciate ligaments and surrounding soft tissue may affect the movement and the 'location of the longitudinal tibia axis of rotation and thcrcby affccl joint load distri- ~-L'-JI Extension \'1 '., ... External rotalion during extension Screw·home mechanism of the tibiofemoral joint. D knee extension. the tibia rotates externally. This mo reversed as the knee is flexed. A. Oblique view of t mur and tibia. The shaded area indicates the tibial solid line axis for knee flexion and extension, dotte internal and external rotation axis of the tibia duri ion and extension. AcJapred from Helier. A.J. (1974j. A and mechanics of movement of rite knee Joinr. In A. He Disorders of the Knee fpp. i -17). Philadelphia.- J B. Lipp bution. A clinical tcst. the Helfet test, is often useel to determine whether external rotation of the tibio- • Distraction A Compression B Surface motion in two tibiofemoral joints with displaced instant centers. In both joints, the arrowed line at right an· gles to the line between the instant center and the tibiofemoral contact point indicates the direction of displacement of the contact points. A. The small arrow indicates that with further fIeldon. the tibiofemoral joint will be distracted. B, With increased flexion, this joint will be compressed. femoral joint takes place during knee eXl thereby indicating whether the screw-home m nism is illlact. This clinical tcst is performe the patient sitting with the kn~e and hip flex ancl the leg hanging free. The medial and borders of the patella are marked on the ski tibial tuberosity and the midline of the pate then designated, and the alignment of the tuberosity with the patella is checked. In a n knee Hexed 90°, the tibial tuberosily aligns w medial half or the patcll" (Fig. 7-121\). The k then extended fully and the movement of the tuberosity is obsel\:cd. [n a normal knee. the tuberosity moves laterally during extensio "Iigns with the latent! h"lf or the patella at tension (Fig. 7-128). RotatOl')' motion in a n knee may be as great as hair the width patella. In a deranged knee, the tibia may not externally during extension. Because of the surface motion in such a knee, the tibiof joint will be abnormally compressed if the k forced into extension, and the joint surfaces m d.amaged. patellofemoral Joint The surface motion of the patellofcmoral joint in the frontal plane may also be described by means or the instant center technique. This joint is shown to have a gliding motion (Fig. 7-13). From full extension to full flexion of the knee, the patella glides caudally! approximately 7 em on the femoral condyles. Both the medial and lateral facets of the femur articulate with the patella from Full extension to 140' of flexion (Helme, 1990) (Fig. 7-14). Beyond 90' of flexion, the patella rotates externally, :md only the medial femoral facet articulates with the patella (Fig. 7-14B). At full flexion, the patella sinks into the intercondylar groove (Goodfellow et aI., 1976). The contact area of the lateral facet joint of the patella is larger than the medial contact areas and ranges from 0.5 to 2.5 cn1~ and less than 0.5 to 2 em"' ) respectively" Contact areas increase \vith an increased amount of flexion of the knee joint and increased pulling force of the quadriceps muscle (I-lehne, 1990). -r -r -.§ ro 30 -0 a: ro ·x 20 -- -- r-~ r-~ - -- -- ............. Tibia Plateau " .. .. ' .......... ACL • 0·'0 \. 5' Lateral '. ::..... '. .... Fibula 30 ". •••••••• ..45" 6t?~L,. . o • ••••••• . Axis locallon an standard deviat location of the axis of tibia axial rotation. The location the axis was close to the tibial insertion of the anterior cruciate ligament (ACL) at A" of flexion and gradually moved toward that of the posterior cruciate ligament a 45 and 90" of knee flexion. The axis then moved anteri again and was approximately at equal distance from th two insertions of the cruciate ligaments at 90" of knee flexion. ACL, insertion of the anterior cruciate ligament PCl, tibial insertion of the posterior cruciate ligament. Reprinted v'/ith permission from MatsurnolO, H., Seedhom, B Sue/a. Y, et £11. (2000). Axis of tibial rotation and its change flexion angle. Clin Orthop, 371, J78-182. 'r"o :0 i= ro 10 -0 >- ! o o IL - Kinetics 15 30 45 60 75 . 90 Knee Flexion Angle jIBm'------The total tibial axial rotation (y-axis) plotted against the magnitude of knee flexion (x-axis) in fresh-frozen specimens tested under unloaded axial conditions, The magnitude of tibial rotation was below 9" at A" of knee flexion. Maximum rotation (31.]0) was measured between 30 to 45" of knee flexion. Reprinted with permission from Matsumoto, H., Seedhom, B.B., Suda, Y, et al. (2000), Axis of tibial rotation and its change vi/ith flexion angle. Clin Orlhop, 371, 778-182. Kinetics involves both static and d,vnamic ana sis of the forces and moments acting on a jo Statics is the stud~/ of the forces and moments a ing on a body in equilibrium (a body' at rest moving at a constant speed). For a bod)' to be equilibrium. two equilibrium conditions must met: force (translator:v) equilibrium, in which sum of the forces is zero, and moment (rotato equilibrium, in which the sum of the moment zero. Dynamics is the study of the moments a forces acting on a body in motion (an accelerat or decelerating body), In this case, .the forces not add up to zero, and the body displaces and the moments do not add up to zero and the bo Knee flexed 90c Knee lully extended Helfet test. A, In a normal knee flexed 90°, the tibial tuberosity aligns with the medial half of the patella. B, When the knee is fUlly extended. the tibial tuberosity aligns with the lat· eral half of the patella . • After the instant center (lQ is determined for the patellofemoral joint for the~ motion from 75 to 90° of knee flexion, a line is drawn from the instant center to the con· tact point (CP) between the patella and the femoral condyle. A line drawn at right angles to this line is tangential to the surface of the patella, indicating gliding . • Superior Lateral M Medial Lateral .,~ ~ '\ (U - - - -30'-90" I I Inferior 20' 45' 90' Superior , ---120" Lateral Me I Inferior 135' ;0.. A IiR'm A, The position of the patella at different ranges of knee I flexion motion. a, Contact areas during different degrees of flexion. Beyond 900 of flexion, the patella rotates slightly outwards. Adapted from GoodfelJo'J'l, I. Hvngerford, D.5., & lin· del, M (1976), PateJlofemoral joint mechanics and pathology. I. B Functional anatomy of the patellofemaral joint. J 80ne Joint 58B. 287; and from Helme. H.I (1990). Biomechanics of the parellofemoral joint and its clinical relevance. Clin Onhop. 25 73-85, • rotates around an axis perpendicular to the plane of the fOl'ces producing the moments. Kinetic analysis allows one to determine the magnitude of the moments and forces on a joint produced by body weight, muscle action, soft tissue resistance, and externally applied loads in any situation, either static or dynamic, and to identify those situations that produce excessively high moments or forces. In this and subsequent chapters. the discussi stalics and dynamics of the joints of the skelela tem concerns the magnitude of the forces and ments acting to move a joint about an axis maintain its position. It does not lake into acc Ihe deforming effect of Ihese forces and mom on the joint structures. This effect is indeed pre but the discussion is not within the s~ope of this STATICS OF THE TIBIOFEMORAL JOINT Static analysis may be lIsed to determine the forces and moments acling on a joint when no motion takes place or at one instant in time during a dynamic activity such n5 walking, running, or lifting an objecl. 1t can be performed for any joint in any position and under any loading configuration. In such analyses, eHher graphic or mathematical methods may be lIsed to solve for the unknown forces or moments. A complete static analysis involving all moments and all forces imposed on a joint in three dimensions is complicated. For this reason. a simplified technique is oflen lIsed. The technique utilizes a frecbody diagrarn and limits the analysis lo one plane, to the three main coplanar forces acting on the frccbody. and ro the main moments ~cting about the joint under consideration. The minimum magnitudes of the forces and moments arc obtained. When the simplified Free-body technique is used to analyze coplanar forces, one portion of the body is considered as distinct rrom the entire bod.\', and all forces acting on this free-body arc identified. A diagram is drawn of the free-body in the loading situation (0 be analyzed. The three principal coplanar forces acting on the free-body are identified and designaLcd on the free-body diagram. These forces afC designated as vectors if four characteristics are kno\vn: magnitude, sense. line of application, and point of application. (The tcrm '"direction" includes line of application and sense.) If the points of application for allthrec forces and thc directions for two forces arc kno\vn, all remaining characteristics can be obtained for a force equilibrium situation. \Vhen the free-body is in equilibrium, the three pl-incipal coplanar forces are concurrent; that is. they intersect at a common point. In other words, these forces form a closed system with no resultant (i.e., their vector sum is zero). For this reason. the line of application for one force can be determined if the lines of application for the other two forces arc known. Once the lines of application for all IIwee forces are known, a triangle of forces can be constructed and the magnitudes of all three forces can be scaled from this triangle. An example will illustrate the application of the simplified Ii'ee-body techniquc for coplanar forccs to the knee, In this case, the technique is lIsed to estimate the minimum magnitude of the joint reaction force acting on the tibiorcmoral joint of the weightbearing leg when the other leg is lifted during stair climbing. The lower leg is considered as a free-body. distinct from the rest of the bod.". <'lIlt! a diagram o this free-bod." in tilt..' stair-climbing situation is drawn (Calculation Bo,,,; 7-1). I":'nlrn all rorces acting on thl..: I"rct..··body. the three main coplanal- I"o('(,:(:s an idenlified a~ the ground rL'action force (equal to body weight I), the tensile force through the patella tcndon exel"ted h.v llw quadriceps Illust..:ll' , and th\. joint reaction force on the tibial plateau. The ground n.:action force (\rV) has a known m<'lgnitude (L'qu;;llto body wt..'iglu). S~IlS('. line or <'lpplicalion. and poinl o application (poii'll or contact between the foot and the ground). The palellnr tendol1 force (P) has a known SC'IlSL' (awa.'; from the knee joint), lint..' of ap plicaliun (along lhe patellar t('ndon). and point application (point or insertion of thc patellar tendol on till' tibial tulx.·rosit."), but an unknown l1lagni tude. The joint n:action force (J) has a known poilU of application OJl the slIrl"acL' or the tibia (the COJ1lac point of tilL' joint surfaces betw(.'en the tibial and femora! cond,vlc:s, estimated from a roelllgenogram or the joint in the pn)!x~r loading configuration), bu an unknown magnitude, sense. and line of applica tion. Using \'L'ctors cakuh\lions <:\11('\ triangles law the joint rL'action force (.1) and thl' pall.'lIar tendon force (P) call be c,dculated (Calcul"tion Bo\ 7-1) It call bt..· seen thaI the main musck forct..' has llluch gre<.ltL'1' inlhICIll:L' on the magnitude of" th .joint reaction force than dot..·s the -~round !"(.'actio force produced b." bod~· \vcight. Note lhat, in this ex amplc (Calculation Box 7·1), only til(' minimum magnitude of the joint rt..'~lction force has been calculated. If other muscle I"oret-'s arc considered such as tht..' foret..' produced b~' tht..' contraction of th halllslring Inusc!es in stabilizing the knee, the jClin reaction force inCl·cast..·s_ Th\.' ne:'.;t Slt.·p in the static analysis is ~\I1;:d."sis o the moments acting around the centcI- of motion 0 the libiofcrnoral joint with the knee in the same po sition and lht..' loading. configuralion shown in Ca culation Bo:'.; Figure 7·1-1. The 1ll01llL'11I <'lIlal.\·sis i Llsed to cstimak' the minimulll ll1agnitudL' of th mOIl1L'nt produced through the patellar h:ndon which counterbalances the 1l10ll1Cnt on the lowcr le produced b.'; the weight or the hod." as the subjec ascends stairs (Calculation Box 7·2). o C:lSl', the ground I"l'~ll"lillll rurn' is ;u.::lu;t1ly L'qllallo hod mintl:- (he \\"..::i!..:hl of till' IO\\"L'r k·t:. BL'l";III:-l' tit...· \widll o the 'it'Hwr k~ is lllillilT~;ll {1L'SS (hall Olli.' 1:'l1(h or Ihl' bOtl\-), i( CI lin lhi:, \\"L'idlt l.k' di,.. rl·~aranl. ;llld the li~url' for tlll:tl hod\- \\"l'i!.!hl l\\·U bl' tu lizL'd ill Illl' l·akul:lIion. . .- ~ !, Free-Body Diagram of the Knee Joint Force P .., ~ I ,.. \.\ .< Known: Point of applicalion ;':" \Q', Known: • ~ Sense The three main coplanar forces acting on the lower leg: ... Force J Line of application Point of application Unknown: Magnitude UnknO'.·',n: Magnitude Sense \ Ground reaction iorce (W). patellar tendon force (P), and joint reaction force (J) afe designaled on a free-body diagram of th lower leg while climbing stairs (Calculation Box Fig 7-1·1), Line of application \\ .,•..,-..__ ....\ ". -'. \. <\. ',-.\ I '. \ "- :i~ /\J o:__ ~~ I · I---==~-r"'iKnown: ~- /' Magnilude Force W Sense Line of application !Point of application . Because the 100\'cr leg is in equilibriulll, the lines of applicatio .' Force J " "'i v~ .:.~ for all three forces intersect at one point. Because the linesof l~~:\ T ibialemoral Force P \ ( \ ..... i contact point \~ \. ,, [I,.VO forces (Wand P) are known, the line of a plication for the third force (;) can be determined. The lines o application for forces Wand P are extended until they jnter~ '.'. .\:;.... sect. The line or application for J Gill then be drawn from its , \\........ '. \ point of application on the tibial surface through the intersec \ tion point (Calculation Box Fig. Intersection ~\ pOint~ I application for 7-1-2). -----·-·-~·=::::·:"~~W-1J.-·'""'---~---- Now that the line of application for J has been determined, it is possible \'.,, I Force J 4.1W Force P ,, 3.2W . ,, .... \, '. , '. , '. , '. , '. , '. '. '. '. _ drawn from the head of ve([or W. Then, to close the triangle force J is drawn from rhe head of vector W. The point at whic Now that the length of all three vectors is known, the magni tude of forces P and J can be scaled from force W, which is . . .J . constru([ a triangle of forces (Calculation Box F forces P and J intersect defines the length of these vectors. Force W {, (0 7-1 ~3). First, a ve([or representing W is drawn. Next. P is equal to body weight. It is determining the number of limes the length of force W can be aligned along the' force P and J, respectively. In this case, force P is 3.2 times body weight, and force J is 4.1 times body weight. Free-Body Diagram of the Lower Leg During Stair Climbing The two main mornents acting around tt1e center of motion the tibioiemoral JOlfH (solid dOl) are designated on the free body diagram of the lower leg during stair climbing (Calcul ( tion Box Fig. 7-2-1). The fleXing moment on the lower leg is the product of t \\ weIght of the body:- (W, the ground reaction force) and its \/\ \ fever arm (a), 'Nhich is the perpendicular distance of the for Force P \ W to the center 0; motion aT the tibiofemoral joint. The co terbalanCIng extending moment IS the product of the quad ceps muscle force through the patellar tendon (P) and its le arm (b), Because tile lo""er leg is in equilibrium. the sum of these two moments must equal zero (~M = 0). In this example. the counterclockwise moment is arbitrari designated as positive (W x a - P ~< b = 0). Values for lever arms a and b can be measured from anatomical specimens o on soft tissue imaging or fluoroscopy (Kellis & Baltzopoulos. 1999; Wretenberg et aI., 1996), and the magnitude of W ca be determined from the body weigt1t of the individual. The magnitude of P can then bE' found from the moment equilib :::M = 0 num equation: W;.: a p~_.- b W>:a=Pxb p =W !Again the welghl of the lower leg IS disregarded because it is lESS ,han Oile \Cntil of body vveight xa b I Calculation Box Figure 7-2-1. e ------..- - - - - DYNAMICS OF THE TIBIOFEMORAL JOINT Although estimations of the magnitude of the forces and moments imposed on a joint in static situations are llseful. most of Ollr activities are of a dynamic nature. Analysis of the forces and moments acting on a joint during motion requires the use of a different technique for solving dynam,ic problems. As in static analysis, the main forces considered in dynamic analysis arc those produced by body weight, muscles, other soft tissues, and externally applied loads. Friction forces arc negligible in a normal joint and thus not considered here. In dynamic analysis. two factors in addition to those in static analysis must be taken into account: the acc tion of the body part under consideration an mass moment of inertia of the body part. (The moment of inertia is the unit used to expres amount of torque needed to accelerate a bod depends on the shape of the body.) (For mo depth studies of dynamics, sec Ozkaya & No 1999.) The steps for calculating the minimum m tudes of the forces acting on a joint at a particu stant in time during a dynamic aClivity arc as fo I. The anatomical structures arc identified: nitions of structures. anatomical landma point of contact of articular surface. and 2. 3. 4. 5. 6. :f 4\rl11s involved in the production of forces for the biol11cchanical analyses. The angular acceleration or the moving body part is determined. The mass monlent of inenia of the moving bod~" part is determined. The torque (moment) acting abollt the joint is calculated. The magnitude of the main muscle force accelerating the body pan is calculated. The magnitude of the joint reaction force at a particular instant in time is calculated by static analysis. In the first step, the structures of the body involved in producing forces on the joint are identified. These are the moving body part and the main muscles in that body part that arc involved in the production of the motion. Great care must be taken in applying this first step. For example, the lever arms for all rnajor knee muscles change according to the degree of knee Ilexion and gender (\Vretenberg et aI., 1996). In joints of the extremities. acceleration of the body part involves a change in joint angle. To determine this angular acceleration of the moving body part, the entire movement of the body part is recorded photographically. Recording can be done with a stroboscopic light and movie camera. with video photogrummetry, with Selspot systems, with stereophotogrammctry, or with other methods (Gardner et aI., 1994; Ramsey & Wretenberg, 1999; \Vintel~ 1990). The maximal angular acceleration for a particular motion is calculated. Next, the mass moment of inertia for the moving body part is determined. Anthropometric data on the body parl can be used for this determination. As calculating these data is a complicated procedure, tables are commonly used (Drillis et aI., 1964). The torque about the joint can now be calculated using Newton's second law of motion, which stales that when motion is angulat~ the torque is a product of the mass moment of inenia of the body part and the angular acceleration of that part: T = Ie< , where T is the torque expressed in newton meters (Nm) I is the mass Illoment of inertia expressed in newton meters x seconds squared ( m sec~) a is the angular acceleration expressed in radians per second squared (rlsec'). , 'i; : .. , .'" Th..: torque is not only a product or the mass ment of inertia and the angular acceleration o body part but also a product or the main Ill force accelerating the body part and the perpe ular distance of the force from the center of m of the joint (level' arm). Thus. T = Fe! where F d is the force expressed in newtons (N) is the perpendicular distance expressed in m leI'S (01). Because T is known and d can be measure the body part from the line of application o force to the center of motion of the joint, the e tion can be solved fOl" F. vVhcn F has been c lated, the remaining problem can be solved l static problem lIsing the simplified free-body nique to determine the l11inirnum magnitude o joint reaction force acting on the joint at a ce instant in time. A classic example will illustrate the usc o narnic analysis in calculating the joint reaction on the libiofemoral joint at a particular in during a dynamic activity (c.g., kicking a foo (Frankel & Burstein, 1970). A stroboscopic fi the knee and lower leg was taken, and the an acceleration was found £0 be maximal at the in the foot struck the ball: the lower leg was almos tical at this instant. From the film. the maxima gular acceleration was computed to be 453 ! From anthropometric data tables (Drillis e 1964), the mass momenl of inertia for the lo\v was determined to be 0.35 Nm sec~. The to about the tibiofemoral joint was calculated ac ing to the equation; torque equals mass mome inertia times angular acceleration (T = (0), 0.35 Nm sec' x 453 rlsec' = 158.5 Nm After the torque had been determined to be Nm and the perpendicular distance fran) the ject's patellar tendon to the instant center fo tibiofemoraljoint had been found to be 0.05 m muscle force aCling on the joint through the pa tendon was calculated using the equation to equals force times distance (1' = Fd), 158.5 Nm = F x 0.05 m F .= 158.5 Nm -=--=-0.05 m F = 3170 N Thus, 3,170 N was the maximal force exerted by the quadriceps muscle during the kicking motion. Static analysis can now bt: performed to determine the minimum magnitude of the joint reaction force 011 the tibiofcrnoraI joint. The main forces on this joint arc id':lltificd as the patellar tendon force (P). Ihe gravilaliomtl force of Ihe lower leg and the joint reaction force (J). P and T arc known vectors. J has an unknown magnitude. sense, and line of application. The free-body technique for three coplanar forces is L1sed to solve for J, which is found 10 be only slighrly lower Ihan P. As is evident from the calculations, the two main factors that influence the magnitude of the forces on a joint in d)'namic situations are the acceleration of the body part and its mass moment of inertia. An increase in angular acceleration of the body part will produce a proportional increase in the torque about the joint. Although in the body the mass moment of inertia is anatomically set, it can be manipulated extenl~l)Jy. For example. it is incrc.:ascd whcn a weight boot is applied (0 the foot during rehabilitative exercises of the extensor muscles 01" the knee. Normally, a joint reaction force of approximately 50°;(; of body weight resulls when the knee is slowly (with no acceleration forces) extended from 90° of flexion to full extension. In a person weighing 70 kg, this force is approximately 350 N. If a IO-kg weight boot is placed on the foot, it will exert a gravitational force or 100 N. This will increase Ihe joint reaclion force by 1,000 N, making this force aimosl rour times greater than it would be without the bool. Dynamic analysis has been L1sed to investigate the peak magnitudes of the joint reaction forces, muscle forces, and ligament forces on the tibio~ femoral joint during walking. Morrison (1970) calculated the magnitude of the joint reaction force transmiucd through the tibial plateau in male and female subjects during level walking. He simultane~ ously recorded muscle activity elcctromyographi~ cally to determine which muscles produced the peak magnitudes of this force on the tibial plateau during various stages of the gait cycle (Fig. 7-15). Just after heel strike, the joint reaction force ranged from two to three times body weight and \vas associated with contraction of the Iwmstring muscles, which have a decelerating and stabilizing effect on the knee. During knee llexion in the beginning of the stance phase. the joint reaction force waS approximately lWO rimes body weight and was associated with the contraction of the quadriceps muscle, which acts 10 prevent buckling of the knee. The peak joint reaction force occUlTed during the en, .. late s(;Jncc phase just before loe-ofr. This ranged from two to four times body weight. v among the subjects tesled, and was <:Issociatcd contraction of the gastrocnemius muscle. In th swing phase, contraction of the hamstring m resulted in a joint reaction force approxim equal to bod)' weight. No significant dilTcrcnc found between the joint reaction force magn for men ancl women when the values were no izcd by dividing them by body weight. Andriacchi & Sirickiand (1985) sludied Ihe n moment patterns around the knee joint during walking for 29 healthy volunteers (15 women a men wilh an average age of 39 years). Figure depicts the flexion-extension, abduction~addu and internal-external moments during the s and swing phase of level walking. The momen normalized (0 the individual's body weigh height and are presented as a percentage. The f 4 en I 3 .cen 'Q; ~ '" 'C 0 e ~ ~ 0 u. 2 fl ./\ ,/\ \/ \! \ .' /-..., l V,..' ' ' 1\ ' \ ' ., . .,-. o1~O:::O-c.._-'-'..J·~---~"'6~O~--'-"'------'~--:1 Percentage of cycle Joint reaclion ..... Hamslrings HS =. Heel strike . - -- Quadriceps le TO =.: Toe-ofl -- .. Gastrocnemiu Joint reaction forces expressed as body weight trans through the tibial plateau during walking. one gait (12 subjects). The muscle forces producing the peak tudes of this force are also designated .. Adapted from Morrison, lB. (1970). The mechanics of the knee joinr in rion lO normal w<lfking. J Biomech, 3, 51. • '!'· extension moments during the stance phnse are approximately 20 to 30 tilTICS larger than the moment produced in the frontal (abduction-adduction) and transverse (intcrnal-extcl-nal) planes. An increase in knee joint flexion-extension mo- A_ Flexion-Exlension Moment Pallerns (Exlernal) ment amplitude has been reported at increased walking speeds (Andriacchi & Strickland. 1985; Holden et aL. 1997). An increase in the production of adduction knee joint moment during stair climbing compared with level walking was reported b.\' Yu et al. (1997). During the gait cycle, the joint reaction force shifts from the medial to the lateral tibial plateau. In the stance phase, when the force I'caches its peak value, it is sustained mainly by the medial plateau i. (adduction moment); in the swing phase, when the force is 111inimal. it is sustained primarily by the lat· eral plateau, The contact area of the medial tibial plateau is approximately 50% larger than that of the lateral tibial plmeau (Keuclkamp & Jacobs, 1972). Also, the cartilage on this plateau is approximately three times thicker than that on the lateral plateau. The larger surface area and the greater thickness of the medial plateau allow it to more easily sustain the higher Forces imposed on it. In a normal knee, joint reaction forces are sustained by the menisci as well as by the articular cartilage. The function of the menisci was investigated by Seed hom and coworkers (1974), who examined the distdbution of stresses in knees of human autopsy subjects with and without menisci. Their results suggest that in load-bearing situations, the magnitude of the stresses on the tibiofemoral joint when the menisci have been removed may be as much as three times greater than when these structures are intacL Fukuda et al. (2000) studied in vitro the load-compressive transmission of the knee joint and the role of menbci and articular cartilage. The load simulated was stalic and d).'namic impact loading. The testing was done in neutral. varus, and valgus alignment of the knee joints in 40 fresh-frozen pig knee speclmens. The compressive stress on the medial subchondral bone was lip to five times higher with the menisci removed. This study points to the importance of the menisci as a structure to absorb load and protect the canilage and subchondral bone under dynamic conditions. In a normal human knee, stresses are distributed over a wide area of lhe tibial plateau. If the menisci are removed. lhe stresses are no longer disll"ibutcd Over such a wide area but instead arc limited to a contact area in the Center of lhe plaleau (Fig. 7-17). Thus, removal of' the menisci not only increases the It Pattern 1 ¥ >~ ;!: "' rL .s0 , 3.0 'xID u:: 1.0 0 I I I I c a 1.0 '00 c 2.0 X '"c w'" ID I I I c a 2.0 '" 3.0 - B. x 0 ;!: 1.0 c .~ a .2 ID ID C "0 "0 , '" Abduction-Adduction Moment Patterns (External) 1.0 I "'~ Stance --l-Swing-- 13 0 « 2.0 3.0 4.0 5.0 C. Internal-External Rotational Momenl Patterns (External) "" I X ;!: OJ 1.0 ;;; c ;;; 0.5 E l, c ;;; ID ;;; x , '0 '"c '" c w 0.5 I I I 1.0 ..- Stance - .. I'-Swing-" Flexion-extension (A), abduction-adduction (8), and internal-external rotation (C) moments produced durin one gait cycle in normal subjects. The moments are no malized to each individual body weight X height and pressed as a percentage. Reprinted with permission from Andriacch( IP. &. Strickland. A.B. (1985). Gaie analysis as a to assess joint kinetics. In N. Berme, A.E. Etlgin. O.A. Correis et al. (Eds.). Biomech(lnics of Normal and Pathological Hum Articulating Joints. (NATO ASI series. Va/93. 'pp. 83-I02)' Orodrechr. Netherlands: Marr;nus Nijhoff. • • Force Force II I l \\ ... III III \\\ tIl \ I ~ tt tt ... Menisci removed Normal Stress distribution in a normal knee and in a knee with the menisci removed, Removal of the menisci increases the magnitude of stresses on the cartilage of the tibial plateau and changes the size and location of the tibiofemoral con· tact area. With the menisci intact, the contact area encom- passes nearly the entire surface of the tibial plateau. With the menisci removed, the contact area is limited to the center of the tibial plateau. magniwde of the stresses on the cartilage and subchondral bone at the ceiller of the tibial pl~llcau but also diminishes the size and changes the location or the contact area. Over the long term, th~ high stresses placed on this smaller contact area may be harmful to the exposed cartilage, which is usually soft and fibrillated in that area. The menisci are 7.1 thought to carry lip to 70<}(; of the load across knee. ,Vlovcmcnt during knee Ilcxion of the meni would therefore protect the articulating surfa while avoiding injury to it. Vedi et al. (1999) studied menisci movement in young football players with normal knees w MRI. The knee flexion movement was scanned fr full knee extension lo 90° of knee flexion. The im ing technique allowed for both standing (weig bearing) and sitting (non-weight-bearing) and w performed simultancollsl~t in the sagittal and tra verse plane. Figure 7-18 shows the movements the transverse plane of the medial and lateral me sci expressed in millimeters (mean) from full ext sion to 90° Ocxion of knee joint motion. Movem was significantly greater in weight-bearing than non-\Veigllt~bearing for both lateral and med menisci. The contributions of the menisci art: lhe fore not only to protect the articular cartilage a subchondral bone but also to contribute aClively knee joint slabilit~:. STABILITY OF THE KNEE JOINT The key to a healthy knee joint is joint slability. T osseous configuration, the menisci, the ligamen the cnpsulc, and the muscles surrounding the k joint produce joint slability (Fig. 7-1, A & B). If a Ant 5.4 Ani 9.5 0 6.3 , Flexion 3.6 I 3.3r i \ N , \ \ \ 90 "- f I ,, • -J_3.9 Medial menisci "- Post \ ., -L_- , "- 5.6 , 3.8 Lateral menisci , -4- Medial menisci Post . -~4.0 -- I ,/ Lateral menisci B A Simplified diagrams showing the mean movements of the medial and lateral menisci from full knee extension to 90° knee flexion during two conditions. A. Erect and weight~ bearing. B. Sitting, relaxed. and bearing no weight. Adapted with permission (rom Vedl, V. WiJjiams. A., Tennant. S.J.. el aJ. (1999). Menisca! movemenr. An In-vivo srudy u~in9 dynamic M Bone Join! Swg, 818(1), 37-41. .ACL Injury i. A J;_ 30·year-old male suffered an external rotation trauma in his right knee while downhill skiing. Following the ~~;,: trauma, he experienced sharp pain, progressive joint effu,~,iqn,_and subjective instability. ,During careful examination Ii' ,by a specialist, an anterior positive drawer test was dlag- <";'n,osed, and the Lachman test and pivot shift test were ': found positive. An MRI confirmed the ACL rupture (Case j' Si~dy Fig 7-1-1) -,. The rupture of the primary stabilizer of the knee joint (ACl) leads to a progressive structural alteration of the knee. A primary objective of the treatment is the prevention of fe-injury of the knee in the hope of preventing additional ligamentous Case Study Figure 7-1-1. of these structures arc malfunctioning or disturbed, knee joint instability will occur. The ligaments arc t.he primm)' stabilizer for anterior ancl posterior translation, varus and valgus angulation, and internal and external rotation of the knee joint (Case Study 7-1), F,; et aL (1994) sUlllmarized the functions of the knee ligaments, The ACL is the predolllinant restraint to anterior tibial displacement. The ligament accepts 75% or the anterior force at full extension and an additional 10% (up to 90") of knee nexion, The posterior crudate ligament 1S the primary restraint to posterior tibial translation: il slIstains 85 to 100% of the posterior force at both 30 and 90" of knee flexion. The latera) collateral ligament is the primary restraint to varus angulation and it resists injuries. meniscal injuries. and possible cartilage degenera· lion. In this case. the patient firs! completed a course of conservative treatment with physical therapy. After 6 months. the subjective inS(clbilily was present during SPOrts and daily activities such as gail and stair climbing. To com· pensate for the ACL deficiency, the patient altered his gait patterns, presenting quadriceps avoidance gait to prevent the anterior translation of tl1e tibia when the quadriceps contracts at the midstance phase of the gait (Andriacchi & Birac, 1993: Berchuck et aI., 1990). The patient went for surgical treatment. The MRI below (Case Study Fig. 7-1-2) shows the ACL status after patella bone-tendon-bone autograft was performed 10 months post-trauma. Case Study Figure 7~1·2. approximately 55% of the applied load at full ext sion. The role of the lateral collateral ligament creases with joint flexion as the posterior structu become lax. The 1l1edial collateral ligament (supe cial portion) is the primary restraint to valgus ( duction) angulation and resists 500/0 or the exter valgus load. The capsule. the anterior and poster cruciatc ligaments, share the remaining valgus lo Internal rotational laxity seen in the 20 to 40° ran of knee nexion is restrained bv the medial collate ligament and the ACL Finally, external rotation l ity seen in the 30 to 40° range of knee flexion is strained by the posterior cruciate ligaJllcnt at 90° knee Ilexion, In vivo measurement or the normal ACL has b performed by Beynnon et aL (1992), They place strain transducer arthroscopic'll)y in the ACL. The results showed that strain in the ACL was related to knee Oexion (with the most strain occurring ncar full extension) and increased with quadriceps contraction. Less strain occurred with co~conlraclions of both the quadriceps and the hamstring muscle groups and at greater degrees of knee flexion. This indicates that muscle contraction and co-contraction contribute to the stabilitv of the knee joint bv incrcasing thc stiffncss of the joint. Kwak et al. (2000) studied in vitro the effect of hamstrings and iliotibial band forces on the kinematics of the knee. At various knee Oexion angles, human knee specimens were tested with different muscle-loading patterns. The quadriceps muscle force was always present, and the test was performed with and withollt han)string muscle force and with and without iliotibial band force. \AJith loading of simultaneous quach·i. eeps and hamstring muscle force. the tibia translatcd posteriorly and rotatcd externally. The effecI """ similar for thc iliotibial band simulated forces but thc cffcct was smallcr. Many in vitro studies suggest that the hamstrings are important anterior and rotational stab.ilizers of the tibia. In vivo studies have shown that co-contractions of the quadriceps and hamstring muscles arc highly present in normal knee joints and dail.'", activities (Baratta et a1., 1988; Solomonow & D'Ambrosia, 1994). Thc co-contraction mcchanisms also increase the knee joint stability in vivo (Aagaard ct al .. 2000; Markholf et al .. 1978; Solomonow & D'Ambrosia, 1994). Howevel~ the complex mechanism in vivo of muscle activity as a knee stabilizer, the extent of protection, and the biomechanical and clinical imparlance needs futther research (Grabiner & Wcikcl', 1993). tendon and it contributes the least to the leng the quadriceps muscle force lever arm (app mately 10% of the total length). As the knee i tended. the patella rises from the intercon groove, producing significant anterior displace of thc tcndon. Thc length of thc quadriccps lever arm rapidly increases with extension up to at which point the patella lengthens lhe lever by approximately 300/0. With knee extension bc"ond 45°, thc Icngth o lever arm is diminished slightly. \,Vith this dec in its lever arlll, the quadric.:cps muscle force increase for the torquc about the knee to rcrnai same. In an in vitro study of normal knees. Licb Perry (1968) showcd that thc quadriccps m force required to extend the knee (he last 15 creased by approximately 60% (Fig. 7-19). If thc patella is removed from a knec, the I)a tendon lies closer to the center of motion o tibiofemoral joint (Fig. 7·20). Acting with a sh lever arm, the quadriceps llluscle must pro even more force than is normnlly required for a tain torque about the knee to be maintained du 75 FUNCTION OF THE PATELLA The patella serves two important biomechanical functions in the knee. First, it aids knee extension by producing anterior displacement of the quach-ieel's tcndon throughollt the entire range of motion. thereby lengthening the lever arm of the quadriceps muselc force. Second, it allows a wider distribution of compressive stress on the femlll- by incrensing the area of contact between the patellar tendon and the fcmur. The contribution of thc patella to the Icngth of the quadl"iceps muscle force lever arm varies from fullllcxion 10 full extcnsion of thc knee (Lindahl & Movin, 1967; Smidt, 1973). At full Ilcxion, whcn the patella is in the intercondylar groove, it pro~ duces littlc anterior displacemcnt of the quadriccps ""',' o 90-60 60-30 30-15 Knee Motion (degrees) Flexion - - - - EXlension ~~------ Quadriceps muscle force required during knee motio from 90" of flexion to full extension. Adapted from L f1. & Perry, 1. (J 968). Quadriceps funccion. An dflacomica mechanical study using dmplJtMed limbs. J Bone Joint Su 50A, 1535. , ,. ~~===== ------------- "- \\ \ --------- ~- Instant center Normal -L=-After Patellectomy Quadriceps muscle lever arm (represented by the broken line) in a normal knee and in a knee in which the patella has been removed. The lever arm is the perpendicular distance between the force exerted by the quadriceps muscle through the patellar tendon and the instant center of the tibiofemoral joint for the last two degrees of extension. The patellar tendon lies closer to the instant center in the knee without a patella. Adapted from Kaufer. H. (1911). Mechanical function of the parella. J Bone Join! Surg. 53A. 1551. knee is almost directly above the center of rotatio of tho patellofemoral joint. As knee Oexion i creases, the center of gravity shifts further awa from the center of rotation. thcrcb.y greatly increa ing the Ilexion moments to be counterbalanced b the quadriceps musclc force. As the quadricep muscle force rises. so docs the patellofemoral joi reaction force (Hungerford & Barry, 1979; Reilly J\>larlons, 1972). Knee flexion also influences the patellofemor joint reaction force by affecting the angle betwee the patellar tendon force and the quadriceps te don force. The angle of these two force comp nents becolTlCs more acute with knee flexion, i creasing the magnitude of the patellofemoral joi reaction force (Calculation Box 7-3). Reilly an Martens (1972) dctcrmincd the magnitude of th patellofcmoral joint reaction force during sever dynarnic activities involving val·ying amounts knee flexion. During IC\'e1 walking, which requir relatively little knee flexion, the reaction force w low. The pcak valuc. ill the middle of the stan phasc when flexion was greatest, was one-ha body weight. The joint reaction forcc was much greater durin activities that require greater flexion. During kn bends to 90'. this force reached 2.5 to 3 times bod wcight with the knee Oexed 90' (Fig. 7-21). Throug 3000 the last 45° of eXlen~ion. Full active extension of such a knee may require as much as 30% more quach'iceps force than is normally required (Kaufer. 1971). This increase in force may be beyond the capacity of the quadriceps muscle in some patients, particularly those who have intra~articular disease or are advanced in age. 1 ...... Palellofemoral joinl reaction force --- Quadriceps muscle force :s 2000J " !! 0 "1000 STATICS AND DYNAMICS OF THE PATElLOFEMORAl JOINT During dynamic activities, the magnitude of the muscle forces acting on a joint directly affects the magnitude of the joint reaction force. In general, the greater the muscle forces, the greater the joint reaclion force. In the patellofemoral joint. the quadriceps muscle force increases with knee Oexion. During re~ (axed upright standing, minimal quadriceps muscle forces are required to counterbalance the small nexion l1lornents about the patcllofemoral joint because the center or gravity of the body above the o 20 40 60 80 Knee Flexion (degrees) Patellofemoral joint reaction force and quadriceps muscl force during knee bend to 90° (three subjects). Adapted from Reilly, D. T. g. Martens, M. (1972). Experim.ental analysis o the quadriceps muscle force and patello/emaral joint reaction force for various activities. Acta Orthop Scand. 43. 126 Joint Reaction Forces at the Knee in Flexion Knee flexion influences the patellofemoral joint reaction force by changing the angle between the patellar tendon and the quadriceps tendon (Calculation Box Fig. 7-3-1, A & B). The angle between the patellar tendon (P) and the quadriceps tendon (Q) is 35° with the knee flexed 5° (left top) and 80° with the knee flexed 90° (left bottom). Values for the tendon angles are from Matthews and Associates (1977), who determined the angle roentgenographically after placing two metal wires along each of these tendons. The pateJlofemoral joint reaelion force with the knee in 5 and 90° of flexion is oblained by constructing a parallelogram of forces for each situation and using trigonometric calculations. The pateJlofemoral joint reaction force (J) is the resultant of the hvo equal force components through the patellar tendon (P) and the quadriceps tendon (Q). As the angle between these force components becomes more acute with greater knee flexion, the resultant joint reaction force (J) becomes larger. Adapted lrom Wikrorin Cv.H. & Nordin, M. (1986). Introduction to Problem Solving in Biomechanics (pp. 87-129). Philadelphia: Lea & Febiger. o ,,/ p ,,'" // P------------/ 1000 N J ~ 601.409 J2 '" Q2 .... p2 _~ 2PQ. cos 35~ Tibio femoral flexion 5" J = )361695 ~ 601.41 J o I 1000 N !---J--""-'r: 80" I I I I I I p I I I 10;0 N ---------j·::-;285_61 J2 I Tibio femoral flexion 90" Calculation Box Figure 7·3·1A. -------_._---- ~~ Q2 -;.. p2 J ~ 11652703 J ~ .~ 2PQ • cos 80~ 1285_57 Calculation Box Figure 7·3-1B. --------_..._------_._--- --_._.--- -._------ '" Extensor Mechanism Injury 30-year-old basketball player had a forceful knee flexion while coming down from a jump. A strong eccen- of the quadriceps produced abnormally high " ,0n",lp loads in the patella, leading to a fracture at the infeIn this case, the patella fracture occurred because forces of the quadriceps overcame the osseous of the patella. The weakest link was the patella. picture shows a fracture at the patella accompaf)ied by a significant fracture separation that resulted from the quadriceps traction force. Because of the fracture. the extensor mechanism is un=;'. able to function and extend the knee. It will directly affect the stability of the pate!lofemoral joint and the distribu- Co; tion of the compressive stresses on the femur. At the same time, the impaired function of the quadriceps decreases the dynamic stability at the knee joint (patellofemoral and tibioiemoral joints) that is necessary :J. for daily activities such as gait and stair climbing. contact surface area increases in size somew (Goodfellow et aI., 1976). To some extent, this crease in the contact surface with knee nexion co pensates for the larger patellofemoral joint react force, 11' a tight iliotibial band is present, patcllofemoral joint force may shih laterally, ca ing abnormal patellar kinen1atics and load-bear (Kwak et aI., 2000), The quadriceps muscle force and the tor around the patellofemoral joint can be extrem high under certain circumstances, particula when the knee is flexed-for instance. whe basketball plaver suffers a patella fracture a result of indirect forces played by an eccen contraction of the quadriceps (Case Study 7 Another extreme situation was observed durin stud~; of the external torque on the knee p duced by weight lifting: one subject ruptured patellar tendon when he lifted a barbell weigh 175 kg (Zernicke et aI., 1977). At the instant tendon rupture, the knee was flexed 90°; torque on the knee joint was 550 Nm and quadriceps muscle force was appro:dt11a 10.330 N. Because of the high magnitude of quadric muscle force and joint reaction force during tivities requiring a large amount of knee flex patients with patcllofemoral joint deran ments experience increased pain when perfo ing these activities. An effective mechanism reducing these forces is to limit the amoun knee flexion. Summarv 1 The knee is a two-joint structure that is co posed of the tibiofel11oral joint and the pate femoral joint. out knee bend, the patellofemoral joint reaction force remained higher than the quadriceps Illuscle force. During stair climbing and descent, at the point when knee flexion reached a maximum of approximately 600, the peak value equaled 3.3 times body weight. When the knee is extended, the lower part of the patella rests against the femur. As the knee is flexed to 90°, the contact surrace between the patella and femur shifts cranially in vivo and under weightbearing conditions (Komistek et aI., 2000). The ?:>: In the tibiofemoral joint, surface motion curs in three planes simultaneously but is grea by far in the sagillal plane. In the patellofem joint, surface Illotion occurs simultaneously in planes, the frontal and the ll-ansverse, and is gre in the frontal plane. 3 The surface joint motion can be described w the use of an inslant center technique. vVhen formed on a non11al knee, the technique reveals following: the instant center for sllccessive inter of motion of the tibiofcmoral joint in the sagi plane follows a semicircular pathway, and the di tion 01" displacement of the tibiofemoral con points is tangential to the surface of the tibia, indi eating gliding throughollt the range o[ motion. w The scrc\v-homc mechanism of the tibiofemoral joint adds stability to the joint in full extension. Additional passive stability' to the knee is given by the ligamentous structure and menisci and the dynamic stability by' the muscles around the knee. S The tibiofemoral and patellofemoral joints are subjected to great forces. Muscle forces have the greatest influence on the magnitude of the joint rc action force, which can reach several times body weight in both joints. Tn the patcllofcmoral joint, knee flexion also affects the joint reaction force, with greater knee flexion resulting in a higher joint reaction force. w Although the tibial plateaus arc the main loadbearing structures in the knee, the cartilage. menisci, and ligaments also bear loads. The menisci aid in distribt~ting the stresses and reducing the load imposed on the tibial plateaus. The patella aids knee extension by lengthening the lever arm of the quadriceps muscle force throughout the entire range of motion and allows a wider distribution of compressive stress on the femur. REFERENCES Aagaard. P., Simons~n, E.B., And~rsen, J.L., et a1. (2000). Antagonist mtlscl~ co-'lctiv<ltion during isokinetic extension. Scali J J1ed Sci Sports, 10(2),58-67. Andriacchi,1. P. & Birac, D. (1993). Functionalligamellt tcstin" in the anterior cruciat~ li!.!alll~nt deficient knee. C!ill Or~JlOp Rei Res, 288 (Alarch), 4~0-47. Andriacchi, T.P., Kramer. G. ;'vl., & Landon, G. C. (1979). Three-dimensional coordinate data processing in human motion analysis. J Biolllcch Eng, /01, 279-283. Andriacchi, T.P. & Strickland, A.8. (1985). Gait analysis as a tool to assess joint kinetics. In N. Berme, A.E. Engin. D.A. COl'reis, et al. (Eds.). Biollli'clulllics o{ Norll/a! and Pathological J1l111WII Articulating .loints (NATO ASI series, Vol 93, pp. 83-102). Drodrecht. Netherlands: :',·lartinus Nijhoff. Baratta, R., Sololllonow, ivi., ZhOll, 8.1-1., Letson, D., Chuinard, R., D'Ambrosia, R. (1988). Muscular co-activation: The role of the antagonist musculature in maintaining knee stability. Alii J Sports Alcd, 16, 113-122. Berchuck i\'1., Andriacchi, TP., Bach, B.R.. Reider, B. (1990) Gait adapt ions by patients who h,\\'e a deficient anterior cruciate ligament. J BOlli' Joillt Surg. i2A, 871-877. Beynnon, 8., Howe, J.G., & Pope, rVI.H. (1992). The measurements of anterior cruciate ligament strain in vivo. lilt 01'lhop, 16. 1-12. Dra!!,\nich. L.D .. Andriacchi, T.P.. & Andersson, G.B.J. (19 I~nlcraction betwccn intrinsic knec mechanism and knee extensor mech'lllislll. ) OnhoJ} Res. 5, 539-547. Drillis, R., Conlini. R., & Bluestein, \1. (1964). Bod~' segl par'lmc[(·rs. A stln'ey of 1l1l..'<lStll"elllent tec!miqut.'s. Lil/lbs, 8. 44. Frankel. V.H. & Burstl'in, A.I-I. (19701. OnJlOptldic Bi cfUlllics. Philadelphia: 1.t.'a &- Fcbigcr. Fr'Hlkel, V.H., Burstein, A.H., l\c Brooks, D.B. (1971). Bi chanics of internal dl'rangeTw:nt of Lht.' knee. Patho chanics as dl'tcrmincd b.v an~d~'sis of the instant ccnte motion. J BOlle Joillt SlIrg. 53..\, 945. Fu, ElL Harner, CD., Johnsoll, D,I.., ct <11. (1994). Bi chanics of the kncl' ligamenLs: Basic concl'pts and cli application. II/Sfr COllrsi' Lt.'ell/re, -t3, 137-148. Fukuda, 'I{., Takai. 5., Yoshino, N., et al. (2000). Impact Lransmission of the kncc joint~infltlence of leg align and the role 01" IllL'niscus and articular carlilage. Cli Bioll/L'ch, 15, SI6-S2l. Gardner. T.R .. Ateshian, G.A .. Grclsarner, R.P., et al. (199 6 DOF knec testing de\'ice to d~termine patellar trac and patellofellloral joint contact area \'ia stereo lOQramrnl'trv. :Idl' Bio('II'.,; :IS.\1E BED. 28, 279-280 Gr)(ldl"dlow, .I.', llungcrl"o'rd, D.S., & Zindd, \1. (19 Patl'llofenloral joillt llH.'chanics and paLhology. I. F tional ,uwtorn.v of the patellofL'lllO!';t1 joint. J Bonc Surg. 5813, 287. Grabine!', .\-I.D. & Weiker, G.G. (1993). :\ntl.'rior cruciate ment injury and hamstring CO-<lcli\'ation. elillical IIlt'ch, 8. 216-119. Grood, E.S. l\: Suntav. \V.J. (1983). A coordinate svstell clinical descriptio;l of three dimensional motion~ App tion to the knee. J Biollll!ch Ellg. 105, 136-144. He!llIL', 1-1.1. (1990). Biomechanics of lhc patellol"emor.d and its clinical relevance. Cfill Orthop, 258, 73-85. Hell"eL, A..J. (1974). AnatolllY and mechanics of mO\'ellle the knec joint. In A. He]fct (Ed.), Disorders o(the Knc(' 1-171. Philadelphia: J.B. Lippincott. Holden, J.P., Chou. G.. &- Stanhope, S.J. (1997). Change knee joinL function o\"er a \~'ide range of walking spc Cfillica! Bio/llL'ch,12(6L 375-382. Hungerford, D.S. &- BarT\', ;\:1. (1979). Biomechanics of patdlofelTloral joint. Cli,l OnhoJ!, Ic+c+, 9-1 S. Kapandji, l.A. (1970). Thc knl'c. In LA. Kapandjii (Ed.l. Physiology oFthe Joillts (Vol 2, pp. 72-135). Paris: Edi \1<1 loi ne. Kaul"er, H. (1971). \Iechanical function of tllt' patella . .l Joill! Surg. 53A. ISS I. Kellis, E. & BHltzopCllJlUS, V. (1999). In \'i\'o delerminatio the patella and Iwmstrings rnonh.'nt arms in adult m using videofilloroscopy during submHxinwl knee l':\tl'n and flexion. Cfinical Biolliech. I -t, 118-124. Kcttelkamp, D.B" Johnson, R.1., Smidt, G.L .. Chao, Walker, rv1. (1970). An c1ectrogoniomctric stud.\" of motion in normal gait. J BOlli': Joilll Sur.>;, 52:\, liS. Kettelkamp, D.B. &- Jacobs, A.\V. (1972). Tibiofclllora! con area-dL'termination and implications. J BOIIL' Joil/t S 5-tA. 349. Komistek, R.D., Dennis, D.A., ;\labe, J.A., d HI. UOOO). A vh'o determination of pHtellol"ernora! c(ll1tact posit Cfillical Biol/lcch, 15,29-36. Kroelller. K.I-I .. M;llTas. \V.S ...\1cGlothin. J.D .. 1..'1 al. (1990). On the IlH.'ilsun:nh.'nts of human stn:ngth. 11111 Imll/strial rr"OIUl111ic!I·. 6. 199-210. Kwak~ 5.0.. ,\hmad, C.S .. Gardner. T.R .. 1..'1 al. (2000). Ham· Sirings and iliotihial forces affect knee ligamellts and contact palh:rn.l O,.,ltop ReS. IS, 101-IOS. LaTllorc;IU.\:. L.. (1971). Kinematic IllC:lsurt:TllcnIS in Ihl.· stlldy of hurn:tll w;:lIking. Biomcch;lnics Lah. Uni\'t.'rsiIY of (;tliforni:l, S'lll rr~ncisco. B//II Prostlteth' Res, Sp1971, 10-15. L:ltlbenth;d. !\..N .. Smidt, G.L., & !\.ctLdkamp, D.B. (1971)."\ quantitative analysis of knee motion during activities of dail\' living. I'II\'S Til!!,., 52. 3-L LC\'I.·ns: '\,S"Yllll1l~lI1. V.T., & Blosser. J.'\. (1948). Trans\'crsc rot;ltion of the seglll"fHs of the lower cXlr('mity in locomotion, 1 BOlli! 10iut Sur..!.:. 30A. S59. Licb. FJ. ~ I)erry. J. (I968). Quadriceps function. An ;UHllOI1lical ;lnd mcch:lllical sludy using ;lIl1putaled limbs. 1 Bmw loil/t Surg, 50~\, 1535. LiIH.bhl, 0 . .& Mm·in,.-\. (1967). TIJ(,' 1lH.'chanil':-; of l'xt...'nsion of the kncc joint. .·\eta 01'1/1011 Scaml, 38, 226. ~;:;II'kholr. K., Gr:lff-R;l<lford, f\., &. Amstul'I.. H. (19iS). In \'i\'O knce slllbility. 1 BOlli' 10ill/ SlIl'g. 60.'1. 664-674. ~laISllmot(), 1-1.. Si..'i..'dholll, B.B.. Suda. Y., C"t <:II. (2000). Axis of tibial rot.Hioll and its change with flcxion angle, C/ill Or· . '),Op, 371. 178-182. ',\l<1ltl1(:\\,s, L.S .. Sonstc:gard. D.A .. &. I-lenke, .I.A. (1977). Lond h"arillg characterislics of lhe palellofemor;d joint. ,·\era Onhop S,·a//(I. 48.311. M'HTis('n.J.B. (1970). The mechanics of the kIWi..' joint in relation [0 nonn:'l! walking. J Bicol/cch. 3. 51. !'\i1.P., Droughl, :\.B .. !\.ory. R.C. (1964). \Valking 1':11ll'::l'nS of l1orl11:d men. 1 130111..' Joillt S!l!·.~, -I6A, 335. N. & Nordin. rio'\. (1999). FlIlId(/I!U'IIUJ!S oj' 13iOlllt'Eqllilihl'illlll, JIOlio/l, (/lId lJeji>rlllatioll (2nd cd.). Ncw York: Springcr-V"r1ag. ~crr.v, 1.. Norwood, L., & House. K. (1977). Klll'C po.o;tUl'\.' ,llld biceps and semimembranosus muscle action in fUllning <'Ind cutting (an EMG study). hailS Ortlwl) UL'S S(lC 2, 258. Ramsi..'y, O.K. & Wn.'tl'llhcrg, P.E (1999). Biolllcchallit,:-; of tht:' kn~c: Melhodological consid('r;ltions in Ihl' in \'i\'o kinl'matic an:dysis (~f II\(' tibiofl'moral and patellofclIlor;d joint. RC\'i(:w pnp('r. Clillical Biomecll. 1-1, 595-611. Reilly, 0.1'. &. "'Iancus. ,\'1. (19i2). E.\:pel'imclll;.'I1 ;Inalysis thl' qlwdriceps musclc force and patcllof"moral joint rea tion force for variolls 'lctivitic-s. :\ew Orrhop Scalld. 4 126. Reulc;1ux. r. (1876). In The KillCl1lalio uf Jlachillc:ry: Outli of II Theory of .'-1achillcs. London: ~l[lcmi lIan. Sccdhom. B.s.. Do\\'son, D., &. Wright. V. (19i4). The 10; bl:'aring function of Ihl' menisci: r\ prdiminary study. 0.5. lngwl'rsl'n, ct :II. (Eds.), Thc h//CC loill/: f?en.'1/t .' \'(l/ICC.~· ill Basic Research (wd Clinica! .'lsp(~rlS (I'p. 37-4 r\mS!l:rdam: Exccrpta Medica. Seh'ik. G. (1978). Roentgcn slC'l'cophotogr.lllllllclry in Lun Sweden. In .-\.M. Coble!l'l. & R.E. j·Jl'lTon (Eds.), Applic /jolls flUlIltll1 Biosten:mJII..' lru:s. fIroe 81'/£ (166) (p 184-189). Sdvik, G. (1983). Roentgen stcrl.'ophotogramllleiry in 0 thop;:ledics. In R.E. HelTon (Ed.), BioslereOluetric:s 'S Pro" SP/E (36/) (pp. li8-18S). Smidt. G.L. (1973). Biomechanical :-nnlysis of knee nexi ;lUd c.\:wnsioll_ 1 Bioll/celt, 6. 79. Solomonow, M. &. D'r\mbrosia. R. (1994). Nellr:!1 l'1'f1ex ;I ;tnd muscle control of knl'c sl;\bilil~' :lIld motion. In W Scott (Ed.), The Knce (pp. 107-120). New York: Moshy. Vl'di, V., Williams, A" Tennant, S.L ('I :II. (1999). Mcnisc rnon::mcnt. An ill-vivo study llsing dynamic i\-lRL J BO loim SlIrf:!.. 8IB(J), 37-41. Winll'l', D.A. (1990). In Bio!llee/ullIics flud Motor CO/ltrol 1'11111/1111 Be!wviour (2nd ed.). Ncw York: John Will'Y or SOilS. Wiktorin. C.v.H. &. Nordin, ~\l. (1986). 1,II/,o(//I(:t;OI/ [() Prohll. So/l'illg ill Biollll..'challics (Pl'. 87-129). PhihlClclphia: Lea Fchig('l'. Wilson. S.A., Vigorila, V.J., &. SCf.HL. WN. (1994). Anatomy. N. SCOIt (Ed.), The Kllee (po 17). Philadelphia: ~'Iosby. Wrclenberg, P., Nemeth. G., Lamontagne. M.. ct <II. (199 P;lssivc knee muscle momcnt <trillS meas.ured in vivo w MRI. Clilliea! Biol1lcclt, I 1(8}, 439-446. Yu, B., Stuart, 1\-1.1 .• Kicnb;ll'chcr, T., et .d. (1997). Valg \·'II·US motion of the knee ill norm.1I lc"d w;dking [lnd st climbing. Clillical Biolllcch, 12(5), 2S6-293. Zcrnickc, R.E, G;lrhammer. J., & Jobe. FW. (1977). Hum pntdlar Icndon rupture. 1 801lt' loillt S1Irg, 59;\, 179-18 Biomechanics of the Hip Margareta Nordin, Victor H. Frank Introduction Anatomical Considerations The Acetabulum The Femoral Head The Femoral Neck , /f Kinematics Range of Motion Surface Joint Motion Kinetics Statics Dynamics Effect of External Support on Hip JOll1l Reaction Force Summary References Ie ~, Introduction The hip joint is one of the largest and most SLable joints in the body. In contrast to the knee, the hip joint has intrinsic stability provided by its relatively rigid bali-unci-socket configuration. (t also has a mobility, which allows normallocomogreat deal lion in the performance of daily activities. Derangements of the hip Can produce altered stress distributions in the joint cartilage and bone, leading to degenerative arlhritis. Such damage is further po~ tcntiated by the large forces borne by the joint. or -~ Anatomical Considerations The hip joint is composed of the head of the femur and the acetabulum of the pelvis (Fig. 8-1). This articulation has a loose joint capsule and is surrounded by large. strong muscles. The construction of this stable joint allows for the wide range of 1110tion required for normal daily activities such as walking, sitting, and squatting. Such a joint must be precisely' aligned and controlled, THE ACETABULUM The acetabulum is the concave component of the ball-and-socket configuration of the hip joint. The acetabular surface is covered with articular cartilage that thickens peripherally (Kempson et aI., 1971) and predominantly laterally (Rushfeld et aI., 1979). The cavity of the acetabulum faces obliquely forward, olitwarcL and downward. The osseous acetabululll in the hip is deep and provides substantial static stability to the hip. A plane through the circumference of the acetabulum at its opening would intersect with the sagittal plane at an angle of 40~ opening posteriorly and with the transverse plane at an angle of 60~ opening laterally. The acetabular cavity is deepened by the labrum, a flat rim of fibro~ cartilage, and the transverse acetabular ligament (Fig. 8-2). The labrum contains free nenle endings and sensory end organs in its superficial layer, which may participate in nociceptive and proprioceptive mechanisms (Kim & Azuma, 1995). The unloaded acetabulum has a smaller diameter than the femoral head (Greenwald & Haynes, 1972) (Fig. 8-3). The acetabulum deforms about the femoral head when the hip joint is loaded. It undergoes elastic deformation to become congruous with the femoral head, and contact is made abollt the periphery of the anterior, sllperi()l~ and posterior The hip joint (front view) 1. External iliac artery. 2. Ps major muscle. 3. Iliacus muscle. 4. Iliac crest. S. Gluteu medius muscle. 6. Gluteus minimus muscle. 7. Greate trochanter. 8. Vastus lateralis muscle. 9. Shaft of femu 10. Vastus medialis muscle. 11. Profunda femoris vess 12. Adductor longus muscle. 13. Pectineus muscle. 14. M circumflex femoral vessels. 15. Capsule of the hip join 16. Neck of femur. 17. Zona orbicularis of capsule. 18. Head of femur. 19. Acetabular labrum. 20. Rim of etabulum. Reprinted with permission (rom McMinn, R.H. Huchings. R.H.R. (988). Color Atlas of Human Anatomy ed., p. 302). Chicago: Year Book Medical Publishers, Inc. arLicular surface of the acetabulum (Konrath et al.. 1998). Load distribution of the acetabulum was studied in vitro in human specimens (Greenwald & I-(aynes, 1972; Konrath et 'II., 1998). Joint reaction forces were simulated to physiological levels. The loading pattern of the acetabulum is shown in Figure 8-3. Removal of the transverse acetabular ligament and labrum sequentially did not affect the loading pattern of the acetabulum significantly (Konrath et 'II., 1998). Superior Anterior Intact THE FEMORAL HEAD The femoral head is the convex component of the ball-and-socket configuration of the hip joint and forms two-thirds of a sphere. The articular cartilage covering the femoral head is thickest on the medi,al-central surface and thinnest toward the periphery. The variations in the cartilage thickness result in a different strength and stiffness in various regions of the femoral head (Kempson et 'II., loading pattern of a human acetabulum in vitro with tact labrum and transverse acetabular ligament. Note: pattern was grossly unchanged after removal of the tr verse acetabular ligament. or the labrum, or both, and therefore those patterns are not displayed. Adapled fro Konrath. GA.. Hamel. A.I.. Olson. S.A.. et elf. {l998). The ro the acetabular labrum and the trdnSverse clCeiabtJ/dr ligame load transmission of tilt, hlp. J Bone Joint Surg. 80/'.,(12), .--I Transverse acetabular ligament 1781-1788. J 971 l. Rydell (1965) suggested I hat most load transmitted in the.: femoral head through the su rior quad,-ant. VOll Eisenharl-Rothe et al. (19 demonstrated in an in vitro study that the ma tude or load influenced tlte loading patlerll on femoral head. At smaller load:s, the load-bea area was concentrated at the periphery of the nate surface of the femoral head. and at hig loads to the center of the lunate and the ante and posterior horns. It is still not known exa how the stresses in vivo on the normal fem head arc distributed. bUl indications from in measurements with an instnll11Cnted prosth head show that the anterior and medial lunat transmitting most 01" the load during daily ac ties (Bergmann et al .. 1993. 1995). THE FEMORAL NECK Schematic drawing showing the lateral view of the acetabulum with the labrum and the transverse acetabular ligament intact. Adapred (rom KOflrarh, G.A., Helmel. Ai.. Olson, s.A., et al. (998). The role of (he acetabular labrum and tile transverse acetabular ligamenr in (oad transmission of the hip. J Bone Joint Surg. 80A(12), 1781-1788. The fen10ral neck has two angular relations with the femoral shaft that arc important to joint function: the angk of inclinalion of the nec the shaft in the frontal plane (the neck-to-shaft gie) and the angle of inclination in the transv 'plane (the angle of anteversion). Freedom of rno of the hip joint is facilitated bv the neck-to-shaft gle, which offsels the femoral shaft from the p laterally. in 11'10$1 adults. this angle is approxima 125~, but it can vary' from 90 to 135 . An angle exceeding 125' produces a condition known as coxa valga; an angle less than 125'" results in coxa vara (Fig. 8-4). Deviation of the femoral shaft in either \vay alters the Force relationships about the hip joint and has a nontrivial effect on the lever arms to muscle force and line of gravity'. The angle of anteversion is formed as a projec~ l.ion of the long axis of the femoral head and the transverse axis of the femoral condyles (Fig. 8~5). In adults, this angle averages approximately 12"', but it varies a great deal. Anteversion of more than 12" causes a portion of the femoral head to be uncovered and creates a tendency tc}\vard internal rotation of the leg during gait to keep the femoral head in the acetabular cavity. An angle of less than 12" (rclroversion) produces a tendency toward ex~ ternal rotation of the leg during gail. Both anteversion and retroversion arc fairly common in children but arc usually outgrown. The interior of the femoral neck is composed of cancellous bone \vith trabeculae organized into medial and lateral trabecular systems (Fig. 8-1, 86), The fact that the joint reaction force on the femoral head parallels the trabeculae of the medial Neck-to-Shaft Angle i Coxa vara angle <.: 125" 1DIIlI Normal 125C, angle Coxa valga angle> 125' _ The normal neck-to-shaft angle (angle of inclination of the femoral neck to the shaft in the frontal plane) is approximately 125". The condition in which this angle is less than 125' is called coxa vara. If the angle is greater than 125·, the condition is called coxa valga. 0------------------ Neck Head Greater trochanter Lateral condyle Medial condyle Top view of the proximal end of the left femur showing the angle of anteversion, formed by the intersection of th long axis of the femoral head and the transverse axis of the femoral condyles. This angle averages approximately 12' in adults. system (Frankel, 1960) indicates the importance the s}'stem for supporting this force. The epiphy seal plates are at right angles to the trabeculae the medial system and are thought to be perpen dicular to the joint reaction force on the femor head ([nman. 1947). It is likely that the lateral tr becular system resists the compressive force on th fcmoral head produced by contraction of the ab ductor muscles-the glutcus medius, the gluteu minimus, and the tensor fasciae latac. Thc th shell of cortical bone around the superior femor neck progressively thickens in the inferior region \Vith aging, the femoral neck graduall.v undergoe degenerative changes: the cortical bonc is thinne and cancellated and the trabeculae arc gracluall}! r sorbed (see Fig. 2-50). These changes may predi pose the femoral neck to fracture. It is noteworth that the fernoral neck is the most cOllllllon fractu site in elderly persons (Case Study 8-1). Kinematics In considering thc kinematics of the hip joint, it useful to view the joint as a stable ball-and-sock configuration wherein the femoral head and aceta ulum can move in all directions. three planes. Measurements in the sagittal during level walking (Murray, 1967) showed the joint was maxirrwlly tlexed during th swing phase of gail, as the limb moved forwa heel strike. The joint extended as the body m fOl\vard at the beginning of stance phase. ll)um extension was reached at heel-oil. The reversed into flexion during swing phase and reached maximal Ocxion, 35 to 40', prior to strike. Figure 8-81\ shows the pattern of hip motion in the sagittal plane during a gail cyc ---------- Femoral Intertrochanteric Fractures A ., n 80-year-old woman falls from her own height af losing her balance. She presented with sharp pain her hip and an inability to stand and walk by herself. Sh is transported to the E.R. and after a careful examinatio and x-ray evaluation, a right intertrochanteric fracture is d,iagnosed. Roentgenogram of a femoral neck showing the medial and lateral trabecular systems. The thin shell of cortical bone around the superior femoral neck progressively thick· ens in the inferior region . • RANGE OF MOTION Hip motion takes place in all three planes: sagittal (flexion-extension), frontal (abduction-adduction), and transverse (internal-external rotation) (Fig. 8·7). Motion is greatest in thc sagittal plane, where the range of ncxion is from 0 to approximatcly 140" and the range of extcnsion is from 0 to 15~. The range or abduction is fTom 0 to 30", whereas that of adduction is somewhat less, from 0 to 25'. External rotation ranges from 0 to 9W degrees and internal rotation from 0 to 70' when the hip joint is flexed. Less rotation occurs when the hip joint is extended because of the restricting function of the soft tissues. The range of motion of the hip joint during gail has been measured e1ectrogoniometrically in all ;' _~o_, ~:,,;:h- The radiograph illustrates a right femoral intertrochanteric unstable fracture with separation of the lesser trochanter. The image shovvs osteoporOlie changes characteristic of the aging process. The decrease in the bone mass at the femoral neck leads (0 reduced bone strength and stiffness as a result of diminution in the amount of cancellous bone and thinning of conical bone increases the likelihood of a fracture at the weakest level During the fall. the magnitude of the compressive forces at the femoral neck overcame its stiffness and strength. In addition. the tensile forces produced by pro tective contraction of muscles such as the iliopsoas gen ated a traction fracture at the lesser trochanter level. :t: ./ i;: " Extension Abduction B Adduction External rotation c D Internal rotation E Movements of the hip joint. A. Flexion-extension. B, Abduction. C. Adduction. D. External rotation. E. Internal rotation . • 't. 'T·- .. , allows a comparison of this motion with that of the knee and ankle. Motion in the frontal plane (abduction-adduction) and transverse plane (internal-external rotation) during gait (Johnslon & Smidt. 1969) is illustrated in Figure 8-8B. Abduction occurred during swing phase, reaching its maximum just after toeolT; at heel strike, the hip joint reversed into adduetion, which continued until late stance phase. The hip joint was externally rotated throughout the swing phase, rotating internally just before heel st.-ike, The joint .-emained internally rotated until late stance phase, when it again rotated externally. The average ranges of motion recorded for the 33 normal men in this study' \vere 12' for the frontal plane and 13' For the transverse plane. As people age, they use a progressively smaller portion of the range of motion of the lower extrem~ ity joints during ambulation. Murray and co\\,/orkers (1969) studied the walking pallerns of 67 normal men of similar weight and height ranging in age from 20 to 87 years and compared the gait patterns of o.lder and younger men. The diffcrences in the sagittal body positions of the two groups at the instant of heel strike are illustrated in Figure 8-9. The older mcn had shorter strides, a decreased nmge of hip flexion and extension, decreased plantar flexion of the ankle. and a decreased heel-to-noOJ" angle of '-. ,. the tracking limb; they also showed reduced do nexion of the ankle and diminished elevation of toe of the forward limb, The range of motion in three planes during c mon daily activities such as tying a shoe, sit down on a chair, rising [Torn it, picking lip an ob from the floor, and climbing stairs \vas measu electrogoniomctrically in 33 normalmcn by John and Smidt (1970), The mean motion during th activities is shown in Table 8-1. Maximal Illatio the sagittal plane (hip flexion) was needed for t the shoe and bending down to sqUal to pick lip object rrom the noor. The greatest motion in frontal and transverse planes was recorded du squatting and during shoe tying with the foot ac the opposite thigh, The values obtained for t common activities indicate that hip flexion o least 120 abduction and external rotation of atl 20" are necessary for caITying out daily activitie a normal manner. G SURFACE JOINT MOTION Surface Illotion in the hip joint can be considere gliding of the femoral head on the a<;etabulum. pivoting of the ball and socket in three planes aro the center of rotation in the femoral head (estim at the center of the femoral head) produces this --Hip joint 70 ----. Knee joint ........... Ankle joint 60~ 50 '" w w :s c 0 1 / 20 ~ 10 u: C , ~ c 10 ... 0 '.'!l" c x LU " " .••.... , 5 •, 1 I . .. w w , Adduction w :s c . \ 5 Q, \ .., / ..... (j) \ I' ........ o i==""1--~ , '" Q Internal rotation ;; ~ , "'i...---' I 5 ...- External rotation 100 Stance phase 60, Swmg phase --.l 51:::::=~....1 £Q. 04 ...... ................ 20 30 100 Abduction \ 'I ./ 0 '0 \ 1 ,, 30~ c ,2 'x w ;; : ,,• 40~ ~ C> . I. ,-', ,, ,, ,, , 100 Stance phase 60 Swing phase 100 Percentage of Cycle Percentage of Cycle B A A. Range of hip joint motion in the sagittal plane for 30 normal men during level walking, one gait cycle. The ranges of motion for the knee and ankle joints are shown for comparison. Adapted from ivlurray, ~J1.p. (1967). Gait as a tolal pattern of movement. Am) Phys iVIed, 46, 290. B, A typical pattern for Differences in the sagittal body positions of older men (left) and younger men (right) at the instant of heel strike. The older men showed shorter strides, a decreased range of hip flexion and extension, decreased plantar flexion of the ankle, and a decreased heel"to-floor angle of the tracking limb; they also showed less dorsiflexion of the an" kle and less elevation of the toe of the forward limb. Reprinted with permission from Murray, MP., Kory, R,C, & Clarkson, 8,H. (1969). Walking patterns in flealthy old men. Gerontal, 24, /69-178. range of motion in the frontal plane (top) and transver plane (bottom) during level walking, one gait cycle. Ad from Johnston, Pl. C. S Smiclt, G.L. (J 969). /vleasuremcn[ of h joinr morion during waf.l-:.ing_ EvaluMion of an electrogoliiom method J Bone Joint Sur~J, 51 A 1083 ing of the joint surfaces. If incongruity is prese the femoral head, gliding may not be parallel or gential to the joint surface and the articular cart may' be abnormally compressed or distracted. In center analysis by means of the Reulcaux me cannot be performed accurately in the hip join cause motion takes place in three planes simul ously. Locating the center of rotation of the hip is essential for prosthetic surgery' of the hip to re struct an optimal lever arm of the gluteus me muscle (Fess)' et aI., 1999), Kinetics Kinetic studies have demonstrated that substa forces act on the hip joint during simple activ (Hurwitz & Anclriacchi, 1997, 1998), The faclo volved in producing these forces must be lI stood if rational rehabilitation programs arc Mean Values for Maximum Hip Motion in fhree Planes During Common Activities Plane of Motion Recorded Value (Degrees) Tying shoe with foot on floor Sagittal Frontal Transverse 124 19 1 Tying shoe with Sagittal Frontal Transverse 1 foot across opposite thigh Sitting down on chair and rising from sitting Frontal Transverse Activity Stooping to obtain object from floor Squatting Sagittal Sagittal Frontal Transverse Sagittal Frontal Transverse J'>.scending stairs Sagittal Frontal Transverse Descending stairs Sagittal 23 33 104 20 117 21 122 28 26 67 16 18 36 i'.ile,)1l for 33 normal men. Data hom Johnston, R.C. & Smidt. G.L. i.i970), Hip motion measurements for selected activities of daily living; (lin Orrhop, 72, 205 developed for patients with pathological conditions of the hip. The abductor muscle group (the glutcus medius and minimus muscles) is the main stabilizer during one-legged stance (Kumagai et aI., 1997; University of California, 1953) STATICS During a two-leg stance, the line of gravity of the superincumbent bod:v passes posterior to thc pu~ bic s)'mphysis, and, because the hip joint is stable, an erect stance can be achieved \vithout muscle contraction through the stabilizing effect of the joint capsule and capsular ligaments. With no muscle activity to produce moments around the hip joint, calculation of the joint reaction force becomes simple: the magnitude of this force on each femoral head during upright two-legged stance is one half the weight of the superincumbent body. Because each lower extremity is one sixth of body weight, the reaction force on each hip joint will one half of the remaining two thirds, or one thi of body weight; however, if the muscles surroun ing the hip joint contract to prevent swa,ying an to maintain an upright position of the body (e. during prolonged standing), this force increases proportion to the amount of muscle activity. vVhen a person changes from a two-leg to a singl leg stance, the line of gravit:v of the superincumbe body shifts in all three planes, producing momen around the hip joint that must be counteracted muscle forces and thus increasing the joint reacti force. The magnitude of the moments, and hence t magnitude of the joint rcaction force, depcnds on t posture of the spine, the position of the non-wcigh bearing leg and upper extremities, and the inclin tion of the pelvis (McLeish & Chamley, 1970). Figu 8-10 demonstrates how the line of gravity in t frontal plane shifts with four different positions the upper body and inclinations of the pelvis: stan ing with the pelvis in a ncutral position, standi with a maximum tilt of the upper bod)' over the su porting hip joint, sianding with the upper body ti ing away from the supporting hip joint, and standi with the pelvis sagging away' from the supporting h joint (Trendelenburg's test). The shift in the grav line, and hence in the length of the lever arm of t gravitational force (the peq)endicular distance b tween the gravity line and the center of rotation the fernoral head), influences the magnitude of t moments about the hip joint and, consequently, t joint reaction force. The gravitational forcc lever ar and the joint reaction force are minimized when t trunk is tilted over the hip joint (Fig. 8-108). Two methods arc used for deriving the magnitu of the joint reaction force acting on the femor head: the simplified free-body technique for cop nar forces and a mathematical method utilizi equilibrium equations. The simplified free~bo technique for coplanar forces was described in d tail in Chapter 7, in Calculation Box 7.1. This tec nique is used in the hip to cstimatc the joint rea tion force in the frontal plane acting on the femor head during a single-leg stance with the pelvis in neutral position (Calculation Box 8-1). The seco method is a mathematical calculation of the jo reaction force on the femoral head using equil rium equations for a single-leg stance with t pelvis level (Calculation Box 8-2). To understand and solve the equations it is ne essaI)' to indicate first the location of the extern forces acting on the body during the single-l Roentgenograms utilizing a plumb line (black line) show that the line of gravity shifts in the frontal plane with differ· ent positions of the upper body and inclinations of the pelvis. A, The pelvis is in a neutral position. The gravity line faUs approximately through the pubic symphysis. The lever arm for the force produced by body weight (the perpendicu· lar distance between the gravity line and the center of rotation in the femoral head) influences the moment about the hip joint and hence the joint reaction force. B, The shoulders are maximally tilted over the supporting hip joint. The grav· ity line has shifted and is now nearest the supporting hip. Because this shift minimizes the lever arm, the moment about the hip joint and the joint reaction force are also minimized. C, The shoulders are maximally tilted away from the supporting hip joint_ Again the gravity line has shifted toward the supporting hip, thus decreasing the joint reactio force. D. The pelvis sags away from the supporting hip join (Trendelenburg's test). The shift in the gravity line is simila to that in C. (Courtes}' of 10hn C Baker. IvI.D., Case vVesrem R serve Unive(sir}~ Cleveland, OhiO} Note: In B, the antalgic gait illustrated, which lowers the load on the head of the femu but alters the load line to a more vertical position. Followi arthroplasty for arthritis, the abductor muscles are weak a atrophic as a result of the disease process and the surgery. External support such as a cane should be used until the a ductor muscles are rehabilitated. The best indication for a rehabilitated abductor muscle is the lack of limping. on a free·body diagram (Calculation Box Fig. 1)_ Because the body is in equilibrium (i.e" the of the 11''lOmenls and the sum of the forces both zero), the ground l'eaction force is equal to the 'Q:l-,witatwllal force of the body, which can be divided 1WO components, the gravitational force of the leg (equal to one-sixth body weight) and the force (equal to five-sixths body weight). body is divided at the hip joint into two h-"e-DC)(lIeS- The main coplanar forces and moments on these free-bodies must be determined. upper free-body is considered first (Calculalion ;1j~IJOX Fig. 8-2-2). In this free-body, t\\'o moments are required for stability. The moment arising from th superincumbent body weight (equal to % 'IN) mu be balanced by a moment arising from the force the abductor muscles_ The rorce produced by the s perincumbent hody weight (% W) acts al a diSlan or b rrom the center of rotation of the hip (Q), thu producing a moment of % Vt/ times b. The force pr duced by the principal abductor, the glute medius, designated as A, acts at a distance of c fro the center of rotation, producing a counterbalan ing rnomcnt of A limes c. For the body to remain moment equilibrium. the sum of the moments mu equal zero. In this example, the moments actin -----------------Simplified Free-Body Technique for Coplanar Forces The stance limb is considered as a free-body, and a free-body lion, simplifying assumptions are made in determining the di- diagram is drawn. From all of the forces acting on the free- rection of this force (Mcleish & Charnley, 1970). Furthermore. body, the three main coplanar forces are identified as the forces produced by other muscles aCiive in slabilizing the hip force of gravity against the foot (the ground reaction force), joint are not taken into account. The joint reaction force (]) \,.'/hich is transmitted through the tibia to the femoral has a known point of application on the surface (lunate) of condyles; the force produced by contraction of the abductor the femoral head bur an unknown magnitude, sense, and muscles; and the joint reaction force on the femoral head. line of application. The ground reaction force ('IN) has a known magnitude equal to five-sixths of body weight and a known sense, line of ap- .'.; The magnitudes of the abductor muscle force and the jOint reaction force can be derived by designating all thr~e plication. and point of appliCc1tion. The abdu([or muscle force forces on the free-body diagram (Calculation Box Fig. 8~_1· (A) has a known sense, a known line of application, and and constructing a triangle of forces (Calculation Box Fig: point of application estimated from the muscle origin and in- 8-1-2). The muscle force is found to be approximately two n sertion on a roentgenogram but an unknown magnitude. Be- times body weight, whereas the joint reaction force is some- cause several muscles are involved in the action of hip abduc- what greater. Force J Intersection point J' // Force A I 2W/ / , 1 ! t Force W Calculation Box Figure 8-1-1. ,._wt/ I I I Force J 2_75 W Calculation Box Figure 8·'-2. ------------------------------------ ----- &--------------------------------------~---------- ,.- . , -:-.-_ .. External Forces Acting on the Body in Equilibrium During a Single-Leg Stance Calculation Box Figure 8-2-1 shows external fOKes acting on the body in equilibrium during a single-leg Slance. The ground reaction force is equal to body weight (W). The gravi!ational force of the stance leg is equal to one-sixth of body weight; Calculation Box Figure 8-2·1. the remaining force is equal to five-sixths of body weight. w The internal forces acting on the hip joint are founel by sepa- rating the joint into an upper and lower free-body; the Lipper free-body is considered first. In this free-body, two moments b 5/6 W : : -_._" are required for stability. Moment equilibrium is attained by the prOduction of two equal moments. A moment arising Calculation Box Figure 8-2·2. from the force of the abductor muscle (A) times abductor force lever arm (e) counterbalances the moment arising from the gravitational force of the superincumbent body (5/6 W) times gravitational force lever arm (b), which tends to tilt the pelvis away from the supporting lower extremity. Q, center of rotation of the hip joint. A = 2W Ax =: A ·sin30~ Ax = O.5A Ay := ~ W A·cos30'" A y = 0.8 A = 1.7 W Ay 30" Calculation Box Figure 8-2-3. Force A is equal to two times body weight and has a direction of 30' from the vertical. The magnitudes of its horizontal (A,) and vertical (A) components are found by vector analysis. Perpendiculars are drawn from the tail of A to a horizomal and a vertical line representing A. and A" respectively. A, and A, can then be scaled off. Allernatively. trigonometry is used to find the magnitudes of the components. .._--------._, ._--~._._-_._-_ Ax clockwise arc arbilraril~' considered to be positive and lhe counterclockwise moments arc considered be negative. Thus, (Y.,W x b) - (A x c) = 0 A= YoW x b -'---- c To solve for A it is necessary to find the values of .~.' and c. The gravitational force lever ann (b) is roentgenographically. Because the cenler of gravity must lie over the base of sUppOrl, a plumb line intersecting the heel can be extended upward; a line drawn from the center of rotain the femoral head (Q) to the line represents """'\IlOe b. The muscle force lever arm (c) is similarl~! found by identifying the glutcus medius muscle on a roentgenogram (Nemeth & Ohlsen, 1985, 1989) and drawing a perpendicular line from the center of rotation 01" the femoral head to a line approximating the gluteus mcdius muscle tendon. In this example, a value for A of two times body weight was derived from thc static free-body diagram and confil'med by instll.llnentcd in vivo measurements (English & Kilvington, 1979; Rydell, 1966). The direction of force A is found from a roentgenogrnrn to be 30 from the vertical. The hOI'izontal and vcrtical components of this force are found by vector analysis (Calculation Box Fig. 8-2-3). The hOl-izonlal component (AJ is equal to body" weight; the vertical component (A y ) is approximately 1.7 times body \veight. Attention is then directed to the lower free-body (Calculation Box Fig 8-3-1). The gravitational forces (Wand II. W) are known. The joint reaction force (force J) has an unknown magnitude and direction but originates from the most narrow joint space in the radiograph and musl pass through the estimated center of rotation in the femoral head. The magnitude of force J is determined by finding the horizontal and vertical force components and adding them (Calculation Box Fig. 8-3-2). The value of J is found by vector addition (Calculation Box Fig. 8-3~3), and its direction is measured on the parallelogram of forces. The joint reaction force on the femoral head in a single-leg stance with the pelvis leveled in the horizontal plane is found to be approximately 2.7 times body weight, and its direction is 69· fTom the horizontal (Calculation Box Fig, 8-3-4). A key factor influencing the magnitude of the joint reaction force on the femoral head is the ratio of the abductor muscle rorce lever ann (c) to the gravitational force lever arm (b) (Calculation Box 0 I I L Fig. 8-2-2). This is particularly important in pros thetic replacements of the hip joint (Delp & Maloney. 1993; Free & Dell', 1996; Lim et al .. 1999 Sutherland et 'II., 1999; Vasavada et 'II., 1994). The center of motion can be altered by the prosthctic de sign and the lever arm for the abductor muscles can be slightly changed by slu-get)! techniqucs. A change of the center of location of the hip joint can de crease the abduction force by more than 40(1'0 and thereby the generated abduclor moment b~} almos 50% (Dell' & Maloney, 1993), Figure 8-11 illustrate the relationship of this ratio to the joint reaction force. ;\ low ratio (i.e., a small muscle force leve ann and a large gravitational force lever arm) yield a greater joint reaction force than docs a high ratio A short lever arm of the nbduc(ol- musclc force, a in coxa valga (Fig. 8-4), rcsults in a sTllall ratio and lInls a sornewhat e1cvated joint reaction force. Mo\' ing thc greater lrochanter latcrally during lata! hip replacement lowers the joint reaction force. as it in creases the lever arm ratio by increasing the muscl force lever arm (Free & Delp, 1996). Inserting a prosthetic CLIp deeper in the acetabulum reduce the gravitational force lever ann, thereby increasing the ratio and decreasing the joint rcaction force. I is difficult. howc\'cl: to change the lever an11 ratio in such a way as to reduce the joint rcaction force sig nificantly been use the curve formed from plouing the ratios becomes asymptotic when the ratio of c to b approaches 0.8 (Fig, 8-1 I). DYNAMICS The loads on the hip joint during dynamic activi ties have been studied by several investigator (Andriacchi ct aI., 1980; Draganich c{ al.. 1980 English & Kilvington. 1979: Rydell, 1966). Using a force plate system and kinematic data for the nOI" Illal hip, J. P. Paul, 1967. (Forces at the human hip joint. Unpublished doctoral theses, University o Chicago) examined the joint reaction force on th femoral head in normal men and women during gait and correlated the peak magnitudes with spe cine muscle activity recorded electromyograph i cally. In the men, two peak forces were produced during the stance phase when the abductor mus cles contracted to stabilize the pelvis. One peak o approximately four times body weight occurred just nfter heel strike, and a large peak of approxi matel:v seven times body weight was. reached jus before toe-ofr (Fig. 8-12A). During foot flat, th joint reaction force decreased to approximately ~. I Free-Body Diagram of the Lower Extremity : In Calculation Box Figure 8~3·1, the supporting lower ex- tremity is considered as a free-body, and the forces acting on the free-body are identified. A, abductor muscle force; J, joint reaction force; 1/6 W, gravitational force of the limb; W, ground reaction force; Q, center of rotation Tana Jy of the hip joint. c·· 2.5W Tan a In Calculation Box Figure 8-3-2, the forces acting on a A Jx 2.5 69 the lower free-body arE> divided into horizontal and vertical components. Because the body is in force equilibrium, the sum of the forces in the horizontal direction must equal zero and 50 must the forces in the vertical direc- tion. The horizontal and vertical forces are added and the magnitudes of J.: and J( are found from force equilibrium Jx W Calculation Box Figure 8~3-3, equations: A,~-J,~O A, ~~J.- '/W"," W ,- a J I.7W-J.-'/W W~O A,~W 2~7 W .I. A J..., 1.7 W J,. - W ~,~ /. W From the Calculation Box 82 A. - I.7W Calculation Box Figure 8-3-4. Addition of the horizontal and vertical components J..- and J, is performed graphically in Figure 8-3-3, and the joint reaction force (J) is scaled off. A parallelogram is constructed and the diagonal of the parallelogram indicates the inclination of the force J. Its inclination in relation to the horizontal plane is measured (n) on the parallelogram, Alternately, trigonometry is used to find the direction of J using tangent equations (Fig. 8-3-4). The joint reaction force has a magnitude of approximately 2.7 times body weight and acts at an angle 69" from the horizontal, w Calculation Box Figure 8~3~1. w Calculation Box Figure 8-3-2. Men 7.0 7 6.0 6 'la> 'ffi :ca> 'ffi ;; >1S 5.0 ~ >- "e0 ~ ~ 0 5 4 @. 3 ~ ~ 4.0 a u. 2 u. 3.0 > Upper and lower limits of 10< angle of 2.0 ~..,._.,......,_..,.._.--..,._.,.......;5,,0:..'.., inclination o 0:2 0:4 0:6 0.8 i of abductor !1Bm I, I I I Ratio of c to b 100 60 Percentage of CyCle Women muscle force line _ The value of the ratio of the abductor muscle force lever arm (e) to the gravitational force lever arm (b) is plotted against the joint reaction force on the femoral head in units of body weight. Because the line of application of the abductor muscle force (its angle of inclination in the frontal plane) has finite upper and lower limits (10 and 100 60 Percentage of Cycle 50°), the force envelope is plotted. The curve can be uti· lized to determine the minimal force acting on the femoral head during a one-leg stance if the ratio of ( to b is known. AdiJpced from Frankel. VH. (960). In The Femoral Muscle AClivity During Slance Phase of Gait Gll/leus ma~,mus Sem:teM'llOSUS Semimem!)r,lllOSUS Biceps tomor,s Neck: Function, Fracture Mechanisms, Internal fixation. Springfield: Charles C. Thomas, I ---------------8 "====== r ~~~~~~~F1'"'' Tensor [osciac S,lrtOri.us _ Gr,lc,bs Reclus femoris Gluteus med,us Gluleus minimus IIi",,", E~tellsors 1------!-- .\dduc!or magous Adductor IOr1(Jus Adductor brevis Abductols .-==:==~AddUCIOIS 100 body weight because of the rapid deceleration of the bodv's centcr of ~ gravilv. Durin t>lT the swing~ .' .' phase, the joint reaction force was influenced by contraction of the extcnsor muscles in decelerating the thigh. and the magnitude remained relatively low. approximately equal to body weight. In the women, the force pattern was the same but the magnitude was sorncwhat loweI~ reaching a maximum of only approximalely four times body weight at late sfance phase (Fig. 8-128). The lower magnitude of thc joint reaction force in the women may have been the result of several factors: a wider female pelvis. a difference in rhe inclination of the fcmoral neck-to-shaft angle, a difference in footwear. and differences in the general pattern of gait. c 60 Percentage of Cycle Hip joint reaction force in units of body weight during walking, one gait cycle. The shaded area indicates variations among subjects. A. Force pattern for normal men. 8, Force pattern for normal women. Adapted from Paul, IP (1967). Forces at the human hip joint. Unpublished doctoral the· sis. University of Chicago. C, Muscle activity during stance phase of gait. The first peak corresponds mainly for the ex- tensor and abductor muscles. The last peak. is for the flexor and adductor muscles. Adapted from the University of California. (1953). The pattern of muscular acrivity in rhe lower excrem i!'/ during ,·valking. Univ Cal Prosthet Dev Res Rep, 2(25). 1--4/. • Walking (0.9 mtsec) - - Ground reaction force .........u". Force on prosthesis 1400 1200 1000 ~ ..e 0 lL t\ 800 600 400 200 , 4 A Walking (1.3 m/sec) - - - - Ground reaction force ............." .... Force on prosthesis 1800 1600 In \"ivo ml..'asurl..'ll1enls of lhe forces acting instrumented hip joint prosthesis demonstra lower joint reaction force or the femoral head ing the stance plwse or gail compared with ex mcnsuremcnts and calculations (Rydell, 1965 8-13;\). At a faster cadence. the forces •.\Cling o prosthesis grcatl~' increased because of an inc in muscle ac(ivi(~' (Fig. 8-138). At both cade the magnitude the forces during swing phas arproxirnatel~: half that during stance phase. l~lble 8·2 summarizes the typical peak joint on the hip joinl load e."pressed as bod~1 weight different studies and with differ·ent methods pattern of loading for \valking is similar for all ies, but the magnitude of joint peak load differ ternal measurements gencraHy yield higher lated peak force on the hip joilll while inslrum implant in vivo measuremcnts ~'ield lower forces. There are man)' reasons 1'01' the diffe for example, the melhod and instn1l11cnwtio normal hip "crsus the "abnormal" instrulllcnte plum. the gait velocity, and age. Activities othe \\·alking, such as stair ascending/descending, loads of around 2.6 to 5.5 body wciglll Illea or 1400 1200 ~ , '"e0 lL 1000 ,ImI!DJI 800 600 ;!- Range of Typical Reported Peak Hip Joint Forces From Selected Studies 400 or-, ,1 200 SlanceDphase Activity i i 1 2 i 3 Reported Peak Force BW Instrumentation ... _--_.,.._,._-- -_.,-.'.'..._-_._---_ Walking Swing phase 2.7-4.3 Insirurnented implants Bergmann 1993, 19 2.7-3.6 Instrumented implantS Kolzar et a 1991 2.7 Instrumented implants English et 1979 1.8-3.3 Instrumented implants Rydell et a 1966 4.9-7.0 EIviG /force plate Paul, 1967 45-7.5 EIviG fforce plate Crowninshi et al.. 19 5.0-8.0 ElvtG fforce picHe Rohrle et a 1984 2.2-2.8 accelerometers van den Bo et aI., t9 Time (seconds) B Forces on an instrumented hip prosthesis during walking. The broken line represents the force on the prosthesis, and the solid line represents the ground reaction force. A. Walking Walking speed 0.9 m per second. B, Walking speed 1.3 m per second. An increase in muscle activity at the faster cadence resulted in higher forces on the prosthesis. Adapted from Rydell, tV. (1965), Forces in tfle hip-joint. Parr If: Intravital measurements. In R.M. Kenedi (EdJ. Biomechanics and Related Reference Sio-Engineering Topics (pp. 351-357). Oxford: Pergamon Press. BW. body \'1elght; EMG. electromyography ~--------'--'---'-"---'------"'-----r Fatigue Fracture of the Hip '1\.", 64-year-old, very active retired man experienced a ·M femoral neck fracture after changing his training ,'regime to prepare for a marathon. The fracture was ~. '.> classified as a faligue fracture caused by overload of -, the hip joint. ;~ of a femoral neck fracture allowed a subseq determination of the forces acting on the imp during activilies or daily living (Fig. 8-14) (Fra el aI., 1971). Although the device measured fo on the implant and not on the hip joint, it was sible to determine the proportion of the transmitted L1nough the device and to calcu the total load acting on the hip joint by mean static analysis. In the case illustrated in Fi 8-14. the nail plate transmitted one fOllrth of lOtalload. Strong forces aCling on the nail plate were countered during such diverse activities as mo onto a bedpan, transferring to a wheelchair, walking. The magnitude of the forces was gr modified by skillful assistance frol11 lhe nurs therapist to control the patient's movement. Fo of lip to fOllr limes boely weight acted on the joint when the patient used the elbows and hee elevate the hips while being placed on a bed (Fig. 8-15), but these forces were greatly red through the lise of a Lrapeze and assistance from attendant (Fig. 8-158). A 5-kg ex.tension tractio the hip had little effect in modifying the forces ing on the hip joint. Exercises of the foot and a increased these forces. Case Study Figure 8-2-1. The figure shows an MRI (frontal view) of the pelvis and both hip joints. The fracture is seen in the left femoral neck distal to the femoral head. The fracture is believed 10 have occurred during running and after an extensive change of training program. Because of the high repetitive I?ading, muscle fatigue, and the change in the load pattern, on the hip joint and femoral neck, the bone fractured. with an instnlmentcd hip implant (Bergmann et al., 1995; KOlzar et aL. 1991). The highest magnitudes of load during daily activities are measured dllling stair climbing and getting up from a low chair when the hip is flexeell1lore than 100" (Catani et aI., 1995; Johnston et aI., 1979). Co-contraction of the biarticular muscles was evident during these activities. Running and skiing lIsing accelerometers yielded calculated forces up to eight limes body weight in middle-aged and older people (van den Bogert et al., 1999) (Case Stlldy 8-2). Insertion of an instrumented nail plate in the p'-oximal femur arter osteotomy or during fixation .., ...,: .. An instrumented nail plate in the proximal end of the mur was used to determine the forces acting on the i plant during the activities of daily living following fra of the femoral neck. In this ca5e, the nail plate was fo to transmit one fourth of the total load on the hip jo Usc of the instrumented nail plate demonstr that, for a bedridden patient with a fractu femoral neck, the forces on the fernoral head ing activities of daily' living approached those iog walking with external supports. These stu support clinical protocols for carly mobilizatio patients and decreased bed rest for patients A B A, When the patient used elbows and heels to elevate the hips while being placed on a bedpan, the force on the tip of the instrumented nail was 670 N. With a spica cast, the force on the tip of the nail was 190 N. B, The use of a trapeze and assistance from an attendant reduced the force to 190 N without a cast and to 70 N with a spica cast. Reprinted "vith permission from Frankel, VH. (7973). Biomechanics of the hip. In R. G. Tronzo (feU Surgery of the Hip Joint (.0,0. 105-125). Philadelphia: Lea & Febiger hip fractures. The magnitude (approximatel Nlll) of the moments acling on the nail-plate . lion in the transverse plane (i.e .. during inte and extcrn~ll rotation) was only approxirnatel)! half the magnitude (18 Nm) of the moments ac in the frontal plane (Le., during abduction) many activities. EFFECT OF EXTERNAL SUPPORT ON THE HIP JOINT REACTION FORCE Static anal.\·sis of the joint reaction force on femoral head during walking with a cane dem strates that the cane should be used on the side posite the painful or operated hip. Neumann (1 High and low load on the hip joint during daily activities. Raising from a low chair produces approximately 8 times body weight (A). Walking with a cane on the ipsilateral side of the affected hip produces approximately 3.4 times body weight (B), and walking with a cane on the contralateral side of the affected hip reduces hip joint load substantially to 2.2 times body weight (C). This figure illustrates how load on the hip joint can be manipulated by simple means (X denotes affected hip). -:i:::~C::I::J::r:S:~ t:~::r::e~h:i::~re:~~o<~Oint Rea~:~::e:::C:uSCle I faKe acting on the femoral head in Ihe late swing phase of I the gait cycle for an 8·year-old boy INeighing 24 kg and momenl relationship T ::::: Fd, i I where F is the extensor muscle force and d is the perpendicu~ lar distance from the center of rotation of the femur to the wearing a )ong-Ieg brace. The main muscle force was produced by contraction of the gluteus maximus muscle and identified through electromyography. The torque about the middle of the gluteus maxim us muscle. Distance d was mea., hip joint \ivas calculated according to the formula sureej from a roentgenogram and found to be 3.2 em, From H'e equation E;::;: TId, the muscle force on the normal Side', . . " as calculated to be 338 N, and on the braced side, 600 N. T ~ la, where T is the torque expressed in newton meters (Nm) I is the The joint reanion force on the femoral head (J) is equal to mass moment of inertia expressed in newton meters times seconds squared (Nm sec!) is the angular acceleration in late swing phase, expressed -;,- in radians per second squared (rlsec 1). In the case of (he braced side. I = I , .;. I;; where is the mass moment of inertia of the leg is the mass moment of inertia of the brace. On the normal side, I "" 0.45 Nm se(' n=: 24 rlsec'. Thus, I = 0.45 Nm sec:' .: ~ force (E) was then found from the x 24 T ~ On the braced side. 1 0.45 Nm sec; + 0.35 Ntn sec' mated to be 40 N. On the normal side, On the braced side, J = E -\iV, J = 338 N - 40 N J = 298 N. J = E -\iV, J=600N-40N J = 560 N. Thus, tile Joint reaction force on the femoral head in the br~1Ced limb v"as over 80% higher than the force in the non~ ;.cc n = 24 rlsee" braced limb, reaching more than two times body weight. Thus, T = (.45 Nm sec) + 35 Nm sec!) x 24 rlsec~ 10.8 Nm. the muscle force (E) minus the gravitational force produced by Ihe weight of the limb (W,). In this example, W L \·\las esti- T ~ rlsec~ 19.2 Nm. I 0~--- ,;:: studied the effects of cane use in 24 subjects with a mean age of 63 years. During walking, the e1ectromyographic activity of the hip abductor muscles was measured, Neumann found that usc of a cane on the contralateral side of the affected hip joint, with careful instructions to use with near maximal effort, could reduce the muscle activity by 42% (Fig. 8-16). This calculates to a reduction of approximately one times bod~; \veight from 2.2 body weight with a cane, compared with 3.4 body weight withOut a cane. These studies give important inforllla~ lion to the clinician about ways to moderate the load for the patient with hip problems. Such use reduces the force on the femoral head of the painf-ul joint without necessitating an anlalgic body position. A cane used on the side of the painful hip works through a shoner levcr arm and thus an even greater push on the cane is needed to decrease the joint reaction force. For the olcler patient, such a large push may not be possible because of a of strength in the upper extremities. The use of a brace on the leg may aller the f on the hip joint but may not always reduce the reaction force on the femoral head. An ischial leg brace used in the treatment of Perthes' dis raises the joint reaction force during late swing p because the large mass moment of inert ia o brace results in a higher extensor muscle force ing this part of the gait cycle (Calculation Box 8 SUl1lmary 1'; The hip JOll1t is a ball-and-socket joint posed of the acetabulum and femoral head, 2 The thickness and mechanical properties o on the femoral head and acetabulum from point to poinl. cartilage -_.~, 3 Hip llexion of at lca~·a 120", abduction of at least 20", and extenwl rOl<.l.lion of at least 20" are necessary for can}'ing out daily activities in a normal manner. \4 A joint reaction force of apprOXill'latc1y three timcs body weight acts on the hip joint during a single-leg stance with the pelvis in a neutral position; its m;gnitudc varies as the position of the upper body changes. _~'> The magnitude of the hip joint reaction force is infiuenced by the ratio of Ihe abductor muscle force and gravitational force lever anns. A low ratio yields a greater joint reaction force than does a high ratio_ ;;_~ The hip ,joint reaction force during gail reaches levels of three to six limes body weight or more in stance phase and is approximately equal to body weight during swing phase. j.7 An incrense in gait velocity increases the magnitude of the hip joint reaction force in both swing and stance phase. ,'8 The forces acting on an internal fixation device during the ~\ctivitie; of daily living vary greatly' depending on the nursing carc and the therapeutic activities undertaken by the patient. /9 The lise of a cane or a brace on the leg can altcrthe magnitude of the hip joint reaction force. REFERENCES Andriacl:hi. T.P.. Andersson. G.B .. Famil:r. R.W., Stern. D.. G,t!anil', J.O. (1980). r\ study of to\\'er-limb medlanics during stair·dimhing. J BOllI.' Johu Surg, 62.-1. 749. Bergmann, G., Graichl'll. E, & Rohlm=lIl1l. A. (1993). Hip joint 1~'H.ling during \V,lIking i.lIld running llh..· .ISurcd in IWo pa· tknts. 1 Bio/lw:h, 26(8),969-990. Bergmann. G.. Graichen. E. & Rohlmann, A. (t995). ts slair.. cast: walking a risk for tht: fi.X<lli(lll of hip implants?} 13ioIlIe,''', 28(5), 535-553. Catani, F.. Hod!!(', r\ .. ~'talln. R.W.. et ::d. (1995). The 1"011.; of museul"r co:eontr<lClion of the hip during mon.·men!. Chi,. Orgalli ;\1ov, 80(2), 227-236. Crowllinshidd, R.D .. Johnston, R.C.. Br<lnd. R.t\. (1978). The eHccts of walking no'locity and agl' on hip kint:m<ltics and kinetics. Clill Or/hop Rd Ues, 132,140-14-\. Ddp, S.L. &. Maloney, \V. (1993l. Effects of hip Centcr locatioll on the momcnl-generating cilpacity of the l1luscks.) 8'0mi'd,. 26(5). 485-499. Drng::lllil·h. L.F.. Andrincchi. T.P.. Strongw::llcr. :\.1\'1.. ct ~\1. (l980). Electronic tlll.:aSLlrcrnent instantaneous fOOLfloor conl;:\ct paltl'rns during gait. 1 BiOlIICdl. 13. S75. English. T.A. & Kilvington, M. (1979). In \'ivo re('ord~ of hip IO:.lds using ::1 fl'llloral impblll with lelemetric output (a or preliminary l't:ponl. 1 Biollled ElIg, 1(2), 111. Fl·ss.\', ~1.I-1.. ;\'Diaye, A.. CalTd. J.P.. et :II. (1999)_ Lo tht..· l·{.·lltl-r or rOlation of IIII.' hip. Sl/r.g nmliol A//at, 24;-250. FI'~lnkd, V.H. (19;3). Biollll'challil:s the hip. In R.G. T (Ed.), Slll'.~t:ry o( I!lt' Hip }Oil/l (pp. 105-125). Phil~ldl. lea &. Fdlif!l'r. Frankd. V.H. (1960). In nUt FI.'J/ulI"iI} Nt'd: {-'Wlclioll. F .\-1f.'(:IUllli:'/II.\, "/lenla! FixaTioll. Springfidd: Char or TlwllWS. Fr'lnkd. V.H .. BurSil'in. A.H .. L~·grl·. L., l't al. (1971). Th L.lil: nail. ) BOIH' 10illl SIII"~. 53:1. 1232. Free, S.A. & Oelp, S.L. (1996). Trochantt.'rll.: transfer i hip n:pJal:i.:IlIl.:nl: Effecls on Ilit' lTlOfllL'llt arms nnd gellt..'ralillg c'lpadtic:;s of Ihe hip abductors. J Onlw 1-I{2L 1-15-250. Gn.·em\'~a1d. A.S. & 1-I<I\'llI,:S, D.W. (1972). Weight-hcaring ill lhe human hip j{~Jinl..! BOIlt' .!Oilll Siltg. 5-1B( I), 15 Hurwitz. D.E. & Andri.tt:("hi. '1'.1>. (1998). Biomechanics hip. III J.J. Cdbgh'111. A.G. Rosl'nberg. &. I-I.E. R (Eds.1. The: Adult flip (pp. 73-85). Philaddphia: Lipp Raven Publishers. Hllr",it;',. D.E. & Andri~\(:("hi. T.P. (1997). Biolllcl:hanks hip .tlld the knl'l'. In ~'1. Nordin, G.B.J .•\ndnsson. & POPl' (Eds.) . .\lusculoskclewl Disorders ill {Itt' Irork Principles (Jud Prauice (pp. -186--196). Phila(k']phia: M Year Book. Inrn;lIl. V.T. (19-47). Flll1l:lional aspects of t!ll' abductO c1es of thl' hip. J BOllI.' loim Sll"~. 29:\. 607. JohnslOn, R.C., Br~Hl(I, R.A .. &. Crowninshicld. R.D. ( Re('onslnH:lioll of tht' hip. } BOI/t' Join, Smg. 6 646-652. Johnston. R.C. & Smidt. C.L. (1969). !\leasur(,llll'nt o join! motion during walking. Evaluation of 'tll I'll' niOllll'tric Illt.'lhod. 1 BOll...' loiJ/! Sll}".~. 51:\, 1083. Johnston, R.C. & Smidl. C.L. (1970). Hip motion tlll m('Il{S for sell-tted .\Ctidties of daily li,·ing. efill O 72. 105. Kcmpson, G.E., Spin.'y. c.J .. Swanson. S.A.V.. cl at. ( Patk'rns of C<lrtila~1' stiffnt..'ss on normal and dC~Cl human femoral he;ds, J liiO/IIt'ch, -I, 59;. Kim, Y.T. &: Azurll.l, H. (1995). The lH.'I'\'e t,;ndings of etabular "tbrum. CfiJl Or/hop, 320, 1;6-181. Konrath, G.A., Hamel. A.1 .. Olson, S.r\ .. d .d. (19981. T of the aCI'I.dHlbr labrum and the (r~\IlS,'e..SI' :lcL'(ablll ;l11lt.:lIt in load transmission of the !lip. J (JOllt~ loill 80A(I2I. t781-1788. Kotzar. G.~L Davy. O.T.. Goldberg. V.;\1., cI al. ( TcJemilriZL'd in "j\'o hip joint forcc dilta. A report o ll<ltit.:IHS aftcr 10iai hip sllrgery. 1 Ort/wp Nt's, 9. 621 Kumagai. iv1.. Sliiba, N., Higuchi, F.. et a!. (19971. FUll t.:\'aIU:llion of hip abductor mllSC!<;S with llSI' of m. resonance imaging. 1 Orlhop Rt,S, 15(6),888-893. lilli. l.A .. CIl"mi(;h;lel, S.W.. &. C<lbancla, M.E. (19991 llH:chanics of lOUt! hip ;lrlhroplnsly. AI/ot UI:C, 2 t 10-t 16. j\'kLl'ish, R.D. & Chilrnh:y. J. (1970). Abduuion forct's one-!l.'ggcd st~lnc('. J B;oll/cch. 3, 191. ;\'ld,linn, ~R.H. &: Huchil1!.;s, R.H.R. (1988). In C%}" Il HI/lllml AWIlO11/Y (2Il1.! cd., p. 302). Chic,lgO: )\:.11 ,\-kdic<ll Publishl'rs, Inc MlllT~I\'. ~LP. (1967).. G.dt <lS ,I lOt~t1 pattern of moveme 1 pjlYs .lled, -16, 290. \-!urray, i'",1.P., Kor.\', R.C., &: Clarkson, 13.11. (1969). Walking patterns in he:dthy old nH:n. } Geroll/ol, 2-1, 169-178. Nemeth, G. &: Ohlsen, I-l. (1985). In vivo moment arm lengths for hip extensor muscles at difTerent angles of hip fle\ion. } Biolllech, 18, 129-149. Nemeth, G. & Ohlsen, H. (1989). rvlolTlel1t arms or hip abductor and adductor muscles in vivo by computed tomogra· phy. Cfill 13iolllccll, 4, 133~136. Neumann, D,A, (1998). Hip abductor JllLlscle activity as subjects with hip prostheses walk with dirrerL~nt methods of 'using a cane. Pllys Ther, 78(5),490-501. IWhrle, H., Scholten R., Sigolo!to, C., et al. (1984). Joint forces in the human pelvis-leg skeleton during walking. } Biolllcch, 17, 409-424. Rushfcld, P.D., \tann, R.\V., &: Harris, W.H. (1979). Influence of cartilage geometry on the pressure distribution in the human hip joint. Sciellce, April 27, 204(4391), 413-415. Rydell, N.W. (1966). Forces acting on the remoral head prosthesis: A study on strain gauge supplied prostheses in living persons. Acta Orr/lOp S("(lIld, SlIppl88. 1-132. Rvdell, N. (1965). Forces in the hip-joint. Part If: Intravital . measurements. In R.rvl. Kenedi (i~d.), l3iolllt'cllilUics (jlld Bio-LII~il/ecrilig Topics (pp. 351-357). Ox f\:rgamon Press. Sutherland, A.G., D'Arcy, S .. Srn:lrt, D., et al. (1999). Ab tor muscle weakness and stress around acetabu!:lr co nents of total hip arthroplasty: A finite dement ana Re!uh'd 1111 Orthop. 23(5), 275-278. University of California. (1953), The pattern of muscula tivity in tht: lo\\'er extremity during walking, (jllll Pros/hd DcI' Rcs Rep, 2(25). 1-41. van den Bogert, A.J., Read, L., &. Nigg, B.rv1. (1999). An ; sis of hip joint loading during walking, running and ing . .lIed Sci Sports, 31(l), 131-142. Vas:lvada, A.N., Delp, S.L., \.laloney. \V.J., et a!' (1994). pens;:lling (or changes in Illuscle length in total hip <l pbst~,. Effects on tllL' llHHTll'nl gt:nt:rating capacity o muscles. Cfill OnllOp. 302. 121-·133. VonF.isenhart-Rotht:, R.. Eckstein, r.. ~luller-Gerbl. rlit., (1997). Direct comparison or contact <.In:as, contact and subchondral mineralization in Illunan hip joint s mens . .. \I/IJ{ f:'lIlbryol ([kr/), 195(3), 279-288. Biomechanics of the Foot and Ankle G, James Sammarco, Ross Todd Hockenbu Introduction Growth of the Foot Kinematics of the Foot Foot and Ankle Motion During Gait Causes of leg Rotation During the Gait Cyde Muscle Action During Gait MotIon of the Tarsal Bones Sublalar Joint Motion Transverse Tarsal Joint Motion Tarsometatarsal and Intertarsal Motion Motion of the Hallux Motion of the Lesser Toes The Medial longitudinal Arch Muscle Control of the Foot Kinetics of the Foot Soft Tissues of the Foot Ankle Joint Biomechanics Kinemaiics Range of Motion Surface Joint Motion Ankle Joint Stability Kinetics of the Ankle Joint Statics Ankle load Distribution Dynamics Effects of Shoewear on Foot/Ankle Biomechanics Summary References Introduction The biomechanics of the fool and ankle arc complex and intricately associated with each olher. The fOOl is an integral mechanical part of the lower cxtrcrnity necessary 1'01' a smooth and stable gaiL The ankle transfers load frolll the lower extremity to the foal and closely inllucnccs foot orientation with the ground. The fOOL is comprised of 28 bones (including scsamoids) whose motions arc closely inlcn-c1alcd (Fig. 9-1). Besides acting as a structural sllpponing platform capable of withstanding repetitive loads or mulliples of body weighl. the fOOL/ankle complex also must be able to adjust to different ground surfaces and vat)'ing speeds of IOCOl11CHion. The unique qualities of the 1'001 allow it to be rigid when ncccsSal)', as in ballet dancing on point, or quile flexible, as in walking barefoot on the sand. The transition from shock-absorbing platform to rigid lever capable of forward propulsion occurs wilh cach slep of lhe gail cycle. The ankle is compriscd of lhrec boncs Ihal form the ankle mortise. This joint complex consists of the libiotalar, flbulotalar, and tibiofibular joints (Fig. 9-2). The ankle is a hinge joint whose stability depends on joint congruency and the medial. lateral, and syndesmotic ligaments. This chapter discusses the motions that occur in the foot and ankle during the various phases of gait as well as during extremes or motion. The close interplay between lower extremity rotation and Forefoot orientation is explained. The ground (foot-tofloor) reaclion force and dislribulion of forccs on lhe planlar aspeci of the fOOL are explored. The 10calion of forces as they pass from the tibiofibular complex into the dome of the talus and then into the foot is disclissed. Vve also discuss the roles of ligaments and muscles in the support of the medial longitudinal arch. Finally, ankle motion and ligamentous stability is outlined. A discussion of sophislicated electromyographic activity during walking is not within lhe scope of this chapter; however, the activity of certain extrinsic and intrinsic muscles is by necessity presented to allow a beller understanding of foot and ankle control during gait. The moments produced about joints by muscle action and resultant effects on foot and ankle position are detailed. Joint axes and instant centers of joint motion are described. Refer Lo Chapler 18 for information aboul the application of biomechanics to gail. Any pathological change in root or ankle structure or motion, however subtle, may have a pro- found impact on the foot and ankle's shock-abs ing, propulsive, and stabiliZing roles. Clinical c lation of alterations in biomechanical Functio presented in case studies. Footwear in \Vestern ciety may VaJ'y from a rigid ski boot to a soft casin. These externally restrictive materials ma ter normal foot and ankle biomechanics ultimately innucncc the development or s pathological conditions, slich as hallux valgus. Growth of the Foot The foot is formcd when the limb buds develop ing the eighth week of gestation. Foot length width increases linearly from age 3 to 12 in and age 3 to 15 in boys at an average of 8 to 10 per year, followed by a plateau in growth (Che al.. 1997). Blais and associales (1956) showcd the foot appe«rs to be closer to the adult size a times during normal development of the child arc other parts or the limb. On average, at a )'ear in girls and 18 months in boys, the leng the foot is one half the length of the respective a foot (Fig. 9-3). This situation contrasts with th the femur and tibia, which do not attain their ture length unlil 3 years later in both boys and g The relativel.\' large size or the fool, then, is im tant for providing a broad base on which the ch body is supported. and Lhis base may at tilTlCS c pensate for the child's lack of muscle strength coordination. Kinematics of the Foot Gross motion of the foot is complex and oc m'ound three axes and on three planes (Fig. Flexion-extension occurs in Lhe sagittal p abduction-adduction occurs in Lhe horizonta transverse plane, and inversion-eversion occu the coronal or frontal plane. Supination pronation are terms commonly used to desc positioning of the plantar surface of the foot occur primarily at the subtalar (talocalcan joint. During supination the sale faces medi and during pronation the sole faces lnter Supination is a combination of inversion, flex and adduction. Pronation is a combinatio eversion, extension, and abduction (Fig. 9-5). motion includes Oexion, extension: adduction. abduclion. For practical purposes. foot motion can be sidered to be of two distinct types; non-we Base\ " \ 'KJ'"'R,.. Os trigonum Calcaneus Head Phalanges \ Sesamoids Tuberosity MEDIAL VIEW Trochlea Os Metatarsals trigonum Tuberosity Point or insenion 01 the carcaneolibular ligament calcan~-~~"""""'''~ -.;...·..><'·:hlea ('ubercle) / Groove lor . TuberOSIty peroneus longus Head TU!)fHc!e LATERAL VIEW Calcaneus Latera! process Talus -"'- ~~,. Medial malleolus lateral Cuneilorms { Middle Lateral malleolus Medial r"rsomelalarsal joint Tibiofibular articulation (of Lisfranc) HaUul( -Great 100- DORSAL VIEW Top, View of the medial aspect of the foot. Middle. View of the lateral aspect of the foot. Bottom left, Superior view of the foot. Bottom right, Anterior view of the ankle mortise, Med. La!. Interosseous membrane L Fibula t . . ~istal llbloflbular •t -~-'-......: .< z:f, ....z.. 0 0~, Ii I (~¥J~Tjbia f&t~~: ;:;if4~. ...-;&"4.~ ::~.,-:"~r-7: •.• Tibiotalar joint II ~ / :R'f1oJ-~- Calcaneus ft ~, ~ f I Ankle joint complex composed of the tibiotalar, fibulotalar. and distal tibiofibular joints. f beal"ing and weight-bearing. Passive. nOI1-wdgh bearing Illotion may be tested with the patien seated and the foot and ankle hanging free. Subtala motion is evaluated by grasping the tibia with on hand and inverting and cvcrting the heel with th other hane!. Abduction and adduction of the fore foot can be tested if the heel is held immobile Supination and pronation of the forefoot may als be tested with the heel fixed, as may flexion and ex tension of the tarsometatarsal joints and [Des. Active, weight-hearing 1TI00ion of the foot differ from passive motion because the forces produced b body weight and by muscle contraction act to stab lize the joints. Generally, functional active foot mo tion during gait tends to be less than passive foot mo tion. Active subtalar inversion can be demonstrate by viewing heel orientation from behind while askin the patient to raise lip on his 01' her toes. Extel'nal ro tation of the leg while bearing weight on the foo causes the hed to invert and the forefoot to pronate therefore raising rhe arch. Rotating the leg internall has rhe opposite effect: it lowers the arch. ~' F ~;: g i t f f I~ ;' ~ I i ~ F ~' LENGTH OF NORMAL FOOT Girls Boys 30 ,----,----,----,----,----,----,----,----,----, 28 +-+-+-f-+-+-+-+-+--I .1. !.! 26 +-+--+--+-+-+-++'-+--1 V 24\ · E 22 W-+-~~·~····· 30 ~ 28 I I I I I I I I.! . l I Iii ' I 26 +1-+1--h-t-:·,pv';l"/~····--4 I .. i<;V i I I·l/;{. § 18:/ § i lOX)· i 16 t-+--flfC'fJ+'-I-+--T.:-'-...L-j 18 \ all· I I 1,0 P.'cenmes 14l-l!f+-++-+--1·······..· · · · / - 1 +1 24+- 4+=1EE 22\ ./...' ::.:& 20 I ::.~ 20 0 50 0 I 16\ i/.'./I/!· 'I 14 +-fff!l+-+-+-+--l---'f---H I \ 10'O±jjt±±t::±j lot,±l=l:±jj=±jj ... ~59~5 .f. 12 o ·i 2 4 6 8 10 12 14 16 18 Age (years) %' • I 12 H'l-+-+-+-+--l-f---H 0 2 4 6 8 10 12 14 1618 Age (years) I. lengths of normal girls' and boys' feet derived from serial measures of 512 children aged 1 to 18 years. Left, Foot length versus age in girls. Note plateau in growth after age 12. Right, Foot length versus age in boys. Note plateau in growth after age 15. Adapted with permission from Blais. M.M.• Green. W T.. & Anderson. M. (1956). Lengths of the growing (oot. J Bone Joint Surg. 38A. 988. thlllst on the subwlar joill!. In the middle or phase and at push-ofr. the entire lower cxtrcll Dorsiflexionl planlarflexion Eversion! inversion ~i~lS t(~ reverse and ,:otatc L·.'(t:nlall~' as ..1he. ~ JOint Sll1lllltancollsl~' II1vcrts (Fig. 9-9). \\illh IIlV of the subtalar joint and supination 01" the fo foot is transformed into a rigid stntclurc cap propulsion. Olerud and Rosendahl (1987) and berg Cl al. (1989a-d) Im"e experiml'ntallv me lhe coupling 01 tibial rot~lliOIl to subtalar m They h<:\ve shown that lhe foot supimues I" fo 0.2 to 0.44" 01" libial external rotation. Causes of Leg Rotation During the Gait Cycl Malln (1993) has described the coordination o Foot motion occurs around three axes. ~~-------- and subtalnr rl'lotion in an ekg::HH model c mitered hingL" (Fig. 9-10). As the tibia interna lales. the subtalar joint evens (pronatcs). COl1 external rotation or the tibia causes in (supination) of the subtalar joint. He attribu internal rotation or th(: lower cx.t1\:l11it~: in stance to the obliquit.\' 01" a general ankk joi (see Ankle Joint l\'totion). According [0 the " axis model the ankle joint, the ankle joint angkd downward and postl.'riorly I'rorn Ilk·dia eral (Fig. 9-1 I). Betau:-ic of the obUquily of th or FOOT AND ANKLE MOTION DURING GAIT The gait cycle consists or a stance phase and a swing phase. The stance phase encompasses 62(1'0 of the gait cycle and the swing phase makes up the remaining 38%. The stance phase is subdivided into heel strike, foot flat, heel rise, push-orr. and toe-off. Swing phase is divided into acceleration, toe clearance, and deceleration phases (Fig. 9-6). The part or stance phase spent with both feet on the ground is termed double limb support and occurs through the flrst and last 121'10 of stance phase (Fig. 9-7). N0l111al men have an average gait velocity of 82 m/min and 58 heel strikes/min (Waters et aI., 1978). Running is defined as a gait speed pasl 201 m/min. At this speed, double stance disappears and a noat phase develops in which bOlh reet are orr or the ground (Fig. 9-8). During normal walking, the entire lower ex.tremity (including the pelvis, femUl~ and tibia) rotates internally through the first 15% of stance phase. From heel strike through foot flat the subtalar joint everts, tbe foot pronates. and the rorefoot becomes flexible to absorb shock and adapt to irregularities in the· ground 11001' surface. The subtalar joint everts in part because the point or contact or the heel is lateral to the center of the ankle joint, thus producing a valgus A, During foot supination the sole faces medially. ing foot pronation the sole faces laterally. • I I Heel stnke ~=K\~ 7TTcr 5l~ ll~l/~ 1~lf~1 I ! I Heel slnke I L 0°" I /1 : Fool flat I 15~o l~~l-p-"-Sh-O-lI~'-T-o-e-o-I-,-1 30~'o I 45~'o I 1 IAcceleration [Toe c1em""e I Decelemtion _~~~~ Stance phase Swing phase <l~--------_~ .~----_~ 60~o Double Double limb support limb Single limb support ~ <1-+ support ~ <1-+ I SWing Phase 70% Heel st II////II! Stance Phase ~I2J~]~~ I Toe ot! iH~e~ 100~,, l i IIBII .. 62 Percent of Walk Cycle '--~~~~~ _ Stance phase consists of two periods of double limb su port and one period of single limb support. 62% of the normal gait cycle is spent in stance phase and 38% is spent in swing phase. I &------------------ A Stance ~65' : o 10 Right heel Q RighI twel slrike ; If: 20 30 :' II 40 , 50 : j SWina(?5~o) 70 60 Mid I. osl"rt~k_e A AA A Ji Walking 80 90 100 Left heel ,-st-a.L:c-eT--T--,-,-F-O-'t-O-'-t'---S-1lr~ke • Mid slance + + ~ + 'I''I' Toe Double limb ~:~tl Mid Double limb Right heel off unsupported strike stance unsupported slnke Running Stance (40%) o 10 20 30 Swing (30'%) 40 50 60 70 80 90 100 { ~ftJz~~~~ I i IBDI'-i • 1 Comparison of walking and running cycles. In the running cycle, stance phase decreases, 9p float develops: _d_O_U_b_l_e_l_im_b_S_U_P_P_O_rl_d_iS_a_p_p_e_a_rs_,_a_n_d_a_d_o_U_b_l_e_l_im_b_U_n_s_u_p_p_o_r_te_d_o_r S_W_in__phase _h_a_se_in_c_r_e_a_se_S_' __ _ .. .. .. Toe-off Heel contact Heel contacl Ankle _. rotation i ! ' !.......-:-- Stance phase ,6~T .:t; :.... _: ,I FOOl immobile ....: Subia la, ~_.'-_..__ rolation :- B A f,' I ° ,._L_~_~.. -----,----;c;;~' m ~ I -----'_._~ W W I I 100 I Percent of Walking Cycle I I l Ankle motion and subtalar rotation during normal walking. Maximal subtalar eversion occurs at foot flat in early stance phase. Maximum subtalar inversion occurs at toe-off. • joint axis, the leg internally rotates with ankle dorsi~ flexion and externally rotates with ankle plal1larflexion. Additional mechanisms by which leg external rotation occurs during late stance include the swing of the opposite leg that causes external rotation of the planted leg and the obliquity of the metatarsal break (Fig. 9- I 2). The metatarsal break is an oblique axis of 50 to 70" with respect to the long axis of the [oat formed by the centers of rotation of the metatarsophalangeal joints. With push-off. the foot and lower extremity externally rotate with respect to the sagittal plane because of this oblique axis. c o Mitered hinge model of leg, ankle, and subtalar mo A. Outward rotation of the upper stick causes inwa tion of the lower stick. 8. Inward rotation of the up stick causes outward rotation of the lower stick. C. nal tibial rotation causes supination of the foot. O. tibial rotation causes pronation of the foot. Number of specimens MUSCLE ACTION DURING GAIT 20 Although the motions of the foot and ankle during the walking cycle OCClIr primarily as a result of the passive constraints of joints and ligaments. electrom.yography has shown that rnuscle activity does occur during normal gait (Fig. 9-13). At heel strike, the pretibial musculature fires eccentrically to slow down the descent of the forefoot and prevent a [oat slap. At midstance, the calf musculature contracls to slow down the Fonvard movement of the body over the foot and prevent a crouch gait. The intrinsics also contract during midstance to toe-off to aid in rigidity of the forefoot. Toe-off is primarily a passive event. The pretibial musculature again contracts during swing phase to ensure (hat the fOOl clears the Ooor during swing~lhrough. 15 r r I 10' { 74 78 82 86 90 94 Angle (degrees) Variations in the angle between the midline of the and the empirical axis of the ankle. The axis is angle obliquely and inferoJaterally 82°. The histogram sho variability among specimens. Metatarsal break version and eversion. The sublalar joint is respo sible along with the transverse tarsal joint (consis ing of the talonavicular and calcaneocuboid joint for transforming tibial rotation into forefo supination and pronation. Because the ankle Joi is to some degree a single-axis joint. subtalar m tion reduces the rOLalol y ' stresses to the ankle join Congenitally blocked motion of the subtalar joi may result in the formation of a ball-aod-sock ankle as a result of the increased rota1011' stress the joint. Manter (1941) determined Ihe subtal axis of rotation to be oriented upward at an ang of 42" fmm the horizontal and medially 16" fro the midline (Fig. 9-14). The sublalar facets resem ble segments of a "spiral of Archimedes," a righ handed screw in the right foot, so that the calc neus actually translates anteriorly along t subralar axis as it rotates clockwise during the m tion of subtalar inversion (Fig. 9-15). A\·erage su talar motion is 20 to 30° inversion and 5 to 1 eversion. Functional subtalar joint motion duri gait is 10 to 15°. During the gait cycle, the he strikes the ground in slight inversion, followed rapid eversion to a maXimUI11 of 5 to 10° at 10% the gait cycle (Fig. 9-9) (Sarrafian, 1993a,b). Transverse Tarsal Joint Motion The transverse tarsal joint, Chopart's joint, consists the talonavicular joint and calcaneocuboid joi Manier (1941) described two axes of motion in t I I Muscle activity I $ ,;.: The muscles of the lower extremity are more aclive dudng nmning. The gluteus maximus and hamstrings arc active in midstance through toe-off and increase their activity 30 to 500/0 1O decelerate the stance phase limb. Dorsillexors of the foot and ankle are active in 700/0 of the running cycle. The intrinsics, plantar flexors, and peroneals are important stabilizers of the plantar surface and hindfoot during the foot nat phase (Aclelaar, 1986). I I Pretibial muscle Triceps (calf) I -.1.,Sw;n9....1 phase I 1 I phase I I .......INlVNm'fM/'lir -NjWl~---- .L--../MlWWJW'J,IIlWil,J----...L I 1 Foot I I '"""-1 li Jl Inilial MOTION OF THE TARSAL BONES I 1__ Slance floor contact Lift off Initial floor contact Subtalar Joint Motion The joint between the talus and calcaneus is lermed the subtalar joinl. Its complex motion in three planes produces the motions of supination and pronation, c1inicnll.y referred to as subtalar in- Schematic phasic activity of leg and foot muscles during normal gait. in it study enlailing eXlkTinh..'lllal sL'kClive fusio thl:sC joints. Sliblala,· joint arthrodesis redllcL'd ta ~l\·iclilar Illolion to 26% or its normal mol ion and duced calcaneocuboid mOlion lO 56~~ normal. Ca neocuboid arthrodesis n.:duccd sublala,· motio 92 % normal and talonavicular motion to 67':'/c mal. Selective fusion of the lalomwicuhu- joint the most profound effect on the remaining joillls ducing their relT'laining Illotion to only 1" each. B doin (1991) showed thal "xperimelllal slIbtala sion resulled in Significant reduction in talonavic joil1l contact and nlso reduced ankh.: joil1l contac A Tarsometatarsal and Intertarsal Motion B Subtalar joint axis. A, Sagittal plane (lateral view). The axis rises up at a 4r angle from the plantar surface. B, Trans- verse plane (top view). The axis is oriented 16° medial to the midline of the foot. Reprinted with permission from Manrer, J. T. (7947), Movements of the subia/af and transverse tarsal joints. Ani!! Rec. 80, 397. transverse tarsal joint: a longitudinal axis and an oblique axis. The longitudinal axis is oriented lSo upward from the horizontal and go medially frolll the longitudinal axis of the foot. Inversion and eversion occur about thc longitudinal axis (Fig. 9-16). The oblique axis is oriented 52° upward from the horizontal and 57(> anlcromcdially. Flexion and extension OCClIr primarily aboLit this axis (Fig. 9-17). Ouzonian and Shereff (1989) determined in vitro talonavicular motion to be 7° in nexion-extension and 17u in pronation-supination. Calcaneocuboid motion is 2{' flexionextension and 7° pronation-supination. The motions of the subtalar joint and transverse tarsal joint interrelate to produce either foot flexibility or rigidity. Elftman (1960) showed that the major axes of the calcaneocuboid joint and talonavicular joint are in parallel when the subtalar joint is everted, thus allowing motion of the transverse tarsal joint. As the subtalar joint inverts. the axes of these joints are convergent, thus locking the transverse tarsal joint and providing rigidity to the midfoot (Fig. 9-18). During the period of midstancc to toe-off, the foot therefore becomes a rigid lever through inversion of the subtalar joint and locking of the transverse tarsal joint. Astion (J 997) showed the close interplay between the subtalar, talonaviclllaJ~ and calcaneocuboid joints The joillls between lhe three clIneiforrns, cub and five melatarsals produce little motion. The terlarsal joints are c1oscl~· congruent and exh minimal gliding motion bct\\'cL'n OIlL: anolher. tarsometatarsal joints, known as Lisfranc's jo , ,~ /' / /.'\. ,, :, ,, , Comparison of the posterior calcaneal facet of the righ talar joint with a right-handed screw. The arrow repres the path of the body following the screw. The horizont plane in which motion is occurring is hh'; n' is a plane p pendicular to the axis of the screw; s is the helix angle o the screw, equal to s', which is obtained by dropping a pendicular (pp') tram the axis. As the calcaneus inverts, tates clockwise and translates forward along the axis. Reprinted '..vith permission from M<1fller, J. T. (1941). Moveme of thl? subta/ar and lransvers£' tarsal joints. Ana! Rec. SO, 391 Talonavicular axis A 9;t~'--~d ..-:--- Lat. Navicular Calcaneus B Normal Longitudinal axis of the transverse tarsal joint. Inversion and eversion occur about this axis. A. lateral view. B. Top view. Reprinted wirh permission from lv/anter. J. T. (194/). Anteroposterior view of the transverse tarsal joint of th right foot. The anterior articulations of the talar head a calcaneus are shown. The major axes of the talonavicul and calcaneocuboid joints are shown in the neutral pos tion (parallel) and with the heel in varus (convergent). Movements of the svbtalar dnd transverse tarsal joints. Anat Rec. 80, 397. Q~~---------------- are intrinsically stable because of their arch-like configuration best seen in cross-scClion. The sec- ond metatarsal base is recessed into the mid foot, forming a key-like configuration with the intermediate cuneiform (Fig. 9-19). A strong ligament Varus • known as the Lisfranc's ligament connects the s ond metatarsal base to the medial cuneiform. T motion of the first three metatarsocuneifo joints is minimal compared with the fOLlrth a fifth melatarsocuboid joints, Ouzonian and Sh efT determined the first metatarsal-med cuneiform motion to be 3.5 0 f1cxion-extension a 1.5 0 pronation-supination, while the fourth a fifth metatarsotarsocuboid joints \Vcre 9 to flexion-extension and 9 to 11° pronation-supin tion. An in vivo study or first mctatarsocuneifo motion found a mean sagittal motion of 4.4(' (Fr & Prieskorn, 1995), An in vilro sludy of fi metatarsocuneiform motion showed that on llOlo of lOa specimens demonstrated an,Y addu tion-abduction (vVanivenhaus & Prcttcrklieb A 1989), There has been recent conccrn that hyp mobility of the first metatarsocuneiform jo may lead to offloading of the first ray and sub bar. ! ------------------- Oblique axis of the transverse tarsal joint. Flexion and ex~ tension occur about this axis. A. lateral view. B, Top view. Reprinted wirh permission from Manter. 1. T. (1941). Movemenrs of the subta/r1r and transverse tarsal foints, Anat Ree. 80, 397. quent hallux valgus deformity (Klaue, Hans & l\IIasquelel, 1994), Mizcl (1993) described plantar first metatarsocuneiform ligament as major restraint to dorsal angulation ancl sllb quent dorsal displacement of the distal fi metatarsal head. Motion of the Hallux The hallux must accommodate a \vide range of m lion of the foot to perform a great variety of tas Dorsal Phalanx Lateral O' 40 70~ Plantar Med Llslranc's JOInt Dorsal Melatarsal Medial B A Plantar A, Top view of the tarsometatarsal joints, known as Lis· franc's joint. Note the recessed second metatarsal base. B, Cross-sectional view of lisfranc's joint seen on computed tomography scan. Note the arch-like structure. ! ! Late Contact distribution of the first metatarsophalangeal in 0<> (neutral), 40° of extension. and 70<> of extension Joint contact of the proximal phalanx. Bottom, Joint tact of the metatarsal head. With increasing extensio 1 .__i_o_in_'_'_U_rf_a_,_e_,_o_n_,a_'_ls_'h_i_f'_d_o_,_,a_1_IY_O_n_'_h_e_m_e_,_a_'d_,_,a_l Normal Weight Bearing Normal Weight Bearing 7 Medial Lateral collateral J;:~",,~'--1~rcollateral ligament - ligament Instant center A B A, Instant center and surface motion analysis of the metatarsophalangeal joint of the hallux in the sagittal plane. Each arrow denoting direction of displacement of the contact points corresponds to the similarly numbered instant center. Gliding takes place throughout most of the motion except at the limit of extension, which occurs at toe·off in the gait cycle and with squatting. At full extension, joint compression takes place. The range of motion of the hallux is indicated by the arc. B, Instant center analysis of the metatarsophalangeal joint of the hallux in the transverse plane during normal weight·bearing. Gliding (denoted by arrows) occurs at the joint surface even though the range of motion is small. One nced onl.\' consider the dorsif1exed positio the gr.:at toe or n baseball catcher crollching be home plate to appreciate the degree of motion p ble in the great lOe. The first metatarsophalan joint has a range of motion from 30" plantarllcxi 90" dorsiflexion with rL'slk'Cl to the long axis o first metatarsal shaft. The first metatarsal is inc 20" with respect to the noor; thercfore, range of lion or the hallux is 50" plantarllc:don to 70" d llexion referenced to the lloor surface. During (oc·off phase of nanna I walking, maximum dors ion of the nrst l1lt.:tatarsophalangeal joint is requ Anal~'sis of motion of the hallux in the sa plane reveals that inslallt centers of motioll fall within the center the metatarsal head, minimal scatter (Fig. 9-20). The surface motio the first metatarsophahll1gcal joint is chan\clC' as tangential sliding from maximum plantarlle to IT'lOdcrate dorsiflexion. with some joint com sion <:\t maximurn dorsillexion (Sammarco, Shc,dL Bcjahi, & Kummc,; 1986). Aim el aL (1 determined the first metatarsal head surface co area to be 0.38 l'm~ in the neutral position, w dec"eases to 0.04 em' in full dorsiflexion (Fig. 9 The metatarsal head surface contact area shi fts sail\' with full c!xtension and is associated with con~prcssion. This explains the characteristic nwtion of dorsal osteophytcs and limited dors ion or the proximal phalanx in cases or h rigidus (Fig. 9-12). or A 50-year-old woman who has been wearing shoes with a narrow Ioebox for <llmost 35 years. By compressing the forefoot medially and laterally, these abnormal forces can lead to a hallux valgus deformrltion. In this way, the proximal pha- lanx shifts laterally and pronates on the first metatarsal head. This abnormal position of the proximal phalanx decreases its ability to depress the metatarsal . lateral view of hallux rigidus. Note osteophytes on the dorsal aspect of the metatarsal head, which limit joint extension . ------------------ The great toe provides stability to the medial aspect of the foot through the \vindlass mechanism of the plantar aponeurosis (sec The Medial Longitudinal Arch). As the body passes ovcr the foot in toeoff. the proximal phalanx passes over the metatarsal head and depresses it. This has been conflrmcd by force plate analysis in the lutc sLancc phase, which shows thai pressure under the first metatarsal head increases in this phase of gait (Clark, 1980). In hallux valgus, the proximal phalanx shiFts laterally ancl pronates on the first metatarsal head. Thb abnOl-· Illal position of the proximal phalanx decreases its head during toe-off (Fig. 9-23). A view of the plantar surface of the foot of a patient with severe hallux valgus. Note calluses underlying the second and third metatarsal heads (transfer lesions), which indicate a transfer of plantar forces away from the first metatarsal head to the lesser metatarsal heads. The altered joint mechanics produced by hallux valgus is evident in analysis of instant centers of rotation, which demonstrate joint distraction and jamming where gliding normally occurs (Fig. 9-24). Abnormal Weighl Bearing (Bunion) ability to depress the metatarsal head during toeoff. Any clinical situation that affects the normal depression of the metatarsal head may t1'ansfer plantar forces laterally to the second and third metatarsal heads and rcsulL in the formation of painful plantar calluses called transfer lesions (Case Study 9-1 and Figs. 9-23 and 9-24). Motion of the Lesser Toes The lateral four loes arc analogous to the digits of the hane!. The lesser tocs have three phalanges. the motion of each being controlled by extrinsic muscles. which originate within the leg, ;;lnd intrinsic muscles, which originate within the fool. Normal motion of the mctatm·sophalangeal joint is approximatelv 90" eXlension to 50" flexion. The extrinsic and intrinsic muscles contribute to the toc extensor hood. which controls motion of the metatarsophalangeal and interphalangeal joints - ' ' ' - - - - - - - - - - - - - - - Ftoor Altered instant centers caused by the presence of a bunion, seen through analysis of motion of the metatarsophalangeal joint of the hallux in the sagittal plane. .I Each arrow denoting direction of displacement of the contact points corresponds to a similarly numbered in· stant center. The arc indicates the range of motion of the hallux, which is more limited than that in a normal foot. Extensor digilorum longus relaxed ! Extensor digitorum longus tendon Interosseous muscle I Extensor Sling hood Flexor digilorum longus tendon insertion site Extensor sling Extensor digitorum longus contracted Lumbrical tendon Deep transverse ligament Flexor tendon sheath 1 Flexor digitorum Sling traction brevis tendon Lateral view of the lesser toe illustrating the extensor Lateral diagram showing the action of the extensor s When the extensor digitorum longus contracts (botto the proxima! phalanx is lifted into extension through sagittal bands. hood with contributing muscles and ligaments. (Fig. 9-2S). The e'trinsics consist of the long toe nexors and extensors. The lumbricals and interossei are the main intrinsic contributors to the extensor hood. The intrinsics act to l1ex the metatarsophalangeal joints and extend the inter~ phalangeal joints (Fig. 9-26). The long toe cxten~ sors extend the metatarsophalangeal joint through the action of the sagittal bands by lifting the pmximal phalanges into e'tension (Fig. 9-27). The flexor digilorum brevis is rhe primary Oexor of the proximal interphalangeal joint. The ne,or digitorum longus is the primary flexor of the distal interphalangeal joint. Neurological conditions, such as diabetic neuropathy or Charcot-l\tlarie-Tooth disease, initially affect the intrinsiL: muscles or the fOOl and res an intrinsic minus condition. The extrinsiL' lllu overpower the intrinsics and a claw !C)C dcform produced with extension of the mt.:tatars langeal joint and flexion of the interphala joints (Fig. 9-28). Interossei Extensors I Extensors - I Flexor digitorum longus Flexor digitorum Lumbricals ~brev!s Interossei Lumbricals The intrinsic muscles (interossei and lumbricals) act to flex the metatarsophalangeal joint and extend the interphalangeal joints. I ------------- A claw toe deformity is produced by an imbalance o extrinsic and intrinsic muscles. Relative weakness of t terossei and lumbricals with overpull of the extrinsic extensors and flexors produces an intrinsic minus def mity of metatarsophalangeal extension and interpha langeal joint flexion. c W 2 W 2 GE'-----------The beam model of the longitudinal arch. The arch is a curved beam consisting of interconnecting joints and supporting plantar ligaments. Tensile forces are concentrated on the inferior beam surface; compressive forces are generated at the superior surface. I ---------------e THE MEDIAL LONGITUDINAL ARCH models e~ist to describe the medial longitudinal arch of the foot: the beam model and the truss model (San-anan, 1987). The beam model slates that the arch is a cun'cd beam made up of interconnecting joints whose structure is dependent on joint and ligamentous interconnections for stability. Tensile forces are produced on the inferior surface 01" the beam and compressive forces are concentrated on lhe superior surface of lhe beam (Fig. 9-29). The truss model Slales that the arch has a triangular structure wilh two slnlls connected at the base by a '[\1,10 tic rod. The SlI-uts arc under compression and tie rod is under tension (Fig. 9-30). BOlh mod h(:t\·c validity and can be demonstrated clinically. The structure analogous to the tic rod in truss model is the plantar fascia. The plantar fas originates on the mediallubcrosily of the calcane and spans the transverse tarsal, tarsometatars and metatarsophalangeal joints to insert on metatarsophalangeal plantar plates and collate ligaments as well as the hallucal sesamoids. Do nexion of the metatarsophalangeal joints pla traction on the plantar fascia and causes elevat of the arch through a mechanism known as "windlass effect" (Hicks, 1954) (Fig. 9-31). Dur toe-off in lhe gail cycle, lhe loes are dorsiflexed p sively as the body passes over the foot and the pl tar fascia lightens and aels to shorten the distan between the metatarsal heads and the heel, th elevating the arch. The traction on the plantar f cia also assists in inverting the calcaneus lhrou its altachment on the medial plantar aspect of calcaneus. The arch has both passive and active suppo I-luang et al. (1993) performed an in vitro study the loaded fOOl and foundlhat division of the pl tar fascia resulted in a 25°M decrease in arch st ness. They found the three most important sta contributors to arch stability in order of imp A o B A, Schematic of a truss. The far left wooden segment re resents the hind foot, the middle wooden segment repre sents the forefoot. and the far right wooden segment is the proximal phalanx. The rope is the plar"!tar fascia. B. Dorsiflexion of the proximal phalanx raises the arch through traction on the plantar fascia. tance were the plantar fascia, the long and shon platHar ligaments, and the spring ligament (cakaneonavicular ligament). The spring ligament forms a sling for the talar head, which prevents medial and plantar migration of the talar head, and therefore provides stalic arch support (Davis ct aI., 1996). Basmajian (l963) demonstrated e1ectromyographi-cally that the muscles of the calf did not contribute to support of the medial longitudinal arch when a load was applied to the leg of a seated individual; however, his experimental model did not simulate normal walking or running, when arch integrity is under grealCr challenge. Thord~l.I·­ son et al. (1995) performed a dynamic study of arch support by simulating stance phase of gait by applying proportional loads 10 tendons while the foot was loaded. [n this study, the plantar fascia contributed the most to arch stability through toe dorsinexion. The posterior tibialis contributed lhe most lO dynamic arch SUpPOrl. In a similar cadav~ eric study, Kitaoka ot al. (1997) demonstrated a 0.5 mm decrease in arch height and an angular change in lhe bones of the arch when tension on the posterior tibial tendon was released during simulated stance. A clinical study of 14 feet postplantar fasciotomy at greater than 4 years follow-up showed a decrease in arch height of 4.1 111m, thus supporting the truss model of arch stability (Daly et aI., 1992). Neuropathic joint changes or trauma may disrupt the bone and joint support of the arch, leading to . arch collapse and a resultant rocker bottom foot deformity (Fig. 9-32). This sequela of joint destruction lends credence to the beam model of arch stabilitv. A lateral radiograph of a rocker bottom foot, dem ing loss of the bony and Iigamento\ls support of th .-._._----_._-----_._-- or bod,v weight forward ~l1ullg the il:ods prC)f (F;ig. 9-33). Thc mOll1cnt produced by cach tendon unit call be pn:dicted b,\' their rclati to the ankle ~HH.l subtahir a,'\(.'s (Fig. 9-34). The soleus and gaslnK!lr.:-l1lius combine t the ,Achilles tendon, which insens onto the neus and is the strongest fk',,\or of the ankle. ; ematical Inodel has prcdicted peak Achilles forces 10 be 5.3 to 10 til1lt.·s bod.\' wt.·iglll durin ning (Burdett, 1982). Firing or the ankk' plan aI's during midstancc acts to slow tht.' fon\'a lion of the tibia on:r lIw fool. MUSCLE CONTROL OF THE FOOT Twelve of the thirteen extrinsic and nineteen intrinsic muscles control the foot and ankle. The plantaris muscle is an extrinsic muscle that generally has no contribution to muscle control or the foot or ankle. The extrinsic muscles are the strongest and most important in providing active control during gait. According to Fick's principle (1911), the strength of a muscle is proponional to its cross-sectional area. Accordingly, Silver et al. (1985) have weighed and measured muscle fiber length to determine the relative strengths of muscles acting on the foot and ankle (Table 9-1). The muscles of the leg fire in a pattern during normal gait to ensure an efficient lransrcr of mw;de force to the floor and smoolh progression of" The Relative Strength5 of MU5Cles Acti on the Foot and Ankle Plilntarflexor Strength Percentage Dorsiflexor Strength Percentage Soleus 29.9 TibIaliS anterior 5.6 Gastrocnemius 19.2 Extensor digitorum long Flexor halluc is longus 3.6 Extensor hallucis longus Flexor digitorum longus 1.8 Peronel.Js tertius 0.9 Inverters Everters Tibialis posterior 6.4 Peroneus longus 5.5 Peroneus brevis 2.6 ------------ Walking Gait 8lance phase Swing phase ~---------I~--~---' Muscle Heel slrike Foot flat Heel all Toe off I I I Tibialis anletior Ext. digitorum longus Ex!. hallux longus Gastrocnemius Tibialis posterior Flexor digitorum longus Flexor hallux longus Peroneus longus Peroneus brevis Abductor hallux Flex. hallux brevis Flex. digilorum brevis Abd. digit minimi Interossei Ext. digitorum brevis I I I I I I I I I I I I i I I I I I I I I I ! ! ! I , , 10% 20% 30% 40% 50~;' I 60% , 70~o 80~o 90% 100% Electromyography ollhe Foot During Walking Electromyography of the musculature of the foot and ankle during one normal gait cycle (heel strike to heel strike) . • The strongest extensor of the ankle is lhe tibialis anterior, which is 1110st active during stance phase from heel strike 10 foot nat. The ankle and toe extensors fire eccentrically to slow the descent of the foot and prevent foot slap. They also are necessary to allow fOOl clearance fTom lhe noor during the swi ng phase. The strongest inverter of the foot ancl ankle is the posterior tibialis muscle. The postcrior tibialis is a dynamic supporter of the medial longitudinal arch. Il flmctions to invert the subtalar joint during midand latc stance, thereby locking the tranSVel'se tal'Sal joint and ensuring rigidity of the foot during toe-all Loss of this muscle results in acquit'cd pes planus with Ilattcning of the arch, abduction of thc forefoot. and evcrsion of thc heel (Fig. 9-35). Patients with posterior tibialis tcndon dysfunction usually are unable to activcly invert their heel while attcn1pting a single toe rise. They have difficulty performing a sin· gle toe rise because of their inability to form a rigid platform on which to support their weight. The primary evertcrs of thc foot and ankle arc th peroneals. The peroneus longus inserts on the ba of the first metatarsal and Illcdial cuneiform an acls to depress the metatarsal head. Injury or para ysis of this muscle may allow elevation of the fir metatarsal head and decrease loads borne by th first metatarsals and can rcsult in the developme of a dorsal bunion. Thc peroneus brevis stabiliz the forefoot laterally by resisting inversion and w found by Hinlermann and associates (1994) 10 b the strongest everter of the foot. Loss or perone muscle strength can result in varus of the hindfo (Sammarco, 1995). The interosseus muscles are active during la stance and are thought to aid in stabilizing the for foot during toe-olT. An imbalance between the i Irinsics and extrinsics will lead to toe deformiti such as hammer toes, claw toes, or mallei toes. ,Both intrinsic and extrinsic muscles; mediate th positional control of the great toe. A cross·section the proximal phalanx shows the relative position Longitudinal axis Sublalar axis I • I. I I EHL I I EDL I I I I I I Ankle axis TP. . FDL. Loss of the medial longitudinal arch in an acquired a flatfoot secondary to posterior tibial tendon deficien '------------------ Subtalar and ankle axes in relationship to extrinsic muscles. EDL, extensor digitorum longus; fHL, extensor hallucis longus; FDL, flexor digitorum longus; FHL, flexor hallucis longus; PS, peroneus brevis; PL, peroneus longus; TA, tib· ialis anterior; Te, tendon calcaneus; TP, tibialis posterior. man absorbs 63.5 tons on each foot while wa Running one mile \\'ould produce 110 tons per that same 150-lb man (Mann, 1982). ivlanter measured the cOr'npl'cssive loads static loading in cadaveric feet lo determine th lribution or forces through the joinls of th (Fig. 9-37). The highest pan or the longit arch, the talonavicular and naviculocune Normal Tendon Position .......... Extensor hallucis long the nexors, extensors, abduclors, and adductors (Fig. 9-36). The tibial and fibular sesamoids lie within the toe tendons of the flexor hallucis brevis muscle, beneath the head of the first metatarsal. Similar to the patella, they increase the lever arm distance of the pull of the newr halluc is brevis muscle and enable greater flexion torque to be generated at the metatarsophalangeal joint. They also act to transfer loads from the ground to the first metatarsal head. Kin.etics of the Foot The magnitude of loads experienced by the foot is astounding. Peak vertical forces reach 120°10 body weight during walking, and they approach 275°10 during running. It is estimated that an average 150-lb ?-'""-"'--.. Extensor hallucis br Lateral Medial Abductor hatlucis Adductor hallucis Flexor hallucis ----{)-r~-.;:_cJ_-_ brevis Flexor hallu brevis I I l Flexor hallucis lonQus ~matiC cross·section of the proximal phalanx hallux showing normal positions of the various tend relation to the bone. The distribution of loads under the foot durin stance has been the subject of intense investigatio for the last half centul)'. Ini1iall:\', the concept of "transverse metatarsal arch" was promoted, i which loads were borne primarily by the heel, firs and fifth metatarsal, as if the foot were a tripod This concept was disputed by Morton (1935), wh thought the forefoot had six contact points tha shared equally in weight distribution, namely, th two sesamoids and the four lesser metatarsal heads Recent plantar pressure studies b.y Cavanaugh et a (1987) of subjects standing barefoot have deter mined that the distribution of load in the foot is a follows: heel 6W/o, midfoot 8(Ye, forefoot 28%-, an toes 4% (Fig. 9-38). Peak pressures under the hee are 2.6 times greater than forefoot pressures (Fig 9-39). Forefoot peak pressures occur under the sec ond rnetatarsal head (Fig. 9-40). Static foot radiographic measurements fail to pre dict 65(Jr) of the variance found among d~'namic pres sures measured in various subjects. Therefore, th dynanlics of gait exert the primary' inOuence o plantar pressure during walking (Cavanagh et a! 1997). Hutton et al. (1973) studied the progression o the center of pressure across the sole of the foot du ing gait (Fig. 9-41). During barefoot \valking, th center of pressure is initially! located in the centra heel and accelerates rapidly across the midfoot t reach the forefoot, where the velocity decrease Peak forefoot pressures are reached at 80% stanc phase and are centered uncleI' the second metatarsa 15 25 40 50 • -'---------4.5 i Compressive forces of the foot after a 60-1b load is applied to the talus. The majority of the force passes through the I talonavicular joint and into the first through third ~ __m_e_t_at_a_rs_a_'s_. _ joints bear the majority of the load through the tarsal joints. The medial column of the foot, consisting of the talus, navicular, cunei forms, ancl first through third metatarsals, bears the majority' of the load. The lateral column, made up of the calcaneocuboid joint and lateral two metatarsals, transmits the lesser load. I Mean regional weight distribution expressed as a percen age of total load carried by the foot in barefoot standing Over 60% of the weight is distributed in th~ rearfoot, 8% in the midfoot, and 28% in the forefoot. The toes have l tle involvement in the weight-bearing process. 132.6 At toc·oIT, the center 01" pressure is located unde hallux. The metatarsal heads arc in contact wit 1100r at least 50% of Slance phase. Soames (1985 termined that the highest peak pressure and gre foot-floor impulse during barefoot walking wa del' the third metatarsal head insLCad of the sec The distribution of plantar pressures changes 138.9 Lateral 53.4 £ t: shoewcar. Shocwcar reduces peak heel pressu producing a 1110re even distribution of pressure u the heel. With shoes, forefoot load distribution medially with maximum pressure under the firs second metatarsal heads. The pressures unde lOes also increase with shocwear (SoalTIes, 1985 The distribution of plantar pressure during ning has identified two t~!PCS of runners chara ized by their first point of contact with the gro rcarfoot strikers and mid foot strikers (Fig. 9 Rcm-root strikers make initial ground contact the posterior third of the shoe. The initial conta the midroot strikers is in the middle third o shoe. In both groups, first contact occurs along later,al border of the foot. Peak pressure does no fer between runner types. The cellter of pressu in the dista!·most 20 to 40°;(; of the shoc in both tact groups for I110St of contact timc, indicating time is spent on the forefoot (Canuulgh ('t al.. 1 27.8 51.8 . ,. Mean regional peak pressures during standing measured in kilopascals (kPa). The ratio of peak rearfcat to peak forefoot pressures is approximately 2.6:1. 80 x· Med 70 60 " 50 5 40 <l. ~ ~ "'"' Ii: x ~ Lat 30 20 10 0 1000;/" M 0% Lal Foot Width B A Metatarsal head pressure distribution during standing. A. A line (XX') drawn in the contour plot between the approximate locations of the first and fifth metatarsal heads. B. The distribution of pressure along the metatarsal head (XX') indicating maximum pressure under the second metatarsal head. ",'jiI. l ,!t .. 55% \50% 45o~,\40% 35'0 ,30% ,, ,, 25% , t, , ,: 20% 15% +, ,, t 10% : r, 5" , 10 ;2% I 1mIIIIL-.I _ The progression of the center of pressure along the sale of the foot during normal walking is expressed as a broken line. Each point on the sale corresponds to a percentage of the gait cycle. Note the rapid progression across the heel and midfoot to reach the forefoot. where most of stance phase is spent. It then progresses rapidly along the plantar aspect of the hallux. During walking and running. several forces arc acting between the foot and the ground: vertical force. fore and aft shear (anteroposterior shear), 111edial and lateral shem; and rotational torque (Fig. 9-43). The vertical ground reaction force exhibits a double peak following the initial heel strike spike, The first peak follows heel strike in early stance and the second peak occurs in IDle stance prior lo toe-olT. The fore and aft shear forces dcmonstt'atc initial braking by the foot as the foot places a fonvard shear force on the ground. followed by a backward shear on the ground as it pushes off in late Slance. Most of the medial-lateral shear is directed laterally because the body's center of gravity' is oriented medially over the rool. Medial (internal rotation) torque is generated early in stance as the tibia internally rotates and the root pronates, followed by laleral (external rotation) torque as the leg externally rotates and the foot supinates. lached as is evident by the somctirncs dramatic do sal foot swelling found during trauma or infection o the foot or ankle. The plantar skin is firmly attache to the underlying bones, joints, and tendon sheaths o the heel and forefoot by specialized extensions of th plantar fascia, This function of the plantar fascia is e sential for traction between the floor and the foot weight-bearing skeletal structures to occur. Durin extension of the metatarsophalangeal joints. thes plantar fascial ligaments restl;ct the movement o skin of the forefoot and pia mar metatarsal fat pa (Bojsen-Moller & Lamoreux, 1979). The heel pad is a highly specialized struclure de signed to absorb shock. The average heel pad area 23 cm~, For the average 70-kg man. the heel loadin pressure is 3.3 kg/cm~. which increases to 6 kg/em with running, At a repetition rate of 1,160 heel im pacts per mile. the cumulative effect of running impressive. These cumulative forces would no mally result in tissue necrosis in other pans or th body (Perry, 1983). The heel pad consists of comma shaped or U-shaped fat-filled columns arrayed ven cally. The septae arc reinforced internally with cla Mid Rear Fore A Midfoot strikers n=5 o (~r:J(_~~)t x- -. x '------1...:..:.. Rear -i.,-" ) Mid Fore B SOFT TISSUES OF THE FOOT The soft tissues of the foot arc modified to provide traction. cushioning. and protection to the underlying structures. The dorsal skin of the foot is loosely at- Two types of runners characterized by initial ground contact. A, Rearfoot strikers. H, Midfoot strikers. Percenlage of Cycle Force (0, ot bOdy wDi('hn o 15 30 60 HS FF HO TO I 100 60 Vertical - -/" I'\. \ 20 1\0. 0 -Fore t Shear AFT Medial 20 t 10 j Lateral Torque (Nm) t j Lateral 0 j Stlear Medial 20 0 ,- .... IA .... ..... 10 HS , \ ......... V I I FF HO I i TO 9 0 9 Ground reactive forces acting on the foot during the gait cycle. HS, heel strike; FF, foot flat; HO, heel-off; TO, toe-off. Reprinted ~vitfJ permission from Mann, R.A (7982). Biomechanics of running. In AAOS Symposium on the Foot and Leg in Running Sports (.0.0.30-44). St. Louis: C V Mosby Co. • tic transverse and diagonal fibers to produce a spiral honeycomb effect (Fig. 9-44). The multiple small closed cells arc arranged to most effectively absorb and dissipate force. \,Vith age, septal degeneration and fat atrophy occur, which predispose the calcaneus and foot to injury (Jahss et a!., 1992a,b). Ankle Joint Biomechanics KINEMATICS The ankle mortise forms a simple hinge consisting of the talus, medial malleolus, tibial pia fond, and lateral malleolus. The talus is shaped like a truncated cone, or frustunl, with the apex directed medially (tnman, 1976). The talus is 4.2 mm wider anteriorly than posteriorly (Sarrafian, 1993a single ankle joi.llt axis has been described as ing just distal to the mcdi,il malleolus and jus tal and anterior to the lateral malleolus (In 1976). This empirical "general" ankle axis ca estimated by palpating the tips of the ma (Fig. 9-45). The single ankle axis is angulated terolaterall y' in the transverse plane and infe erally in the coronal plane. Several authors disputed the theory of a single axis of ankle tion and have described multiple axes of moti the ankle moves from dorsiflexion to planta ion (Barnett & Napier, 1952; Hicks, 1953; H mann & Nigg, 1995; Lundberg et aI., 198 Sammarco et a!., 1973). Barnett and Napier ( describe a dorsiflexion axis inclined down Structure of a normal heel pad as seen on magnetic resonance imaging (MRI), A, Lateral view. Note vertically oriented fat-filled columns. B, Top view of the heel pad demonstrating the spiral structure of the septae, which separate the fat-filled cells . • and laterall.\' and a plantarflexion axis angled downward and medially (Fig. 9-46). The ankle joint axes for dorsillexion and plantarflcxion cliffeI' by' 20 to 3D" in the coronal plane but remain parallel in the transverse plane. A small amount of talnI' rotation occurs during ankle motion, which varies with axial load. Lundberg X' X ~ ,~ Y-~Y ~Z' Iam _ The empirical axis of the ankle joint estimated through palpation of the malleoli. The axis angles downward and posteriorly, moving from medial to lateral. Ankle joint axis variation. Top, In dorsiflexion (DF) the a of motion XX' is inclined downward and I~terally. Midd In neutral the axis of motion YY' is almost horizontal. B tom, In plantarflexion the axis ZZ' is inclined downward and medially. el al. (I 989a-<l) used slereophologrammell)' 10 measure talar rotation during motion of the weightbearing ankle in normal volunteers. The talus externally rotated 9° from neutral to 30° dorsiflexion. From 0 to 10" plantarflexion, the talus internally 1'0tatcd 1.4", followed by external rotation of 0.61> at 30" planlarllexion (Fig. 9-47). An in vilro slud~' of loaded ankles demonstrated 2.5<" external rotation in 25(0 dorsiflexion, and < I" internal rotation at 3Y' plantarllexion (Michelson & Helgemo, 1995.) c .2 x " S! 20 0 10 ~ ;;; "9 0:; 0:; " " c 0 ·x ~ ~ 10 B 6 4 ca> c 2 0 a: - -2 "0 -4 -6 "iii -B E -10 w rl--~i-~--~-~I--~i -~i -12 PF 30 20 10 0 10 20 30 DF x" Input foot position Horizontal rotation of the talus around the vertical axis at different positions of ankle dorsiflexion·plantarfJexion. Moving from plantarflexion to dorsiflexion, the talus ini· tially internally rotates slightly, then externally rotates markedly. &.$ I 10 20 60 Percentage of Cycle 100 100 _ Range of ankle joint motion in the sagittal plane durin level walking in one gait cycle. The shaded area indica variation among 60 subjects (age 20 to 65 years). Rr:pr '..viti! permiSSion from Stiluffer. P. N.. Chao. E- YS., S Brew5:e norma!. d,seas RL {1971i. Force (Jocl molion analySIS 01 the I • c1fl(/ pro5the;;c anNe JOIn!s Ci:~l Onhcp. 127. 189 ankle planlarJlc.'\ion is actual!." occurring dista the ankle itself. 'rhis midfoot motion e.'\plains apparent ability the foot to dorsillcx ~llld p tarnex following ankle fusion. It also explains abilil~' dancers and gymnasts to align the with the long a:"is of the leg during toe point. S marco and associates (1973) found ave non-weight-bearing ankle mol ion measured r ographically 10 be 2-1." dorsiOexiol1 and 24" p tarlk·xioll. The normal pallcrn or ankle motion has b studied extensively (Lalnon'aux, 1971; Murra aI., 1964; Staullel· et aI., 1977; Wright et aI., 19 At hed strike, the ankle is in sligh I plantarflc: Plantadlexion increases until foot llat. bu( (he tion rapidl~' reverses to dorsilk:xion during stance as the bod.\' p,ISSCS O\'el' the root. The mo thel1 returns to plant<:lrllexion at we-ofr. TIl(' a again dorsiflexcs in lhe middle or swing phase changl,.·s to slighl plalltarflexion at heel strike (F 9-9 and 9-48). Ankle molion during normal wal ;:lverages 10.2" dorsiflexion and 14.2" plantarflex with a total motion or 25". Maximum dorsifle occurs at 70(;'1(.. stance phase and nwximlllll p larJlexion occurs at loe-ofr (Slaufrer el aI., 1977 or or 12 ~.~ " 0 f- rDm'-- siflexion (extension). A wide range of normal mo- lion for Ihe ankle has been reported and depends on whether the Illotion is measured clinically with a goniOll1cter or whether it is measured radiographically. Goniomctric mCasurements yield a normal motion of 10 10 20" dorsiflexion and 40 10 55" plantarflexion. Lundberg el al. (1989a-d) found thal the joints of the midfoot contribute 10 to 41% of clinical plantarflexion from neutral to 30" planlarflcxioIl. Therefore. whal appears to be clinical 0:; 0 a": Ankle mol ion occurs primarily in the sagittal plane and is described as planlarflexion (flexion) and dor- ~ ;;; '0 6 I I C RANGE OF MOTION " ~ ·0 SURFACE JOINT MOTION Sammarco el al. (1973) pcrronncd analyses o slant centers or rOlation and surface velocitie both normal and diseased ankles. They round ,,~ -0... the instant centers of rouuion fell within the tali of normal ankles but that their positions changed ,,'ith ankle motion (Fig, 9-49), This confirms that the ankle axis of rotation does not remain constant with motion, Surface motion from full plantarllexion to full dorsiflexion was also determined. Beginning in full plantarncx-ion. the ankle joint ;howecl early distraction as dorsillexion began. Joint gliding then took place until full dorsillexion waS reached and jamming of the joint occurred. It is possible that distraction and jamming of the .Otihiotalar joint playa role in lubrication the joint. In arthritic ankles the direction of displace- or , , , \ \ \ I I \ I " 1 f f ,I \ ' I ,I ,," / , ' , I, I I' I I ,< " ,' \ I I 1 " J 5 " _.....--1-...... " ,' _4 I I Anterior f \ I \ 1 I - - _ .J Posterior '- ment of the conLact points showed no consistent palLcrn. The tibiotalar joint sllrfaces distracted in an unpredictable manner, and they jammed when the joint was in neutral position rather than at the end of dorsillexioll (Fig, 9-50), ANKLE JOINT STABILITY Stability or the talocrural JOIllt depends Oil both joint congruency and supporting ligamentous structures. The lateral ankle ligaments responsible for resistance to inversion and internal rotation are the anterior taloflblilar ligament, the calcancoflblilar ,, ,, ,, ,, , ,, ,, ,, ,, ,, Tibia ,, AnI. Post. Talus Instant center pathway for surface joint motion at the tibiotalar joint in a normal ankle from full plantarflexion to full dorsiflexion. All instant centers fall within the talus. The direction of displacement of the contact points shows distraction of the joint surfaces at the beginning of motion (points 1 and 2) and gliding thereafter (points 3 and 4). 1 Reprinred with permission from Sammarco, G.J., Burnstein, AH., & Frankel, VH. (1973). Biomechanics of rhe anklE': A kinemarie sWdy. OrttlOP Clin North Am, 4, 75. Instant center and surface velocity analysis in an arth ankle. The instant centers vary considerably. Joint co pression occurs early in motion and distraction occurs dorsiflexion (velocity 4). ligamenL. and the posterior talofibular ligam (Fig, 9-51), The superfiCial and deep deltoid ments arc responsible for resistance to eversion external rotation stress. The ligaments respons for maintaining stability between the distal fi and tibia are the syndesmotic ligaments. The desmotic ligaments consist of the anterior tibio lar ligarnent, the posterior tibiofibular ligament transverse tibiofibular ligament (also referred a deep portion of the posterior tibiofibular), and interosseous ligament (Fig. 9·52). The lateral ankle ligaments are the most c monly injured and thercroh.~ the 1110st frcquc studied. The anterior t41lofibular and calcancofib ligaments form a I OSlO angle with one another 9-53). They act synergistically to resist ankle in sion forces. The anterior talofibular ligament is del- greatest tension in plantarnexion and [he c neofibular ligament is under greatestlcnsion in a dorsillexion (Cawley & France, 199\; Inman, \ Nigg et aI., \990; Renstrom et aL,1988), The anle taloribular ligament therefore resists ankle inver in plantarOexion and the calcancofibular ligamen sists ankle inversion during ankle dorsiflexion. accessory functions of the anterior talofibular ment are resistance to anterior talar displacem from the.: m()J"tisc, c1inicall.\' refclTt:d lO as anlcrio drawel: and resistance LO int<:rmtl rotation of (h talus within the mortise (Fig. 9-54). The calcnn ofibular ligament spnns both the lateral ankle joi and lateral subtalar joint. thus conlributing lo sltt)l Jar joinl stabilitv (Stephens & S'lllll11arCO, 1992). Th posterior talo(lbular ligamenl is under greatest strai in ankle dorsiflexion and aCls lo limit posterior (al displacement within the monise.: as \\'ell as limit tal eXlernal rotation (Fig. 9-55) (Sarrafian, 1993a). I vitro tesling of unloaded ankles subjected to alllcrio drawer testing del1lonstrated that the anterio talofibular ligamcill was most important in pla tarnexion and that the calcancofibular and postel"i taloflbular ligaments were most important in ank dorsiflexion (l3l1ll1clI el aI., 1991 l. Groove lor ,,:-'- Tibialis Posterior - Tendo calcaneus _ Groove for FI. Halfucis Longus Bursa n/f.J Navicular bone j 1 Tibia-navicular fibres· ) Plantar calcanea-navicular .I Jig. and associated fibres Calcaneus I LMediallubercle of talus l Posterior tibio:,alar fibres Sustentaculum tall Tibia-calcanean fibres Top, Lateral side of the foot and ankle. Bottom, Medial side of the foot and ankle_ From Anderson, 1. (Ed.) (1978). Gram's Atlas of Anatomy. Baltimore, MD. Lippincott, Williams & Wilkins. 20 D Posterior View Anterior View "5 ro "E ~ ~ c "E 'u "c. Interosseous membrane 10 ~ "0 Anterior tibiofibular ligament ci z 0 %I!P--\t-lnterosseous ligament 80 Posterior tibiofibular ligament Components of the ankle syndesmosis. £ I 120 140 Degrees Inferior transverse (tibiofibular) ligament as I 100 Average angle between the cafcaneofibular and talofibu lar ligaments in the sagittal plane. The average angle is 105° with considerable variation from 70 to 1400 among measured subjects. < Sprain Injury o ATF A basketball player with an injury that results from a filII on a plantarflexed and inverted ankle position during a game (Case Study fig. 9·2·1). IR B func.tion of the anterior talofibular ligament (AIF). A. The ATF lim· its anterior shih (AS) of the talu~ or posterior shift (PS) of the tibia- fibula. B. The ATF limits internal rota lion OR} of the talus or external rotation (ER) of the fibula. PTF. posterior talofibular ligament; 0, deUoid ligament. Clinically, lite most commonly sprained ankle ligament is the anterior talofibular ligament, followed by the calcancofibular ligament. These injuries most commonly occur as a result of landing or falling on a plantarflcxcd and invcl·ted ankle (Case Study 9-2). During periods of ankle unloading, the ankle rests in , position of planlarflexion and inversion. rr the ground is met unexpectedly, lateral ligament injury occurs. Ataarian ct al. (1985) tested the strength of the ankle ligaments by loading cadaver ligaments to failure and found the strength of the various ligaments from weakest to strongest to be anterior ., Case ,.. Study Figure 9-2-1. An abnormally high road in conjunction with the loading :: rate produces the injury (Case Study fig. 9·2-2). The sprain inversion injury produced by high stress (load per unit of area) in the plantarflexion and inversion direction will most commonly aff~ the antero talofibular ligament (failure PTF load -139N). This produces lateral instability in ttle ankle joint an abnormal anterior talar displacement from the mortise. decreasing the resistance to internal rotation of the talus within the mortise. Abrupt rupture Microtears [ 139 N .. A I sprain InJury B ~-----­ Function of the posterior talofibular ligament (PH). A, The PTF limits posterior shift (PS) of the talus or anterior shift (AS) of the tibia·fibula. B, The PTF limits external rotation (ER) of the talus or internal rotation (lR) of the fibula, ATF, anterior talofibular ligament; D, deltoid ligament. f k..""....... ,_..~ :00""" .j. associa~e~ ~vilh "",.. ",,,,L _I!!""'l'-"""" -~ -...:.<.. Elongation of the Antero Talofibular Ligament (:ase Study Figure 9-2-2. _~...~~=. . . . . . . . . . . . . . ------,---~.~.~'-7.".7.=.' ..._ ... ,:,,:,.~.4,~,~~. TI14'" (Medial talar lill) LS 1.9 mm AS 5.6 mm IT 0 .... Transection of both superficial and deep components of the deltoid ligament resulted in an average 14 0 of valgus talar tilt (TT) among 24 cadaver specimens. lalofibular 139 N, posterior talofibular 261 N, calca, neofibular 346 N, and deltoid 714 N. Therefore, incidence of ankle ligamentous injury tends to match both mechanism of injury and ligamentous strength. The deltoid ligament acts to resist eversion. external rotation, and plantarOexion of the ankle joint (Fig. 9-56) (Harper, 1987; Kjaersgaard-Andersen et aI., 1989; Nigg et aI., 1990). It also resists lateral talar shift within the mortise when the mortise is widened by distal syndesmotic ligamentous injUl)1 or distal fibular fracture (Fig. 9-57) (Michelson, Clark, & Jinnah, 1990; Harper, 1987). Under physiological load. the ankle articular surface congruency takes on more imporlance (Cawley & France, 1991; Stiehl et aI., 1993; Stormont et aI., 1985). Stormont et al. found that in a loaded slate the ankle articular surfaces provided 30% of rotational stability and 100% of resistance to inversionl eversion. They hypothesized that during weightbeming. the ankle ligaments do not contribute to an· kle versional stability. although rotational instability may still OCClll: Cawley and France showed that the force to cause ankle inversion and eversion increased by 91 % and 80%, respectively, with loading. Stiehl et al. (1993) found lhal loading of Ihe ankle resulled in decreased range of motion (especially plantarncx.ion), decreased anteroposterior drawer. as well as increased stability against version and rotation. Cass LS 1.9 mm AS 5.6 mm IT 0 LS 3.8 mm AS 8mm IT 0 The lateral malleolus is excised to simulate a fibular ture or distal syndesmotic ligament injury. A lateral rected load is applied to the talus in an unloaded a modeL The talar lateral shift (LS). anterior shih (AS) valgus talar tilt (TI) are measured. S~etioning of the deltoid (DO) doubled the lateral talar shift from 1.9 mm. SD, superficial deltoid. • and Settles (1994) performed CT scans of loaded cadaveric ankles in an apparatlls that did not constrain rotation and demonstrated that talar lilt or an average 20 still occllITcd in loaded ankles aftcr sectioning both anterior talofibular and calcaneonbular ligaments. They did not feel thal the articular surfaces prevented inversion instability during ankle loading. Most studies agree that loading of the ankle results in increased stability as a result of articular surface con· grueney. especially in ankle dorsiflexion. Syndesmotic stability is dependent on the integrity or both malleoli. the syndesmotic ligaments, and the deltoid ligamentous complex. During ankle dorsinexion. there is approximately I rnm of mortise widening and 2" of external rotation of the fibula (Close, 1956). The normal distal fibular migration with loading is I rnm (\'\fang et aI., t 996). This distal fibular mignllion serves to deepen the ankle mortise for addcd bony stability (Scranton, McMaster, & Kelly, 1976). With disruption of the mortisc in an external rotation injury, the s~'n­ desmotic ligaments and deltoid ligaments are tOI'n, the distal fibula fmclures, and the talus displaces laterally. A study of cadaveric ankles by OlgivieHarris et al. (1994) deflI1ed the conlribution to resistance to lateral talar displacenlcnt by the syndesmotic ligaments to be 350/0 for the anterior tibiofibular ligament, 40% for the postcrior tibiofibular ligan1cnt, 22% for the interosseous ligarncnt, anel less than t 00/0 for the interosscous membrane. The deltoid ligament appears to be key in preventing lateml talar shift. Burns et al. (1993) found only minimaltalar shift in a loaded cadaveric ankle study with sectioning or the syndesmotic ligaments until the deltoid ligament was sectioned. Michelson and colleagues (1990) simulated 4 mm of lateral fibular displacement in a cadaver study and placed the ankle under a 100-lb load. Lateral lalar shift doubled from I to 2 mm following sectioning of the deltoid ligament. Pereira et al. (1996) simulatcd a laterally displaced fibular fracture of 4 mm and placed cadavcric ankles under a 500-N load in vm"· ious static positions of ankle dorsiflexion and plantarnexion. Cutting the deltoid ligament in this study did not result in significant lateral talar shift or alteration in joint contact area or pressure. They hypothesized that under static loading the talus moves to a position of maximum congruence within the mortisc rather than displacing laterally with the distal fibula. Most studies agree that the deltoid ligament and medial malleolus arc most important in resisting talar external rotation and latcral talar shift. 1 .:' /,' r ) 5ubtalar joint and ankle joilll inversion are ofl difficult to separate clinically. The calcaneoribular l ament provides stability to inversion and torsion stresses to both the ankle and sublalar joints. Stephe and Sarnmarco (1992) provided an inversion stre (0 cadaveric ankles and sequentially sectioned t anterior talofibular and calcnncofibular ligamen They found that lip to 500/0 of the inversion observ clinically was coming from the subwlar joint. T structures that contribute to stability of the subia joint afC the calcancofibular ligament, the cervic ligament. the interosseous ligament, the lateral ta calcaneal ligament, the ligament of ROtlvicrc, a the extensor retinaculum (Harpel: 1991) (Fig. 9-58 Kinetics of the Ankle Joint The reaction forces on the ankle joint dudng g arc equal to or greater than the hip and knee join respectively. The following static and dynam analyses give an estimate of the magnitude of the action forces acting on the ankle joint during stan ing, walking, and running. Ligaments of the lateral ankle and subtalar joints: 1, ant rior talofibular ligament; 2, posterior talofibular ligame 3. calcaneofibular ligament; 4, lateral talocalcaneal liga· ment; 5, fibulotalocalcanealligament, or ligament of Ro viere; 6, cervical ligament; 7, ligament of the anterior ca sule of the posterior talocalcaneal joint; '8, interosseous ligament, STATICS In a static analysis of the forces acting on the all klc joint, the magnitude of the force produced by contraction of the gastrocnemius and the soleus muscles through the Achilles tendon, and consequently the magnitude or the joint reaction force. w The Free-Body Diagram of the Foot A, On a free-boely diagram of tile fOOl, including the talus, the lines of clpplication for 1/1/ anel A are extended until they intersect (Intersection point). The line of appli can be calculated through LIse of a free-body dia- Hon for J (dotted line) is then delermlned by connecting gram. In the following example. the muscle force transmitted through the Achilles tendon and the its point of application. the tibiotalar contact point, wit the interseCllol1 point for VV and A (Calculation Box Fig. reaction force on the ankle joint are calculated for 1- 1). B, A tnangle of forces is constructed. Force A is 1 a subjecl slanding on liptoe on one leg. In Ihis example, the foot is considered a free-body with three main coplanar forces acting on it: the Ilmes body ,-,veight and force J is 2.1 times body weight (Calculation Box FIg. 9-1-2). ground reaction force (\V), the muscle force lhrough the Achilles tendon (A), and the joint reaClion force on the dome of the lalus (J) (Calculalion Box 9-1). The ground ," reaction force (equaling body weight) is applied under the forefoot and is directed upward venicallv. The Achilles force has an unknown magnitude but a known point of application (point of insertion on the calcaneus) and a known direction (along the Achilles tendon). The talar dome joint reactive force has a known point of application on the dome of the talus, but the Illagnitude and line of direction are unknown. The magnitude of A and J can be derived by dc:signating the forces on a free-body diagram and constructing a triangle of forces. Not surprisingly, these forces are found to be quite large. The joint reactive force is approximately 2.1 times body weight. and the Achilles tendon force reaches approximately 1.2 times body weight. The great force required for rising up on tiptoe explains why the patient with weak gastrocnemius and soleus muscles has difficulty performing the exercise 10 times in rapid succession. The magnitude of the ankle joint reaction force explains why a patient with degenerative arthritis of thc ankle has pain while rising on tiptoe. An in vilro sludv bv Wang el al. (J 996) found lhal lhe fibula Iransmits 17% of Ihe load in the lo\\'er extremity. \Vith the ankle positioned in varus or plantarncxion, the [-Ibular load decreased. During ankle valgus or dorsiflexion, fibular load transmission in~ creased. Cutting the distal syndesmotic ligaments decreased fibular load transmission and incrc:ased distal fibular migration. Cutting the interosseous membrane had no effect on fibular load transmission. The distal syndesmotic ligamcnts arc therefore importanl for pn;venting distal migration of the fibula and maintaining Ilbular load. Tibiotalar contact point t ForceW A Calculation Box Figure 9~1-1. ~a '\ :,\ \ ~ ,: : " Force J ~ Force A 1.2 W \ \ \ - 2.' W : B Force W ~ Force W = Ground reaction l Force A = Muscle force thro Achi11eslendon Force J = Joint reaction forc the dome of the la Calculation Box Figure 9-1-2. ANKLE LOAD DISTRIBUTION The ankle has a relatively large load-bearing surface area of II to 13 CIl1~, resulting in lower stresses across this joint than in the knee or hip (Greenwald. 1977). The load dislribution on lhe talus is determined by ankle position and ligamentous integrity. During weight-bearing, 77 to 90('/0 or the load is transmitted through the tibial plafoncl to the talar dome, with the remainder on the medial and lateral talar facets (Calhoun et aI., 1994). As lhe loaded an- A B c kle moves into inversion, the medial taJar facet is loaded morc. Ankle eversion increases the load on the lateral tabH facet. The centroid of contact area moves from posterior to anteriOl: frolll plantarOexion to dorsiOexion, and from medial to lateral during the motion frorn inversion to eversion. The total lalar contact was greatest and the <.werage high pressure was lowest in ankle dorsinexion (Calhoun et al.. 1994) (Fig. 9-59). TalaI' load distribution is also determined by ligamentous forces. Sectioning of the tibiocalcancal fascicle of the superficial deltoid ligament in a loaded cadaver model resulted in a 430~) decrease in talar contact area, a 30 fYo increase in peak pressures, and a 4 mm lateral shifl of the centroid (Eadl el aI., 1996), Dynamic studies of the ankle joint are needed to appreciate the forces that act on the normal ankJe during walking and running. Stauffer et al. (1977) used force plate, high-speed photography, radiogntphs, and free-body calculations to determine ankle joint compressive and shear forces. The main compressh'c force across the normal ankle during gait is produced -i·i.by contraction of the gastrocnemius and soleus mus"··,des. The pretibial musculature produces mild com,>;'pressive forces in early stance of <20% body weight. '/~ compressive Force of five times body weight was ';produced in late stance by contraction of the poste,'rial' calf musculature (Fig. 9-60A). The shear force ·~:.:"reached a maximum value of 0_8 times bodv weight "dm'ing heel-off (Fig. 9-608). Proctor and Pa~11 (1982) also measured ankle compressive forces during gait and found peak compressive forces of four times body weight. In COIHrast to work by Stauffer et al. (1977), they found substantial compressive forces equal to body weight produced by contraction of the anterior tibial muscle group. The pattern of ankle joint reactive force during gait c1iffers Wilh different walkin!! cadences (Fi!!. ,?-61). In a faster cadence, lhe paltern showed I":;' '--J- __ ."-_ .. / '-:C o E F Schematic demonstration of prints on pressure-sensitive film representing high pressure contact areas on the left talus. A, 490-N load in eversion; note lateral shift of talar contact area. B. 490-N load in neutral version. C. 490-N load in inversion; note medial shift of talar contact area. 0, 490-N load in 100 dorsiflexion; note anterior shift of ta lar contact area and an increase in contact area. E. 490-N load in 30° plantarftexion; note posterior shift of talar co tact area. F. 980-N load in neutral; note increase in talar contact area with increased load. • peak fOI-ces of three to five times body weight, on in early slance phase and the other in late stan phase. In the slower cadence, only onc peak force approximately" five times bod.y weight was reach during late slance phase (StaL\ffer et aI., 1977). Du ing running, localized ankle forces I11ay be as hig as 13 limes body weight (Burdett, 1982). Effects of Shoe wear on Foot/Ank.le Biomechanics \¥estern society places great importance on the a pearance of footwear, especially among wome vVomen's footwear is designed to make the foot a pear smaller and the leg appear longer by narrow ing the locbox and elevating the heel. A narro J~-"-'."" toebox compresses the forefoot medially' and laterally, thus contributing to the development of hallux valgus, hammer toes, and bunionettes. A study of 356 women by Frey ot al. (1993) found that 88% of women \\'ith foot pain wore shoes that \vere on average 1.2 em narn)\ver than their foot. Women who \vore shoes on average 0.5 em wider than the foot had no symptoms and less deformity. Shoes with elevated heels increase forefoot pressure compared with standing barefoot (Snow, \Villiams, & Holmes, 1992). A 1.9 cm heel increased forefoot pressure by 22%, a 5 em heel increased peak pressure by 57%. and an 8.3 CIll heel increased peak pressure b:v 76%. An elevated heel can cause pain "n" Ankle joint reaction force expressed in multiples of weight in a normal ankle during the stance phase o at two velocities. In the faster cadence, there were peaks of three to five times body weight, one in ea stance and one in late stance phase. In the slower c only one peak force of approximately five times bo weight was reached during late stance phase. Repr with permission from Stauffer, R.N., Chao, E.Y.S., & ster, R.L. (1977). Force and motion analysis of the n diseased and prosthetic ankle joints. (lin Orthop, 12 Normal subjects Preop. pIs. ,,_. Postop. pIS. 5 ~ 4 '0; ;:, 3 '8 £ 2 "'" o u. 100 A 20 40 60 under the metatarsal heads and mav also tribute to interdigital neuroma formation. tion of thc heel also Illa}' ovcr time res Achilles contracture, limited ankle dorsif and an altered gait. The amount 01" ankle joi tion in the gait cycle decreases as heel hei creases (Murray ct aI., 1970). 0.4 i" :E C> 0 'Q5 lL. ~ >. 0 D 0.2 0.2 £ '" " 0 u. ""..,...•"..,. 0 t;: « 0.4 0.6 0.8 '"<;;'" '0 'iD '" I Summarv 100 60 Percentage of Cycle (stance phase) B A, The compressive component of the ankle joint reaction force expressed in multiples of body weight during the stance phase of normal walking for five normal subjects and nine patients with joint disease before and after prosthetic ankle replacement. B, The fore-aft shear component produced in the ankle during the stance phase of walking for the same subjects. Reprinted with permission from Stauffer, R.N., Chao, E. Y5" & Brewster, R.L. (1977). Force and motion analysis of the normal, diseased and prosthetic ankle joints. (fin Orthop, 127, 789. 1 The I"oot alternates in form and functi t\veen shock-absorbing flexible platform and propulsive lever during different phases of t cvcle. 2 The ankle and subtalar joint act like a m hinge. Ankle dorsiflexion and tibial internal ro are associated with subtalar eversion (pron ankle plantarllcxion and tibial external rotati associated with subtalar inversion (supinatio 3 SlIbtalar motion is screw-like and infl the flexibility of the transverse tarsal joinl. lar inversion locks the transverse joint and the foot to become rigid; subtalar eversio locks the transverse tarsal joint and allow flexibility. "4 The Lisfranc's joint (tarsornctt\tarsal joints) is intrinsically stable and relatively immobile as a result of its arch-like configuration and the key-like structure of the second tarsometatarsal joint. 15 The deltoid ligament prevents ankle ever sion, extcTnal rotation, and lateral talar shift. It i kc~" in maintaining the integl-it,Y of the syndes mosis. ,~ The first metatarsophalangeal joint exhibits a wiele range of motion, with gliding throughout Illost or its range and jamming at full extension. 16 The fibula bears approximately one sixth of th force exerted through the lower extremity. <,§:,(' The medial longitudinal arch acts like both a beam and a truss. The arch is elevated through the windlass mechanism of the plantar fascia. The posterior tibial tendon provides c1ynarnic support to the arch. ? Foot muscle action during standing is relatively silent, but sequential firing of both extrinsic and intrinsic muscles is necessary to produce a normal gait patlcrn. The anterior tibial musculature fires during early stance to slow fOOL plantarOexion and prevent fOOl slap. The posterior cnlf musculature fires during mid- and late stance to control pro~ grcssion of the body over the foot. ~8 During barefoot standing, the heel bears 60% or the load and lhe forefoot bears 28(Yr!. Forefoot peak pressures occur under the second metatarsal head. <,9 During walking, the ccnter of pressure moves from the posterolateral heel rapidly across the midfoot to the forefoot with peak pressures under the second 01· third metatarsal head. At toe-ofr. the hallux bears the most pressure. ·10 The heel fat pad is specificallv designed to absorb shock dUI'ing heel strike. The plantar fascia attaches the skin of the hecl and forefoot to Ihe underlying bony and ligamcntous strUClures_ 1:1 The ankle joint has muhiple axes that change during motion. Minimal talar rotation occurs during dorsiflexion and plantarflexion. ',Tt: Ankle joint inslant centers of rotation fall within the talus during range of motion. In 1ll0VCrncnt from plantarnexion to dorsiflexion, joint surfaces first distract, then glide and eventually jam at the end of dorsiflexion. 13 Ankle joint stability is determined by joint congruency and ligamentous integrity. Ankle stability increases and depends marc on articular surfacc congnlency during weight-bearing. 14 The anterior talofibular Hnd calcancofibular ligaments synergistically provide stability againsl inversion during ankle motion. JZ The distal syndesmotic ligaments preven separation of the distal fibula and tibia and hel lransmit force through the distal fibula on wcight bearing. 18 Ankle joint centroid (cenler of pressure) pos tion changes wiLh ankle flexion-cxtension and inver sion-eversion. TalaI' surface contact is maximize and joint pressure is minimized in dorsiOexion. 19 The rorces acting on the ankle can rise to level exceeding five tin1es body weight during walkin and thirteen times body weight dudng running. 20 Narrow shoes and high heels can ad\'ersel~' af fect foot mechanics, leading to forefoot deformities heel pain, and Achilles contracture. REFERENCES Adt::l:wr. R.S. (1986). Tht:' pr;ll.'tical biomechanics of running :\111 J Sporls Jlcd. I -I. -197. Ahu. T.K .. KilJoka. I·I.B .. 1..110, Z.I)., el at. (1997). KincmaLic and Conl.acl c1lilracl(.'ristics of thL" first mCI:lI'.lrsoph langt';d joint. Foot AI/kit' 111/. 18. 170. Astion, D.J., Debnd, J.T., Otis, J.e., l.'1 <II. (1997). Motion o Ihe hindfool after simulated arthrodesis. J 8011t' Joil1 S/lI'g. i9A. 241. Ataarian, D.E .. t\1cCrac:kin. 1·1.1.. De\·ito. D.P., CI ill. (1985 Bionll.'chanical char:ICIl.'ristics of human ankll.' lig<lItll.'lll Foot .·lllklc', 6. 54. BaruNI. C.J. &. N<lpier, l.R. (1951). The axis of rol<llion at th ankle joinl in Illall. Its influl.'llCe upon [hi.' form of Ihl.' wlu :lIld the mobility of the fihula. J fhulf, 86, I. Basmaji<1Jl. J.V. &: Stedo, G. (1963). The role of flHI:>l:!l.'S <\rch supporL of Iht:: fool. J BOllI: Joint Sllf~. 45..\. 1184. Beaudoin, A.J., Fiort::, S.~·l., Knillse. \V.R .. eL al. (1991) Effe of isolated talocalcaneal fusion on conlact ill Ihe :Ink and talonavicular joints. Foor AI/Uf, 12, 19. Bbis, i\'1.~'1.. Green, \V.T., I$.: . \IHll.'rson, M. (1956). Ll.,tlgths the growing fool. J BOllI: JOiHl S/lrg. 38:\. 988. 13ojsen-j\lolIl.'l'. F. &: LllllOI'<..'II.\:. l.. (1979). SignifiC:lIlCc of fr< dorsifk'xion of 1!Jl.' IOt:'S in walking. ..tcUl 0,.11101' Sc:al/ll, 5 471. 6111111':U. C.. Thonl<ls, l\..A., Hakorson, T.L., l.'1 :11. (1991). Bi l1l('c1wnical ('\'31ualioll of the anlt..'rior drawer test: Ti contrihution of Ihe lah.'ral ankle ligamenLs. FOUl ..Il/kll'. I 389. Bunk,lt. R.G. (1982). Forces prl.'dicll..'d :tl Ihe :tllkle durin rulllling . .\led Sci Sporls EXn"c, Iol, 30S-310. Bllrll:'i. W.e. II. Prakash. K., ,\debar. R.. l.'t al. (1993). Tihiol "II' j()inl dynamics: Indit.::ltions for lhl.' syndl.'Slllot :'CI'<..'W-:\ l::ldaH'r study. FOOl .-I,,"-Ie, 14. 153-158. Calhoun. J.I-I .. Eng. :\1 .. li. E. ('1 ;d. (199·0.:\ compr... hL'nsh·c study of !1t"l..'S:iUI"I.' dislribulion in th~' :lnkk joint with inn:rsion and cn:rsion. FOOl AJlk/~ 1111. 15. 123-[33. Cass, J.R. &. s~,tlh.':-. H. (199-0. Auld ... in stahility: In vitro kinL'nw.tics in r...sponsl· (0 :lxi:llioad. FoOl ..\Ilk/c II/I. 15(3). 13~-I~O. C.l\·anagh, P.R .. Morng. E., BOU!lOIl, :\.J.I\-I.. L'I a!. (1997). Thl,,' rl'latiollship of SI;\lic foot :-tnll.:tUl"i· to dyn;ll1lic fOOl func· lioll. J BioJ/ll.!t:I/(lIIic_.;, 30, 243-250. Ca\"anagh. P.R .. Rodgi.:l's. rvl.\l.. &. lib()~hi. A. (1987). PressLln:.' distribulion undt.=r s,\"mptom . frcc !"L't:! during bardont standing. FOOl AI/kle. 7, 262. Cawli-.'\', P.W. &. Fmncc, E.P. (1991). BiolllC'(:h:lnio; of thc latera'i ligal1\('llls of lhl..' ankle: An e\"alll:llioll of thL' effects of <lxialload ;lIld single plane motion 011 Iig;UHI'IlI slr;lin p,lttcrns. FOOl ..\Ilkk 12. 92. Cheng. J.CY.. LL"ung. S.S.F.. lcung, .-\.K.l... et al. (1997). Ch'lllgt:.' of foot sizL' with wcighthe:lring. r\ sludy of 2829 childr~'n 3 to 18 YC,lr~ of age. Clill OnJwp. 3-12. 123. Clark. T.E. (19S0J. The Pn'Hllre lJi.";lrihl/licHl UI/{!cr fh .. FOOl Duril/g 8(/1'4001 H'alki,l,li. Uni\'l'rsity Park. f'.... nnsyh'ania: Pcnn Sl.:lli.' Uni\'crsity. Close. J . R. (1956), SOIllI' applications of thi.' fUIH:tion;:tl ,lnalomy of the <lnkk joint. J BOlle Jnillf Slfr~. 386, 761. Daly, P.J., Kitaob, I·I.B .....\: Chao. E.Y.S. (1992). Pbnl:ir f:ts· ciotomy for intnlct:tblc planlar fasciillS: Clinical !'csults and biomt.'dl:lnil.:~\1 c\'altmtion, FOOl AI/kIt'. 13. 188. D:t\'is, W.l-! .. Sobl:1, M .. Diarln, E.F. ..... t :tl. (1996). Gross. his . tological. and mil.:nwascubr analolll~' ;lIld bioll1cchanicd tesling of the spring ligamt.·111 complex. FOOl ..II/kIt' IlIl. 17. 95. Earll. :\1., \Vaync. J.. Brodrick. C. ct 411. (1996). Contribution of the ddlOid ligamellt to ankle joint cont,lct characteris· tics: r\ GH!<IH'l' study, foot AIlk/e Illf. Ii, 317, ElfllHan, H. (1960). The transverse tars.t! joint and its contro!. Clin OnllOp, 16. 41. rick, R. (191 I). J-/wulbuch dcr '-\/Hl{OJl/il , /fI1i! .\It'cJHlllik da Ct'.. b,ke, Jena Fischer. Frey. Thompson, F.. Smith. J .. ~'I al. (1993). Am~'1"ican Or· thopaedic Foot ;llld Ankle- Society wOlllen's ~hot: stirn:.... FOOf ..iI/kit:, 14. 7S-SI. Fritz, G.R. & Pricskorn, D. (1993). Fina IIk'lat'lrsocunciform motion: :\ radiographic and ~tatislical an~tlysis. FOOf Allklc llIl, /6, 117. Grecnwilld, S. (I 97i'). Unpllhlish~d d<ll;:t dtl'd in R.N. Stauffer, E.Y.S. Chao. ~\nd C. Brewster. F()l'c~' alld mOlion analysis of the lIorll\;d. diseased, and pr(lslh~'lic ;Inkle joint. C!i!1 Onhop. 127. [89. Harper. ;\1. (1987). Deltoid ligamcllt: An t:Xpl.'l'illlClltal cvalua . tion of funclion. FOOl Allkle, 8. 19. Harpcr. '\'1. (1991). The lateral ligamentous support of tht.' subtalar joint. FOOl Allklt', II, 334. Hil-ks, J,H. (1954). Thc mechanics of the foOl 11: The plantar aponeurosis and the arch, ) ..lllal, 88. ~5. Hinterm<tnn. B, Nigg. B,M. (1993). In vitro kinem.. tics of the .. xi .. lly loaded ankle cOJnple.'\ in r('sponsl' to dorsiflcx· ion and pl:IIH:ll'flexion. fOOl AI/kl.: 1'11. 16. 514, Huang, C.K .. Kitaoka, 1·1.8,. An. K,N .. et :11. (1993). Biom~'· chanical c\'aIUalion of longitudinal <In..-h sl~lhili[y. fool AIl- c.. '* k!c, I{ 333. Hutton, W.C" SCUlL J.R.R" 6: Stokes, J.:\.F. (1973) The llll'chanies of thl! fool. In L. Klentrlllan (Ed.). The FoOl tIIul Il.' Di_\orda\ (p. 41). O'\iord: BIa . : b'l'!1 li(H1S. Sd~·I1!i"il,.' P or dlt' AI/kit-. l1u Williams 6: \\·ilkin:!o. J~lhs:-- ..\1.1-1.. KUlllllh·r. F.. &. ,\lidldsoll. J.D. (199hJ. g~ltion into lill' bt p:lds of Ih...· sok- of th~' foo!: H~ sure s!lldi~·s. Fl)OI .. llIklt'. 13.22/. J;dlS~ . .\1.11 .. :\-1 ichd . . l.lI'l. J.D .. Desai. P.. r..'l :11. (1~91h). g:ltions into lhe bt p~lds of Illl' sole of I Ill' foot: :\ and histulogy, FOOl .. II/Ide. 13.133. Kil:\Oka. I'I.B .. Lun, 70.1'.. \,\: :\n. K.N. (1997), Efl'l· ... l oi p libia! I~'ndoll on th ... an.:h 01' the fOOL during sirnul:llcd hearing: Bionl~'ch:tnk,d ;uwh'sis. FOil! .. \nk/,' 1111. 18,4 l\..ial'r:-.ga:IHI-:\nd~·rsl.'n. P.. \Vdh ... lund. J.O .. I!('lmig. P {19~9J. Stahilizing dfl'Ct (If the tibiOl:ak:lll ...· al bsc thL' ddlOid on hindfoot joillt mO\'l'lll~'lltS: :\n ~·.·qwr ~tml.'-". FOl!! .. Iukll'. /(), 30. "-hiliI..'. K.. H:IIlSI.'Il, S.T.. c.:. \Iasqlll·kl, ".c. (1994 L C quanlitali\"l;:' aSSn;Sllll'lll of firsl larsolll.... till,lrs.11 m (Ill Ihe s,lgitl;d p1;Il\~' ;lnd il~ r.. . lation 10 h.dlu.'\ ,·:d formily. [-'OOf Allkll' 1111. 15( [J, 9-13. l.all1<lr~·all.\, L.\\·. (1971). Kin~'lIlatiL' ITh:;:lsurl'IlII,,'nb ~tlldy of hlllll:1I1 w'ilking. Bilil Prosthl'f Rl·.,. 10. I. Lundh...-rg, ..\ .. Goldil'. I .. K:ilin. B., ~'l :11. (19~9a). Kin~· (If IhL' :lnkkffolll ('(Illlpk-\: PI;llItarfll·,ion :Ind dl,rsi FOol .. InUt'. 9. 194-100. l.undbl'l"g . ...\ .. S'·l'ns:-.l.ln. O.K .. B.\"Itllld. C.. ~'I :11. ( Kinel1latics of I Ill' :lnkldfoOI cOlllplc-x,--P:lrt ~: Pr :Illd slIpin:llio!l. rool ..\I/h!('. 9. 248. l.undh(·r!! ...\ .. S\·.. ' llsfln. O.K .. N"'111('[1I. G.. (.[ :'d. (1%9i...' lllati(~ 01' Ihe :lllkk,/foOI ~·()Jllpk·.\--Parl 3: lnfllll'IK n:Il:llioll. roO! :llIkll'. V. 304. Lundherg, A., S\"('llS~lll. O.h .. N(:llli..'th, G., l'l :11. (1989£ axis of rotation of thc :llIkk .ioint. J BOlI l ' Joilll Sll lnmall. \'.T. (19761. Til. JO;IH_' 94-Y9, .\l~lnn, R...\. (1982>. Biollll·t.:h:1I1ics of rUllllill~. In plIsilllJl 011 Ihl' FO//l /llId I.l'~ ill R/IlIII;II~ Sp{/1'(S ..l,.\O (pp. SL Louis: C\· .\Iosh\" Co. '\!anll. R ..-\. lI97')). Bi~md:hJni\'·s of th ... rnO!. In ..\ ..\O (It" Orr!lolit·s: Bioll1n:lltIlJint! PrillciJJk,; tllld :l{lp/iclIt 257-266). Sl. Louis: C\', .\Io:!ob.\' Co. .\bnn. R..-\. (1993)' Biollk'l:h:lllics of thl' foot :Illd au SlIl";.!ay o{ llli: FpPI illld ..\/lk/t· (PP. 3-431. 51. L.ol \Iosby Co. :\lallkr. J.T. (19411. '\\()\"l'r1ll'IHS of thl' sllhwlar Hnd Ir;l lars;l! joinb..··1))111 Rl'(.'. 80. 397. Mit.:helsnn. J.D .. Clal'h'. 1-1../ .. &. Jinnah. R.I-!. (1990). f~'cl 01' loading on libiol:dar alignlllent in (,;,Id;\\'er fOOl .-\111:11.'. 10. 280. \lidl~·lson. J.D. 6.: Hl.'lgl'Il\C'. S.L. Jr. (1995). Kill~'lllati( ;l.\i:tll~· loaded :Inkk. rHiIl ..\l1klt' 1111, 16,577, :\tizl.:I. ,\I.S. (1993). Th~' role of the plallt~tr first me l.·ul\~'ir()rll1 ligallwnt in w('iglllhearing 011 11 Illl'tatarsai. fOOl .-\lIkle. I-l, 82. .\I()rlon, D,J. (l935L TlI(' /lumlill Fool. liS Emlmiol/. (1';:.'" 111111 FII//dio//al J)i...;on/as. ;'\('\\. York: COllllllhi:1 Sil.\· Pr...· ss . .\ Illrra.,·. .\-J.P., Drought, .-\.11 .. &. Kor.\'. R.C (1964). W paul'rlls in nnnll:ll Ill~·n. J BOllt' JOill1 51/1':-:, ,lti.-\, 33 ~\Iurray, ,\I.P., Knry, R.C, 6.: Sepie, S.B. (19;0). \V'llki terns normal woml'n . .-II'"h I'hys ,\fed Udwhil. 51 or Nigg, B.:'V1., Skorran, G., Frank, e.8., l't .d. (1990), Elongation and forces of ankle ligamellts in a ph:,siological range 01 motion. Fool ..\/ll:Ie. I J, 30. Okrud, C. &. Rosendahl. Y. (1987). Torsion-transmitting properties of the hindfoo!. Cfill On/wI', 2/4, 285. Ohdrie-Harris, D.1 .. Reed, S.c.. &. Hedman, T.P. (1994), Dis~rllplion of the ankk syndesmosis: Biomechanical stud~' of ligamentous reslrainb. /\rlhroscopy, 10, 558~560. Ouzoniall, T.1. & ShL'I'l'IT, \.Lf. (1989l. In \'tIro dClL'rrnination of midfool motion. 1"001 Allkte, !O, 140. Pereira, D.S .. Koral, K.J., Resnick, R.B .. ct al. (1996), libiotalar contact area and pressure distribution: Th(' effect of mortise widening and syndesmosis fixation. FOOl AI/kle 11Il, 17, 269. Perry, J. (1983), Anatomy and biomechanics or the hindfoot. Cfill OrtflOP, 177, 9. Proctor, P. &: Paul. J.P. (1982). Ankle joint biomechanics. J Bio!llcch, /5, 627. Renstrom, P., \Venz, \1., Inc<l\'o, S., et aL (1988). Strain on the lateral ligaments of the ankle. FOOl Ank/e, 9, 59. Sammarco, G.J. (1980). Biomechanics of the fool. In \'.1-1. Frankel &: \-1. Nordin (Eds.), Basic Biolllcchallics of Ihe AllisclIloskdCf(/! Systelll (2nd ed., pp. 193-2(9). Philadelphia: Lea &. Febiger. Sammarco, G.J. (1995). Peroneus longus tendon tears: Acute and chronic. Foot Ankh, lilt. /6(51,245-253. Sa tTl lTIan:o , G.J .. Burnstein, A.H., &. Frankel. V.I-I. (1973). Biomechanics of the ankle: A kinematic study. O/"lhop Cli/I Norlh Alii, -t, 75. Sarrafian, S.K. (1987), Functional characteristics of the foot and plantar aponeurosis under tibiotalar loading. FOOl .'111l:lc, 8, 4. Sal'l'afian, S.K. (1993a}. FUllctional anatomy of the foot and ankle. In An(ltomy or the Foo! ami Anl:lc (pp. 474-602l. Philadelphia: Lippincott. Sarnlfian, S.K. (1993b). Retaining systems and compartments. In A/UlIOIllY the Fool (lilt! Auk/e (pp. 137-149). Philadelphia: Lippincott. Scranton, P.E., ;V\c,\'laster, J.I-I., &. Kelly, E. (1976). Dynamic fibular function: A new concept. Clill OrfflOp, 118,76-81. or SherefT, ,\l.J., Ikjahi, F.J .. &. Kummer, F.J. (1986). Kinern or the first met~llarsophabngeal joint. .I BOlle Joint S 68..1, 392. Silver, R.I.., de la G~lrz~l, .I., l\: Rang, M. (1985). The my llluscle balance: i\ stud~ of 1"t.'lati\'e strengths ~lnd ex sions about the foot and ankle. J BOlle Joiu! Surg, 432. Snow, R.E., \Villiams, K.R., & Holmes, G.B. Jr (1992). Th reets of wearing high heeled shoes on pedal pressure women. Foo! Allkle, /3. 85-92. So~\nll's, R.\\'. (1985), Foot pressures during gait. J 13io Eng, 7, 120-126. Stauffer, R.N., Chao, E.'{S., &: Brewster, R.I.. (1977). F and !'ll(Jti(lll ~lnal.\'sis of till: n()1'mal, disl'<lsl:d and thetic ankle joints. Clin Orthop, /27, 189. Stephens, :\1.\1. & Sammarco, G.J. (1992). The stabili role of the lateral ligalllt.'nt complex around the ankle subtalar joints. FOOl AI/klc, 13, 130. Stiehl, J.B., Skrade, D.A., Needleman. R.I.., et al. ([ 993) fect of a.'\i~ll load and ankle position on ankle stabili Ort/lOp 7/"al/l1l(/, 7, 72-77. Stormont. D.\I., :\10rrey, B.F.. An, K.N., et al. (1985). St ity of the loaded ankle. Relation between articula straint and prinwr.\' and sec()[ldar~' restrainis . ..1111 J S Met!, 13, 295. Thordarson, D.B.. Schmotzer, H . Chon, .I .. et al. (1995) namic support of the human longitudinal arch. Clill lhop. 316, 165. Wang, O.W., Whiltle, .\1.. Cunningham . .I .. et al. (1996). F ~Hld its ligaments in load transmission and ankle joint bilit.\". Clill Ort/lOp. 330. 261. Wanin:nhaus, A. & Pn:Llcrklieber, \1. (1989), First sometatarsal joint: Anatomical and biomeehanical s FOOl Allkfe, 9, 153. \Vaters, R.I.., Hislop, I-U., Perry, .I., et al. (1978). Energc Application of the study and m,Hl~lg('mcnt of locom disabilities. Orlhop Clill North Am, 9, 351. \Vright. D.G" Desai, S.\1.. 6: Henderson, \V.H. (19M), A of the sublalar and ankle joilll comrle.'\ during the st phase of walking . .1 130111.' Joil/I Sur,!!" 46.'1, 361. Biomechanics of the lumbar Spine Margareta Nordin, Shira Schecter Weiner adapted fro Margareta Lin Introduction The Motion Segment: The Functional Unit of the Spine The Anterior Portion of the Motion Segment The Posterior Portion of the Motion Segment The Ligaments of the Spine Kinematics Segmental Motion of the Spine Range of lvIotion Surface Joint Motion Functional Motion of the Spine The lvIuscles Flexion and Extension Lateral Flexion and Rotation Pelvic Motion Kinetics Statics and Dynamics Statics Loading of the Spine Dunng Standing Comparative loads on the Lumbar Spine During Standing, Sittin and Reclining Static Loads on the Lumbar Spine During Lifting Dynamics Walking Exercises lvIechanical Stability of the Lumbar Spine Intra·Abdominal Pressure Trunk Muscle Co-Contraction External Stabilization Summary References Introduction The human spine is a complex structure whose principal functions are to protect the spinal cord and transfer loads from the head and tnmk to the pelvis. Each of the 24 vertebrae articulates with the adjacent ones to permit motion in three planes. The spine gains stability from the intervertebral discs and from the surrounding ligaments and muscles; the discs and ligaments provide intrinsic stability and the muscles provide extrinsie support, This chapter describes the basic characteristics of the various structures of the spine and the interaction of these structures during norrnal spine function. Kinematics and kinetics of the spine afC also examined. The discussion of kinematics covers both the thoracic and lumbar spine. but that of kinetics involves only' the lumbar spine because it is subjected to significantly' greater loads than is the rest of the spine and has received more attention clini· cally' and experimentally. The information in the chapter has been selected to provide an understanding of some fundamental aspects of lumbar spine biomechanics that can be put to practical use. 11 12 13 14 2 3 10 4 5 6 8 7 Posterior portion ~,_,-,~ Anterior portion Schematic representation of a motion segment in the lum bar spine (sagittal view), Anterior portion: 1, posterior lo gitudinal ligament; 2, anterior longitudinal ligament; 3, vertebral body; 4, cartilaginous end plate; 5, intervertebr disc; 6, intervertebral foramen with nerve root. Posterior portion: 7, ligamentum flavum; 8, spinous process; 9, inte vertebral joint formed by the superior and inferior facets (the capsular ligament is not shown); 10, supraspinous lig ment; 11, interspinous ligament; 12. transverse process (t intertransverse ligament is not shown); 13, arch; 14, verte bral canal (the spinal cord is not depicted). Molion segment The Motion Segment: The Functional Unit of the Spine I ~----I I 1 I!l Anteroposterior (A) and lateral (B) roentgenograms of the lumbar spine. One motion segment, the functional unit of the spine, is indicated. The functional unit of the spine, the motion segmen consists of two vertebrae and their intervening so tissues (Fig, 10- I). The anterior pOl,tion of the se ment is C0111posed of two superimposed intervert bral bodies, the intervertebral disc. and the longit dinal ligaments (Fig. 10-2). The c9ITcspondi vertebral arches, the intervertebral joints formed the facets, thc transverse and spinolls processes, an various ligaments make up the postcrior portion. T 25 Anterior Vertebral body Iliopsoas muscle Arch Spinal canal Transverse with cord process Interspinous Facet ligament Spinous process Erector spinae muscle Posterior The nucleus pulposus Iic's directl~' in the cente all discs e\cept tllOSL' in tilL' IUlllhar segments, wh it has a slightl~, posterior position. This inner m is surrounded b,\' ;:\ tough outer covering, the ann lus fihrosus, composed 01" fibr()cani];:I!~:e. The cri cross arrangcrnent 01" the coarse collagen fiber b dles within till' fibrocartilage all<)ws the annu fibrosus to withstand high lk'!l{ling and torsicJ loads (se(' Fig, II-II). Discs with annular tears d pia." increased [,(ltational 1T'IOlllents during k,ad compared with IHlndegenerated discs (Haughton aI., 2000). The end phlte, composed of hyaline ca lage, separ(a~s the disc from the vertebr<:11 bod~' (F 10-2). The disc composition is simihlr to that ticular cartilage, described in detail in Chapter 3 During dail.v activities, the disc is loaded in a co pIc.\: manner and is usually subjected to a combi tion ()I" comp]'ession, bending, and lo]·sioll. Fle.\:i extension, and lateral lle.\:ion or the spine produ rnainl~' tensile ,Ind compressi\'(' stresses ill the d \\'hereas rotation produces mainly shear stress. \\-'hen <:1 motion segment is tr<:lllsected vLTtica the llucleus pulposus 01" the disc protrudcs, indic ing that it is under pressure. !\iIcasurement (If the tracliscal prcssure in normal and slightl." degenera cadavcr lurnbar nuclei pulposi has shown an intr or Transverse section of a motion segment at the L4 level viewed by computed tomography. The vertebral body, arch, spinal canal with spinal cord, and transverse processes are clearly seen. The view is taken at a level that depicts only the tip of the spinous process, with the interspinous ligament visible between the spinous process and the facets of the intervertebral joints. Directly anterior to I .. the transverse processes and adjacent to the vertebral body are the iliopsoas muscles. Posterior to the vertebral body, the erector spinae muscles can be seen. _---------- arches and vertebral bodies form the vertebral canal, which protects the spinal cord (Fig. 10-3). THE ANTERIOR PORTION OF THE MOTION SEGMENT The vertebral bodies are designed to bear mainly compressive loads and the.\,' arc progressively larger caudally as the superimposed weight of the upper body increases. The vertebral bodies in the lurnbar region are thicker and wider than those in the thoracic and cervical regions; their greater size allows thenl to sustain the larger toads to which the lumbar spine is subjected. The intervertebral disc, \vhich bears and distributes loads and restrains excessive motion, is of great mechanical and functional importance. It is well suited for its dual role because of its location between the vertebrae and because of the unique composition of its inner and outer structures. The inner portion of the disc, the nucleus pulposus, is a gelatinous mass. Rich in hydrophilic (\vater-binding) glycosaminoglycans in the young adult, it diminishes in glycosaminoglycan content with age and becomes progressively less hydrated (Urban & McMullin, 1985). Distribution of stress in a cross·section of a lumbar disc der compressive loading. The compressive stress is highe in the nucleus pulposus, 1.5 times the externally applied load (F) per unit area. By contrast, the compressive stres on the annulus fibrosus is only approximately 0.5 times externally applied load. This part of the disc bears pred inantly tensile stress, which is four to five times greater than the externally applied load per unit area. Adapted ,vith permission from Nachemson, A. (!975j, Tmvards a berr understanding of back pain: A reviev'/ of the mechanicS of th lumbar disc RheuI"T1illol Rehabi!, j'l. 129 sic pressure in the unloaded disc of approximately \0 N per square centimeter (Nachemson, 1960). This intrinsic pressure, or pre-stress, in the disc results from forces exerted by the longitudinal ligaments and the ligamentum flavum. During loading of the spine. the nucleus pulposus acts hydrostatically (Nachemson. t 960), allowing a unifonn distribution of pressure throughout the disc; hence, the entire disc serves a hydrostatic function in the motion segment, acting as a clishion between the vertebral bodies to SLare energy and distribute loads. In a disc loaded in compression. the pressure is approximately 1.5 times the externallyay)plicd load per unit area. Because the nuclear material is only slightly compressible, a compressive load makes the disc bulge laterally; circumferential tensile stress is slIst"ined by the annular fibers. In the lumbar spine the tensile stress in the posteriol- part of the annulus fibrosus has been estimated to be four LO five times the applied axial compressive load (Galante, 1967; Nachemson, 1960, 1963) (Fig. 10-4). The tensile stl·ess in the annulus fibroslis in the thoracic spine less than that in the lurnbar spine because of diffe ences in disc geometry. The higher rario or disc d ameter to height in the thoracic discs reduces the ci cumferential stress in these discs (Kulak et aI., 1975 Degeneration of a disc reduces its protcoglyca content and thus its hydrophilic capacity (Fig. 10A-C). As the disc becomes less hydrated, its elasti ity and its ability to store energy and distribu loads gradually decrease; these changes make th disc(s) more vulnerable to stl-csses. THE POSTERIOR PORTION OF THE MOTION SEGMENT The posterior portion of the Illotion segment guid its movement. The t~'pe of motion possible at an IC\'e1 of the spine is determined b~" the oriental ion the facets of the inlervenebral joints to the tran verse and frontal planes. This orientation chang throughollt the spine. Human intervertebral disc composed of an inner gelatinous mass, the nucleus pulposus (NP), and a tough outer covering, the annulus fibrosus (AF). A, Normal young disc. The gelatino nucleus pulposus is 80 to 88% water content (reprinted with pe mission from Gower. WE. &Pedrini, V. 1969. J Bone Joint Surg, 51 1154). Age-related variations in protein-polysaccharides from human nucleus pulposus. annulus fibrosus, and costal canilag are easy to distingUish from the firmer annulus fjbrosus. B. No mal mid-age disc. The nucleus pulposus has lost water conten a normal degenerative process. The fibers on the posterior pa of the annulus have sustained excessive stress. C, Severely degenerated disc. The nucleus pulposus has become dehydrated and has lost its gel-like character. The boundary between the nucleus and the annulus is difficult to distinguish because the degree of hydration is now about the same in both structures ~!!II!JIIII!~~~~~~~~~~-----'---~~;~"_'.<"~ -~_.,__ ;- ... _~-<C-'t'" Except for the facets of the two uppermost cervical vertebrae (CI and C2), which arc parallel to the transverse plane, the facets of the cervical inten.1crtcbral joints are oriented at a 45° angle to the transverse plane and arc parallel to the frontal plane (Fig. 10-6A). This alignment of the joints of C3 to C7 allows flexion, extension, lateral flexion, and rotation. The facets of the thoracic joints arc oriented at a 60 angle to the transverse plane and at a 20° angle to the frontal plane (Fig. 10-68); this orientation allows lateral flexion, rotation, and some flexion and extension. In the lurnbar region. the facets arc oriented at right angles to the transverse plane and at a 45° angle to the frontal plane (Fig. 10·6C) (White & Panjabi. 1978). This alignment allows Oexion, extension, and lateral Oexion, but almost no rotation. The lumbosacral joints differ from the other lumbar intervertebral joints in that the oblique orientation of the facets allows appreciable rotation (Lumsden & Morris, 1968). The above-cited values for facet orientation are only approximations, as considerable variation is found within and among individuals. The facets guide movement of the Illotion segment and have a load-bearing function. Load-sharing between the facets and the disc varies with the position of the spine. The loads on the facets arc greatest (approximately 30% of the lotal load) when the spine is hyperextended (King et aI., 1975). Because the facets arc not the primary SUPPOI1 structure in extension, if total compromise of these joints occurs, an alternate path of loading is established. This path involves the transfer of axial loads to the annulus and anterior longitudinal ligament as a way of supporting the spine (Haher et aI., !994). High loading of the facels is also present during forward bending, coupled with rotation (EI-Bohy & King, 1986). The vertebral arches and intervertebral joints play an important role in resisting shear forces. This function is demonstrated by the fact that patients with deranged arches or defective joints (e.g., from spondylolysis and listhesis) are at increased risk for fOl'\vard displacement of tbe vertebral body (Adams & Hutton, 1983; Mi11er et aI., 1983) (Case Study 10·1). The transverse and spinolls processes seniC as sites of attachment for the spinal muscles, whose activity initiates spine motion and provides extrinsic stability. 0 THE LIGAMENTS OF THE SPINE ,,- The ligamentous structures surrounding the spine contribute to its intrinsic stability (Fig. 10-2). All spine ligaments except the ligamentum Ibvulll have ;;;: ~_Q- ' .. ; - a high collagen content, which limits tht.:ir ext bility during spine motion. The ligalllL'llllllll fla which connects two adjaCL'1l1 \'(,I"lcbral arches l (lldinall~·, is an exception. having a lal'gc percen of dastin. The clasticit.Y of (his ligamclll <.lIlows contract during c.'\h..:nsion of the spine and 10 gate during nc.'\ion. Even when the spine is in a tral position, the ligamentum flavul11 is uncleI' Slant tension as a result of its elastic prope Because it is located at a distance from the cent motion in the disc, il pre-stresses the disc; th along with the longillldinal ligaments, it create irllradiscal pressure and thus helps provide intlsupport to the spine (Nachemson &. EV<.H1s, Rolander, 1966). Research suggests that with de erative changes such as spond~"lolisthesis, tra spurs, and disc degeneration, which rna}' lead t stability, altered mechanical stress \vill increas load the ligamentum fluvul11 and callSt: hypertr (FlIkllyama el aI., 1995). The amount of strain on the various ligam differs with the type of motion of the spine, D llc.'\ion, the interspinous ligaments arc subject the greatest strain, followed by' the capsular ments and the ligamcnturn flavum. During e sion, the anterior longitudinal ligarnent bear greatest strain. During latcral fle:don, the contr eral lransverse ligament sustains the hig strains. followed by the ligament nuvulll and capsular ligaments. The capsular ligaments o facet joints bear the most strain during rot (Panjabi et aI., 1982). Kinematics AClive mol ion of the spine as in any joint is duced b.'/ the coordinated interaction of n and muscles. Agonistic n1uscles (prime mo initiate and carry out motion and antagon muscles control and modif." the motion, whil contraction or both groups stabilizes the s The range of motion differs at various levels o spine and depends on the orientation of the f of the intervertebraljoinls (Fig. 10-6). Motio tween two vertebrae is small and docs not o independently; all spine movements involve combined action of several motion segments. skeletal structures that influence motion o lrunk are the rib cage, which limits thoracic lion. and the pelvis, which augments trunk m ments by tilting. Cervical (C3-C7) A Thoracic B I ,, Lumbar I c Orientation of the facets to the transverse plane Orientation of the facets of the intervertebral joints (approximate values). Reprinted with permission from White, A.A. & Panjabi, M.N. (1978). Clinical Biomechanics of the Spine. Philadelphia: J,8. Lippincorr. A. In the lower cervical spine, the facets are oriented at a 4S~ angle to the transverse plane Orientation of the facets to the fronlal plane and are parallel to the frontal plane. B, The facets of the thoracic spine are oriented at a 60° angle to the transvers plane and at a 20° angle to the frontal plane. C, The facet of the lumbar spine are oriented at a 90° angle to the tran verse plane and at a 45° angle to the frontal plane. ~ ,'Sponaylolisthesis: Anterior Slippage of One Vertebra in Relation to the Vertebra Below It 'A 30-year·old male gymnast complains of severe back pain that lies betyveen the inferior and superior facets). This bilat· periods of strenuous training and the symptoms decrease eral defect leads to an (lnterior displacement of the vertebra ~ith rest or restriction of activity, After a careful examination lS onto S1. As the l5 vertebra begins 10 slip forvllard, the cen b'ya specialist and ter of gravity of the body is displaced anteriody. To compen- MRI films. a diagnosis was made of spondylolisthesis at the level L5-S1 (Case Study Fig, 10-1-1), a disease continuum. the abnormal forces placed on the inter Case Study Figure 10-1-1. vertebral disc leads to herniation into the neural foramina, producing moderate stenosis. of both l5-S I nerve rootS. Case Study Figure 10-1-2. SEGMENTAL MOTION OF THE SPINE The vertebrae have six degrees of freedom: rotation about and translation along a transverse, a sagittal. and a longitudinal axis. The motion produced during nexion. extension, lateral nexion, and axial rotation of the spine is a complex combined motion resulting fTom simultaneous rotation and translation. ~:" pan of the trunk is displaced backward. Because thi) is upper (Case Study Fig, 10-1-2), -' Physiological loads during repeated flexion-extension mo- ~,. sate, the lumbar spine above the lesion hyperextends and the with concurrent bilateral pars interarticularis defects of L5 tiOn, of the lumbar spine caused a fatigue fracture of the pars ,i; •t" interarticularis (aspect of the postenor arch of the vertebra '''''..' , with radiation to both legs. The pain is associated with Range of Motion Various investigations using autopsy material or radiographic measurements in vivo have shown di~ vergent values for the range of motion of individual motion segments, but there is agreement on the relative amount of motion at different leve the spine. Representative values from \Vhite Panjabi (1978) are presented in Figure 10-7 1 lc)\v a comparison of motion at various levels o thoracic and lumbar spine. (Representative v for motion in the cendcal spine are included comparison.) lnvestigations of the thoracic and lumbar s show that the range of l'lexion and extension i proximately 4° in each of the upper thonlcic mo segments, approximately 6° in the mid thoraci gion, and approximately 12° in the two lower racic segments. This range progressively incre in the lumbar Illotion segments. reaching a m mum of 20' al Ihe lumbosacral level. o I 10' ' 20 I 0' I 10 ' 20 I 0' I 10' ' OC- Cl Cl - 2 2 -3 3-4 4 -5 5 -6 20" I '-'47' 6-7 C7 - Tl Tl - 2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9 - 10 10 - 11 11 - 12 T12 - Ll L1 - 2 2-3 3-4 4-5 L5 - 51 Flexionextension Lateral flexion Rotation A composite of representative values for type and range of motion at different levels of the spine. Reprinted with permission from White, AA & Panjabi, M.N (978). Clinical Biomechanics of the Spine. Philadelphia: J.B. Lippincorr. Lateral flexion shows the greatest range in each of the lower thoracic segments, reaching 8 to go. In the upper thoracic segments, the range is uniformly 6°. Six degrees of lateral flexion is also found in each of all lumbar segments except the lumbosacral segment, \vhich demonstrates only 3° of 111otion. Rotation is greatest in the upper segments of the thoracic spine, \vhere the range is 9°, The range of rotation progressively decreases caudally, reaching 2° in the lower segments of the lumbar spine. It then increases to 5° in the lumbosacral segment. Surface Joint Motion i\''Iotion between the surfaces of two adjacent vertebrae during flexion-extension or lateral flexion may be analyzed b:v means of the instant center method of Reuleaux. The procedure is essentially the sanle as that described for the cc,·vical spine in Chapter 11 (see Figs. 11-18 and 11-19). The instantaneous center of flexion-extension and lateral flexion in motion segment of the lumbar spine lies within t disc under normal conditions (Fig. 10-8A) (Cosse et aI., 1971; Rolandel~ 1966). With abnormal conc tions such as pronounced disc degeneration, t instantaneous center pathway will be altered (F 10-8B) (Gertzbein et aI., 1985; Reichmann et aI., 197 FUNCTIONAL MOTION OF THE SPINE Because of its complexity,') the motion of a single nl tion segment is difficult to measure clinically. A proximate values for the normal functional range motion of the spine can be given. Variations amo individuals are large and show a Gaussian distrib lion in the three planes. The range of motion strongly age~dependent, decreasing ·b~v approx mately 30% from youth to old age, although \vith a ing, loss of range of motion is noted in flexion a lateral bending while axial rotation motion is mai I Normal G-J iJ A ~ rained with cvidcnc~ or increased coupled m (iV!cGill et aI., 1999). Differences have also noted between the sexes: men have greater m in flexion and extension whereas women arc mobile in later,,1 Oexion (Bie.-ing-Sorensen, Moll & Wright, 1971). Loss of range of motion lumbar and/or thoracic spine is compensat mainly by motion in the cervical spine and hi Moderate disc degeneration J ~ (. ' B ~ "-'",7 L5 THE MUSCLES Instant center pathway for a normal cadaver spine (A) and a cadaver spine with moderate disc degeneration (B). Instant centers were determined for 3~ intervals of motion from maximum extension to maximum flexion. In the normal spine, all instant <enters fell within a small area in the disc. In the degenerated spine. the centers were displaced. and hence the surface motion was abnormal. Reprinted wirh per· mission from Gerr£bein, S.D., er al. (1985). Centrode patterns and segmental instability in degenerative disc (Jisease. Spine, 10,257. The spinal muscles can be divided into flexo extensors. The main flexors arc the abdomina cles (the reClus abdominis muscles, the intern external oblique muscles. and the transver dominallllllsclc) and the psoas muscles. In g muscles anterior to the vertebral column act ors. The main extensors arc the ercCLOr spina cles. the multifidus muscles. and the intcnra sarii muscles auached to the posterior eleme general, the muscles posterior to the ve columns act as extensors (Fig. 10-9). The ex Anterior Rectus abdominis muscle Transversalis fascia Transversalis Internal oblique Vertebral body External oblique Lumbar fascia: Anterior Middle Posterior Quadratus lumborum muscle Erector spinae muscle Posterior An MRI transverse cross-section of the body at the l4 level of a normal adult human spine. The major trunk ffius<les are (R, right; l, left). By courtesy (rom Ali SfJeikzahed, PhD., Hospilal for Joint Diseases, Mr. Sinai, NYU Health, New York, NY, USA. muscles bridge between each vertebrae and motion segment as well as over several vertebrae and motion segments. \Vhen extensor muscles contract symmetrically, extension is produced. "'!hen right and left side flexors and extensor muscles contract asymmetrically, lateral bending or t\visting of the spine is produced (Andersson & Lavender, 1997). Flexion and Extension During unloaded flexion-extension range of motion, the first 50 to 60° of spine flexion occurs in the lumbar spine, rnainly in the lower motion segments (Carlsoo, 1961; Farfan, 1975). Tilting the pelvis forward allows for further flexion. During lifting and lowering a (oad, this rh)rthm occurs simultaneously, although a greater separation of these movements is noted during lifting than during lowering (Nelson et aL, 1995) The thoracic spine contributes little to Forward flexion of the entire spinal column because of the oblique orientation of the facets (Figs. 10~6 and 10-7), the nearly vertical orientation of the spinolls processes, and the limitation of motion imposed by the rib cage. Flexion is initiated by the abdominal muscles and the vertebral portion of the psoas muscle (Anders~ son & lavendel~ 1997; Basmajian & Deluca, 1985). The weight of the upper body produces further flexion, which is controlled by the gradually increasing activity of the erector spinae muscles as the forward-bending moment acting on the spine increases. The posterior hip muscles are active in con~ trolling the forward tilting of the pelvis as the spine is flexed (Carls!j", 1961). It has long been accepted that in full flexion, the erector spinae muscles become inactive once they are fully stretched. In this position, the fonvard bending moment was counU:Tacted passively by these muscles and by the posterior ligaments, \vhich are initially slack but become taut at this point because the spine has fully elongated (Farfan, 1975). This silencing of the erector spinae muscles is known as the flexionrelaxation phenomenon (Allen, 1948; Andersson & Lavender, 1997; Floyd & Silver, 1955; Morris et a!., 962). However, Andersson et al. (1996), using wire CICCII'()('"'' inserted in the trunk extensor muscles ~l1icl("cI by ultrasound or MRI, showed that in the flexed position the superficial erector spinae muscles relax, while the quadratus lumborum and lateral lumbar erector spinae muscles become '''''V<lIeu (Fig. 10-10). In forced flexion, the superfiextensor muscles become re~activated. From l Electromyography of the quadratus lumborum (QL) and erector spinae superficial (£5-5) and deep (£S-d) muscles Wire electrodes were inserted in QL and ES-d; surface e trodes were used for ES-s. Five positions (a-e) of trunk f ion are depicted. In full nonforced trunk flexion (e), the ES-s activity is silent; however, the ES-d and QL are very tive to counterbalance the trunk flexion movement. (Co tesy of Eva Andersson, MD Ph.D, Karofinska Institute. Stockholm, S~veden.) full flexion to upright positioning of the trunk, pelvis tilts backward and the spine then exten The sequence of muscular activity is reversed. T gluteus maximus comes into action early toget with the hamstrings and initiates extension by p terior rotation of the pelvis. The paraspinal musc then become activated and increase their activ until the movement is completed (Andersson Lavender~ 1997). Some studies have shown that the concentric ertion performed by the muscles involved in rais the trunk is greater than the eccentric exertion p formed by.' the muscles involved in lowering trunk (de Looze et a!., 1993; Friedebolcl, 19 Joseph, 1960). However, this finding has been c tradicted in several studies (Reid & Costigan, 19 Marras & Mirka, 1992). Creswell ancl Thortens (1994) support the finding that less electrom graphic (EMG) activity is noted during eccentric tivity, as in lowering, despite high levels of fo generated. The compressive load of the sp caused by the muscle exertion produced by low ing the trunk with a load or resistance can appro the spinal tolerance limits, putting the back greater risk for injury (Davis et aI., 1998). When the trunk is hyperextended from the right position, the extensor rnuscles are active d ing the initial phase. This initial burst of activity creases during funher extension, and the abdominal muscles bccorne active to control and modify the motion. In extreme or forced extension. extcnsor activity is again required (Floyd & Silver. 1955). joints function mainlv as shock nbsorbers a important in protecting the intervertebral (Wilder et al.. 1980). When loaded in vitro. the Sljoint exhibits dimensional movement with joint opening r ranging from 0.5 to 1.2 and sacrum an 0 Lateral Flexion and Rotation During lateral Oexion of the trunk, motion may predominate in either the thoracic or the lumbar spine. In the thoracic spine, the facet orientation allows for Imeral Oexion. but the rib cage rcstricts it (to vmying degrees in different people); in the lumbar spine. the wedge-shaped spaces between the intervertcbral joint surfaces show variations during this motion (Reichmann. 1971). The spinotransversal and transversospinal systems of the erector spinae muscles and the abdominal muscles arc active during lateral I1cxion; ipsilmeral COnlractions of these muscles initiate the motion and contralateral contractions modify it (Fig. 10-11) (Andersson.& Lavender. 1997). Significant axial rotation occurs at the thoracic and lurnbosacral levels but is limited at other le\'els of the lumbar spine, being restricted by the vertical orientation of the facets (Fig. 10-6C). In the thoracic region, rotation is consistently associated with lat· end flexion. During lhis coupled motion. which is most marked in the upper thoracic region, the vertebral bodies generally rotate toward the concavity of the lateral curvc of the spine (White. 1969). Coupling of rotation and lateral nexion also takes place in the lumbar spine. with the vertebral bodies rotating toward the convexity of the clIlve (Miles & Sullivan. 1961). During axial rotation. the back and abdominal muscles are active on both sides of the spine, as both ipsilateral and contralateral muscles cooperate to produce this movement. High coactivation has been measured for axial rotation (Lavender et al.. 1992; Pope et aI., 1986). Pelvic Motion Functional trunk movements not only involve a combined motion of different parts of the spine but also require the cooperation of the pelvis because pelvic Inotion is essential for increasing the range of functional motion of the trunk. The relationship betwecn pelvic movements and spinal motion is generally analyzed in terms of motion of the lumbosacral joints. the hip joints. Ol' both (Fig. 10-12). Load transfer From the spine to the pelvis occurs through the sacroiliac (SI) joint. Biomechanical analysis of the sacroiliac joints suggests that these Left 0 L3 0 L5 80 60 40 20 0 20 VFRA (pV) \( 1\ 40 60 80 \\ FIG. 10-11 ------ Example of eleetromyographic activity of the erect spinae muscles collected with surface electrodes du side-bending of the trunk. The figure illustrates tru bending to the right and muscle activity at the ll, lS level of the lumbar spine. Substantial contralate cle activity (left) of the erector spinae muscles is re when bending to the right to maintain equilibrium duced wirh permission from Andersson. G.BL Orrengr Nachemson. A. (1977). fIHradiscaf pressure. intra-abdom pressure and myoelectric back muscle dcrivily related to and foading. Clin OrtllOp, 129. 156, lIsed. This involves measuring the m.voelectric aCli it~,. of the tnlllk muscles and correlating this activit with calculc:tlcd values for muscle contractio forces. The values obtained correlate well with thos ,I, { ',. The pelvic ring with its linkage to the spine and the lower extremities. The antero-posterior view of these structures on film gives a hint of the irregular shape of the sacroiliac joint 5urlaces. but an oblique projection is required for an accurate view of the joints. , ,, , posterior rotation ranging from 0.3 to 0.6°; translation ranged from 0.5 to 0.9 mm (\Vang & Dumas, 1998). In vivo analysis of the 51 joint utilizing roentgen slereophotogrammetr~!shows joint rotation mean at 2.5 0 and translation mean at 0.7 mm with no differences between symptomatic and asymptomatic joints (S,uresson et aI" 1989), Muscle forces acting on the 51 joint have a stabiliZing effect. helping to attenuate the high stress pelvic loading (Dalstra & Huiskes. 1995). or obtained through inLradiscal pressure measuremen and can thereFore be lIsed to predict the loads on th spine (Andersson & Lavendcr, 1997; bnengren aI., 1981; Schuhz et aI., 1982). Another method is the lise of a mathcnHuic model for force estimation that allows the loads o the lumbar spine and (he contraction forces in th trunk muscles to be calculated for various phy'sic activities. The models are useful as predictors o load, for load sharing analysis under different con ditions, to simulate loads, and in spine prosthet and instrumentation design. The precision of th model depcnds on the assumption L1sed for the ca culations. Two categories of models currently use arc the EiV1G-c1l'iven model based on c!cctromyo graphic trunk mllscle rccordings and the morc tr ditional biomcchanical model based on trunk mo mcnts and [OI'ces (Chaffin & Anderson, 199 Lavender et al.. 1992; Marras & Granata, 199 Sheikhzadeh, [1997]. The eflixi pllre alld COl billed loading OIl the recruitment pattern ten s lecled Intlll< IIlllscles, Unpublished doctoral thesi New York University. New York). or or STATICS AND DYNAMICS In the following section, static loads on the lumb spine are examined for common postures such a standing and silting and also for Iif1ing. a commo activity involving external loads. [n the final sectio the dynamic loads on the lumbar spine during walk ing and common strengthening exercises for th back and abdominal muscles arc discllssed. Statics Kinetics Loads on the spine are produced primarily by body weight. muscle activity, pre-stress exerted by the ligaments, and externally applied loads. Simplified calculations of the loads at various levels of the spine Can be made with the usc of the [Tee-body technique for coplanar forces. Direct information regarding loads on the spinc.,(ll the level of individual intervertebral discs can bq/~btained by measuring the presSure within the d(scs both in vitro and in vivo. Because this mel~od is too complex for general ; application. a semidircclmcasuring method is often '."..... .' .•.. ' The spine can be consich.:red as a modified elast rod because of the nexibility of thc spinal colum the shock~absorbing behavior of the discs and ve tebrae. the stabilizing function of the longitudin ligaments, and the elasticity of the ligamen flavum. The two curvatures of the spine in the sagi tal plane-kyphosis and lordosis-also contribute the spring-like capacity of the spine and allow th vertebral column to withstand higher loads than if were straight. A study of the capaci.ty of cadav thoracolumbar spines devoid of muscles to resi vertical loads showed lhat the critical load (th point at which buckling OCCUlTed) was approx ',: ,. FIG. 10-13 The line of gravity for the trunk (solid line) is usually ventral to the transverse axis of motion of the spine and thus the spine is subjected to a constant forward-bending moment. cles, the abdominal muscles an.: often interm ,-lclive in maintaining tlte neutral upright p and stabilizing the trunk (Cholc\\'icki el aI., However, this activit~· is readily reduced by th mand to stand relaxed (Hodges &. Richardson The venchr;:ll portion of the psoas muscles invol\'ed ill producing postul'al s\\'a~' (Bas 1958; Nachemsoll, 1966). The level of aCl these muscles varies considerably among in als and depends to some extent on the shape spine, for eX<:\J11plc, on the magnilllde of h kyphosis and lordosis. The pelvis nlso plays a role in the muscle and resulting loads on the ~pine during s (Fig. 10-14). The base of the s<lcrum is inclin ward and downward. The angle of inclinat sacral angle, is approximatcl~1 30° to the tr~ plane dudng relaxed standing (Fig. 10-148). or the pelvis abollt the tnmSVL:rse axis betw hip joints changes the angle. \Vhell the p tilted backward, the sacr<ll angle decreases lumbar lurdosis flallens (Fig. 10-14..1). This ing afrects the thoracic spine. which extends to <'Idjust the ccnter of gravity or th<.' trunk the cncrg,\' expenditure, in tcrms or muscle e /' ....... )i ~ : (( \ ;,; ! \ i 'J .~ A LOADING OF THE SPINE DURING STANDING When a person stands. the postural Illuscles are constantly active. This activity is minimized when the body segments arc well aligned. During stand· ing, the line of gravity of the trunk usually passes ventral to the center of the fourth lumbar vertebral body (Asmussen & Klausen, 1962). Thus, it falls ventral to the transverse axis of motion of the spine and the motion segments are subjected to a forward-bending moment, which must be counterbalanced by ligament forces and erector spinae muscle forces (Fig. I 0-13). Any displacement of the line of gravity alters the magnitude and direction of the moment on the spine. For the body to return to equilibrium, the moment must be counteracted by increased muscle activity, which causes intet"mittent postural sway. In addition to the erector spinae mWi· ~ r'" !'. \~\, mate!y 20 to 40 N (Gregersen & Lucas, 1967; Lucas & Bresiel; 1961). The critical load is much higher in vivo and varies greatly among individuals. The ex· trinsic support provided by the trunk muscles helps stabilize and modify the loads on the spine in both dynamic and static situations. B j o -30 Effect of pelvic tilting on the indination of the ba sacrum to the transverse plane (sacral angle) durin right standing. A, Tilting the pelvis backward redu sacral angle and flattens the lumbar spine. D, Duri laxed standing, the sacral angle is approximately 3 Tilting the pelvis forward increases the sacral angl centuates the lumbar lordosis. • minimized. \Vhcn the pelvis is tilted ronvard, the angle increases, accentuating the lumbar 101'and the thoracic kyphosis (Fig. 10-14C). Forand backward tilting of the pelvis influences activit)! of the postural muscles by affecting the loads on the spine (Floyd & Silvel; 1955). Values of Intradiscal Pressure for Differe Positions and Exercises As a Percentage Relative to Relaxed Standing in One Subject (Chosen Arbitrarily As 100'1',) i Position/Maneuver vu~""u LOADS ON THE LUMBAR SPINE STANDING, SIDING, AND RECLINING position affects the magnitude of the loads on spine. As a result of in vivo intradiscal pressure studies conducted by! Nachemson 1975), it was shown that these loads are minimal well-supported reclining, remain low during re"",";u upright standing, and rise during sitting. A recent in vivo investigation of intervertebral disc pressure utilizing more sophisticated technology, based on only one subject, suggested that in relaxed, unsupported sitting, interdiscal pressure is less than in standing (Wilke et aI., 1999). Aclditional pertinent pressure measurements can be seen in Table 10-1. Sato et al. (1999) have verified Nachemson's (1975) findings, showing an increase in spinal load from 800 N in upright standing to 996 N in upright siuing. The relative loads on the spine during various body postures, as described by Nachemson and \Vilke, are presented in Figure 10-15. During relaxed upright standing, the load on the third and fourth lumbar disc is almost twice the weight of the body above the measured level (Nachemson & Elfstrom, 1970; Nachemson & MOiTis, 1964; Wilke et aI., 1999). Trunk flexion increases the load ancl the fOl"\vard-bending moment on the spine. During forward flexion, the annulus bulges ventrally (Klein et aI., 1983) and the central portion of the disc moves posteriorly (Krag et aI., 1987) (Fig. 10-16). iVlore than trunk extension, trunk flexion stresses the posterolateral area of the annulus fibrosus. The addition of twisting motion and accompanying torsional loads further increases the stresses on the disc (Andersson et al.,1977; Shirazi-Adl. 1994; Steffen et aI., 1998) (Case Study 10-2). The loads on the lumbar spine are lower more during supported sitting than during unsupported sitting. During supported sitting, the weight of the upper body is supported by the backrest, which reduces the muscle activity, relieving intradiscal pressure (Andersson et aI., 1974; Wilke et aI., 1999). Backward inclination of the backrest and the use of a lumbar support further reduce the loads. The use of a support in the thoracic region, ho\vevel~ pushes the thoracic spine and the trunk fOl\vard and n1akes the lumbar spine move toward kyphosis to remain in lying supine Percent ------20 Side-lYing lying prone lying prone. extended back, supporting elbows laughing heartily, lying laterally Sneezing, lying laterally Peaks by turning around 24 22 50 30 76 140-1 Relaxed standing 100 Standing, performing Valsalva maneuver 184 Standing, bent fOIVvard 220 Sitting relaxed. without back rest 92 110 166 86 Sitting actively straightening the back Sitting with maximum flexion Sitting bent forNard 'vvith thigh supporting the elbows Sitting slouched into the chair Standing up from the chair Walking barefoot Walking with tennis shoes Jogging with hard street shoes Jogging with tennis shoes Climbing stairs, one at a time Climbing stairs, two at a time Walking down stairs, one at a time Walking down stairs, two at a time lifting 20 kg, bent over with round back lifting 20 kg as taught in back school 20 kg close to the body Holding 20 kg, 60 em away from the chest Holding Pressure increase during the night rest (over a period of 7 hours) 106-1 106--1 70-19 70-17 100-1 60-24 76-12 60-18 460 340 220 360 20-48 Adapted '.vith permission from Wilke, H.J .. Nee!. P. Caimi, M.. et (1999). Ne\lv in vivo measurements of pressures in the interverteb disc in daily life. Spine, 2";, 755. -----------_._- ~ 500 0 Nachemson, 1975 "(;9.- 400 0 ro" ,~ '1" i~ ] 200 E 0 z ,~ " 1 100 A ~ ~~~~ .~ ~ ~, % ',Z ;) ~ ;, " ~ c ~ #~ ter working a 12-hour shift, when he lifted \-vhile twistin ~ an unusually large, yet lightweight box, During the iirst J.! week of pain, he visited his physician, who prescribed pa c!ll ,;I ,.i • 35-year-old male presents complaints of low back pain with radiation lO the posterior aspect of the le thigh, not past the knee. His pain staned 3 weeks ago, % 300 '" Nonspecific Low Back Pain ~ Wilke, 1999 '" iii £ 'C . ".~ ,S en c '5 c ~-----'-----'-'-'-- ~ , '1 medication. Currently, he is still in pain, panieufarly durin sitting or standing for loog periods. During a iollmv-up physician visi!, a careful examination showed the patien to be someV'ihat ovel\veight. ",vith weakness in his abdo ~ ina I and back muscles and poor flexibility in hiS ham'1 strings. psoas, and back muscles. Neurological tests were normal and diagnostic x-rays were normal as well, leadin to a diagnosis of nonspecific low back pain (Case Study Fig.10-Z-1). Data from two studies using intradiscal pressure measurements. The relative loads on the third and fourth lumbar discs measured in vivo in various body positions are compared with the load during upright standing. depicted as 100%. Adapred wirh permission {rom Nachemson. A. (1975). TOWMds a berrer unrJersfc1nding of back pain: A review of rile mechanics of the lumbiJr disc. RheumaroJ Rehabil, 14, /29 and {rom Wilke, H.J.. Nee!. P.. (aimi. M., ei al. (999). Ne~·'/ in vivo medSuremen!s 0; preSSlJreS in rhe inrerve(rebral disc in daily life. Spine, 24, 755. • ) Tension Compression , Case Study Figure 10-2·1. Combinations of different factors have resulted in th injury. From the biomechanical point of view, although the load lifted was considered light, the vastness of the The forward-bent position produces a bending moment on the lumbar spine. The moment is a product of the force produced by the weight of the upper body (W) and the lever arm of the force (lw). The forward inclination of the upper body subjects the disc to increased tensile and compressive stresses. The annulus bulges on the compressive side and the nucleus is shifted posteriorly. package and the resultant large lever arm (the distance from the center of gravity of the person to the package crea'ted a larger than expected load on the lumbar spine In addition, weakness in the abdominal and extensor muscles of the spine led to an additional mechanical dis advantage in stabilizing the lower back. Tight psoas and hamstring muscles place restrictions on the mObility of pelvis, stressing the range of motion jn the lumbar regio and affecting the normal loads and motions at this leve ·i· .,. -~ .-; conwct with the backrest. increasing the loads on the lumbar spine (Andersson et aI., 1974) (Fig. 10-17). Loads on the spine arc minimized when an incliassumes a supine position because the loads b)' the bod;!s weight arc eliminated (Fig, With the body supine and the knees ex>tenden, the pull of the vertebral ponion of the psoas muscle produces some loads on the lumbar spine. the hips and knees bent and supported, howthe !umb4\r lordosis straightens out as the psoas >'n'luscle relaxes and the loads decrease (Fig. 10-18). ! '"-~--.~. I ----,. B FiG."'10.18 STATIC LOADS ON THE LUMBAR SPINE DURING LIFTING The highest loads on the spine are generally pro!t~>;.,S7 duced by c:'\ternalloads, such as lifting a heavy object. Just how mllch load can be sustained by the spine before damage occurs continues to be investigated. Pioneering swdies by Eie (1966) of lumbar ii ~ fil II'I III ,1J i /j: !i))i r~1 '-'...J j ! ' A B I~Di~sc1P;re~ss~u~re=========~;I:'===1~ ~ ;, : Influence of backrest inclination and back support on loads of the lumbar spine. in terms of pressure in the third lumbar disc, during supported sitting. A, Backrest inclination is 90" and disc pressure is at a maximum. e, Addition of a lumbar support decreases the disc pressure. C, Backward inclination of the backrest is 110", but with no lumbar support it produces less disc pressure. D. Addition of a lumbar support with this degree of backrest inclination further decreases the pressure. E, Shifting the support to the thoracic region pushes the upper body forward, moving the lumbar spine toward kyphosis and increasing the disc pressure. Adapted with permission from Andersson, G.8.1.. Ortengren, R.. Nc1CiJemson, A.. et at. (1974). Lumbar disc pressure and myoelectric back muscle activity during sitting. 1. SHldies on cln experimental chair. Scand J Rehabil Med, 6. 104. A, When a person assumes a supine position with legs straight, the pull of the vertebral portion of the psoas muscle produces some loads on the lumbar spine. B, Whe the hips and knees are bent and supported, the psoas mu cle relaxes and the loads on the lumbar spine decrease. I • vertebral specimens from adult humans showe that the compressive load to vertebrae failur ranged from approximately 5,000 to 8,000 N. O the whole, values reponed subsequently by othe authors correspond to those of Eie, although va ues above 10,000 N and below 5,000 N have bee documented (Hullon & Adams, 1982). The appl cation of static bending-shearing moment on lum bar motion segments revealed that bending mo ment of 620 Nm and shear rnoment of 156 Nm were tolerated before conlplete disruption of mo tion segmel1l occurred. The flexion angle befor failure was recorded as 20° with 9 mm of horizon tal displacement bct\veen the two vertebrae (0$ valder et aL, 1990). 80th age and degree of disc de generation innuence the range preceding failure Although the vertebral body strength is relative t Ihe bone mass, with aging ~ the decline in bon strength is more pronounced than is the decline i bone mass (iVlosekildc, 1993). Eie (1966) and Ranu (1990) observed that durin compressive testing the fracture point was reache in the vertebral body, or end plate, before the inte vertebral disc sustained damage. This finding show that the bone is less capable of resisting compres sian than is an intact e1isc, During the testing, a yiel point was reached before the vertebra or end pla fractured. \Vhen the load was removed at this poin the vertebral body recovered but was morc suscep lible to damage when reloaded. Evidence exists that the spine may incur microdamage as a result of high loads in vivo. T. I-Jansson (1977, The hone mineral contenl (Int! biomec!tanica! properties lumbar vertebrae. All ill vitro study based on dllal p1W101/ absorptiometry. Unpublished thesis, University of Gothenburg, Sweden) observed microfractures in specimens from "noI'mal" human lumbar vertebrae and interpreted this microdamage to be fatigue fractures resulting from stresses and strains on the spine in vivo. In vitro examination confirmed the cxistance of microdamage near the end plated with compression loading (Hasegawa et aI., 1993). Lifting and carr.ying an object over a horizontal distance arc common situations wherein loads applied to the vertebral column may be so high as to damage the spine. Several factors influence the loads on the spine during these activities: (t) the position of the object relative to the center of Illotion in thc spine; (2) the sizc, shape, weight. and density of the object; (3) the degree of flexion or rotation of the spine; and (4) the rate of loading. Holding the object close to the body instead of away from it reduces the bending moment on the lumbar spine because the distance from the center of gravity of the object to the center of motion in the spine (the lever arm) is minimized. The shorter the lever arm is for the force produced by the weight of a given object. the lower the magnitude of the bending moment and thus the lower the loads on the lumbar spine (Andersson et al., 1976; Nachemson & Elfstriim, 1970; Nemeth, 1984; Wilke et aI., 1999) (Calculat ion Box 10-1). Even when identical and nonfatiguing repeated lifting tasks are perFormed, variability in lifting technique of the same subject has been shown in trunk kinematics, kinetics, and spinal load (Granata ct aL, 1999). ,",Vhen an individual repeatedly performs an identical lift, great variability is recorded, which indicates that the brain may have several motor strategies to do a task. It also indicates the sensitive responsiveness of the muscle system to subtle changes to maintain the performance despite fatigue. ,",Vhen a person holding an object bends forward, the force produced by the weight of the object plus that produced by the weight of the upper body create a bending moment on the disc, increasing the loads on the spine. This bending moment is greater than that produced when the person stands erect while holding the object (Calculation Box 10-2). Health professionals generally recommend that lifting be done with the knees bent and the back relatively straight to reduce the loads on the spinc. or Influence of the Size of the Object on the Loads on the Lumbar Spine The size of the object held influences the loads on the lum- bar spine. If objects of the same weight, shape, and density but of different sizes are held, the lever arm for the force produced by the weight of the object is longer for the larger object, and thus the bending moment on the lumba spine is greater (Calculation Box Fig. 10-1-1). In these t\VO situations (Calculation Box Figs. 10-1-1 ,md 10-1-2). the distance from the center of motion in the disc to the front of the abdomen is 20 em. In both cases, the object has a uniform density and weighs 20 kg. In the case of Calculation Box Figure 10-1-1, the \·vidth of the cubic object is 20 crn; in the case of Calculation Box Figure 10-1-2, the width is 40 ern. Thus, in Case 1 (Calculation Box Fig. 10-1-1), the forv·.wd-bendlng moment ,Kung on the 10'/'Jest lumbar disc is 60 Nrn, as the force of 200 N produced by the weight o the object acts \vith a lever arm (U of 30 cm (200 N ;< 0.3 m).ln Case 2 (Calculation Box Fig. 10-1-2). the forvvardbending moment is 80 Nrn, as the lever arm (L;.) is 40 ern (200 (\J x 0.4 rn).IConsidering 1Kg 10N.] 20 em 200 N Forward-bending moment 60 Nm Calculation Box Figure 10-1-1. Forward-bending moment 80 Nm Calculation Box Figure 10-1-2. I ~ :./.;~; :, !ji/! I ~:I~~;Cto~::haet~h~e:u~::r~;i~~on lii}i!, I ;:~ n~ ~O~; ,:;: dSehn~: ~1 i:b~:~~::;h~nBgO;~i~~r~s 10- ~: lifted, In Case 1 (Calculation Box fig. 10-2-1) (upright f I standing), the lever arm of the force produced by the.- f i' '/ bending moment of 60 Nm (200 N weight of the objecr (lp) is 30 em. creating a forward- r x 0.3 m), The forward-bending moment created by the upper body is ~', 9 Nm; the length of the lever arm (LW) is estimated to t',; body is 450 N. Thu5, the tolal forward-bending moment • in Case 1 is equal to 69 Nm (60 Nm + 9 N(m). b d In Case 2 (Calculation Box fig. 1O-2 -2J upper 0 y ~ 2 em, and the force produced by the weight of the upper J ~.' flexed forward), the lever arm of the force produced by f I the weight of the object (l..) is increased to 40 em, creat- Ii ing a fon,.vard·bending moment of 80 Nm (200 N x 0.4 m). furthermore, ,he force of 450 N produced by the weight of the upper body increases in importance as it acts with r a lever ann (L...) of 25 em, creating a forward-bending moment of 112.5 Nm (450 N x 0.25 m), Thus, the total I ff, ~: forward-bending moment in Case 2 is 192.5 Nm (1125 ~'M"" t I I " However, this recommcndalion is valid only if this technique is used correctly and optimally, with the load positioncd betwcen the feet and thcrcby reducinn the lever arm of the extcrnal load (van Dieen ct al~ 1999) (Calculafion Box 10-3), A literature review revealed no significant differcnce in spinal compression and shear computed forces betwcen SlOOp or squat lifting (van Dicenc ct at., 1999). However, it was suggestcd lhat loss or balance is more likely during squat lifting. which in turn m,,)' ,-,dd addilional stresses on the lumbar spine. In thc following examplc, thc frcc-body technique for coplanar forces will be used to make a simplified calculation of the static loads on the spine as an object is lifted (Calculation Boxes 10-4; and 10-4B). Calculations made in this way for one point in lime during lifting arc valuable for demonslrating how the lever arms of the forces produccd by the wcight of the uppcr body and by the weight of the obj;ct affect the loads i~,posed on the spine, The usc or the salllc calculations to compule the loads produced when an 80 kg objcct is lifted (representing a force of 800 N) yields an approximate load o 10,000 N on the dise, which is likely to exceed thc fracture point of the vertebra. Because athletes who lift wcights can easilv reach such calculated loads without sustaining fr~ctures, other factors, such as intra-abdominal p~-essurc (lAP), may be .involved in reducinrr thc loads on the spine in vivo (Krajcarsk et aI., 1999), Dynamics ~Lw <:r'=i~-A-=-::t=-=-i J 450 N Almost all mol ion in lhe body increases muscle rc cl'uitmcnt and the loads on the spine. This increase is modest during such aclivities as slow walking or casy twisting but becomes morc marked during variou's exercises and the coqlplexity or dynamic movement and dynamic loading (Nachclllson & Elfstrom, 1970), WALKING ,I 200 N I I i Tolallorwardbending moment = 192,5 Nm Talal fo"vardbending moment = 69 Nm 11 Calculation Box Figure 10-2-1. Calculation Box figure 10-2-2. $---------"-~---_. In a study of normal walking at four speeds, the compressive loads at the L3-L4 mOlion segmen rangcd from 0.2 to 2.5 limes body weight (Fig 10-19) (Cappozzo, 1984), Thc loads were maxima around loe-off and increased approximatel~1 lin early with walking speed. Muscle .action was mainly concentrated in the trunk extensors. Incli vidual walking trailS. parlicularly the amount o forward flcxion of the trunk, influenced thc loads The Technique Employed During Lifting Influences the Loads on the Lumbar Spine In the three situations shown in Calculation Box Figures total forward· bending moment of 151 Nm(f200 N 10-3-1, 10-3-2. and 10-3-3, an identical object weighing 20 kg is lifted. Case 1 (upper body flexed forward) (Caleula- + 1450 N x 0.18 m!). Case 3 (Cal(l,lIalion Box Fig. 10-3-3) shows that be tion Box Fig. 10-3-1) is identical to Case 2 in Calculation Box knees per se do not decrease the forward-bending mo x 10-2, v~here ttle total fONJarcl-bending moment is 192.5 Nm. If the object lifted is held out in front of the knees, the In Case 2 (Calculation Box Fig. 10- 3-2), lifting with the arm of the force produced by the weight of the objec knees bent and the back straight places the object closer to creases to 50 the trunk, decreasing the forwclrd-bending moments. The the weight of the upper body (L..:) increases to 25 em. ern. and [he lever arm of the ierce produ lever arms of the forces produced by the weight of the object tile total forwarcj-bending moment created is 212.5 N (U and Lipper body IL,) are shortened 10 35 and 18 em, re- 1{200 N x 0.5 mj ... 1450 N x 0.25m!). spectively. at this point in the lifting process. The result is a .';' 200 N 200 N Total forward·bending moment = 192.5 Nm Total forward-bending moment = 151 Nm Total forward-bending moment =: 212.5 Nm Calculation Box Figure 10-3·1. Calculation Box Figure 10-3-2. Calculation Box Figure 10-3-3. 0'-~~~~~~~~~~~- The greater this flexion, the larger the muscle forces and hcncc the compressive load. Callaghan et al. (1999) corroborated lhese findings and further showed that walking cadence affects lumbar loading, with increased anterior-posterior shear forces noted as speed increased. Limiting arm swing during walking resulled in increased com- ,- pressive jomt loading and EMG outpu creased lumbar spine motions. 1n conc cause of low tissue loaeling, walking is perhaps ideal therapeutic exercise ror low back pain (Callaghan et al" 1999) w tion to speed of walking can further spinal loads (Cheng ct al" 1998)_ Diagram Technique for Coplanar Forces. Calculation the Static Loads on the Spine As an Object Is Lifted loads on cl lumbar disc I,'iill be calculilled for on,! point in tim[' 't/hen a NUJ to be posit]>,'e and counterciochvis(' moments are considered to cer50n who weighs 70 kg lifts a 20 kg obj!'ct The spin£' is flexed ilpprW.i. negeltl'/e.i , Thus, 35 In this example, the three priflclpdl forces ,Kling on the lurnl)ilf Q allhe lurnbosilual le'lei ilfe: (1) HK force produced by \1',(' ','ieight 01 upper body (Wi. Ciilculat.:d to be ':50 N (approxin\ille!y 65",0 of the exerted by lhe IOtal body weigh!); (2) tile fOlce produced b~i the of the object (Pl, 200 N; "neJ (3) trw force produced by contraction the erector spinae muscles (El, which has a kno':m direction anel point of 0 L) \t ;.: 0.05 m) (450 N ;.; 0,25 m) " (200 N >: DAm) E ;.: 0,05 rn 112.5 Nrn -;- 80 0 1~m Solving this equation for E yields 3.850 N ap,al",,'''''' but an unknQl,W1 magnitude (Calculi1lion Box Fig. 1O·4A·l) a distance 0 :::: r,il {\,V :< L<..J· (P >:U - iE >.: from the center of motion Th!: to",1 compre~~!'/e force exe!led on the disc (Cl can now be calcu- in the spine, they crcate moments in the lurntBr spille, Two fOPNard-ber:(j· lawd trigonof1,etricaliy (Calculation Box Fig. 10·cIA-2). in the example, C is in,) moments (\iVl:,_ cl!ld PL,) are the products of (V{: and (Pi and the perpen- lhe sum of lhe compressive forCi':s "cLng O'lel the disc. '.vhich is inclined diculars from the instanl center of rOIMion to the lines of action of these to the triirlSVelSe pldfle These forces are Because these three forces <let at fc,rces (,heir lever Mms), The le'/er arm (L'J for (Pl is 0.<1 III and the !e'/er arm The compressive force produced by the weight of the upper body (L,,) for (W) is 0.25 m. f\ counterbalanCing moment (ELi is the product of (W:-: cos 35°). 2 The force producHJ by the '.veight of the object (Pl, which acts on the disc inclined 35" (P '< cos 35"') (V'/), \'ihich acts on the disc inclined 35'0 lEi and Its lever arm, The lever elm: (l) is 0,05 m, TIl(' fI1agnitude of (E) CiJll b~~ found tilrough the use of the equilibriufI1 equa,ion for moments, For ,he kJdy to be in moment equilibrium, the sum of the moments aCiing on the ;u:nbar spine must be zeo Th;;,: fOlce produced by ,he erector spinae muscles (E), 'i,hich acts ap- 3 (In ihis example. cloch'!ise morm,nts "re consid- proxirn,w.:ly at il right angle to the diSC inclination. The totdi compressive fOlce acting on the disc (C) has a known sense, point of ilppliciltion, dnd line of iKlion but an unknown magnitude. The n:dgnllude of C (an bt": ioundlflrough H\e use of the eq\Jiiibriurn 0quation for forces For the body to be in force eqUilibrium, the sum of the forces must b(, .-,qual to zero Force W (450 N) Thus, a :::: forces f (v') :-: cos Force C (magnitUde unknown) I \, L-! \, \ \ l' C '" 4382 N S 373 N \ ~P Force P (200 N) (';501'-1 ;.: cos ~ C Force S (magnitude unknown) 3S~) .c. cos 35") .. E C ° (200 N ;.: cos 35") -'. 3850 N ." C ;co 0 368.5 N .;. 163,8 N·· 3850 1'J Solving the equation for C yields 'U82 1'1 The shear component for the reaction force on the disc (S) is found in the same'i/ay {450 Calculation Box Figure 10-4A-2. Calculation Box Figure 10·4A~1. 35") ,. (P:· ~J '.,: sin 35") ;, {ZOO N :': sin 35") _. S ,'. ° 373 N ~r----------------------------------------- Free~Body Diagram Technique for Coplanar Forces. Calculation of the Static Loads on the Spine As an Object Is Lifted 8ecause C and S form" right angle (Calculation 'Box Fig 10·48·1). the Forco S j 85 35 35 Forco W • Force P / / vc .;. s;' {R'I R . ; ; Forco R Forco<;!/ . Pythagorean theorem can be used !O fin,J the towl reaction force on the disc (Rt Forco 4398N The direction of (R) is determined by means of a trigonometric function: / sin h) '" CJR FOTeo E (tt) '" archsin (ClR) ;co 85" where (t,) is the angle belween lhe total vecto force on the disc and the disc inc!ina,ion The problem can be graphically solved by COlf5trUCling a v0ctor diagram b"sed on the knQ'.'m values (Calcul,llion Box Fig, 10-48·2), A venicalline repre 35 Calculation Box Figure 10-48-1. senting (V'l) ,-;. (P) is drawn first; Calculation Box Figure 10~4B·2. (EJ is added at a right angle to the disc inclina tion, and (R) closes ,he triangle. The direction of (R) in relation to the disc termined Walking Speed Slow (1.05 mfsec 1) Normal (1.38 m/sec' l ) --------Fast (1.72 m/sec' l ) Very fast (2.16 mfsec 1) :c0> ·ffi 2 ;;: >- "In0 <5 ~ E ~ f- .£ ~ ~ 0 LL 0 RHS LHS LHS ~0l-L---:':--'-l.:-I-:'::-'-:!:::-''---:JL 20 40 60 80 100 Percentage of Cycle Axial load on the L3~L4 motion segment in terms of body weight for one subject during walking at four speeds. The horizontal line (UBW) denotes the weight of the upper body, which represents the gravitational component of this load. Loads were predicted using experimental data from photogrammetric measurements along with a biomechanical model of the trunk. LHS, left heel strike; RHS, right heel strike. Adapted with permission from Cappozzo, A. (1984), Compressive toads in the lumbar vertebral column during normal level walking. J Orthop Res, 1, 292. EXERCISES During strengthening exercises for the erector spinae and abdominal muscles, the loads on the spine can be high. Although such exercises must be effective for strengthening the trunk muscles con~ cerned, they should be performed in such a wa:\' that the loads on the spine are adjusted to suit the condition of the individual. The erector spinae muscles are intensely activated by arching the back in the prone position (Fig. IO-20A) (Pauly, 1966). Loading the spine in extreme positions such as this one produces high stresses on spine structures, in particular the spinous process (Adams et aI., 1988). Although intradiscal pressure in a prone position with upper bod)' support on the elbows is halF that in standing (Wilke et aI., 1999), it is recolllmended for c:\ercise that an initial posi that h:eps the \"e["IL'hr~IL' in a 1l1Orc parallL'1 alignn is preferahle \"hen stn.:ngthcning e:\L'rcises for LTector spinae IllLlscles ~\re performed (Fig. 10-2( The importance of the ahdorninal muscles spinal stahilit.'· and inlL'rpla.\" in the production lAP reinforces the need ror strong ahdorninal l ors. Sit-ups arc a userul exercisc ror ahdcllll rnusclc strengthening, wilh man." vari:'\lions p ticed and cncOllraged by health prol"cssionals, cerlain vari:'l1ions arc viewed as hannflll to the back. Although the most common belief has h that sit-ups with the knees llexed and fed ll chored will emphasize the abdominal contribut while minirnizing psoas acti\"it." (.JukeI' et aI., 19 this is not true. Both bent knee and straight leg ups will producL' comparable le\"L'ls of PSO~\s and domin~d aClivit.\", crealing c(lmpressivc spinal lo ing. Curl-ups, in which the head and shoulders raised onl.\" to the point where the shoulder bla clear the l<.lble and lumb<.lr spinL' motion is m mi!.ed (F'ig. 10·21) and ortell emphasized in rehab tation programs, arc recommended for minimiz compressi\"L' lumbar loading (A,der &. l\-!cGil!. 1 Jukcr L't aI., I (98). This modilic<'ltion of the exen has been shown to be cl"fecti\'c in terms or Illotor recruitment in the muscles (F.kllOlm el aI., 1 Flint, 1965; Partridge &. \\'alters, 1959); all port 01' the external oblique and rcctus abdolninis rnus <.lre activated. Sit-ups with feet unanchored, legs vated, or tOI"SO twisting do not significantl.v incre abdorninal muscle <.lclivit." (AxleI' &: J'dcGill, 1997) lirnit the psoas activity, <.l reverse curl, wherein knees arc brought t()\vard the chest and the butto <.lre raised I"rom the table, aetivalL's the internal external oblique rnuscles :.md the reclus abdom muscle (P<.\rtridge &: \Vahers, 1959), If the rev curl is perrormcd isometricall.v. the disc pressur lower than that produced during a sit-up, bUl the ercise is just as effective for strengthening the dominal rnuscles (Fig. IO~22l. It can he conclu that no single :.Ilxlominal exercise C<.m optimall.v t ~dl tnmk flexors while minimizing intervertebral j loading. Instead, a \'aried program must he scribed, tailored to the training objectives of the dividual (;hler & McGill, 1997). \Vhen designing a back strengthening cxer progr:'llll, the most import:.lnt consider:.ltion is conclusion drmvn bv the Paris Task Fo (Abenh:.lim et al., 2(00). The guidelines set forth clude recommendations that exercise is benefi for subacute and chronic low Inlck pain, No par ular group or t~'pe of exercises has bcen shown t most efTccti\·c. A A, Arching the back in the prone position greatly activates the erector spinae muscles but also produces high stresses on the lumbar discs, which are loaded in an extreme position. B, Decreasing the arch of the back by placing a pillow under the abdomen allows the discs to better resist stresses because the vertebrae are aligned with each other. Isometric exercise in this position is preferable. .> ,? .',. 200 150 100 50 Performing a curl to the point where only the shoulder blades dear the table minimizes the lumbar motion and hence the load on the lumbar spine is less than when a full sit-up is performed. A greater moment is produced if the arms are raised above the head or the hands are clasped behind the neck, as the center of gravity of the upper body then shifts farther away from the center of motion in the spine. A reverse curl, isometrically performed, provides efficient training of the abdominal muscles and produces moderat stresses on the lumbar discs. The relative loads on the thir lumbar disc during a full sit-up and an isometric curl are compared with the load during upright standing, depicted as 100%. Adapted with permission from Nachemson, A. (1975 Towards a better understanding of back pain: ~ review of the mechanics of the lumbar disc. Rheumatol Rehabil, 14. 129. • ........ ... . ".~- ey,. MECHANICAL STABILITY OF THE LUMBAR SPINE i\tlechanical stability for the lumbar spine can be achieved through several means: lAP, co·conlraclion of the tnlllk muscles, external support, and surgery. Surgical procedures for lumbar spine stability will not be covered in this section. Intra-Abdominal Pressure lAP is one mechanism that may contribute to both unloading and stabilization of the lumbar spine. .IAP is the pressure created within the abdominal cavity by a coordinated contraction of the diaphragm and the abdominal and pelvic floor mus- cles. [lS unloading mechanism was first propos Banelink in 1957 and I\-torris et al. in 1961. suggested that lAP serves as a "pressurized ba altempling lO separate the diaphragm and noor (Fig, 10-23, A & B), This creates an ext moment that decreases lhe compl-essio!1 forc lhe lumbar discs_ The extensor moment pro by TAP has been calculaLCd in several biomecha models, with widely varying resulting reductio extensor moment from 10 to 400;' of the ex load (Anderson et at .. 1985; Chaffin, 1969; Eic, Lander et al .. 1986; Morris et aI" 1961). Recent studies using fine-wire EMG o deeper abdominal muscles found that the tran sus abdominis is the primary abdominal musc sponsible for lAP generation (Cresswell, Spine 160 140 Force ." -lbs- --::._. ....... 120 " 100 80 Abdominal cavity 60 40 Respiratory flow Positive values = Inspiralion Negalive values=expiralion 20 r Line of lorce 0 \r- n:--~ Inlraabdominal pressure : -mmHg- ! ",,' lime -msec- Ibs mmHg A B A, Schematic illustration of the effect of intra-abdominal pressure. An increased pressure will create an extension moment on the lumbar spine. B, Intra-abdominal pressure (lAP) (measured by a nasogastric microtip transducer) and respiratory flow (measured through a Pneumotach) during stoop lift of 120 lbs. (approximately 60 kg). Solid line, lAP; dotted line, force exerted in lbs; dashed line, respiratory flow (negative values delineate expiration and positive values delineat ration). Note that the subject inspires before the lift an the breath throughout the lift. lAP increases and peaks gether with the lifting force. helping to stabilize and un the lumbar spine. (ourtesy of Markus Pieuek, MD and Mar Hagins PI: MA. Program of Ergonomics and Biomechanics, N York. UniversiC'j and Hospital for Joint Diseases, New York, Ny Cresswell el aI., 1994a; Cresswell eL aI., 1992; Hodges cl aI., 1999). As the transversus abdominis is ho.-izontally oriented. it creates compression and an incre,lSC in lAP withollt an accompanying flexor moment. A recent experimental study gave the.firsl direct evidence ror a trunk extensor moment produced by eleva Led lAP (Hodges el aI., 2000), It has been demonstrated that the lAP contributes LO the mechanical stability of the spine through a co-activation between the antagonistic lrunk lIexor and extensor muscles (Cholewicki et aI., 1997, 1999a,b; Gardner-Morse & Stokes, 1998). As the abdominal musculature contracts, JAP increases and converts the abdornen into a rigid cylinder that greatly increases stability as compared with the muhisegmenlcd spinal column (ivlcGiII & Norman. 1987; Morris et al.. 1961). lAP increases during both stalic and dynamic conditions such as lifting and lowering, running Hncl jumping, and unexpected trunk perturbations (Cresswell el aI., 1992; Cresswell et aI., 1994b; Cresswell & Thorslcnsson, 1994; Harman et aI., 1988). Current research suggests that the transversus ab~ clominis muscle, together with the diaphragm, plays an important role in stabilizing the spine in preparalion for limb movement, regardless of the direction in which n1QVement is anticipated. Transversus abdominis and diaphragmatic activity appear to occur independently, prior to activity of the primary limb mover or the other abdominal muscles (Hodges eL al.. 1997, 1999). Trunk Muscle Co-Contraction To understand the phenomenon or co-contraction during trunk loading, Krajcarski et al. (1999) studied the in vivo muscular response to perturbations at two rates causing a rapid flexion moment. The results of maximum trunk flexion angles and resulting extensor moments were compared. The results showed that with higher levels, co-contraction, spine compression. and trunk muscle stiffness increase. During unexpected loading, a 70% increase in muscle activity has been noted as compared with anticipated loading, which may lead to injury (Marras et al.. 1987). Further investigation into the loading response has revealed that an inverse relationship (i.e .. the shorter the warning time, the greater the peak trunk muscle response) exists between peak muscle response and warning time prior to loading (Lavender ct a\., 1989). Loss of spine stabilit~. . can be achieved throug repetitive loading. This can be achieved throug repetitive continuous motions that fatigue th trunk muscles. l\tlusclc endurance is mechanicall defined as the point at which fatigue of the mus cles is observable. usually through a change i movement pauern. Parnianpour ct al. (1988) use an isoinenial triaxial device to study force outpu and movement patterns when subjects performe a flexion and extension rnovemcnt of the trunk un til exhaustion. The results showed that with fa tigue, coupled motion increased in the coronal an transverse planes during the Oexion and extensio movement. In addition, torque. angular excursion and angular velocity of the motion decreased. Th reduction in the functional capacity of the f1exion extcnsion muscles was compensated for by sec ondary Illuscle groups and led to an increascd cou pIc motion pattern that is morc injury prone Figure 10-24, A & B shows the increase in axial ro tation (torque and position) during flexion and ex tension of the trunk until exhaustion. In an animal study, Solol11onow ct al. (1999) in duced laxity of the spine in the ligaments, discs and joint capsule by' cyclic repetitive loading of fc line in vivo lumbar spines. The cyclic loading re sulted in desensitization of the mechanoreceptor with a significant decrease or complete elilnina tion of reflexive stabilizing contractions of th multifidus muscle. This may lead to increased in stability of the spine and a lack of protective mus cular activity even before muscular fatigue is ob sen/cd. A IO-minute rest period restored the mus cular activity to approximately 250/0. External Stabilization Restrict.ion of motion at any level of the spine ma increase motion at another level. The usc of bac belts as a means of preventing low back injury re mains controversial. Originally it was believed t assist in increasing lAP as a way of unloading th spine during lifting; however, inconclusive ev dence exists as to the biomechanical effectivenes of these devices (Perkins & Bloswick, 1995). Th National "Institute for Occupational Safety an Health has advised against the use of back belts L prevent low back injuries (NIOSH, 1994). As wel an orthotic worn lo restrict lhoracic and lumba motion may result in compensatory motion at th lumbosacral level (Lumsden & Morris, 1968 Norton & Brown, 1957; Tuong et aI., 1998). , 78.3 ~ 7.9 E J\I\I~I~\~I\ !! T ~ -·32.4 llmlm 1111111 Ill\/II\ !i~i I~i~ -'::======::;:======::;====================~ -127.1 58.6, 40.9 21.6 1/111/1111\ ;111111 III I1II1 2.3 - 15.4 +--.,-'''''1-'--,'--'1-r'-' , 'I'-'I--''--'--Ir-,TI--...,--,1+-i-.....-r-1T,- ,I-"--""I-,rl 0.1 9.3 18.4 27.5 36.6 45.7 54.7 A ROlation Torque and Position vs. Time 4.3,.-------------------------------, , 2.3 E z, " ~ 1.1 !! ....0 -0.6 -2.1 19.1 , w 13.3 'i'" 7.0 "co~ c .Q .~ 0 Q. 0.7 - 5. 1 -!--,----,-.-,.--...,--,-.,..--,----,-.-,.--...,--,-.,..--,----,-'-.-,.--...,--,,-.,..--,----,--I 0.1 9.3 18.4 27.5 36.6 45.7 54.7 B Dynamic (isoinertial) flexion-extension trunk testing until exhaustion for one subject. Torque and position data is depicted for two planes, flexion-extension (A) and axial rotation (B). Note that flexion-extension torque production is diminishing as is the amount of performed extension of the trunk (A). Rotational torque and movement amplitude, increased accessory motion, and torque is shown in B. Data for lateral flexion was similar to axial rotation a not shown here. Adapted with permission .from Parnianp M .. Nordin. M.. Kahclnovitz, N.. er al. (1988). The rriaxial coupling 0; torque generarion of trunk muscles during iso ric exertions and the effect of fatiguing isoinercial movem on the motor outpur and movemeor pc1tterns. Spine. 13(9 982-992 Investigation into the effect of back bells on muscle activity has revealed no significant EMG activity differences in the back extensors during Iifling with or withoul a back bell (Ciriello & Snook. 1995; Lee & Chen. 1999), while McGill el al. (1990) showed slightly increased EMG aClivily in the abdominal (except for the intel'l1al oblique l11uscles) and erector spinae muscles. Thomas et '.11. (1999) have verified a slight increase in EMG activity (2(10) in the erector spinae during symmetric lifting wilh a back belt. Back bells have not been shown to significantly increase Iifling capacity (Reyna el aI., 1995). on the spine to bc minimized during lifting, the d tance bL'lween the trunk and thc object lifted shou be as short as possible. 10 lAP and co-contraction of trunk musculatu increases the slability of the spinal column. 1/1' Trunk muscle fatigue may expose the spine increased vulnerability as a result of loss of' mot control and thereby increased stress on the su rounding ligaments, disc, and joint capsules. ,12 vValking is an excellent exercise thai poses low load on the lurnbal' spine. REFERENCES Summary 1 The lumbar spine is a highly intricate and complex structure_ 2 A vertebra-disc-vertebra unit constitutes a molion segment, the [unctionalunit of the spine. , 3" The intervertebral disc scrves a hydrostatic function in the motion segmcnt, storing energy and distributing loads. This function is reduced with disc dcgeneration. ~:"4 The primary function of the facet joint is to guide the motion of the mOlion segment. The orientation of the facets determines the type of motion possible al any level of the spine. The facels may also sustain compressive loads, particularly during hyperextension. S Motion between two venebrae is small and docs not occur independentl~y in vivo. Thus, the functional motion of the spine is always a combined action of sevel',,1 motion segments. /6;- The instantaneous center of motion for the mo~ lion segments of the lumbar spine lIsually lies within the lumbar disc. The trunk muscles play an important role in providing extrinsic stability to the spine; the ligaments and discs provide intrinsic stability. 8 Body position affecls the loads on the lumbar spine. Any deviation from upright relaxed standing increases the load. Forward flexion and simultaneous twisting of the trunk produce high stresses on the lumbar spine. 9 EXlernally applied loads lhal are produced. 1'01' example. by lihing or carrying objects Illay subject the lumbar spine to very high loads. For the loads Abt.:nhairn, L., Rossignol. ~'t., V~dal. J.P., l't a1. (2000). Thc r of acti\'ily in the tht.:r:q1clllic management of back pa Report of thl: Inll:rn:l1ion~1 Paris Task Force on Back Pa 5pinc. 25. IS. Ad~lillS, M.A.• Dolan, P.. & 1·lutlon W.e. (1988). The lum spin\.' in b~tChnlrd bl·nding. Spille. 13. 1019 Atbms. ~L\. & Hulton, W.C. (19831. The 1lll.:ch;lIlicill fu tions of Ihe lumbar apophyst.\ljoints. Spil/e. 8, 327. Alkn, C.E.L (1948). i\lusdc action potenlials used ill sludy of dynamic analolll~', 13r J Phy.~ Jtcd, 11,66. Anders()Jl, C.K., Chaffin, D.B., Hcrrin, G.D .. ct :11. (1985) biolllcchanical model of the lumbosacral joint during l ing ~\cli\'ities. J BiollH'dllllll·I.:S. 18, 57. :\lHkl'sson. E.A .. Oddsson, L.I.E .. Grundstrom, 1-1., ct (1996). E;\'lG acti\'ities or thl' qlladnHlIs lumborum ~ erector spinae musdes during flexion-relaxation and ot motor lasks. CJinial! Bio1llcdUlllics, J J, 392. Andersson, G.B.J. & l:l\·cnd"r. S.A. (1997). Evalu;ltion Muscle Function. In J.W. Frymoycr (Ed.). Tilt' A(/1I11 Spi Pril1c.:ip{es (lilt! Praclict: (2nd cd., pp. 341-3$0). New Yo Li ppi ncot t-Ra\·l:n. Andersson. G.B.J .. Ortengrcn. R.. 0:: Nachcmson. r\. (197 Quantitati\-c studies or b~lck loads in lifting. Spillt:. I. 1 Andersson. G.B.J .. Oncngren. R., & Nachcmson. r\. (197 lntr'ldiscal pressure. intr~\·... bdomin ... 1 prl.'SSUrl' ilnd m det:tric bad lllusck :ll:tivity rdated to pOSl\lrc <lnd lo ing. Clill OrtJlOp, 129, 156. Andersson, G.B.J .. Onengl'cn, R., Nachclllson, A., et (1974). Lumbardisc pressure and myoelectric b"H:k mus ~\clidty during sitting. 1. Stu,dics 011 an experimen chair. SC(/I/d J Rclwhil ,Ht.:d, 6, 104. ,\sl11usscn, E. & Klauscn, K. (1961). Form and fUllction or (.'rcc( human spinc. eJill OrlllOp, 25. 55. Axlel". c.T. & McGill. S.M. (1997). Low back loads o\·cr'l \' ...' tv of abdomin.;\1 e.'\en.:ises: SeillThing for th..., s;lfcst d~l1lin;l1 challeng(·. .\lcdiefl/I: & Sciem"":c ill Sports & E cise. 29, 804 B:II·tclink, D.L. (1957). Thi.: role of i.lbouminal prC'ssure in lieving the pressure on thl.' lumbar inu:n'crlcbral disc BOlle }oilll Slfr~, 39B. 718. B:ISIlWji.lll. J. V. (t958). Electromyography of iliopso'ls. :l Rcf.:, 131. 127. B~sl1laji~n. J.V. & DeLlll;;l, C.J. (1985). ;\1l1sdc:s Alive. Ba mort: Williams &. Wilkins. ·f. ..;.' , BicrilH!·SorenselJ. F. (1984). Plwsical nh.'<lsurcmenls as ri~k indlcalors for I()\\'-b~\ck lrol;bll.' over a on<:-yl.'ar period. Spill!.'. 9, 106. Callaghan. J.P.. P'IIla, .\..E .. & McGill, S.\\. (1999). Low back thrcc--dimellsional joinl forces. killl.'malics. and kinetics during walking. Clill;cal BioIJlC(://(lllics, I';', 203. Cappozzo, A. (1984). Compressive loads in the lumb'lr vent:br.1! column during llorlllal le\'el walking. J Or/hop Res, l, 192. Carlsoo. S. (1961). Thl.' st .. tic muscle load in differelll work posilions: All dL'ctromyographic study. Ergollolllic.,·. .J.. 193. Chaffin. D.B. (1969). r\ cornputcriz.cd biolllechanical modclDc\'dopmcl11 of and usc in stud~'ing gross body <lctions. ) BiOiI/Cc!wl/ics, 2, 429. Chaffin, D.B. Andersson. G.BJ. (1991). Occup(1tiol/al BioI/wcJ/(lIli<:s (2nd cd .. pp. 171-263". New York: 10hn Wile.\' 6.: Sons. Inc. Cheng. c.1\: .. Ch..:n. H.H .. Chen, C.S., et 31. (1998). Influcnces of w~dking spce:d change' on the lumbosacral joint fon:c distribulion. BiOl1lt'd Mara £11::', S, 155. Choh:wicki. 1.. Juluru. K.. &: ~-IcGill, S.l\-l. (1999a). Intra-abdominal prt:ssun: ll1(.'chanisrn for stabilizing tll(.' lumbar spine.) Bio/lwch, 32. 1-13. Cholcwicki, 1., luluru, K .. Radt:bold. :\., I:t <II. (1999b). Lumb.ll· spinc stability (;'til be :lllgmcnted wilh an abdominal bcli and/or increased inlra-abdominal presslIrt:. f.:'llropcall '* SJlill~ )ollrJ/al. 8, 388. Choh.'wicki, J., P.mjnbi, ~L\L & f\.h;lChntr~·all. A. (199i). Sta· bilizing function of trunk f1e.'l:or·('.'I:tcnsor llHlsch.. s around <I neutral spine posture. Spine. 22(19), 1207. Ciriello, \'_,\I....\: Snook, S.H. (1995). The dfl.-·Cl of b<lck belts 011 lumbar muscle fatiguc_ Spillt', 20, 1271. COSSl.'ltc. J.W., Farfan, H.F.. Robertson, G.H., e[ at. (19il). The instant~1I1l.'OUS <:enttr of rot~l[ion of Iht.' third Illlllb,lf inll.'ITl.'nebral joint. J Biol/m.:It . .;.( 149). 19i 1. Creswell, A.G. (1993). Rl.'sponses of intnl-abdorninal pressure and ,Ibdomin~ll muscle activity during dynamic trunk loading in man.l::ur) Appi Physiol, 66,315. Crt.·sswcll. A.G .. Blakt.'. P.L.. &. Thofstcnsson. A. (1994<1). The drect of an ahdomin.d muscle Iraining program 011 ill[raabdominal pressure. Scant!) RC/Ulbil ;\ld. 26, 79. Cr~ss\\"dl. kG .. Grundstr6m. 1-1 .• 6.: Thorslcllsson. :\. (1992). Observations on intr<I-.llnlomillOlI preS$ur..: and p~lllcrns of abdominal intra-muscular acti\'it~· in man . ..lcta Physiol Scaml, 1';'4. 409. Crcsswdl. A.G .• Oddsson. L.. & Thorstcnsson. A. (1994h). Thl.' influcnce of sudden perturbations on ll'unk muscle: aCli\'ity and intrn.- .. bdominal prcssU!"t' while: sl<\nding. EXJll'l' Brain Res, 98. 336. Cresswell, A.G., & Thorstensson, A. (199-0. Chnngl.'s in intraabdominal pressure, Ll'llnk muscle acti\· ... tion and force during isokinetic Iirting and lowering. Eur ] Appl Physiol. 68, 315. Dalstr~l. M. 6.: Huiskcs. R. (1995). Load lr.lnsfcr 'Kross the pelvic bone. J Biollli:t:It. 28. i 15. Davis. K.G., M ... ITas, W.S .. & Waters, T.R. (1998). Evaluation of spinal IOildillg during lowering and lifting. CliJl Biolllt.'dumics, /3, 141. de Looze, ~LP., Toussainl, 1·1.0.'1.. vall Dieen, J.H., et al. (1993). Joint moments and llluscle activity in the lowel" e.'l:tremitics and lowcr ktck in lifting and lowering tasks. J 8iol/I.:c/',26, 1067. Eie. N. (1966). Load capacity of the low b ..d. J Oslo Cit.\" f1osp, /6. 73. Eklwllll. J....\rhorelith. L·. F~lhlcr:Ill!I: ...\ .. l·t :11. (lYiY \';lli(11I (If :,hd(llHin:lllllu:-clc:- during SOli\\,.' ph.\-sihl!ll, til,.' c\l,.'l"ci .. l,.·!-. SnlJld) I?dwhil.l1t:d. II. 7.~. 101 U"hy. ,1.,.1, " K;n~, ..1.1. (1 'J~6j. Inlcn,'rlchr,d d; f:l("l'! l,.·ollta("\ prl,.·ssufl· in a:\i~tll(lr!-i(1l1. In S.:\. l.anl King (Ed:-.). /9SfJ :\d\'/I/H"".~ ill Uilh·lI:.;illt'I'l'ill.:: (pp. :\l,.'\\' York: . \!l1l'l"icall SOl,.·iL·\\ or \In:h:\lliL':d Ellginl:( r:ad~ln.ll.F. (19751. \ltlsl,.·III~II' Jlh: .... h;llli .. 1lI fir the lUJ11k ~tl\d I Ill,.' P(,:>ilioll of Pfl\\"l'!" and dficil·/lL·..... On/H Xorth . \IlI. 6, 135. Flinl . .\1..\1. (1965J. Ahdominal ll1u:>dL' in\'uln'l1ll'lll thl' pl,.-rfnrm:lllt.·l,.· of \·.lriol1~ forms 01 :.it-up l,.·\l,.·n; l,.·k('tr(Jm~·hgr'lphic :-llld~·. _'\1lI J Phys .\I,·d. ';'4. 114. Fln\'d. \\'.F. <.\: Silv . . . r, P.fl.S. (193.:;). Thl' fUJll,.'linll (Ii lll lors spinal,.' Inllsl·k" in cnl~lill 111()\clll . . . n\S ~llld post man. J NI.niol. 129. 1$4. Fri(·dd)old. G. (195~J. Dil,.· ..\kti\·iL'1 nOI·lIl.da Rih:h·n: rnuskublur illl Ek,ktfom .... ogralllln ullll'r n'rschi !·Iallllllgsik·dillgungl,.·n: l,.·illc Studil' Skl·kllllltlskdnwchanik. I Of/hop. 90. I. rukll\:tlJl:t, S .. :\,tbmur:l, T. nl·d~l. T.. l't ,d. (1995). k,(t of Itlc(!1:ulic.t! slrl's;,,; 011 hvpntroph\' o!" 11k li,\;;II11l'!lIUIll Il:l\'tllll. ) .')pilluIIJl'.-nn/. S. 126. Gatlilk. J.O. (1<)67). h'n:>ilc propl'nil's 01 11K hUIll;11l :lIl1lUIIlS fihrh:->II:-, .. lew O,.lIwJ' S,'lmd SII1'''!. 100, 1G;lrdll~'r·.\I(.tr~l,.·. .\I.G. & S!('Kl,.· .... LA. (19%l. Till,.' dfl'({~ dominal lllu:-dl' CI,·;ll·ti\:ttic.n on lumbar :-pilll' :S/Iill... II::!,'), ~9. G(·rtl.lkill. S.D .. SI.·liglllt,IIl,.I. Hl.lltln, R....... l .d. (1%5 trndl' P~ltlL'l"llS :lnd sl'glllL'Il!,lI il1Sl~d)i1il.\ ill ,kgL'l di:,\.· disl':l:-e. S"i!/,·. 10. 257. Granat'l. K.P. .i: Sanford. :\.1-1. (19991. LUll1har-pd\'k n~tli(ln i:> influl'IH..·l,.·d b~ lifting l~\sk par~\nh.'kr:>. 25(111.1413. Grl,.'gCl":>l,.'Il, G.G . .i Luc~t~. D.l3. (1967). An in \'inl slud.\ lUlllbar a,xi:\1 rot;lliU!l (If thl' hlllll~\n Ilwrac(Jllllllb:\ .l 8011,' .luil!! Sur:: . .,19:\. 2-+7-232. llaht.:r. T.IC. O·llrit:lI ..\1.. Dr.\·a, .1.\\'".. d 'II. (l\FHI. TIll till' IUlllb.\I· f,ll,."l,.'! joinl::- ill spinal :-tahilil~·. IllcntiliGt :illtTn:niH' palh~ of 1(1~\{lillg. Sl'ilh·. /9. 26-67. Hannan. E...\., Fr~"klll:lll. P.:\ .. Clagclt. E.R ....... t ~tl. (19SS ,.hd(llllinal ,tlld in!r:l·[hor~lL'ii.' prt'-,,~uro during lift jumping . .\led Sci S;10/"0 Erac 20. 195. Hasq:;\\\:l. K.. T~\bhashi. IU_:. "Og~l. Y.. l,.'1 :1I. (199 Ch'Hlic:d propl,.,rti(·s or ostc'olwnic \,t:rtdn:d hodk's torl,.·d hy ;Il,.·ou:-;tie l.'llIb:-ioll. /Jeme. /-1-. 7.'ii. l-laugh!oJl, \'. .\1.. Schmidt. T.A .. Kl,.·l,.-k, K .. l,.'1 ;.1. (200()) bility of lumbar spin:d motion :-l'gllll'lltS I..'orrd:lll'd of 1.:.lrs in Ilk' ;lIl11ulus fihro:-llS. J .Vel/rosur.::. 92. S I-Iodgl,.·:-. P.\\'.. Crl,.·s:-\\dl. A.. l\: Thortl,.·l\:-~(JIl. :\. Prl,.·p.\r:llor.\' [rullk lllUlioll ~ICc(lll1p:tnit:s r:lpid uppt mO\·l,.'l1ll'l1l. 1:.\)1 Bmiu Ues. 12..+. 69. Hodgl.'s, P.W. l\: G.l1HII..'\·j,l. S.c. (2000-,...\.t.·ti\·:lliO!l of Ill,tn di:lphr~lgm during:l rt.·lk·tili\·l.' pOSIIlI';llt:lsk . .l fiJI/I.!) . .)22. 165. I-Iodgl,.·$ P. W. ~ Richardson C.A. (199iL Fl,.·cdforw:ml ( lion of tranSH'rSllS ahdomini:-; is nOI infhll,.·lln-d ll\' l"l·...:lion of :lrm nW\·l,.·llH:nl. Ex!, U/"I/ill Ik~, 1/-1-. 36i. I'[utton. W.e. Adam::;, \-1.:\. (1982). Can tIll' lumklr IT\lshcd in ht:a\'~' lifting"? Spill!!. 7. 5S6. '* 10s(·ph. J. 1.1960). "'flJI:': PO... ltll"t·: Ut!ClrolJl.'·fi~m/lhic S Springfil,.·hl: Charll,.·s C Thollla~. Jukn. D., .\!cGil\. S., l\.ropL P., ct ;11. ([ 99:n. OU:II11i[:t tr:lllltl:-(lIbr llIyock·nril.' :H:ti\·it.\ of IlImb.lr porli psoas and lhe abdominal wall during a wid\: \'ariel,v of lasks. }fed S,:i SporlS f.XCfC. 30. 30 I. A.I .. Pr:\sad. P.. ~ Ewing. C,1.. (1975). ~·h·dl<lnism or spinal injur~" du(.' to l:<\udoCl.:phabd accderalion. Onlw/J CIi" North :1m, 6, 19. Kkin. J.A., Hickey, D.S., 0& Huskins. D.W. (I983). Radial bul!:!in!: of lhe annulus fibrosus and the fUlll.:lion and fail· llr~ -of ~he inlt:ITerlL'hr~t1 dbc. J Bi01llt'ch. 16. 211-217. ~'1.H .. S~:mllssi, R.E., Wilder, D.G .. el ,d. (1987). Inh:r· nal displacelll~nl dislribution from in vilro loading of hu· man lhoracic and lumb'll" spinal motion segments: Expl:r· ill1l:ntal results and lhl'orelical predictions. Spill/.', 12, 1001-1007. Krajc;Hski, S.R .. POI\'in, J.R .. & Chi;1I1g:, J. (1999). The in d\'O r~sponse to the pl.:rlttrbations c:lllsing rapid f1cxion: Ef· f~cts of pre-load and step inpul magnitude. Cii,l BiO/lll/dUlU;CS. 14, 54. Kuhlk. R.E. Schuhz. A.B .. BdYlschko. 1., el <II. (1975). Bin· ll1echank;tl charactcristics of \'crICbral motion segments and intern:rtcbral discs. Or/hop Cfill North AI/I. 6, 121. L1nd~r. J.E .. B~ltes, B.T.. & Devita. P.(1986}. Biolllt.'chanics of th\.' squat (.'xen:isc using. a modifil.'d Ci.'Ill(.'r of mass bar. .\lcd Sci Sports E.un:, IS. 469. Ll\'cnder. $.:\., ;\lirka. G.:\ .. Schol.'I1ITI.lrklin. R.\\'., el al. (1989). Thi.' l:rrCC1S of pre\'iew :lnd t<l.sk sym!lll:lry on trunk mllscli.' response to sudden loading. lIwlIlm Fat-tors. J I, 101. [.:\\'coder, S.A .. Tsuang. 1".1-1 .. AIH.li.'rsson, G.B .. ~t a1. (1992). Trunk nlusch.' co·contraction whitt: resisting applied mo· ments in a Iwisted poslure. Ergollolllics. 36, 1145. Lt.'c. Y.H. &. Ch('ll. C.Y. (1999). Lumbar vcrtebr.d angles and b.lck muscle loading with bells. Illd Health. 37, 390. Lucas. D.B. &. Brcsier.~B. (1961). SIal,ilil.'· of ,ht: Ligamt.'I1fOUS SpiltE.'. Biomechanics Laborator~', University of Cnliforni:\, San Francisco and Berkeley. Technical Repon 40. San Francisco: Th~ Ltbor.llor\,. Lumsden. R.M. & Morris. J.~'1. (1968), :\n in dvo study of axinl rotation and immobilization at the lumbos:\cral joint.} Heme Joill/ SlIrg. 50A. 1591. Manas, \V.S. & Granata. K.P. (1995). A biomechanical assess· ment and model of axial t\...· isling in the thoracolumbar spine. Spil1c? 20, 1440. :\larras. \V.S. & Mirb. G.A. (1992). :\ comprehcllsivl.· c\'alll:\lion of trunk respon$c 10 asymmL'tric trunk lllOlion. Spill/!. 17. 318 . .\larras, \V.S .. Rang<lrajuili. S.L.. & Lan'nder. S.:\. (1987). Trunk loading and ('xpectation. Erf,:ollolllics. 30, 351. McGill, S.~vl. & Norman, R.\V. (1987). Reasscssm~nt of the role of inlra·abdominal prcsstlrt.' in spinal compression. Ergollomics. 30, 1565. McGill. S.M .. NOrllmn. R.W., &. Sharratt. M.T. (1990). The d· fret of an 'Ibdominal b~1t on trunk muscle acti\'ity and intra·abdominal prl.'sslln..' Juring squal lifts. Er~ol/olllic.s. 33, In. McGill, S.~1.. Yingling, V.R .. &. Pl..'ach, J.P. (1999). Tlm..'I..'dimcnsion:d kin~rnatics and trunk muscle my()d~ctric ac· ti\'it~' in the ddcrly spinc-<l dawb,lse compared to young people. CJiIl Biolllcclu/llics. 1'+. 389. ,\'111es, M. & Sull ivan. W. E. ( 1961 ). L:lti:nd bi.'llding .11 Ihe hlln· bar and lumbosacral joints. .·\11111 U~C, /39. 38i. ~liller. l.A.A., Hadersp\.'ck. K.:\., 6.: Schuhz, A.B. (1983). Pustl'rior (.'lelllenl loads in lumbar motion segmcnts. Spille. S, 331. ~Ioll. .I.M.H. &. Wright. V. (1971). Normal range of spin;. bility. An objcclh'e clinic:l! study. AIlII UhClIIll Dis, 30 Morri!'O. J.!"..l.. Bl:lllli:r. G., ~ Lucas. n.B. (1962). An tf(Jl1lyographic study of Ihe intrinsic muscles of the in man. J :111t1l l.mul. 96. 509-520. ~-lorris. J.~L, Lucas. D.B .. &. Bn:sier, B. (1961). Role o trunk in st.:tbility of the spin<.'. } BUIIi: JOil1f Slll~~ . .:31\ :\-losekilck. L. (1993). Vertebral struclllr\.' :lIld strength in and in vitro. Cah:i{Tissllc 1111.53.5121. N;'lChl'llIson. A. (1960). Lumbar inlrndisc'll pn:SSllrl.' ..-k/ [hop Sculld SlIp,.,l. n. 1-140. Nachclllson, A. (1963). Thl' influence of spinal rno\'erni..' till.' lumbar intrndiscal pri.'SSUl'l.' and on till' tensile str in IIIi.' ;tnlltllus fibrosus . ..lela Off/WI' SClIIul. 33, 183. Nachcmson, :\. (1966). EIeCiromyographic studies Oil th td}ral portion of the PSO:IS llluscl~. With special rcr to its stabilizing function of the lumbar spine. lkw O Scalld, 37, Iii. Nach<.:mson, A. (1975). Towards a beller understandi back pain: A review or the mechanics of the lumbar RlJellllwlO! RelUlhi!. I.J, 129. Nach(.·m;.;on, A. & Elfstrom. G. (1970). IlIlravital [)\'I Pre:ssure: .\lC(ISllrCJ1h.'lllS ill Llflllh(lr Dis,;:,: ,.\ Sllui.\' r.;{ IlIOIl .\IOl·i'J11CIlfS. :\Ialli:III'crs allll Exercises. Stock Almquisl & \\"iksc-lt. ~achelll;.;on, r\.L. .& E\':lns. J.H. (1968). Some mech propcnies of the third human lumbar intalaminar ment (IigameJltulll flavum). J BiomcclulIIh's. I. 211. Nachclllson, A. & Morris, J.\l. (1964). In \'in) mCaSUri. of inlradiscal pressure. Discolnl'lry. a mcthod iol' th termination of presslll"i.' in the lower lumbar discs. J 1ui1l/ SI/rg, -16:\, 1017. !\alional Institute for Occupational Safety .tnd Health (1 WOl'kplclCe Use S(lck Be/IS. DHHS (NIOSH), Nu 94-122. Ndson, J.M .. \blmsky, R.P.. 6.: St(.'n:nsoll, J.,\'I. (1 RcI'lti,·c IUlllbm' ;tnd pl.'!\·ic motion during londcd fk'\ion/extcllsion. Spinc, 20. 199. Nemeth. G. (1984). On hip and lumbar biomechanics. A of joint load and muscular ;lCtivity. SC(/l/(! J Uelwhi or Stlppl, 10. Nonon, P.L. & Brown. T. (1957). The immobilizing c-ffi of h'ICk braces. Their eHecl on th~ posture and mot thl..' 11l1lloosacr;.\1 spine. 1 BOIIt' }oill1 S/lrg. 39:\. Ill. Dncngrcn, R. And~rsson. G.B.l., & 1'\achcmson. A.L. (1 SllIdies of rd:lliollships bel'\'cen lumb,lr disc pre myoelectric b<'H:k muscle acti\'it~,. and intra-abdominn trag;:\slric) pr~SSllre. Spille, 6. 98. Osval<k'r. A.L.. Ncumann, P., Lo\"sund. P.. et a1. (1990) maII..' strength of the lumb.\r spine ill fle.xion-.1I1 in study.} Bio1l1t.:clull/ics, 23, 453. Panjabi. ,\13\-1., Gael, V.I-\. .. ..\: Tabtn, 1-\.. (1982). Ph~'si strains in the lumbar spinal ligamC'1l1s. An in vilro b chanical slUdy. Spi1/i', 7. 192. Parnianpour. M., Nordin, \1.. K<thnnol"iti'.. N., l.'t al. ( The tri<.lxia! clHlpling or torque gcneration of trunk cles during isometric e.'\cnions and the effect of f:tl isoincrtial mo\,emenlS on till.' motor output and 1ll0\ p~IIt('rns. SpiJlt', 13(9). 982-991. Partridge. M.J. 6.: Walters, C.E. (1959). ~articip.llion abdominal IlHlsdes in \'arious mOVt'lllc-lltS of Ih ..~ tru IlWIl. An L'!cl::trol11yographic study. Phys Tll!'r Rei', 39 Paul.\', J.E. (1966). :\n dcclromyographic '\llaly~is of cl:nain nlO\'('Il1(:nls <lilt! ex('rcises. I. SOllle c!('cp muscles of Lhl' f back. All(/( Ret:. /55. 223. Perkins, ~I.S. & Bloswid. D.S. (1995). Th,· usc of h'ICk belts 10 increase inlr~wbdominal pressure as a means of pn:venting lo\\' back injuries: A survey of lhe lih:r.lIun:. 1m J OCCIlP EUl'imll rlt.:lllih, I, 326. Pope, M.H., Andersson, G.B.J., Brolll.lIl, 1-1., L't .d. (1986). Electromyogr:q1hit studies of the ]lImb'H' trunk muscula. tun: during thl: dcvelopmcnt of a.xi.1I torques. J (Julio!) Res, cJ, 288-297. Ranu, B.S. (1990). Measuremctlt of pr(.'SSllrcs in the nucleus and within the annulus of the hurn'lll spinal disc due 10 l.:X' trcme loadilll.!. Pmc Illst Jleclt Rcichm;t1Hl, S.~ (1971). Ef/J,!, (llJ. 204, 14l. Motion or thc lumbar anicular processes in fh:xiull-cxtcnsion and lateral flexion of the spine. ~1c[(t .\!orp!lo/ Nt'crl SCfllld 8, 261. Reichmann, S., Bl.'rglllnd, E.,..\: Lundgr(,.·n, K. CJ972). Das B(,.·. wegungszcntrum in der LendenwirbclsAule bei Flc.xio[l lind Extension. Z A/llIl EIII\\·ickltlll;;';;;':I!.\·(.·h. /38,283. Reid, J.G. &. Costigan, P.A. (1987). Trunk llluscle balance and muscular force. Spill/!, /2, 783. Reyna, J.R. Jr., Leggdl, S.I-I., Kenney, ~., "I .d. (1995). The ef. feCi of Ilimbar bells on isolated IlImb,lr musl'!e. Slrength and dynamic (;"padt)'. Spilll.', 20, 68. Rolandcr, S.D. (1966). ~'lotion of lhe lumb'lf spine wilh special reference to the stabilizing effect of posterior fusion. An experiment"l study 011 autopsy specimens . .'\elll Ortho}l 8colld, .)11]1]11. 90, 1-144. Sato, K., Kikuchi, S" & '{onez;Jwi1, 1'. (1999). In vivo intradis. cal pressure llle'lsuremCllt in healthy individuals and in p<iticnts with ongoing back probll.'ll1S. Spille, 24, 2468. Schuhz, A., ~l ~\1. (1982). Loads on lhe lumbar spine. Valida. lion of a biolllcchanicaJ analysis by mC<lsurenH.'IllS of in. Ir~ldisc.d prc:sslIn:s <1I1d myol.'!l.·ctric sign ... k J BOIlt' Joil/! Surg. 64.·\,713. Shir4lzi-t\dl. ,.\. (1994). Biomt'ch4lnics of lhe lumbar spine in sagittalJlatcral moments. Spi/h:, /9, 2407. SIlI(ImOIlO\\" \\., ZIHIlI. B.-II.. Bar:llia. R.\'., 1.'1 itl. (19 llh.'l'h:lnic. tlf incrl'a:-I.:d expUSlIl"(,.· to hlilihar inju h.l.· l'~"l'lic loadillg: Pan I. lo;.;s of rdl,:.\i\·(,.· 1Il1l:,(,: liz;llioll. Spillt'. 2·H23). 2-l26-243-l. Sldkn. T., B;II·alllki. II.C .. Ruhin, R.. d :11. (19t.1~J_ inlradisl'al pr(,.'~:-.ur(,.· Ilh:a::-urcd in I hI.' :I11t(,.·riol" itllt bler:!l anllul;,r rcgi(l/IS during :i='.\"Inml.'lricll I(I;l Biol/f(·clu/!/ic.\. 13, 495. Sllln:ssoll, B" Sl'h·ik. Go, & L"dCIl, A. (1'J89). i\!on thl.' s:lcroili:ll' joints. ,\ rt)I.'rllgt·ll sLerl'oplHllOg :lIwlysis. Spi!lc, 1-1, 162. Thomas, ./.5., Lavelltk'r, S.:\ .. Corcos, D.M., 1.'1 :t1. (1 fl'ct 01' lifting bl.'l!:- Ol1 ll'unk Illusck :1l"lil':Lli()[l sudlk'111.\" appli<:d In.td. /l1t'WIII Faclo!"s. ·iI. 6iO. luong, N.H., {),lll:-l,.·r(,.·:HI. J.. \latlr~tis, G.. 1.'1 ~d. (199S dillll..'llsional I.·\·alu.ltion of lumhar urthosis d :-pinal bL'h;l\"i\11". J Uclwbil Il.·... Del'. 35. 3-1. lrb'lIl, J.P,G. & :\kMllllin, J.r. (1985). Swdlillg prl thl' inleJ"l.·crtd.u·al di!'L'; Infllh:llc(,.' 01 pmIL'oglyc;1I1 I:q;l'n COlllt:'IlI:-. Bifl,.h.,o!o.t:..\·, 22. 1-45. \',111 Dici:n. l,tI., llUCt"l.I.'lllan:-, M.L\!.. &. TUllssai (1999J. SH,op or squal: ,.\ rl.'\·it'\\· of bilJlll(,.·<:hallil:a on lifling 1I.:(hniqu(,.'. CliJl Bifl/lft'c}wuic:•. /4. 6$5. \bn~..\1. & DUlllHS. G.:\. (199S). \kdl;lIlh,:al hdl;l\'i fl.~lllak s:l(roilial.· joinl [llld inflllL'nCt' of t!l(,., :l1l11 poskrinr s:II.::roilia<: Iigamcnts under sagittal In Hiw!lccltullio. /3, 193. \\'hi!l.', :\..-\. (1969). An:d.\"sis of thL' Illl'I'!l,lnit:- 01' spint:' ill Illan. An expl.·rillll.'llt,t1 stud.\' (II' :lllIOPS mens . .·le/ll ()rt/wp ,"jcl/ud, SIIp/i! il7, 1-105. Whilt:, A.A. l\:. P:lnjahi. :\I.N. (19701. Cliniud HiulIlt'cI Iht'Spilit'. I'hil:ldl.'lphia: J.B. Lippincott. ""ilkl.', "1.1., Ned. Po, C:lillli. :\1.. <'I al. (19991. ;\I.'W ml.·asur(,.·IIK·nts of pr(,.'ssurl.'S in Ihl.' inl<.'l'\·(,,·rlChra dail~' life. Spill.·, 24. 755. Wilder. D.G., POpl.· ..\1.1-1 .. &. Fr.\"Ino~·L'r. J. W. (1980). T tional lopogr,tphy of the s;Jcroiliac joint. Sp;/It'. 5, Biomechanics of th Cervical Spin Ronald Mosko Introduction Component Anatomy and Biomechanics Anatomy Osseous Structures Intervertebral Discs Mechanical Pmperties Vertebrae Intervertebral Discs Ligaments Muscle Neural Elements Kinematics Range of Motion Surface JOint Motion Coupled Motion of the Cervical Spine Atlantoaxial Segment Subaxial Spine Abnormal Kinematics Spinal Stability Occipitoatlantoaxial Complex $ubaxial Cervical Spine Applied Biomechanics Decompression Arthrodesis Cervical Spine Fixation Biomechanics of Cervical Trauma Airbag Injuries Whiplash Syndrome I· • :1( il1Lh"'d,~.,%>. mLdi. Summary References Introduction Knowledge of spinal biomechanics advanced exponentially during the second half of the twentieth centlll)'. A two-column model of the spine was describcd by Sir Frank HoldswOIth (1963) and, latel; a lhree-column model by Denis (1983), further refining the principle or spinal stability. The computer age has produced powerful methods For modern hiomcchanical modeling, the promise being the ability to assess the slnbility of n construct prior to implantation. Today. the application of cendcal biomechanical knowledge spans many inclustl'ies and supports improved medical diagnoses and treaLment that is more effective. Future technological and electronic advances will continue to build on basic biomechanical principles, many of \\:hich will be outlined in this chnpter. arc five t.'·pical cervical vertebrae, C3-C7, which arc sirnilar in structure and function. The spine has four curves when viewed in the sagiual plane. The celvical and lumbar regions arc convex anteriorly (lordotic), while the thoracic and sacra) regions are convex posteriorly (k)iphotic). The lordotic curves develop after birth as the infant's spine straightens out, which facilitates development of a bipedal posture. Although there is a Cl (atlas) C2 (axis) Facet joint Component Anatomy and Biomechanics C7 ANATOMY The exquisite design of the cervical spine uniquely contributes to the structure of the human body and profoundly enhances its function. The cervical spine supports the skull and acts as a shock absorber for the brain. It also facilitates the transfer of weights and bending moments of the hend. It protects the brainstem, spinal cord, and various neurovascular SU'UClures as they transit the neck nnd when they enter and exit the skull. The vertebral column also provides a multitude of muscle and ligamentous attachments for complex movement and stability. The neuromuscular control afforded by the muscle ntlnChmenls combined with the numerous articulations of the cervical spine allows for a wide range of physiological motion that maximizes the range of motion of the head and neck and serves to integrate the head with the rest of the body and the environment. The spine consists of 33 vertebrae divided into five regions: cervical (7) (Fig. II-I), thoracic (12), lumbar (5), sncml (5 fused segments), nnd coccvgeal (approximately 4). The two most cranial vertebrae, Cl (atlas) and C2 (axis). [Ire atypical. with a unique structural role in the articulation between the head and the cervical spine. The atlanta-occipital joint, bel ween CI and the oecipital bone of the skull, is also a functional part of [he cervical spine. There A Uncovertebral joint Intervertebral disc Tracheal air shadOw C7 B A, lateral roentgenogram of the cervical spine. Note the lordosis. The facet joints are aligned obliquely only to the frontal or <oronal plane; hence, their excellent visualization in the lateral view. a, Anteroposterior view of the <er· vical spine. 287 •> QJ- ._ harmonious progression of these curves from one to another, which may help distribute stresses and strains, injuries occur more commonly at the jllnc~ tional areas because of differences in the relative stiffness of each anatomical segment of the spine. The physical structure of the anatomical elements modulates from the cervical to the sacral region in relation to the segmental function. The lordosis in the cervical spine, like that in the lumbar spine, is maintained predoI1linantl:v b.v slightl)' wedge-shaped intervertebral discs that arc larger anteriorly' than posteriorly. In contradistinction, thoracic kyphosis is maintained largely.' by the vertebral bodies themselves; because the posterior portion of the thoracic vertebral body is larger than the anterior portion, there is a relative kyphosis of the thoracic spine. The conceptual biomechanical building block of the spinal colunln is the Functional spinal unit or motion segment. It consists of two adjacent vertebrae and the intervening intervertebral discs and lig~ aments between the vertebrae. These ligaments arc the anterior and posterior longitudinal ligaments; the intertransverse, interspinous, and supraspinous ligaments; and the facet capsular ligaments. As a re~ suit of the different functional demands of the various parts of the spinal column, segmental variation is expressed by! changes in the size and shape of the vertebrae, the anatomy of the discoligamentous structures, and the alignment and structure of the facet joints. Biological structures behave differently than do common engineering materials. Collagenous tissues exhibit both viscoelastic and anisotropic behavior. Viscoelastic properties are rate~dependent (timedependent) behaviors under loading that are seen in both bone and soft tissues; mechanical strength increases \vith increased rates of loading. Anisotropy is the alteration in mechanical properties that is seen \vhen bone is loaded along different axes. Anisotropic behavior occurs as a result of the dissimilar longitudinal and transverse microstructure. sagittal range of motion of the cervical spine. Cl-C2 articulation is the joint primaril.\-' respon for rotation in the cel'Vical spine. The atlas, or C 1, is a bon.\-' ring consisting o anterior ;:\lld posterior arch that is attached to two !;:\terallnasses of the atlas (Fig. 11~2). Th pcrior surfaces or the latent! rnasses, which crani<:dly <:md inward, form an <:\rticulation with caudally and clutward-facing occipital cond.\-'l the skull (Fig. 11-3). E.xtension ()f the occipit vical joint is limited by' the bon.\-' <:uwt()!ll.v; fle is lirnited prirnaril.\-' by ligarneluous structures tectorial mernbrane, and the longitudinal fibe the crucifonll ligarnent as well as by the post ligaments (Figs. 11 ~4 and 11 ~5). The anterio berek' on the <:llTh of C I serves <:lS <:In attachm for the longus colli muscle, a flexor of the n The posterior arch of the atlas is a modified lar that is grooved on its superior surface for the sage or the vertebral arteries as the.v enter the fClrarnen magnum after piercing tile post ntlanto~(lccipital nlelllbl"ane. Simi!;:\r to the occipitocel'Vical junction, the 11() intervertdJral disc betwecn Cl and C2. Stab at this levcl is thus predicated on intact ossco l1lentclus structurcs. The articulation between and C2 is primaril.\-' specialized for rotation. bod.\-' of C2 projects superiorly to fOlTn the odon process, or dens (Fig. 11 ~6). The projection o bod.\-' or C2 and the dens has a characteristic ob appearance on lateral cervical radiographs an Superior articular facet Orig tran liga Groove vertebral artery Osseous Structures The occipul-CI-C2 complex comprises the upper cervical spine and is responsible for approxilnately 40% of cenrical flexion and 60% of cervical rotation. The occipital condyles articulate with the slightly concave lateral masses of the atlas. The primary Illotion permitted by this articulation is flexion and extension, accounting for a large portion of the Facet for dens. on anterior arch 1cm I on posterior arch DmI'---- 1 Bony architecture of the atlas. Bar " "__" ~ 1 C111. Occipital condyle ,r--__ Occiput·C1 joint Dens Transverse process Cl-C2 facel joint Lateral mass of Cl Lateral mass ofC2 . The open-mouth radiograph demonstrates the occiput-(l and the atlantoaxial articulations. Note the symmetric spacing between the lateral masses of (1 and the dens. Asymmetry or widening of these spaces may occur after rotatory disturbances or fractures of the (1 ring . -------------------------- Anterior I Apical ligament 01 dens Posterior Superficial layer of tectorial membrane Tectoria! membrane ! Ant:~~~ I of alias I ",,","u ~-~ Dens of axis -t;~ffB~ll Transverse ligament ~~~~y.. . ~~ )'., -.:,::(-'\ ~.. Anterior I atlantoaxial I I @0 ligament Flexion interv~~:bral fibrocartilage Ilongitudinal Anlerior 1 ligament Body of C3 Posterior longitudinal ligament Median sagittal section through the occipital bone and the first three cervical vertebrae showing the articulations and surrounding ligaments. Extension Tracings of lateral flexion and extension radiographs ing the occiput, el, (2. and (3. The substantial relati motion between the occiput, (1. and (2 can be seen Large arrows indicate the direction of motion_ Small rows indicate that approximation of the posterior ele ments limit occipitocervical extension. In contradistin maximum flexion is controlled by tautness of the Iiga ments. Reprinred with permission {rom MoskoviclJ, R. & J D.A (1999). Upper (eNica/spine instrlJmellt<11ion, Spine: of tIle Art Reviews. 13(2), 233-253. disc, The Dens (odontoid process) \'crtebral hod,\" is o\'ul-slJaped transverse pnlcL'ssCS Spinal canal Superior Spinous facet process Foramen lransversarium cCryiC~IJ is widl.:l" mediolaLerally Lhan <:Illtcropostcriorl:", Inferior facet lcm or the cervical spin\,.' arl' uni in thaI Lhe,Y all contain a lr~Hlsn:rse foramen for passage of thc vertebral a ncr:', Tl\\"~ transv processes or the suba:xial c(~r\'ical \'('ncbr~\e h two projections, lhe alllerior and p(Jstcl'ior tu cit's. which serve as allilChl11\,.'nt points for ante and poslcriOl- lTluscles, respectively, The large a rior tubercle of C6, rt'fcrrcd to as the carotid tu cle, can be an important surgicallandmal'k, The peric>!' surract;.' 01' the lrans\,\,.'l'se proc~ss provid groove for the exiling nerve root. Each pedicle connecLs the vl".·ncbral bod~! to a eral mass, that portion of bone containing lh~ s rior and inferior facets, The facct joints n.::gulatc or mO\'l".'ll1ents the spine and pla~- a c.:ritical rol spinal slabilit~" Those 01' the cervical spine are orie The axis vertebra, or (2. The superior facet articulation permits multiplanar motion while the inferior facet is aligned to articulate with a more typical cervical facet, which is more constrained. There is a smooth surface on the front of the dens for articulation with the anterior ring of (1. Bar>:: 1 em. at approxirnatd~' 45'~ to the coronal pbne and arc cated in the sagittal plane (Figs, 11-8 and 11-9), orientation allows grc(l(('r anlounts of flexion t does latl".'ral bending or rotation in the cl.;.'lyical sp The facet joims resist most of lhe shear forces and often a helpful anatomical landmark. The dens ar- proximalL'!:' 16l.:!r; of lhe compressi\"c forces acting the spine (Adams & I-lutton, 1980). The laminae arise from the lalcral masses, Tht' IaLeral rnHSS\,.'S h important surgical implications in tht: subaxial ce cal spine because they conwin a relativel:' l ticulates with and is restrained within a socket amount 01" bone and are easil~' accessible for the pl • formed by the transverse ligament of CI and the anterior arch of C1. The transverse ligaments run from the anterior arch of C I. behind the dens, and prevent anterior translation of Clan C2. The other ligaments at the CI-C2 articulation are the alar, apical, and accessory alar ligaments. The alar ligaments, which are symmetrically placed on both sides of the dens, attach the dens to the occiput to prevent excessive rotation. The left alar ligament prevents right rotation and vice versa, To some extent, the alar ligaments also act to limit motion during side bending (Dvorak & Panjabi, 1987). The apical ligament also connects the dens to the occiput. Unlike the 1\\'0 most cranial vertebrae, the anatomy of the third through the sixth cervical vertebrae is similar (Fig. I 1-7). These fOUl" cervical vertebrae consist of a body, two pedicles, two lateral masses, two laminae. and a spinous process, The seventh cervical vertebra is slightly different in that it has a transitional form, It is called the vertebra prominens and has a larger spinolls process that is not bifid like those of C3-C6. The anterior components of a subaxial cervical motion segment are the vertebral bodies and the Inferio facet Groove for spinal nerve Super facet Pos tube Anteri luberc Uncinate process lcm Superior view of a typical cervical vertebra, representa of (3-(6 ((7. the vertebra prominens. diHers slightly in that it has a prominent nonbifid spinous process), C4 CS C6 C7 Orientation of the facets of a typical cervical vertebra in three planes. The facets are oriented at a 45" angle to the transverse plane and the frontal plane, and are at right angles to the sagittal plane. Y indicates the craniocaudal axis. z the anteroposterior axis. and x the mediolateral axis. Adapted '.t'lieh permission from White, A.A. 111 & Panjabi, Lateral photograph of a fourth and seventh cervical ver bra. The facet joint alignment is fairly close to 45° from the transverse plane. Note also the difference in size of the spinous processes, which are a reflection of the size and importance of the muscle attachments_ M.M. (1990). Cfinical Biomechanics oi the Spine. Pili/adelphia: I 1. B. Lippiflcort e menl of screws. as opposed lO the pedicles. which arc difficult to cannulate sarelv in the neck. The superior surfaces of the cervical vertebrae are saddle-shaped because of the uncinate processes, bony protuberances that aJ-ise from the lateral margins or the superior end plates (Fig. I 1-10). The uncovertebral joints Uoints of Luschka) develop during spinal maturation and pla~,/ an important biomechanical role with respect to kinetics and stability. • Intervertebral Discs The intervertebral discs are highly specialized st11.1Ctures that contribute up to one-third of the height of the vertebral column and fmm specialized joints between the cartilaginous end plates of the adjacent vertebral bodies. Activities such as running and jumping apply short-duration, high-amplitude loads to the intervertebral discs, whereas normal physical activity and upright stance resuh in the application of long-duration, low-magnitude loads to the disc. Discs are able to withstand greater than normal loads when compressive forces arc rapidly applied based on the biomechanical principles of viscoelasticity. This propeny protects the disc /i'om catastrophic failure until extremely high loads are applied. 1cm Anterior view of a sixth cervical vertebra. The shorr arro indicate the uncinate processes and the Ic;mg arrow ind cates the pathway of the sixth cervical nerve root. The facet joints are located posteriorly. Bar = 1 em. The nucleus puiposlis is centrally locmed within the disc and consists of almost 90% water in young individuals. The water content is highest at birth and decreases to approximately 700/0 as the disc degenerates with age. The rest of the nucleus pulposus consists of protcoglycan and collagen, which is exclusively type II collagen. Type II collagen fibrils are thought to be able to absorb compressive forces better than type I collagen fibrils. Prolcoglycans consist of a protein core auached to polysaccharide (glycosaminoglycan) chains. The pol.vsaccharicles are either keratin sulfate or chondroitin sulfate. The core protein, with its attached polysaccharides, is aggregated to hyaluronic acid through a link protein. The proteoglycans in th(: intervertcbral discs are similar to those in articular cartilage, except that the protcoglycans present in [he intcl\'ertcbral discs have shorter polysaccharide chains as well as shorter core proteins. The nucleus pulposlls contains more protcoglycan than does the annulus fibroslis. \¥ith increasing age and disc degeneration. the lotal protcoglycan content decreases. The annulus fibrosus is the OtHer portion of the disc. Its water content is slighl1y less than that of the nucleus, being only approximately 78£10 \vater in younger individuals. \-\lith age, the water content falls to approsimalely 70%, like that of the nucleus pulposus in older persons. The annulus consists of collagen that is arranged in approximately 90 concentric lamellar bands. The collagen fibers in these shccts nll1 at approximately 30 0 to the disc or 1200 to each other in the adjacent bands. This unique orientation confers strength to the annulus while permilling some nesibility (Fig. I I-II). The composition of the collagen in the annulus is approximately 60% type II collagen and approsimately 40% type I collagen. As the disc ages, the collagen undergoes irreducible cross-linking and the relative amount of type 1 collagen increases, replacing type [[ collagen in the disc. MECHANICAL PROPERTIES Vertebrae The mechanical properties of the bone and soft tissues diffe!: Strength and stiffness and the relation of stress to strain are the key IT'lCchanical properties. Stress-strain curves are used to determine the relative loading behavior of bone. Stress is the load per unit area of a perpendicularly applied load. Strain is [he change in length per unit of original length, usually expressed as a percenlage. :lj , I .ZiZY>_ .~" ). -~-"'>-- - ~." . ~~~~~~~~?-I) ,./ pulposus Nucleus ( A I Disc - An fibe ' ~~~ I Schematic drawings of an intervertebral disc showing criss-cross arrangement of its fibers. A, Concentric lay of the annulus fibrosus are depicted as cut away to s the alternating orientation of the collagen fibers. B, layers of the annular fibers are oriented at a 30'" ang the vertebral body and at 120 angles to each other. Acf,;pred '.-Vlil! permission l,om W;uw, AA. 111 c~ Pan,iabl. M 0 (l990} Clinical B!Drn~'dl,Hll(<' of the Sp:rl(' Pili/dcle/ph/a: I Lippincor[ ! • Cortical bone is stilll.'!" than cancellolls bone can withstand greater stresses before failure. \V the strain in "h'o excl.;cds 2t:'(; of the originallen conic::d bone l"1"<.tclLtrcs, but c<.tllcdlous bone withSland somewhat greater strains before fra ing. The gr('.. ltl.~r abilit~· to withstand strain i cause of" the stl"uctUrt: of cancellous bon ... : its p it,v varies from 30 t() 9CY>" compared with con bone with valuL's or 5 to 30%; (C~\rtcr & H 1977). Vertebral corn pression strength incre from lhe upper cervical lo th(' lower lumbar k\ Bone strength decreases with age as a resu [he de\'c1oplllcnt of osteoporosis. The mineral tenl (If' vertebr~\e decreases with increasing age relatively con::'1ant rale (Hansson & RODS, Hansson et 411., 1980).;\ 25 cj{; «(..crease in osseou suc results in a 1110re lhan a 500'0 tkcrcasc in strength of the \"Crlchrac (Bell. 1967). BeC'lllSC cortical shell of a venchra is responsible for on1 proximatcl.v 1(Y;',; 01" its strength during com 'sion, good quality c~lIlcellotis bone is criticall~ !Jorlant (McBroom e:t aI., 1985). \Vht,:.~n bone is loatk.'d in "h'o, contraction o musclt:s attached [0 the bone can alt0r [he s distribution in the bone. Bending moments are applied to the vertebral bodies during motions. During lie.""", tensile stresses are applied to the posterior and compression to the anterior cortex of the body. To perform lifting tasks, the back are required to develop considerable forces IS"hltltz et aI., 1982). Stresses in a typical cClvical change from tensile to cornpressive in a region approximately 0.5 to I cn1 anterior to the pos~ longitudinal ligament (Pintar et al., 1995). As is weaker and fails earlier in tension than in compression, posterior paraspinal muscle contraccan decrease the tensile stress on bone by! producing a compressive stress that reduces or neutralizes the posterior cortical tensile stresses, allowing the vertebrae to sustain higher loads than would othenvise be possible. However, bone will often fail prior to damage occurring to the intervertebral disc under compressive loading. Finite element modeling of the cervical spine indicates that the increase in end plate stresses may be the initiating factor for failure of this component under compressive loads ('loganandan et al., 1996a). Intervertebral Discs Intenlertebral discs exhibit viscoelastic properties (creep and relaxation) and hysteresis (Kazarian, 1975). Creep occurs more slowly in healthy discs than in degenerated or herniated discs, suggesting that degenerated discs are less viscoelastic in nature (Kazarian, 1972). Ligaments Clinical stability of the spine depends primarily on the soft tissue components, especially in the cervical spine. The spinal ligaments are functional rnainly in distraction along the line of their fibers. Ligament strength and limited extensibility help maintain stability, especially around the craniocervic'll junction. The alar ligaments have an in vitro strength of 200 N, and the transverse ligaments have an in vitro strength of 350 N (Dvorak et aI., 1988a). The strength of the ligaments is related to both the anatomical demands and the flexibility re~ quired, \vhich is a classic example of form following function. The ligaments all have high collagen content except for the ligamentum flavurn, which is exceptional in having a large percentage of elastin. The ligamentum flavum is under tension even when the spine is in a neutral position or somewhat extended, and th pre-stresses the disc to some degree and provid some intrinsic support to the spine (Nachemson Evans, 1968; Rolandcl~ 1966). The elastic properti also assist in limiting the in\varcl buckling of the ligaments during extension, which could potential compress the neural clements. Muscle Muscular strength and control is imperative maintain head and neck balance. In the cervic spine, muscle strength also has a role in reduci stresses on bones. Bending n10ments are applied the vertebral bodies during various motions. Duri flexion, tensile stresses are applied to the posteri cortex and compression to the anterior cortex of t vertebral body. Substantial loads on the eel-vie spine have been calculated during neck flexion, pa ticularly! in the lower cenrical motion segments. Harms-Ringdahl (1986) calculated the bendi moments generated around the axes of motion the atlanta-occipital joint and the C7-TI moti segment in seven subjects with the neck in five po tions: full tlexion, slight flexion, neutraL head u right with the chin tucked in, and full extensio The load on the junction between the occipital bo and C I was lowest during extreme extension (ran ing from an extension moment of 0.4 Nm to a fle ion moment of 0.3 Nm). It \vas highest during e treme flexion (0.9 to 1.8 Nm), but this was only slight increase over that produced when the ne was in the neutral position. The load on the C7motion segment was low with the neck in the ne tral position but bccame even lower when the he \vas held upright with the chin tucked in (rangi from an extension m0111cnt of 0.8 Nm to a flexi moment of 0.9 Nm). The load increased somewh during extreme extension (ranging from 1.1 to 2 Nm) and substantially during slight flexion (reac ing 3.0 to 6.2 Nm). The greatest loads \vere p duced during extreme extension, with momen ranging from 3.7 to 6.5 Nm. In the same study, surface electrode electromyo raphy was used to record activity over the erec spinae muscles of the cervical spine with the neck the same five positions described above. Intere ingl}', the values obtained showed very low levels muscle activity for all positions, even during treme flexion in which the flexion moment on t C7-Tl motion segment increased n10re than thr fold over the neutral position. The fact that the e1 tromyographic levels over the neck extensors were low in this and other studies (Founlain et aI., 1966; Takebe et aI., 1974) suggests that the nexing moment is balanced by passive connective tissue SlILICtures, such as the joint capsules and ligaments. This phenomenon is seen in many other joints in which passive support is provided by the ligaments. The values for the moments computed by Harms-Ringdahl (I986), howcvcr, are appmximately 10 % of the maximal values measured by lvloroney and Schullz (1985) in 14 malc subjccts who resisted maximal and submaximal loads against the head while in an upright siuing position. The mean maximal voluntary moments were 10 Nm during axial rotation or the cervical spine, 12 to 14 Nm during flexion and lateral bending, and 30 1m during extension. Calculation of the maximum (compressive) reaction forces on the C4-C5 motion segment ranged from 500 to 700 N during flexion, rotation, and lateral bending. and rose La 1,100 N during extension. AnteroposLerior and laleral shear forces reached 260 Nand lION, respectively. Calculated moments and forces generally correlated well with mean~ll1easured m:voelectric activities at eight sites around the perimeter of the neck al the C4 level. Muscles playa cr-iLical role in basic postural homeostasis, as can be observed in both historical and present-day clinical settings. In unique observational studies in lhe 1950s of severclv affecled poliomyelitis patients, improvements in respiratory assistance for those with respiratory paralysis led to higher survival rates and a large number of patients who sustained complete paralysis of the cervical musculature. Patients with completely nail cervical spines were unable to support their heads unless adequate support was provided and actually remained in bed despilc the good function of their cxtrcmities (Perry & Nickel, 1959). Similarly, severc cervical kyphosis occasionally is seen in elderly patients \vho do not have an obvious stntetural ctiol~ ogy when investigated radiologically. Some of these patients were found to have marked cenrical extensor muscle weakness that has been attributed to senile cervical myopathy (Simmons & Bradley, 1988) (Fig. 11-12). Neural Elements Biomechanics of the neural elements have noL been as well studied as the biomechanics of the osteoligamentous vertebral column, but our know!- lateral cervical radiograph of a 68-year·old 1Noma presented with severe torticollis. She denied any h injury and did not have evidence of a suuctural ve abnormality, infection. tumor, or inflammatory dis Pseudosubluxations (arrows) are evident subaxiall consequence of the marked kyphosis. Her neck wa twisted only because she had a severe cervical flex formity. and she was unable to see forward excep turning her head to one side. She was neurologica tact. It was possible to extend her neck to a relativ tral position using gentle traction. Following post sion from (2 to (7, she returned to a normal inde life. Muscle biopsy was consistent with senile myo Reprinred with permiSSion from tvioskovlCh. R. (1997). mSiabihw (rheumarold. d\·varfism, degeneraCtV£!. Oihers Bridwell &- RL DeW,11d fEds), The Te;(lbook of Spln,ll S (D.o. 969- i009). Phdacie!pfJJd. Lippincorr-R,wfdl Publishe edge base is growing. To date, certain bas meters have been established. The cervica undergoes significant changes in length f1cxion and extension (Breig el "I" 1966 1960). Thus, while there is some longitudin ticity to the spinal corel. it tolerates axial tion poorl~". It is the translator~' forces llw cally result in neurological in.illr~·. A comp tolerance betwecn 2.75 and 3.44 kN is cs for the adult c~rvical spine before significa rological injury occurs (i'Vlyers &. \Vink 1995). Spinal cord injuries can also result from extreme or sudden flexion-c:\tcnsion movernCl1ts, especially in the face of a $hallow spinal canal. I-lead Oexion alone has been shown to result in significant increases in the illlramcdullary spinal cord pressure in dogs (Kitahara et aI., 1995). Neurological injuries result from anteroposterior compression of the ,i",··· .. corel and arc I110re common if the spinal is stenotic. Flexion motions can result in inwhen the spinal cord makes contact with cervical osteophytes, and cxtcnsion mOl ions may resull in a pincer-like compression of the cord between (anterior) ostcophytes and (posterior) invaginated ligamentum Oavum. Anterior or central spinal cord injuries may ensue. Although a diagnosis of spinal stenosis may be made based on the absolute size of the spinal canaL imaging of the neuraxis itself may be of greater value. Contrast-enhanced computerized tomography, myelography, and magnetic resonance imaging can demonstrate actual impingement or distortion of the spinal corel. Studies performed in flexion and extension may enhance the value of the information by demonstrating the contribution of any dynamic soft tissue component to the impingement. The exact size of the bony cervical spinal canal and the vertebral body \vas measured in 368 cadaveric adult male vertebrae (Moskovich el aI., 1996). This sludy used well-validated parametric statistical methods to determine that the mean sagittal diameter of the spinal canal for C3-C7 was close to 14 mm (14.07 ± 1.63 mm; N ~ 272) (Figs. 11-13 and 11-14). The mean ratio of the sagittal canal diameter to the vertebral body diameter (canal to body [cob] ralio) was 86.68 ± 13.70. Thirly-one percent of subaxial vertebrae would be diagnosed as having spinal stenosis if a c-b ratio of less than 80% was considered abnormal. Another study also found a high false-positive error rate for the e-b ratio, with 49% of 80 asymptomatic football players having a c-b ratio of less than 80O,b at one or morc cervical levels (Herzog et aI., 1991). Yet another group evaluated the reliability of the c-b ratio using plain lateral radiographs and CT scans (Blackley el al., 1999). Results confirmed lhat a poor con-elation exists between the true diameter of the canal and the ratio of its sagittal diameter to that of the vertebral body. The variab.ility in anatomical morphology means that the use of ratios h'om anatomical measurements within the cervical spine is not reliable in delermining the true diameter of the cervical canal. Spinal cord injul)' without radiographic abnormalities (SCIWORA) have also been described, N 9i . ...l2 ::"':3 " C1 •• 3. Oa N , ~. a_ I': 17 20 23 26 29 mm "e ~ '- . 10 12 14 16 18 20 mm N e. 4; 0.__ 10 12 14 16 18 20 mm N 16: ".Il C4 e· ,. ____ .. 10 12 1': 16 18 20 mm O~. N 1. ':.a. C5 ., 0.. . 10 12 14161820 N 16 ".&.. rnfn C6 e .;. 0.. .___ 10 12 14 16 18 20 mm N : ; ...t..-. C7 81 4' ot. _ 10 12 14 16 18 20 mm Histograms of the C1-(7 sagittal canal diameters. Apart from the (1 plot, the same scale is used on all of the horizontal axes so that the distributions of the diameters can be compared, Reprinted with permission (rom Moskovich, R.. et at. (996). Does rhe cervical canal [0 body rario predict spina stenosis? Bull Hasp Joint Dis, 55, 61-71. especially in children (Dickman el aI., 1991; Osen bach & Menezes, 1989: Pang & Pollack, 1989). Th etiology of such injuries is unknown, but one mech anism may be longitudinal traction. The unusuall elastic biomechanics of the pediatric bony spine a lows deformation of the l11usculoskeletal structure beyond physiological extremes, perm·itting direc cord trauma followed by spontaneous reduction o the bony spine (Kriss & Kriss, 1996). The isolate spinal cord resists tension poorly; axial tensil rotation or translation of a body ~l!ong one a.\ consistently' assclciated with a simultaneous tion or translation ~dong another axis, the mot are coupled. Coupled motions are normally' pressed as displacements in the X, Y, or Z d tions and rotations about the three orthog axes. Further analysis involving whole spine more complex, and while it has y'iclcled interes results, motion segment analysis remains im tant for basic understanding. Posterior longitudinal ligament Superior facet vertebra, which was not part of the study detailed in the Spinous process text. The anteroposterior diameter of the spinal canal measured 13.96 mm in this specimen. • forces to failure For three adult spinal cord specimens arc reported to be 278 N ± 90 (Yoganandan et aI., 1996b). Lower forces may result in direct neural injury or vascular disruption. Kinematics Kinematics is the studv of motion of rigid bodies without taking into consideration other relevant forces. Kinematics of the spine descrihes the physiological and pathological motions that occur in the various spinal units. The traditional unit of stud)! in kinematics is the motion segment, or the functional spinal unit. As described earlier, each motion segment consists of two adjacent vertebrae and their intervening soft tissues (Fig. 11-15). Basic biomechanical testing involves the application of forces to a vertebral body and the subsequent measurement of the movemcnts that occur (Fig. 11-16). Movements can be either rotational or translational. A degree of freedom is def-lned as a motion in \vhich a rigid body can cither translate back and forth along a straight line or rotate around a particular axis. Thus, each vertebral body may either translate or rotate in each of three orthogonal planes, for a total of six degrees of freedom (Panjabi et aI., 1981) (Fig. 11-17). When either l\v.-o-od{L Capsular ligament, Axial computerized tomographic scan of a sixth cervical "'" Interspinous and, supraspinous ligaments II Anterio . / longitud ligame Interverte disc J I I p . . ostenor L' fl Inferior facet J longitudinal Intertransv I ligament Igamentum avum ligamen // // I A Superior facet I I Capsular ligament Anterio longitu ligame Spinal canal "- ' "Spinous process ~ "' Interverte disc Interspinous ~ and supraspinous ligaments ""_~/ Ligamentum flavum ~Ls_te_r_io_r Interlransv ligamen Transv forame J Posterior '!Iongitudinalligament I A_n_t_e_ri_o_r Schematic representations of a cervical motion segmen composed of two typical cervical vertebrae ((4 and (5) intervertebral disc, and surrounding ligaments. The bro line divides the motion segment into anterior and pos components. A, Lateral view. B, Superior view. Adapied permission from White, A.A. 1/1, Johnson, fUvr, Panjabi, lvUv al. (7975) Biomechanical analysis of clmica! stability in rhe c cal spifJ(l. (lin Orthop. \09,85··-96 n , , 9 : i I c..; Diagram of a test rig to evaluate a functional spinal unit using videophotogrammetry. This technique facilitates accurate measurements of motion without the measurements themselves having an effect on the displacements of the mobile vertebra. A. light-emitting diodes (LEOs); Band C. Guide bars for application of tensile and compressive fOfces. D. Pulley for application of torques. Weights are at· tached to guide wires (w), which go around pulleys to produce torque on the upper vertebral body (v). n. intervertebral disc; 9. acrylic cement that attaches lower aluminum plate (m) to test rig, which is rigidly bolted to the support frame. The upper plate (p) and upper vertebral body (v) are the moveable elements to which the loads and torques are applied. LEOs are rigidly attached to the upper plate and their movement is recorded by two video cameras. ;·Fy c1 Reprinted with permission {rom Raynor. R.• Moskovich. R.• Zide/, P. et a/. (1987). Alteration in primary and coupled neck motio1J5 l__ I m aled is the fact that a considerable amount of fle:x ion and extension occurs at the CI-C2 articulation 5 to 20° or flcxion and extcnsion occur with mean of 12" (aetive) to 15° (passive) (Dvorak et aI., 1988b) Approximately 90° of axial rotation takes place in the subaxial eel"vic'll spine (C3-C7), about 45° to cach side of ncutral. Even greater lateral flexion i possible: approximately 49° to each side of ncutral giving a total of about 98°. The range of flexion and extension is approximately 64°: about 24° of cxten sion and 40° of flexion. The motion in each plane i fairly evcnly distributed lhroughollt the motion segments (Lysel!, 1969). Thc mean total range o anteroposterior translation in subaxial spinal mo tion segments is 3.5 ± 0.3 mm divided unequally 1.9 mm for anterior shear and 1.6 111m for posterio shear. Lateral shear loading results in a mean tota range of lateral motion of 3.0 mOl :!:: 0.3 mm, di vided equally between right and left; tension result in 1.1 mm of distraction and compression, 0.7 mIl of loss of vertical height (Panjabi et aI., 1986). The great Ilexibility of the eel'vical spine allow the head to be positioned in a wide variety of ways a_f_,e_'_fa_,_e_,e_,_,o_n_'Y_._N_,e_"_,o_S_"'_9_e_ry_,_2_1(_5_),_6_8_'_-_68_7_. _ l1Y RANGE OF MOTION Measurements of cervical range or motion are based on radiographic studies or postmortem investigations. Inclinometers and various optoelectronic and electromagnctic dcvices L1sed clinically for noninvasive evaluation of cervical spinc motion are not as accurate; in particular, coupled motion is poorly quantified (Roozmon et aI., 1993). The established range of active axial rotation to one side at CI-C2 is 27 to 49° (mean = 39°); passive rotation is 29 to 46° (mean = 41°) (Dvorak et aI., 1987; Dvorak et aI., 1988b; Penning & Wilmink, 1987). These measurements account for approximately 50 r1'o of the lotal cervical rotation. Another stereoradiographic study of neck motion in adult men found a mcan of 105° axial rotation between the occiput and the C7 vertebra. Seventy percent of the total axial rotation OCCUlTed between the occiput and the C2 vertcbra. Eaeh motion segment between the C2 and C7 vertcbrae avcraged fTom 4 to 8° rotation (Mimura et aI., 1989). Less well appreci- +Fz ~ 0i---le. ~Fx A vertebral body showing the three primary Cartesian axes, x, y. and z. Along each axis a positive force, ~ F. is de noted by the direction of the arrow. The curved arrows indicate the direction of a positive torque. +..... Reprinted will permission from Raynor, R.• Moskovich. R., Zidel, P., el af. (/987). Alteration in primary and coupled neck motions aller facete<:tomy. Neurosurgery. 21 (5). 681-687. ; i l-",~~=~,...,.-.,-,~~-=-=-=--~;==;r.----,.--------:=,,~, '.""'" .-.&f "" .. ",-"_",,,,,""£1 - .~"""'"- .-~- ~.~. •__ ' -, permitting one, with equal ease, to gaze at an air~ plane overhead, glance over one's sho111dcI~ or look for an object under a table. An analysis of the com~ bined motion of the cervical spine Llsing an eleclrogoniometer produced a remarkably large range of motion (Feipel et aI" 1999): 122' ;;; 18' of nexion and extension, 144°:!: 20 0 of axial rotation, and 88 0 ::': 16° of lateral nexion, All primal)' motions were reduced with age. Sex had no influence on cervical motion range. The active range of cervical spine motion required to perform daily functional tasks was studied in healthv adults (Bennett et aI" 2000), A cervical spine range of motion device was fastened to the subject's head with a Velcro strap and a magnet was placed on the patient's shoulders to calibrate the instrument for measurement of cervical rotation and motion, Of the 13 daily functional tasks performed, tying shoes (flexion-extension 66.7°), backing up a car (rotation 67.6°), washing hair in the shower (flexion-extension 42.9°), and crossing the street (1'0· tat ion head left 31.7° and rotation he4:KI riglll 54.3°) required the greatest active range of motion of the cervical spine. Of interest, several tasks were not found to produce the degrees of lnotion expected, and these included reading a newspaper (19.9° Ocxion-cxtcnsion), writing at a table (26.2° flexionextension), and reaching 1'01' objects overhead (4.3° flcxion·extension). Side·bending was not found to be a significant movement in completion of the tasks but was coupled \vith rotation in one of the tasks (looking left and right to cross a street). SURFACE JOINT MOTION The motion between the joint surfaces of two adjacent vertebrae Illay be analyzed by means of the in· stant cenlct· techniquc of Rculeaux, described in Chapter 6. The method may be used to analyze sur· face motion of the cen:ical spine during flexion· extension and lateral flexion. In a normal cel'vical spine, the instant center of flexion-ex lens ion is located in the anterior part of the lower vertebra in each motion segment. Instant center analysis indicates that tangential motion (gliding) takes place between the Facet joints as the cel"ical spine is flexed and eXlended (Fig, 11-18), A consequence of these rnotions is that the size of the intervertebral foramina increases with flexion and decreases wilh extension (Fielding, 1957), These alterations have been quantified in a cadaver study that found there were statistically significant reductions of 10 and 13% in foraminal diameter, at 20 and \ ! II C5 Analysis of surface motion of the facet joints of the motion segment during flexion-extension. The sche drawing represents superimposed roentgenograms motion segment in the neutral position and in sligh ion. The upper vertebra ((4) is considered to be the ing body and the subjacent vertebra ((5) is the bas bra. Two points have been identified and marked o moving body in the neutral position (sofid outline o and the same two points have been marked on the roentgenogram with the motion segment slightly f (dashed outline of e4). li"es connecting the twO se points have been drawn and their perpendicular bi have been added. The intersection of the perpendi sectors identifies the instant center of motion (larg dot) for the degree of flexion under study. The per ular bisector (arrowed fine) of a line drawn from th ter of motion to the contact point of the facet join faces indicates tangential motion. or gliding. .-,-~~~~~--~-~~- 30° of extension, r('spccti\'c1~'. Convl'rsel~", in lh~r~ \\'cre statistically significant increases o 10% at 20 and 30° 01" fll.:'xion, l"l.'spcclin~ly (Yoo 1992), One p""clical application of this dala to cervical collars lIsed for the rdief of neck Conventional foam collars tend to place pati slight extension, which IlU\.\' aggravatc the toms. Turning a foam collar around. with thl.:' and narrow parl anterior, puts (he neck in flexion, which may increase the size of the in tcbral foramina and thereb~1 relieve some pressure on an inflamed nerve root. The instant center motion of' the cClvica may be displaced as a result of patho processes such as disc d~gcncrati()n or ligam pairment. In such cases, instant center analys reveal distraction and jamming (compress the facel joint surfaces during flexion·extens sl~ad or gliding (Fig, 11-19), or C4 AnI. Post. C5 I Schematic drawing representative of a roentgenogram of the C4-CS motion segment of a patient injured in a rearend auto collision. The instant center of flexion·extension at this level (represented by the large solid dot) has been displaced from the anterior to the posterior part of (5 as a result of the injury process, which impaired the ligaments (compare with Fig. 11·18). The analysis of surface motion shows compression and distraction of the facet joints with flexion and extension. • there are 2° of coupled axial rotation for every 3 lateral bending, which gives a ratio of 2:3 or 0".67 C7, there is lOaf coupled axial rotation for ev 7.5 0 of lateral bending, which gives a ratio of 2: 1 0.13 (White & Panjabi, 1990). Results of finite ment rnodeling indicate that the facet joints and covenebral joints are the major contributOl"S to c pled motion in the lower cervical spine and thal uncinate processes effectively reduce motion c pling and primal)1 cer'vical motion (in the same rection as load application), especially in respo to axial roLation and lateral bending loads. covcrtcbral joints appear to increase primal)' ce cal motion, showing an effect on cervical mot opposite to that or the uncinate processes (Clau et aI., 1997). Coupling or nexion-exLension w transverse translation may be visualized roentge graphically (Fig. 11-23). During flexion, the VC bral body normally shifts rorward: the racets g up and over one another with pseudosubluxali Changes in normal coupling patterns occur foll iog palhological changes or surgical intervention Tensile load application to the cervical sp occurs in therapeutic traction, but more co monly so during trauma. Deployment of pass COUPLED MOTION OF THE CERVICAL SPINE Atlantoaxial Segment i: The coupling characteristics of the atlantoaxial spinal motion segments arc particularly important, as this area of the neck is extremely' mobile. The dens is constrained within the osteoligamentous ring of the atlas, causing the C l-C2 lateral masses to articulate similarly to the condyles of the knee, with some sliding and rolling during Oexion and extension. The instant centers of both rotation and n~xion-extension lie in the center of the dens itselr. Rotation at CI-C2 is coupled with both vertical translation along the Y axis (Fig. 11-20) and a degree of anteroposterior disphlcenlent (\,Verne, 1957). This implies that the CI-C2 joint is most stable in the neutral position, and, if rotated, allempts should be made to return it to the reduced position when performing an arthrodesis. Subaxial Spine The coupling patterns in the lower cervical spine arc slIch that in lateral bending to the left, the spinous processes move to the right, and in lateral bending to the right. they move to the left (Lysell, 1969: !vloroney et at.. 1988) (Figs. I 1-21 and 11-22). AL C2 . . _----~" [ ~' \ \ A \1' B Coupling of rotation and axial translation is depicted schematically. A, (1 and (2 are in the neutral position. S, C1 rises fractionally on C2 (arrow) as the head is rota away from the midline. Adapted with permission [rom of Fielding, J.W (1957). Cineroentgenography the normal cervical spine. J Bone Joint Surg. 39A, 1280-/288. -:__ I' vehicular n:.'S!l"airH s.\'stCtllS, such as airbags, induce tensile forces in the neck. Isolatcd vertebral discs fail at 569 N 54. and intac man cadaver cervical spines rail at 3373 N , (Yoganandan cl aI., 1096h1. Active muscular traction, 110\\'C\'i..'r. is liki.'I~' 10 raise these fi cons i dl'J"~1 hi v. Abnormal Kinematics Left lateral flexion Neutral Righi lateral flexion Abnonnal kinematics gCl1crall.\" refers to l'.'\c motion within rUl1etional spinal units; howeve norrnal kinClnatics Ina.. . · also refer to at.\vica Coupled motion during lateral bending is depicted schematically. When the head and neck are flexed to the left, the spinous processes shift to the right, indicating rotation. The converse is also illustrated. Adapted with permission from White, AA 1// (~Panjabi, MM (1990). Clinical Biomechanics of the Spine. Philadelphia: 1.8. Lippincott. A A, This diagram illustrates some of the other coupled motions, which occur in response to a torque (liZ) about the Zaxis (lateral bending). lateral translation (Rx) and vertical motion (Ry) occur, as well as horizontal rotation (l!lY), which terns of motion such as abnormal coupling or doxicalmotion. Par~\do.\icallllotionis seen whe overall pattern of motion of one ~lSPCCI of the is in one direction while the local pattern is th posite. For instancc, parado'\ical llc'\iol1 is when llc'\iol1 occurs at a single functional s unit, although the SpillL' as a whole is (''\te B results in motion of the spinous processes to the right left. B, The subject is bending to her right, demonstrat the large range of normal cervical motion possible (ap mately 50°). I I Coupling of flexion-extension with transverse translation of the cervical spine is visible roentgenographically. A, Ouring flexion the vertebral body shihs forward (small white arrow); the facets glide up and over one another with moderate subluxation at full flexion (large white arrow). Up to 2.5 mm of transverse translation may normally occur at the (1-(2 articulation during flexion-extension; no translati evident in this example (black arrow). B, During extensio the reverse occurs, and the spinous processes limit motio they touch at full extension (arrow). The size of the inte tebral foramina increases with flexion and decreases wit extension. • These types of abnormal motions describc a pattern of mOVCIl'lCllt known as instability. Spinal Stability The concept of spinal stability is an intriguing and sometimes confusing notion. Medical practitioners are frcqucnlly asked to look at a sel'ies of radiographs and make a dctermination whether the spinc is stnblc. Exnclly what is stability, how is it determincd, "'ld what happens if it is not present? The term spinal stability has acquired diffcrent mcan~ ings, depending on the setting in which it has been used. White and Panjabi describe the term clinically as the loss of the ability of the spine undcr physio- logical loads to maintain its pattern of displacem so that there is no initial or additional ncurolog dcficit, no major defornlity,' and no incapacita pain (1990). Using this definition, physiolog loads are those incurred during normal activity, pain is not felt to be incapacitating if it can be c trolled by non-narcotic drugs. Stability is detcrmined by many factors. There different anatomical considerntions in different gions or the spine. Certainly, ligamcntous anato plays a large part in the stability or the spine, but muscular and bony clements of the ~pine also important roles. Instability can be analyzed by considering kinem instability and structural or component instabi Kinematic instability focuses on either the quantity of motion (too much or too little) or the qualit..., of motion present (alterations in the nOimal pattenl), or both. Component instability addresses the clinical biomechanical role of the various anatomical components of the functional spinalunil. In this type of instability, los..o:; or alteration of various am.llomical portions detennines the presence or instability (Box II-I). Sir Frank 'Holdsworth's account of a simple twocolumn concept of spinal stability provided a constructive basis for describing and analyzing the basic biornechanics or the spine: The synal1hroscs bClwcen Ihe \'cl1cbral bodies rely for theil' slabilit~, upon the immensely slrong annulus fi~ brosus. ThL: clim'lhrodial apophys(.';:d joints arc slabilised by the capsule, by rhc interspinous and sllpraspinous Iigamenls and the ligamcnta Ila\"a. This group of ligaments [ call the 'postedol' ligament complex.' It is upon this complex (hal (he swbilily of lh~ spine largdy dq:>ends (Holdsworth, 1963). Subsequently, Denis (1983) described a classification system for thoracolumbar fractures that can also be applied to a biomechanical analysis of spinal stability. In this description, the spinal elements are divided into three regions that form three spinal columns: I. The anterior column consists of the anterior longiludinalligament, the anterior annulus fibroslls, and the anterior half of the vertebral bock 2. The middle column consists of the posterior longitudinal ligament, the posterior half of the venebrnt body, and the posterior nnnulus librosus. 3. The posterior column consists or the pedicles, facet joints, lam inn, spinaliS processes, as well as the interspinous and supraspinous ligaments. 4. The anterior and middle columns form the primary weight-bearing column of the spine, with the posterior column providing the guiding and stabilizing elements. Occipitoatlantoaxial Complex The transverse ligament of the atlas completes the socket into which the dens is inserted. The ligament allows the dens to rotale, but limits its anterior translalion. The ligament is inelastic and will not permit more than 2 to 3 mm of anterior subluxation of the first on the second vertebra (Field- Conceptual Types of Instabilit _ Kinematic Instability Motion increased lm.tantaneous ax(,>s of rotation altered CDlipiing characteristics changed Parddo;O:I(d\ mOllon present Component Instability i"raun1C1 TUIl10! Surgery Degenerative changes Developmental changes Combined Instability K:nemiitK Cornponen! ing I..'t aI., 1974). Antcrior displaceml... nt or CIo or 3 to 5 mm is usuall:' indil';:ltive or a ruptu till' tr;:\Ils\\:rse ligament. while disphlCl'Il1ClHS o 10 mill suggest nCCl.:'Ssol'\ lig<.\JlK'llt d~l11lage; plncC'tlll'lll greater than !O lllill occurs with ru or all lhe ligal1lL'lllS (Fil.'lding et :.II .. 1(76). An translations or displacemcnts of Cion ('2 :J, sessed I'adi()graphicall~' by measuring tilt..' dis from lI1L' anteriur ring of the alias In the back o dc'J1' (atlaJ1todent,,1 intcly,,1 or ADI) (Fig. 1 (CasL' Stud.'" 11-1). Poslt.:'rior subluxation of lhe C<.lll onl~' OCCUI" if lhe dens is fractured or if tllL all OS-OdOlllOidL'ulll or hypoplastic dellS. SOl11e diseases can wcakc'll or dL'S{nl~' th\.' n.'l'sc ligament. :\,10S1 nOlahl...-. s.vllo\'itis in rhL !oid arthritis can crL'~ltc a pannus. which hel d(;SII'O,\' the atlantoa.\:ial ~1J'liClllalion as well a ll'ansn'!'SL' Iigamcnt (Figs. 11-25 and 11·26L tients with Do\\'n s.\'ndroll1c <.l1'C also susceptib \\,c;:lketlcd tmnsvcrsc ligaments and must be full." assessed clinicall.\" and radiographicnll .., b bdng allowed 10 pankipalc in sporting events as the Special OI~'l1lpics. Stccrs "Rule of Thirds" is a guidt: 10 the am or allanloaxial displw':cl11cl1t that can OCCllr b spinal cord comprcssioll ensues (Ste~1. 1968) illt~rnal anteropO-SIL'l"inr diameter pl'o.,\jJl1alel~· 01" thL' atlas 3 cm: of thal. the dens occupie proximately I cm and thl' spinal cOI"(1 approxim I cm. kadng I CI11 of space for soh tisSlh.' an normal mo\"cnwnt to OCClll" (Fig. 11-27). subaxial Cervical Spine (1963) stated that the musculature of the spine and the intervertebral discs were the most significant anatomical structures is providing cer'vie'll stability. Holdsworth (1963) emphasized the imparlance or the supraspinous and interspinolls ligaments as well as the nuchal ligament. The nuchal ligament is thought to playa major role in Atlantoaxial Instability Without Fracture f\:},9;Ye'a.r~old woman had a traumatic injllry as a result .'co"J;t!.of a forced flexion movement in a car acodent. Conhinuous and severe neck pain occurred after the accident. (s.,he visited the emergency room and after a careful ex· ~~~~rr,ination and x-ray evaluation, an anterior displacement ~'ZR.f:Cl on (2 was discovered (Fig, cs11-1·1). ~:(E:::;A severe amerior dislocation of the atlas on the axis ~~:/~~as confirmed /':"__i"::",'" '.,1' SAC after measuring the atlantodental interval . (~'rnrn). ,~or this case, no fracture of the atlas or the axis \i\I~~"~:7,:te~,t,~d and thus a defiCiency of the transverse ligalfIent,:'r;na'y, be assumed. Clinical instability of the spine depends mainly on the soft tissue components. The cervical spine is a very mo· ,bile area, especially at the atlanlOaxiallevel. Cervical subluxations and dislocations resulting from injuries of the osteoligamentous complex affect spinal stability and rna~ility at the cervical area. In addition, it may narrow the spin,al canal and cause neurological impairment. Pure lig. am~ntOtJs injuries (as in this case) are less likely to heal and become stable, and therefore surgical procedures are lik,ely to be considered. SAC I I ~ I I, L.- _ The atlantodental interval (AD!) is inversely related to the space available for the spinal cord (SAC), which is denoted by the dotted lines. Anterior atlantoaxial subluxation causes a reduction in the SAC. Normal measurements for the ADI are less than 3 mm in adults or 4 mm in children. Reprinted V'lirh permission from Moskovich, R. (1994). Atlanroaxial instability Spine: Stare of the Art Reviews. 8(3), 531-549. -rease. Study' Figure 11-1·1. lateral radiograph demonstrating . ,,~,?j}.n:~,~!~?sed atlantodental interval of 6 mm after trauma. proprioception and correct functioning of the erector spinae muscles. Experimental section o the ligaments in sequence from either anterior to posterior or posterior to anterior suggests L1lat i a functional spinal unit has all or its anterior ele ments plus one additional structure or all or it posterior elements plus one anterior structure in Flexion lateral radiograph of a patient with rheumatoid arthritis. The dens (0) has been eroded and the transverse ligament is incompetent, resulting in atlantoaxial subluxation. The markedly widened atlanto·dens interval is indicated by the arrowed line. tact, it will probably remain stable under normal physiological loads. To provide for some clinical margin of safety, any malion segment should be considered unstable in which all of the anterior clements or all the posterior elements arc destroyed or are unable [0 function (Panjabi eL al.. 1975; White et al.. 1975). The clinical sLability of various injuries must be assessed individually. The importance of clinical evaluation cannot be underestimated because significant spinal corel damage may occur after trauma even in the absence of fractures or ligamentous injuries (Gosch et al.. J972; Schneider et aI., 1954). Valuable guidelines ror the detenllination of clinical instability in the lower cervical spine have been provided in the form of a scoring system checklist (Box 11-2). Using this scale. the measurement of translation takes into account variations in magnificalions and is based on a tube-ta-film distance of 183 em. The 11 rotation is defined as 11 greater than the amount of rotation that exists at the motion segment above or below the functional spinal unit in question. The 3.5-mm value represents the radio-graphic measurement of the maximum permissible translation when the radiographic magnification is taken into account (Panjabi et aI., 1986). 0 0 1DmI, ! , 1 ....d_e_co_m.pression by transoral resenion of the dens c 1 ,.. ItM'~'"'~"~~M" .' A~_.~' ,~;;;Sk0. •.,••. ·W·,·_<· ,,,-- ',<>'< '. "..... terior stabilization. III C~lS(.' or Cjl1l...·slicJl1able in.iury. when ~Hl(1 extension ill~lI)Cll\'l'rS should Ilol b formed, a stretch Lest l'~lll b~ d01l1:.' 10 asse deal inlq.!rit~·. A la(I.'ral CI.·ITiGll spine ograph is taken al a standardized lube d of 180 em. ~\lld incrclll(,.'nud .i-kg weights plied as traction \n lhe Sklllilisillg LT<.lllia (Fig. 11-28). RadiogT~\phs arc tak~n ark addilion~d /'i,' hl~~ iPi Postmyelogram cervical radiograph in a patient wi standing rheumatoid arthriris and fixed atlantoaxi luxation. The space available for the cord has been duced to 6.5 mm and compression of the proximal cord is evident by examining the subdural space. w outlined in white by the injected contrast medium tient developed cervical myelopathy. which necess .''''''~' wcight has been npplicd. An Dens 1 "3 Space available for the cord (SAC) is approximately twothirds of the anteroposterior diameter of the spinal canal. One-third is taken up by the dens, one-third by the cord itself, and one-third is free space. Reprinted with permission ;rom MoskDvich, R. (994). AtlanrD-axial in5tclbility- Spine: State Df (he Art Reviews. 8(3). 531-5/19. Stretch test. A, Lateral radiograph of the cervical spine o 19-year-old boy who was admitted with multiple injuries and neurological deficit compatible with an anterior cor syndrome. The radiograph shows increased angulation a C3-C4 and a block vertebra at (5-(6. Flexion-extension r diographs were contraindicated for fear of exacerbating his neurological injury. B. The stretch test was performed to ascertain whether significant instability existed. No ab normal distraction occurred at the interspace in question His other injuries and fractures were treated routinely an he was given a soh collar to wear for 6 weeks. He made slow but steady recovery with almost complete resolutio of his upper extremity deficit. with no evidence of instab ity 1 year after the accident. Reprinted with permission (rom Moskovich, R. (1997). Cervical insrability (rheumatoid, c!warfism degenerative, others). In K.H. Bridwell & R.i. DeWald (Eds.), T Textbook of Spinal Surgery (pp. 969-1009). Philadelphia: Lipp • mal Slretch tesl is defined as differences of ·'greater than 1.7 mm 'interspace separation or greater than 7.5 change in angle between prestretch condition and the application of Ot1Cthird body weight. 0 • I I I Checklist for the Diagnosis of Clinical Instability in the Middle and Lower Cervical Spine cott-Raven Publishers. Point Value- Element • ,A,nteri~r elements destroyed or unable to function I fJostenor elements destroye or unable to function I Positive strelch test I ! ! Applied Biomechan.ics 4, Radiographic criteria A thorough understanding of biomcchanical princ Flexion and extension x-rays ples is an important aspect of the treating phy cian's knowledge base because the normal structu and function of the spinal ,column is frequent altered during surgcly. \'''hether treatment is a d c:omprcssive cervical laminectomy, a posteri I-oraminotomy with partial facetcctolllv, or an ant rior cervical fusion, all of these inlen~'cnlions ha certain ramifications with which onc must be fmn iar. This kno\vledgc not only benefits patient care b also is valuable in planning and executing treatmen Sagittal plane translation> 3.5 mm or 20% (2 pts) I Saginal plane rotatiDn > 20° (2 pts) OR I Resting x-rays Sagiltal plane displacement> 3.5 mm or 20%. (2: pts) Relative sagittal plane angulation> 11° (2 pts) I Developmentally narrow spinal canal I .:\bnorrnal dISC narrowing I Spinal cord damage 2' ! 1 Nerve root damage I Dangerous loading anticipated 1 DECOMPRESSION -"' Total of 5 or more == clinically unstable. I blodifiecJ from While. A.A. III & Panjabi. M.M. (1990). Clinical Biome<hdnrcs oi the Spine (2nd ed., p. 314). Philadelph;ej: 1.B. lippincott. • Cervicallaminectom)' often is performed to decom press the spinal corel. The compression may caused by a stenotic process and can result neurological sy'mptoIlls such as racliculopath.v or my'clopath)'. Other posterior decompressive procedures such as partial or full facctcctomics for visualization or decompression of nerve root pathology also arc comnlonly performed. Development of postlaminectol11)' kyphosis is well known in children and may develop in 17 (Miyazaki et a!., 1989) to 25% of adults (Berkowitz, 1988). [,ostlaminectomy spinal deformity may..' occur in up to 50% of children who undergo laminectomies for spinal cord tumors (Lonstein, 1977). Simulated finite element anal.'.'sis on cervical spines indicates that the primary cause of postlaminecto!l1 yr deformity is resection of one or more spinolls processes as well as the posterior ligamentous structures, such as the ligamentum flavulll or interspinous or supraspinous ligaments. The removal of these structures causes the tensile forces normally present in the cervical spine to become unbalanced and place extra stress on the facet joints. Results indicate that either a kyphotic or a hy'periordotic cervical defonnity' may' ensue, depending on the center of balance of the head (Saito et aI., 1991). Although reports exist of patients undergoing multiple-level cervical laminectomy with no evidence of clinical instability or deformity' on longterm follow-up (Jenkins, 1973), most biomechanical studies indicate some degree of instability when the posterior elements arc resected. Multilevel cervical lan1inectom.v induces significant increases in total column flexibility associated with increased segmental flexural sagittal rotations. In a cadaveric laminectomy model. the mean stiffness of the intact ce,·vical column was significantly' greater than the mean stiffness for the laminectomized specimen, and there were consistently greater rotations as compared with the intact specimen (3.6 versus 8.0°) at ever)' cendcal spine level (Cusick et al., 1995). The loss of facet joints alone causes a significant decrease in coupled motions that result from lateral bending. A moment about the anteroposterior axis results in a significant reduction in lateral displacement, a decrease in vertical displacement, and a decrease in rotation about the vertical axis. Partial facetectomy «50 % ) did not, however, significantly alter flexion and extension movements (Raynor et a!., 1987). Another anatomical study demonstrated that progressive laminectomy with resection of more than 25% of the facet joints resulted in signif~ icantly increased cervical flexion-extension, axial torsion, and lateral bending motion when compared \vith the intact spine (Nowinski et a!., 1993). A nUlllber of studies using three-dimens nite element models showed that f::\Cetectom greatcr errc'ct on annulus stress th<:ll1 (Hl in bral joint stillness. Based on these models concluded that a signific<:lnt increase in strcsses and scgmentalrnobility nlaY ()ccur lateral facet resection c:\:ceeds S(Y'·'r (Kuma a!., 1997; VClO et a!., 1997), Deccnnpression cal laminoplasty, in which the facet joints sacrificed and the laminae are reconstructed in maintenance ()f fle:\:ic)n-e:\:tension and bending stability', \vith a nwrginal increase torsion. Iatrogenic injur.v is less likel:' to res capsules of the rernaining facet joints and elements remain intact. Cervical subluxations and dislocations from injul':' may narrow the spinal canal an neurological imp::lirmenl. In SOllle cases, a reduction and re::dignment or the verteb lowed b~' stabilizatkm, may' decompress th elements without resecting bone (Fig. 11-29 ARTHRODESIS Spin::d arthrodesis is indicated in man:' dise cesses such as spinal instability, neoplas traumatic and degcnerative conditions of th Thc goa! of arthrodesis is to achicve a solid sion in which two or more vcrtebrac solidi In 1l1an:' cases, internal fixation is ust::.'d to initial st::\bilization as well as to correct defc necessary. An important principlc regarding v arthrodesis is that the stabilit.\' established nal fixation is a prelude to the biological p fusion. Although fusic)l\ can occur without fi:\ation, its appropriate USl' helps to increas sion rate and maintain structur::ll alignmen ware b:' no means supplants the necd for gecm to perfol'ln a thorough and careful pre of the vertebrae and add bone graft tCl ach sion. \Vith few exceptions, Imrdw::lre that is matel.v supported and protected by a soli will fatigue and fail after a finite number o There is ~l race b." the body to achieve a sol mass before a fatigue failure (If the implan tion device occurs, The choice of a surgical approach !() the spine, as well as whether an anterior, poste combined arthrodesis should be performe pendent on the particular patholog." presen performing a fusion, it is important for the to understand the biomeclwnical propertie Unilateral facet dislocation in a 24-year-old woman who was involved in a motor vehicle accident. The degree of vertebral subluxation is less than half the AP diameter of tbe vertebral body. Her spinal cord was compressed and she presented with an incomplete spinal cord injury. A, Computerized tomographic scans and reformatted images demonstrate the canal compromise at (5-(6. An attempt to realign the spine was made using longitudinal traction of up to approximately one-third of her body weight. B, The lateral radiograph with traction applied shows persistent malalignment. An open reduction was performed using a posterior exposure of the vertebra. Once the spine was realigned, the posterior tension band was recreated using interspinous wiring, and autogenous bone graft was inserted. C, The radiograph taken after the operation demonstrates restoration of normal vertebral relationships. The fixation provided good stability and enabled the patient to mobilize early. She had a good clinical outcome. Reprinced with permission from Mos,l:ovich. R. (1997). Cervical instability (rheumatoid. dwarfism, degenerative. others). In K.H. Bridwell & R.L. DeWc1/d (Eds.). The Textbook of Spinal Surgery (pp. 969-1009). Philadelphia: Lippincott-Raven Publishers. fcrent t)'PCS of fusion constructs. Cervical arthrode- sis also has an affect on adjacent motion segments. Thcorcticall.',/, there is an increase in motion at nearby' unfused levels. Subsequent degeneration of other motion segments requiring additional levels of arthrodesis has been demonstrated in multiple studies (Cherubinc) et aI., 1990; Hunter et aI., 1980). A recent study by Fuller ct al. (1998) evaluated the distribution of motion across unfused cervical motion segments after a simulated segmental arthrodesis in cadaver cenrical spines. The authors simulated 01112-, t\\'o-, and three-level fusions in human cervical spines. They then moved the cervical spines through a nondestructive 30° sagittal range of motion and compared this range of motion with that of unfused cervical spines. The findings of this study were interesting in that sagittal plane rotation was not increased dispropc)J'tionatel)! at the cervical motion segments immediately adjacent to a segmental arthrodesis. Although the authors ackno\vledged certain limitations of the study, they proposed that a cenrical fusion causes a fairly.' uniform increase in motion across all remaining open cervical motion segments; therefore, an increased potential for degenerative change may exist at all cervical levels. Another stud.y was performed descdbing the incidence, prevalence. and radiographic progression of symptomatic adjacent level disease after cervical arthrodesis (Hilibrand et aI., 1999). Adjacent level disease \vas defined as the development of a new radiculopath,Y or myelopathy that was referable to a motion segment adjacent to the site of a prior anterior cervical arthrodesis. The findings revealed that symptomatic adjacent level disease occurred at a relatively constant incidence of 2.9 % per yeac A survivorship analysis revealed that approximately 26(Y£) of patients \vha had an anterior cervical arthrodesis would have new disease at an adjacent Icvel within lO y"cars of the operation. The study also demonstrated that more than two-thirds of the patients who developed adjacent level cervical disease experienced failure of nonoperative treatment and needed an additional procedure performed. tems have become available that can satis stabilize the cervicul spine from either app \Vhen an anterior approach is used, a dis or a vertebrectOlny at one or more levels is co performed, usuaIlv - followed bv. an anterior arthrodesis. The excised disc or bone mu placed with a structural graft or prosthesis t anterior column support to the spine. Oss placement ma:v be in the fm'm of autogenou genic bone, comrnonly from the iliac crest i nous or from t.he fibula or iliac crest if alloge 11-30). The more cortical nature or flbula gr result in dela.yed incorporation compared"" w crest grafts, which, therefore, are preferred c The immediate postoperative strength of an.v bone grafts under axial compression on a testing machine reveals that they \\'ill adequa port the loads required in the cervical sp ph.\'sical properlies of human bank bone se preferable to autologous bone grafts, espe Cervical Spine Fixation Arthrodesis of the cervical spine Illay be indicated for a number of reasons, most commonly for trauma and degenerative diseases. lVluch research hus been performed to analyze the biomechanical advantages of anterior approaches, posterior approaches, or a combined procedure. "Vith the advent of newer technologies, internal fixation s.\'s- Lateral radiograph of the cervical spine of a 35-y male who had an anterior cervical discectomy at interbody arthrodesis using autogenous iliac cres bone graft. Note the maintenance of lordosis at segment and the integration and remodeling of Compressive Strength of Various Interbody Graft Materials Graft Type Mean ~ Standard Deviation Fibular strut 5.070" 3.250 N Anterior iliac crest 1.150" 487 N Posterior iliac crest 667,,311N Rib 452 " 192 N Hydroxylapatite 200 f-l.m pore size 1,420" 480 N Hydroxylapatite 500 f.lm pore size 338" 78 N Fibular strut significantly stronger than crest or rib grafts: P < 0.05 Pore size of 200 ~m signifj· cantly stronger [han 500 j-Lm pore size: P < 0.05 Adapted with permission itQm Wittenberg, R.H.. Moeller. J., P. White, AA Ill. (1990). Compressive strenglh of autogenous (lnd alfogenolls bone grafts for thOf.lcolurnb.lf and cervical spine fusion. Spine. 15(0), 1073-1077. older patients (Wittenberg et aI., 1990). The biological incOIvonuion or allograft appears to be satisfactory and its USe obviates the need to harvest autologous bone from the iliac crest, sparing patients additional surgical trauma and potential complications. Calcium phosphate (bone) ceramics form a strong bond with host bone because of a zone of apatite microcrystals deposited perpendicular to the hydroxylapatite ceramic surface (Jarcho, 1981). Synthetic hydroxylapatite blocks used for interbody cervical arthrodesis in goalS produced similar fusion rates and biomechanical stifFness when compared with autogenous bone (Pintar et aI., 1994). 11 is interesting to note that the goat holds its head creCl, and thus loads the cervical spine similarly to human bipeds. Goat models for cervical spinebiologicallbiomechanical testing are, therefore. populal: A can inc thoracic anterior al1hrodcsis model yielded contrary results when tested biomcchanically: autogenolls iliac crest grafts were stiffer in all motions than \\lcre ceramic graft substitutes (Emery et aI.. 1996). aturally grown coral is also a useful bone graft substitute once it has been processed. Jt is available con1rncl'cially in t\\-'o different porosities (200 ~m and 500 lJ.m pore size). The grafls of lower porosity had a compl"Cssive strength comparable \vith biconical iliac crest grafts, although they were much more briule. They can therefore be considered appropriate for clinical lise with rcspect to their compressive strength (Table 11-1). An increasing variety of prosthctic intcrbody cages ~\re becoming available (Shono ct aL, 1993); long-tellll clinical results arc eagerly awaited. Additional anteJior (Fig. 11 ~31) or posterior intcI11al ft:-.:atio!l using plates and screws for added support ma.\' be lIsed. Posterior arthrodeses of the ccndcal spine commonl:v performed after trauma but may also L1sed to treat degenerative, inflammatory, or n plastic conditions. Unlike anterior fixation devi which arc solel y lIsed in the subaxial cervical sp posterior fixation devices can extend up to the ciput (Fig. 11 ~32). Posterior fi.'\ntion has commo consisted of various wiring methods, which used as a tcnsion band posteriorly (Sutterlin et 1988). Within the past decade. however, techniq utilizing screws in both thc atlantoaxial and sub ial spine have become rnorc popular. Techniques for aLiantoaxial fixation have evol from onlay bone grafting alone to cver~more sop ticated methods of internal fixalion. Interlam clamp fixation is clinically reliable and (iVIoskovich & Crockard, 1992). Screw fixation te niques may confer increased translational and a rotational stability, but with greater surgical ri A biomechanical evaluation of four commonly u techniques for posterior atlantoaxial fixation performed (Grob et aI., 1992\ Cadaver CI-C2 f, tional spinal units were tested in nexion~cxtensi lateral bending, and rotation, while intact as we after a complete ligamel1lous injul)/. The fixa techniques used wcrc: l I. Sublaminar wire with one median graft (Gallie type) 2. \Vire fixation with two bilateral grafts (Brooks lype) 3. Transarticular screw fixation (Mager!) 4. Two bilateral posterior Halifax clamps (Fig 11-33) str<:ltcd similar 01' irnproved fusion rates com with stand-alolle bone grafts. Anolher group <:lssl."s;ed lilt, biollH.:L·hallical s it~· 01" sen.'n different cervical n:construction ods using 24 c(lif cClyieal spinl' segments (KotC al.. 1994). Thl'~' studied Ihn:\:.' posk'rior-nnl,\' niques. anlcrior-onl.\' u...· lyical platL·s. antcrior graft alone. and a combined <:lIlterior plate wilh (erior triple wiring. The results dcmollstnlled for onL'-lc\"L'1 instrlbilit\', the stiffncss \"alues consllliciion with an a;1terior procedure alOlle si!.!nificanth- smaller than those of all the othL' st;ucts tcst~'d. Intcrl'slin!!lv, stiffness w<:\s n.:sto prc-in,iul-y le\'cb in specin'1L'nS \\'itlt lwo-lc\'L'! i patterns reconstrucled using an~' of the three riot" 1l'lL,thocls. TilL' findings do not SlIpport the C':\clusi\'L' ank'rior methods in either posteri An example of an anterior cervical plate applied over two motion segments on a model of the cervical spine (Peak cervical plate, DePuy Acromed, Inc., Raynham, MA). The screws do not penetrate the posterior cortex and are locked to the plate to prevent backing out. The Gallie technique allowed signincanLl.y morc rotation in Oexion, extension, axial rotation, and lateral bending than did the other three fixation techniques, No significant differences were noted in the amount of rotation between the other three fixation techniques, although the Magerl transarticular screw fixation tended to permit the least amount 01' rotation, as might be expected, A variety of anterior cervical plating systems are currently available, and many have been subject to biomechanical analyses. Investigators demonstrated that in a single-le\'e1 procedure, an anterior cel"Vical plate serves as a load-sharing device rather than as a load-shielding device, enabling graft consolidation as observed in other clinical studies (Rapoff el aI., 1999). Bone lInion may be expecled to occur at a lower rate if the bones are shielded from compressive forces; however, clinical experience with anterior cervical plates has generally demon-r EIIEI Posterior occipitocervical fixation using a highly mod stainless steel loop (Ransford loop, Surgicraft; Redditc Worcs" U.K.) wired into place on a skeleton model. T cranial loop is attilched to the occiput Llsing wires tha pass through paired full,thickness burr holes. The wir not passed around the foramen magnum. The loop is bent to conform to the posterior craniocervical angle may be adjtlsted intraoperatively. The limbs of the lo are attached via sublaminar wires at each level. Note the axis laminar wires are tightened below the flare loop, which serves to maintain distraction between th ciput and the axis vertebra. Sublaminar wire use is on many techniqlles available to achieve segmental post fixiltion. Reprin:ec! (''1iIh permission from tvioskovirh R.. e (2000i. OcopitorervJCaJ stabi/izallon for mj'eJopathj' In pat : ','Itil rlleumdtoid (1f!hritI5 ImpliratlotlS 0; /lO! ball£' gr(lf;lflg BOllc JOlni Sur£). a2A. 349-365 ! • C1 C2 Diagram of a skeletonized atlantoaxial motion segment viewed from its posterior aspect. The method of applica~ tion of the interlaminar clamps can be seen. The bone grafts lie under the Halifax clamps and allow the posterior construct to be locked tightly, providing immediate stability. Reprinted with permission from rvloskovich, R., & (rockard, HA (992). Atlantoaxial anhrodesis using inrcr!aminc1{ clamps. An improved technique. Spine, 17, 261-267 threc-colurnn instabilit.y. In three-column or multilevel instability, combined front and back procedures utilizing an anterior plate and posterior triple wiring demonstrated clear biomechanical advantages. Clinical use of anterior plate fixation without posterior fixation for three-column cen/ical injuries, however, resulted in satisfactory clinical outcomes (Ripa et aI., 1991). These results underscore the need for good clinical evaluation and studies that take into account the fourth-dimension-time. Fu sion is a biological process that occurs over time and supercedes the importance of in vitro or computer-simulated biomechanical studies. w equipped vehicle (King & Yang, 1995). Flowever, long after airbag devices becan'lC available, air injuries involving front seat passengers began to described and by late 1997, 49 child deaths and seriolls injuries had been attributed to passeng side airbags (Marshall et aI., 1998). A retrospective review was performed by three diatric radiologists to determine if a pattern of inj was common to this new mechanism of pedia trauma (Marshall et aI., 1998). Their findings, as w as the findings of other studies, demonstra that passenger-side airbags pose a lethal threat children riding in the front seat of an automo (Giguere et aI., 1998; McCaffrey et aI., 19 Mohamed & Banerjee, 1998). Many of the child killed in these accidents \vere front seat passeng involved in low-speed collisions in \vhich the dr sustained no or only minor injuries; the pattern injury' seen is different in the rear-facing infant seat versus the fOl\vard-facing child car seat. In formec the injury' to the infant \Vas often mas skull injury and cerebral hernatoma as a result of proxirnit)' of their heads to the airbag, while in the ter, children sustained man)' more cervical injur '1\\'0 of the older children had autopsy findings o lanto-occipital dislocation and one sllstained a "n decapitation" injury, demonstrating the vulnerab of the pediatric cervical spine to the explosive fo of an expanding airbag that hyperextends the ch fragile neck. Recent guidelines from the NHTSA ( 11-3) were prompted by reports of airbag injurie children and emphasize that children of an}' should be properly secured in the back seat (Natio Highway Traffic Safety Administration, 1996). A mathematical simulation was performed study the potential of head and neck injury to an belted driver restrained by an airbag (Yang et Nai;ional HighwayTraffic Safety Administration .Guidelines Regarding Airbags & Children BIOMECHANICS OF CERVICAL TRAUMA Airbag Injuries Motor vehicle accidents continue to be the leading cause of injury-related deaths in the United States. [n 1984, the National Highway Traffic Safety Adrninistration (NHTSA) required that automatic occupant protection devices (airbags or automatic scat belts) be placed in all automobiles in the 1987-1990 model years. In 1993, the passengelcside airbag was introduced. Studies generally' concluded that front seat occupants are adequately! protected against frontal impact if belts are worn in an airbag- The back seat is the safest place for children of any age to ride. Never put an infant (less than 1 year old) in the front of a c"ar with a passenger-side airbag. Infants must always ride in the back seat, facing the rear of the car. Make sure everyone is buckled up. Unbuckled occupants can be hurt or killed by an airbag. 1992). It was found lhal when the standard 20° angie steering wheel was used, neck joint torques were decreased by 22%. The resultant head acceleration increased 41% fTom the baseline study when a vertical steering \vheel \\"las used. "If the verlical (limensian of the airbag was reduced by 100/0, neck joint torques were increased by l4%, while head acceleration showed a slight decrease of 90/0. Although ideal dimensions and inflation rates for airbags remain elusive, their use has resulted in a significant reduction in head and neck injuries. Whiplash Syndrome Whiplash syndrome is a complex se' of symp,oms ,hal may present after an acceleration hyperextension injury. These injllIies typically occur when a Car is struck From behind, bUl may also be caused by lalcml or Fron'al collisions (Barnslev el aI., 1994) (Case Study 11-2). The acceleration of the car seat pushcs the torso of the occupant forward with the result that the unsupp0l1ed head falls backward. resulting in an extension strain to the neck. A secondaJ)' Aexion injlll)' may occur if the vehicle just struck then strikes another vehicle in front and just as suddenly decelerates again, throwing the occupant forward once more. Crowe coined the tellll 'whiplash' in 1928 in a lecture on neck injllIies caused by rear-end automobile collisions in the United States but latcr reported that he regretted using the term (Breck & Van Norman, 1971) because il describes only the manner in which a head was moved suddenly to produce a sprain in the neck and not a specific injUl}' pattern. Therefore, we refer to this clinical entity as whiplash syndrome, not as whiplash injllry. Allhough whiplash injuries are a common traumatic event, the pathology is poorlv understood. 01'len the severity of the whiplash trauma does not con-elate with the seriousness or the dinical problem, which can include neck and shoulder pain, dizziness, headache, and blurring of vision (Brault et al., 1998; Ettlin Cl aI., 1992; Panjabi el aI., 1998a; Sturzenegger el 'II., 1994). Apart Ii'om a frequently observed loss of physiological lordosis, a radiographic examination of the cervical spine is often normal. Even newer technology such as MRI is not always able (0 I-eveal a soft tissue injul}'. MRI examination of the cerebrum and cervical spinal column perrormed 2 days after a \vhiplash neck sprain inJUI)' in 40 patients did not deli~cl an)' pathology con~ nected lO the injlll)', nor was the MRI able predicl symptom development or outcome (Borchgre\fink d aI., 1997). Injuries thaI have been documented in- '0 elude inlerspinous ligamcllt tears, spinous pr fractures. disc rupture. ligamcntum f1avulll nlp facd joinl disruptiun, and stretching of 11K.' ~\I musdl's. The diagnosis and managcll1clH of lash injuries an.. . oflel1 t.:ol1l"oUIl(!L·d b~r concom psychosoci;':il ,lilt! IlH.'dicolcgal isSlIl..."S':IS well (\ et ,Ii., 1998j. One of the earlv (unpublishcd) biomceha studies or whiplash injury was clone b.v the la Irving Tuell. nil orthopuedic surgeon in Se \V~\shillgton. I-h.: uscd a cinl' camenl to photog himsdf driving as he was rammed from behin his surgical J'csidL'nt driving anothLT car. Fr drawn frolll Lhat movit.' ckarly demonstrate til pen:xtension of his neck over the sl.'alback of h (Fig. 11-34). One C;:\l\ also scc thl' dlect of in l"orces on tht.: mandible as his jaw snaps opt'n the Hcccicratioll forces his !lead backwards. mechanism Illi.l\ explain the tL'mporomandi injuries thal (In,.' a COllllllon accompanilllL'nt o \'ical \\'hipbsh injuries_ r\ recent sLud~' of rcar·end collisions qual !i\'d~' elucidated the m.:tllal neck movements wke place (Castro CI aI., 1997). A1kr ramllled from behind, the Illcan accl'ler<:\tion target \'chicles was from 2.\ to 3_6 g. Maxim tension was n.'ached when the head contacte headrest; the angle b('!wt...·L·n the head and t bOth' varied From 10 to 47" (meall = 20"). In th SCIlCL' or a headrest, tht..., maximal r('cordcd e sion was 80". FollOW-LIp clinical and rvIRI cxam tions were pcrformt.:d_ The sttlcl~! concluded the "limit or hannlessllcss" for stresses ar from rear-end impacts with regard to the ve changes lies bd\\'cen 10 and 15 km/h, In stud reproducible \\-hiplash trauma mode! using w cervical spinc specimens 1ll00Jnh.:d on a belH.: sled was used to simulate rear-end collision increasing horizontal accderations applied t sled (Panjabi ct 'II" 1998a.b). Bo,h sled and kinematics CHIl be measur'cd using potentiom and acceleroll1l.~tcrs. Using this whiplash modt S-shapL'd cun:c was (kscrihed in whiplash in in which the lowt.:'r cervical spinC' h.\'perext ~JIld the upper cCITical spine flexed (Grauer 19(7). The invesLigators r~h that Ihe in.iur~' w currcd during the hyperextension phase i knvcr cervical spinc, Correct positioning or adjustable headn..'s hind the skull. not hehind the Ileck, is impo Tests using a H~'gc sk~d and a Hybrid III dll wcre pcrrorl11L:d by other iJ1\'estigalOrs to deter fi .•,;;;;: _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ------------ ------------------------.- , ;" ;: .. , ' ,~ if ! Whiplash Syndrome ;::',:'i::'/.::,';): ::'»::)/:,:'':-''P,}\ 32-year·old man was injured in a car accident. The trau- ::~9>:~;:_~i~;~ matic event happened after being struck from behind. [n /~~{~:d{t:ih.iS case. the inertia of the head and the flexible spine re- from Fr(1nkel, VH. (1972). Whiplash injuries to the neck. In C. Hirsh & Y Zotterman (Eds.), Cervical Pain (pp. 97~111). New York: Pergamon Press. "?g;'sulted in marked hyperextension movement of the head. The .¥a~celeration of the car seat pushed the torso of the occupant :,~?;~ ;7'''':; $Jo~va{d with the result that the unsupported head 'ft~~ard, resulting in an extension strain to the neck. fell back- lr -;. The palient presented with severe neck and shoulder pain accompanied by sleep disturbance. The right sternocleido- j- mastoid muscle was approximately twice as large as the left sternocleidomastoid muscle. After muscle testing against re- t, >.: ..t. ' f.. ~> sjstance, the pain increased (Fig. cs 11-2-1). 'To decelerate the posterior rotating head, a moment and g>:a.force must be developed by active and passive stabilizers, i.-?int surfaces, and the intervertebral disc. The moment and force give rise to tension, compression, and shear stresses .',: .find strains in various parts oi the neck Cdusing, damage. fl., During the collision, rhe muscle tension increases with ihe , velocity of lengthening. A tension inappropriate for the length tension curve occurs and partial rupture of the sterno4 cleidomastoid muscle is produced. Reprinted with permission Case Study Figure 11-2-1. '.- -.. I I I I'IDIIDI 1~_H.a_nd_-_d_r.aw_n.c.e.I ., • ~. ,; .. ... ,.- B (A) Before and (B).tr.a.c.ed_f'.o.m_t_w_o_f.r_am_e_'_o_f_a.C.in.e.-.m_o_V_ie_o.f.D.r_.I.rV.i_n9_T_u_e_lI_d.,.iV.in.9_h.i'.c.a.raher bein9 rammed from behind by another car driven by hi, re,ident_ _ biomechanical responses for the various conditions observed in normal driving (Viano & Gargan. 1996). Risk of injury was assumed to be proportional to neck extension. A low headrest position carries a relative injl1r~' risk of 3.4 in rear-end crashes. compared with 1.0 for the favorable condition. If all adjustable headrests were placed in the up position, the relative risk would be lowered to 2.4, a 28.30/0 reduction in whiplash injury risk. Significantly more complex acceleration injuries ascribed to high G-force activities are now occurring. There is a report in the Htel·ature of cervical injury in pilots of F- t 6 fighter planes, citing a I-year prevalence of neck injury of 56.6% and a career prevalence of neck injury of 85.4% (Albano & Stanford. 1998). Maintenance of neurological homeostasis and protection of the spinal cord. nerves. and vessels and support and protection of the skull are the ultimate tasks of the cervical spine. An appreciation of the principles presented should afford a greater understanding to physicians and allied health professionals involved in the treatment of cervical spine pathology. 'vVith increasing technological advances in our society, \ve remain vulnerable not only to common types of cervical trauma but also to idiosyncratic methods of injury' related to these new technologies. Acceleration injuries have been OCCUI'l'ing for decades but cervical spine injuries caused by airbags have only recently been described, and aliI' involvement in increasingly greater speeds and higher risk activities places LIS in increasing jeopardy, As our body' of knowledge expands, we continue to discover new methods of injury. It is critical to pursue rational treatments for these disorders based on sound biomechanical principles, 5 11111111 a Iy 1 A functional spinal unit or motion segment consists of t\\'o adjacent vcnebrae and the intervening intervertebral discs and ligaments between the vertebrae. _2 Whcn eithcr rotation or translation of a body along one axis is consistently associated with a simultaneous rotation or translation along another axis. the motions arc coupled. 3,; Intervertebral disks cxhibit viscoelastic properties (creep and relaxation) and hysteresis. (4;' Discs arc able to withstand greater than nor- l1l~r loads when compressive forces arc rapidly ap- plied, which protects the disc from catastrophic lIrL' ulltil e . . . trl"lllcl~· high loads are applied. 5 Vi".'I·lL"bl·al bod~' from upper ct.'n:ical cUlllpr~ssioll str~ngth 10 incre IO\\'L'r lumbar lev-cis. 6 TIlL' mean sagillal diametL'r of the malt.- a spinal canal for C3-C7 is close LO 14 Illm; the sp cord diameter is about 10 111 Ill. 7 LigamclltuJ11 flavlIJ11 is undt..'1' tension when the spine is in a Ill:1Itral position or somew cxtended. prestressing the disc and pro\'iding s intrinsic support 10 tht..' spine. S iVlllscles pla~' a critical role in b~lsic post homeostasis. Patients with paral~'zL'd ct..'lyical m cles arc unable to support thdr heads. 9 The spinal cord has some longilUdinal c1a it.'· blll it tole-ratL's axial translation pond.". It is translatolY forces {hat t.\'picall,\· rcsult in neurol cal injul·.\', 10 Instant c('ntl.'!' anal.\'sis indicates that tange lnotion (gliding) wkcs placl.· bel\\,(.'c·]1 lltt::.' facet jo as lItc cervical spine is flexed alld extended. The or the intelTcrtebr~il !"()ramina inc.TeDscs with fle and deCl'l:ascs \\'ith extension, 11 Kinematic instability' refl'rs to the quantit motion (too much or too little) or till.' quality of tion present (alterations in lile normal pattern both. Component instahilil.\' addrcsses the clin biomechanic~ll role of till.' \'ariolls anatomic st tures of the functional spinal unit. 12 An." motion s~~glllent in which all of the a rior ek'llkllts or all the posterior ekments m·e stn>yed or are unabk' to function should be con ered unstable. 13 A significant incn:asc in annulus stresses scgrnclltal mobility' may' OCCUI' when bilateral f resection exceeds 50(*;. 14 Appropriate use or internal fi:G1lion helps L (:reasc the fusion rate and maintain structural a ment. 15 Front seat occupants are adcqLH.llcl~' prote frontal impact if scat belts arc worn in airbag equipped ,·chick. Passenger-side air pose a lethal threat to children riding in the f scat 01" an automobile. ~lgain.s{ 16 \'Vhiplash s.\"ndrome is a complex set of s~ tOI11S thal may present aher an (\("ccleration hy extension injury. REFERENCES Adams .•\LA. 8.: Hutton, W.e. (1980). The ('freet 01" poslllrl' on the rok of the apophysial joints in resisting in[e!'n:"rlebral compressive forces. J BOIlt' .Ioill! SUfg. 62B, 358-62. Albano, 1.1. &. Stanford, J.B. (1998). Pn.'venlioll of minor neck injuries in F-l6 pilots. ill·juf Space EI/I'iro!l ,lied, 69. 1193-·1199. Bailcy. R.\V. (1963). ObSCIT<ltions of cerdcal inten"crll'bral disc lesions in fraclUres and dislocations. } BOlle Joillf Sutg" ':'/5..\, 46l. Barnsley, L, Lord, S" &. Bogduk, N. (1994). \Vhiplash injun', Pail/, 58. 283-307. . Bell, G.I-!. (1967). Variation in tht stl't.'n~[h of vcrtdH"<! with ;lge and their relation to oSkoporosi:. C(/Ici( Tissl/i.' Res, /, 75. Bennett, S.E., Schenk, RJ., & Simmons, E.D. (2000). Active range of motion utilized in the cervical spine to perform daily functional tasks (Abstract). Associalioll BOlle alld Joillt Sllri~colls 5llld .'lIlII/W! J1,-,<-'tillg. San Antonio. TX. \lay 2000. Blackley, H.R .. Plank, L.D .. & Robertson, P.A. (1999). Determining the sagillal dimensions of the canal of the cervical spine. The reliability of ratios of an~ltomical Tlleasure- or ments.JBollcJ()intSlIrg,8IB.110~112. Borchgrevink. G.. Smevik. 0 .. Haave. I.. et al. (1997). \lRI of cerebrum and cer\'ic<:ll cOIUTlln;:l within two davs after whiplash neck sprain injury. Ilijury. 28, 331-335 .. Brault. J.R .. Wheeler. J.B., Siegmund. G.P.. et al. (1998). Clinical response of human subjects to rear-end automobile collisions [published erratum appears in Arch Ph"s :\led Rehabil 1998 Jun;79(6):723j. Arch Phi'S "II:'(/ RellO'hi! 79 72-80. . ' , Breck. L.\V. & Van Norman, R.W. (1971). \lcdicolegal aspects of cervical spine sprains. C!ill OrtIJop, 7.J, 124~8. Breig. A.. Turnbcll, I., & Hassler. O. (1966). Effect of mechanical stress on the spinal cord in cen'ical spondylosis. J NCllwsllrg, 25. 45-56. . Cart:!', n.R. & Hayes. \V.c. (1977). The compressi\'(.' behm'ior o! bone as a two-phase porous structure. J BOIIC Joint SlIrg, 59:1. 954-962. Castro. \V.H .. Schilgen. :\1., :\Ieyer, S., et a1. (1997). Do "whiplash in,iuries" occur in low-speed rear impacts? EliI' Spill" J, 6, 366-375. Cherubino. P.. Benazzo. F., Borromco, U., et a1. (1990), Degenerati\'e arthritis of the adjacent spinal ioint~ follo\~'irH! anterior cen'ical spinal fusion: Clinicoracfiolouic and sta~ tislical correlations. Iial J Orthop 7/"a/l/lUlIO!. 16. 533-543. Clausell, J.D .. Gocl, \'.K .. Traynclis. V.C., el a!. (1997). Uncin;:tte processes and Luschka joints influence the biomechanics of the ccrvical spine: Quantification using a finite dement model of the C5-C6 segment. J Orrhop ~Rl's. /5. 342-347. Cusick. J.E. Pintar, EA .. & 't'oganandan. N. (1995). Biomechanical alterations induced bv Illultilevel cervical laminectomy. Spille, 20. 2392-2398: Denis, F. (1983). The three column spine and its significance in the classification of acute thoracolumbar ~pinal injuries. Spille, 8. 817-831. Dickman, C.A .. Zabramski. L\1.. Hadley. ,\1.N., et al. (1991). Pediatric spirwl cord injury without radiographic abnor-' malities: Report of 26 cases and review of the literature . .I Spillal Disonl. .J. 296-305. Dvorak, J .. Froehlich. 0 .. Pt.:llning. L., et a!. (1988b). Fun ~ional r:ldiographic diagnosis of the cervical spine: Fle Ion/extension. Spillt'. 13(7).748-755. DnJnlk, .I .. f·la~'ek. J., &. Zehnder. R. (1987). CT-functional d agnostics of the rotatory instability of the upper cervic spine. Part II. An evaluation of healthy adults and patien with suspected insl;:lbilit.v. Spillt', 11(8). 726~731. Dn,lt';lk, J. &. Panjabi. \1.\-1 .. (1987). Functional anatomy the alar ligaments. Spille, /2, 183-189. ' Dvorak, J., Schneider. E.. Saldinger. P.. et al. (1988a). Biom chanics of the craniocervical region: The alar and tran verse ligaments. J Orthop Res. 6,~452-461. Emery'. S.E .. Fuller, D.A .. & Stevenson. S. (1996). Ceramic a {eri(~r spinal fusion. Biologic and biolTlcdwnic<l1 compar son In a canine Illodel. Spilil'. 2/. 2713-2719. Ettlin. T.iVI.. Kischka. U.. Reichmann, S., et al. (1992). Cer bral Sy'lllptoms after \\'hiplash injun' of the neck: prospective clinical and neuropsycl{ological studv whiplash injury. J ""'t'/lro! ,Vellrosllrg Ps\'c!Jiottv.· 5 943-948. ' . . FeipeL V.. Rondelet. B., Le Palkc. J., et al. (1999). Norm global motion of the cervical spine: An declt'O!!oniometr study. Ctill Bill!IICch, /.J, 462-470. Fidding. J.\V. (1957). Cinero<.:ntgenograph~· of the norm cen'ical spine. J BOIlt' Joilll SlIr~. 39:\. 1280-1288. Fielding. J.W.. Cochrnn, G.\'.B .. La~~'sing, J.F. !II, et a!. (1974 Tears of the tranS\'erse ligamenl of the atlas: A clinical an biomechanical study'. J BOlle Joil/t Slitg. 56:\. 1683-1691 Fid~ling,. J.\V.. Hawkins. R.J .. & Ratzan. S.:\. (1976). Spine f sIon lor atlanto-axial instability. J BOIIl' Joilll Slirg. 58 400-407. . Fountain. F.P.. :\linear. \V.L., & Allison. R.D. (1966J. Functio of [on gus colli and longissimus cen'icis muscles in ma Arch Phys Alcd. 47. 665'::-669. . Frankel. \'.H. (1972). \Vhiplash injuries to the neck. (n Hirsh 8.:: Y. ZOlLerman (Eds.). Ct'n'ica{ Paill (pp. 97-111 New 'I'ork: Per£amon Press. Fuller. D.A .. Kirk()alrick. J.S .. Emery, S.E., et al. (1998). kinematic study of the cervical spine before and after se mental arthrodesis. SpillC, 23, 1649-1656. Giguere, J.E. St-Vil. D.. TUrlnel. A.. et 'II. (1998). Airbags an children: A spectrum of C-spine injuries. J Pcdioll': SlIn 33,811-816. Gosch, H.H .. Gooding. E... & Schneider. R.C. (1972). An e perimental study of cerdcal spine and cord injuries. trill/nUl. 12.570-575. Grauer. J.N .. Panjabi, \1.M., Cholewicki. J .. et al. (1997 \Vhiplash produces <In S-shaped cun',Hure of the neck wi hyperextension at [ower le\'cls. Spilw. 22. 2489-2494. Grob, D., Crisco, J.J. 11(, Panjabi, \1.\-1., et al. (1992). Blom chanical nalu,ltion of four different posterior atlantoaxi fixation techniques. Spillt', 17(5). 480-490. Hansson. T. 8.:: Roos. B. (1986), A!!c chan!!es in thc bone mi eral of the lumbar spine in n~}rnwl \~Olllen. Calcit' Tiss llIl, 38, 249-251. I-lansson, T., Roos, B., & Nachemsoll. A. (1980). The bo mineral .content and ultimate compressive strength lumbar vertebrae. Spint', 5. 46--55. ~ Harms-Ringdahl. K. (1986), On assessment of shoulder eXl cise and load-dicited pain in the cervical spine. Bio}// dUlI/ica! AIU/!ysis oj' Load-ElfG-,I/t'thodoiogica! St/U!i u(Poill Pnl\'oked by Ex/reille POSilioll. Karolin~ka Inslitut Universit~' of Stockholm. Herkowitz. H.N. ( 1988). .-\ comparison of ~ln".:rinr cer\'ic~ll fu· sion, (:cr\'ic~t1 bmilll..'ctollly, and cc:rviC<11 hllninopl:lsty for the surgical managi:nH:nt of Illultipli: kvd spondyloti\'" radicul()path~'. 5pillt', 13, 774-780. Herzog, R.J .. Witns. J.J .. Dillingham. M.F.,l'l al. (1991). NorJ1l~d (:elyiGII spinc morphomelry iwd ce.:rviL"ul spin:lI slcno· si:-: in ,lsymplOmatic professional foot hall players. Phlin film r~diography, muhiplanar computed lomogr''11hy, and m,lgn\,.,tk n:sonancc imaging. 5pil/('. 16, S178-S186, Hilibrand, A,S .. Carlson, G.D" Palumbo, M.A .. t:t al. (1999), Rndiculop.l\hy and myelopathy at sl.'gmcnts ;loj.u;l:l1t to lhc SiL\.' of'l previolls anterio!" cc:n'ic~tl anhrodc:.sis, 1 BOllt' 10illl Sllr~, 81A, 519-2:>, Holdswor(h, (1963). Fr<lctures, dislocations, :lud lurc-dislocOitions of thc spinc. 1 BOlle 1uiut SlIrg, r.w. 6-10. Huntl.'r, L., Braunstein, E., & B~li)('y. R. (1980). Radiogr:'lphil; chilng('s following nnterior fusion. Spille. 5, 399-401. hrcho. ~'1. (1981). Calcium phosphale CI.:ramics as h~Il'd tissul.' prosthctics. CJiIl Ort/lOp, 157,259-278. Jenkins, O.H. (1973). Extensiv(' ccrvic~1 larnincclOlll~" LongIeI'm results. Bt'l S/fl"g, 60. 352-$54, h:.~lz.lrl'lll. I...E. (1975}. Creep charactaistics of thl.' human spinal t:oluIII11. Onhol' C!ill North Alii. 6, 3. K"tl.:trinn, L. (19i2). Dynamic response Chan\l;leristi(,;s of Ihl.· hUIlWll inlervt:rlebral column: An cxperirllc:nlal stud,\' of autopsy specimens. Acta Orr/IO{J Sea lui, No, 146, King, A,I. &: Y.Hlg, K.H. (1995), Research in biolncchanic:-I of occupant pn.Hcclion, J hal/IIUI, 33. 570-576. K.il:dl~lra, Y.. Uda, B .• & Tachihana. S. (1995), Effect of spin:d cord stn:Lching uuc to head fk.\ioll on intr~lnH:duliary pn:ssul'c. NI.!/lrol ,VIed Chi,., 35. 285-288. Kot;'llli. Y., Cllllningh~un, B.W" Ahumi, K., el a1. (1994). l3io· mechani<.:.d anaiysis of cervical st:ibiliz<\lioll systems, :\11 i1SSCSSll1Cllt of tl',l(lspedicllhlr screw fixation in the l.,'\":r\'ic;~\l spinc. SpiJlt·. /9, 2529-2539. Kriss, V.M, I.~ Kriss, T.e. (1996). SCIWORA (Spill~ll cord injury without rlldiogr;'lphil; <lhnormality) in infnnts and chil· dren. CJiIl Pt,·dii/II", 35, 119-124. KUlll:ln:san, S., Yo~an;'lIldan. N.. PimaI', EA., et al. (199;). Finih..' ,·lell11.'!H modeling of cervical Iamillt:ctonw with gr'l(k'd f.ICCtcClOI1lY. 1 S~)ill(J! Dison!, 10, -10--16. . Lonstcin, J.E. (1977). Posl.laminc:clomy k.... phosis. CliI, Or· t/lOp, /28,93-100_ Lysdl. E. (1969). ,\-Iotion in thc cer\'ic~\1 spine. An eXIH.'rilllcn· 1<11 sludy on autopsy specimcns . .'lcla OnllOl' S,:cwd, SlIpp! /23. 1-61. Marshall. K.II'.. Koch. B.L.. & Egelhoff. J.e. (1998). Air bag. rcl<\lcd dC<.lths and seriolls injuries in childrl.'n: Injury p;,H· terns ~Ind imaging findings. A'/I J Neuwnlt!;o!, 19, 1599-607. j\·kBroom, R.J., Ha\'t's. W.c., Edwards, W.T.. d nl. (1985). Prl.'diction of \·c'-·Il.'bral body cOl1lprcs~in' fn.lClurc using quantitath'c compuled tomogr'lphy. 1 BOlle loil1f Surf,:, 67A. 1106-1114. McC.11'frcy, M., Gl.'rman, A., Lalondt:, F., (.'t al. (1999). Air bags ~IIH.I children: A pOlcntinlly I!.'lhal (ombin:ltioll. 1 I'l·dillir On/lOp, /9, 60-64. ~'lil1lllnl, M., tvtoriya, I-I.. Watanabe, T., ct al. (1989). Threl.'· dimension,11 motion anal~-'sis of lhe cer\'ic:d Spillt: with spl:cial n.:fcrcnct: to lhe axial rotation. Spille, 14, 113;-1139. .\li.\:uaki, K.. T'I(b. h .. \lahIHb. Y.. l'[ :11. / 19ii9J. I'o:l·\ll·n:>i\'\.· silJlllll:llIl'(JIl~ lllulti:>l'gllll'lll:tl 11I:-lllll fC'I- l'l'!T 1JI.\dl'P~lthy widl l-l'l"\'il':c! ill~l:JhiliIY :tlltl k~phlJlil' :t1 S·:-hapnl ddorlilitil·~. Spi!ll'. 1,1{ 11 L 1160-1170. \I<oh:Il11nl. :\.:\. & Ball\.·rin·. ,\. (199$1. P;lltl:l"lI:> of injun :>oL"i;'lll'd wilh :llltullH,hik ;'lirh;lg Il~l·. J1o.'1.~nl/l .\I•.'d 4:15--i38. .\({lrolll·\. S.P. &. Sl,.·llUh/, :\.B. (19$;). :\n:d.\:-i:- ~llIti Illl',h flIl'llt (,I loads 1111 !Ill' nl·d. hew,' Onh R,'., Soc, J(J, _,1 .\l(Jn,nl'\, S.P., S.:hllllz...\"B .. ,\1ilb', 1.. l't :11, (I%SJ. l.t1:HI placl'llll'lll pnJjh'rtil'~ 01 10\\'\.'1' n'l'\·;l.'al ~pirh' motion ml'IlI~. J HiuIJlt'ch, 21, 769-779. ,\llI~k(l\'ich. R. (1 'J94J. ,\lhllllCI-;t\i:!l in~l:lhili[.\·, Spill,': SflI fhi' .-\1"/ Uni<'l\'.,. S(31. 531-5-i9. .\Io:-kmich, R. (1\)97). Cenk:d ill:>I'lhilit.\ (r!h'HIIl: d\\:lrfi:-IIl, (kt!l'lJl'rati\l·. Illhl:[':-). In 1'..11. Brid\\\.·11 l\: Dl·\bld (Ed:-. .i. Thl' 7~·.r1hfJc,J~ o( Spinll/ S/ft'J,:CIY 969-1009). Phil:tddphi~l: l.ippillCl'tt·Ran·1l Ptlhlislh-"I'~ .\to:>kll\-ich R.. 1,,'1 a!. {2000i. On'ipit(ICl'ITicti :>1:lhili/.'lIi(Jl Illydllpalhy ill Jlatielll~ willi rhl'ulll;'l[oid :tnhrili:>. Im ClliollS (If 1101 hOll\.~ gr:dtil\~. J Built· luhlf Sun.:, 1 349-365. ,\I(I~ko\"idl, R, &. Crod.anl. 11.,\. (199~). :\tlanloa :Irlhmlksis tlsinl~ illll'rblllin:lI' l'lalllp~. :\11 ill1pnj\"l'd Iliqlll'. Sllill,', f7, 261··1C\1. .\loskO\'kh, R. ~\: ./Olll'S, [)".\, (19~~), CpPl:r l'l'n ical spin s[I"t!Il)('nl:ilioll. Spill": SieUc' ('I Iii,' '\0 Un'i,'I\'s, 2.,3·_·2:'3. \losko\il:h. R., ShOll, S". & Zil:\n~. Z,ll. (1006), DOl'S till' \'k,i1 '::1I1,L! to hod\' r:llllJ 11I"l'dil'l ,.;pill:d Sll'llOSi<i Bill! l .luinr Oi:>, S..:;. 01-71. \l.\'l,.'rs. B.S. l\: \Vinkl']sll'ill, 13. . \ , (1995l. Epidl'll1i(ll()g~, l' l'iC:lliu!l, 1l1l'l'hanis1l1, alld lokr~1I1(l' (II' hUIll:1I1 l'('r spinl,.· injuril:s. Crir Ra Biulllni 1-:11,::', 23, 307·_·+09. \:adll,.'IllSUIL A, l\: E\":lllS . .I.H.ll\)tJ$J, S(Jllll' Illl'l.'h:1I1ic:d p l'r1il'S of till' third hunwll IlImh:l!' inll,r!:llllin:lr lig;,u lligallh.'1l11l1lJ f!;'lnlIII J. .ll3iuJllcch. /, 101. \::ltillnal lIigh\\"a~' Trani..: S:tfl't.\" Adlllinbll':llioll. (1996) h:lg akn. Allil LJllc'I~!,: ,\It'd, 2S, 14~. \owinski. G.P.. Vi:-ariu:-. II.. \olk, I..P.. l·t al. (1993).:\ lIIl"l-klilicll comp:lri:»un or Cl'l"\'k:t1 lalllinopbsly :lIal \·il·:III:tlllin\.'i,·lomy with pn1grl'::.:-i\·l' bn·I\,·d(llll.\. Spine 1995-1004. Os,,-·nbach. R.t\.. l\.;. .\kne"lt.'~, A.II. {1"JS9). Spinal l'(}rd in WilllllUI I"i\diog:raphit:: ... hl1()rfll<.llit~' ill (hildrl'n. !'e .\"('IIHhCi. 15. 168-174, Pang, D. &: f'ollad, I.E (19S~)' Spill:t!l.."ord ill.i\lr~· withou diogr:lphil' ,dmorlll'llil.\ in L"lIildrl'JI-Thl' SCJ\\·OR.,\ drollll·.l li"lllllllll, 19. 6~-I-(l6-1,' P:lni"bi, ,\I ..\L Cholc\\"id,i, J.. i\ihu, K.. l't;,)1. (I99~h). S lation (If \\"hipla:.;h Ir:tlllll~t using w!l(,k (l'rdr:,1 :.;pinl' S illll·n:.;. Spillc', 23. 17-14. Panj:lhi. ,\1.:\1.. Chn]\.'wil"ki. J.. \!ihll, K .. l'1 :t1. {199:-bi. Ca br lig~1Il1l'1l1 strt'lchl's during ill \'Itro whiplash :o;i li(IIl:-.l Spimd Dison/. 1/, 21i-l.'1. Panjabi. \L\1., I\.r~lig. .\1.11 .. & God. V.I\.. (19~I). A 1l'.... hn for m\.'i.t:'url·IlH.'llt alld dl'sl"l'iptioll of lhrl:('-diml·ll::.iolla dl..'g:rl'I.'-of-frl'l'dolll motioll 01':1 body joint \d[h :1Il app lion totltl' hum,1l1 ~Jlinl:. 1 WIlIJh'dltJlJics. 1-1. -i-l7--i60 P:llIjahi, M.M,. SUlllllll'r:-, 0.1.. Pdkt'r, R.R .. l't al. (19 Thr('l..'-d;l1ll'lISion:,lloitd-displ:'I.'l,.'l1ll'lll CIll"\"('S dui..' LO fo (111 till' l,.'l'l"\·il'al spilll' . .f On/lOp Uf.", -I, IS~-161. Paniabi, ,\1.,\-1., \Vhile, A.A. III. & Johnson, R.,\1. (1973). Cer\:ical spine mechanics ,IS a I"unction 01" transection of cornpOneJllS. 1 HiOlllt'c!/(lIIics, 8, 327-336. Penning, L. & Wilmink, J.T. (1987). Rotation of the cCl'vical spine. A study in normal people. Spille 12, 732-738. Perry, J. & Nickel. V.L. (1939). Total cervical-spine fusion for neck par,t1ysls. 1 BOIIt, loilll SIIJ'g, .Jl:\, 37-60. Pintar, FA., \himan, D.J., Hollowell. J.P., L't 'II. (1994). Fusion rak and biorHechanical stillness of hydroxylapatite versus autogenous bone grafts for antnior discectorny. An in vivo animal study. Spille, 19, 2324-2328. Pintar, F.A., Yoganandan, N., Pesigan, ~'1., L'l a!' (1995). Cervical vertebra! strain llleasuremellts under axial and eccentric loading. 1 Bioll/cch Ellg, /17, 474--478. RapolT, A.J., O'Brien, T.J., Ghanayem, A.J., L'l al. (1999). Anterior cervical graft and plate load sharing. J Spillill Dis, 12,45-49. Raynor, R.B., Moskovich, R., Zidel. P., et a!' (1987). Alteration in primary and coupled neck motions after fncell.'ctomy. NellnlslIl'.i.iery, 2l, 681-687. Ri:ill. J. (1960). Effects of flexion-exti:lIsion movement of thi: head and spine on the spinal cord and nerve roots . .I ,VCIIml ,VClII'OSlIr.i.i Psychi(/IJ', 23, 214-221. Ripa, D.R., Kow,dl, \I.G., "'!eyer, P.R. Jr., et al. (1991). Series of ninety-t\\·o traumatic cervical spine injuries stabilized with anterior ASIF pl,lte fusion technique. Spille, 16(3), 46-55. Rolander, S.D. (1966). \-lotion of the lumbar spine with special refi:rence to thi: stabilizing (.'ffect of posterior fusion. An experimental study on autopsy specimcns. /lcw Orthop Snllld,90, 1-144. Roozmon, P., Gracovetsky, S.A., Gouw, G.J., ct a!' (1993). Examining motion in the cervical spini:. I: Imaging systems and measurement ti:chniqui:s. 1 Bio/lled Ellg, 15,5-12. Saito, '1'., 't'amamuro, '1'., Shikata, J., et a!' (1991). Analysis and prevention of spinal column deformit.;.' following cervical laminectomy. I. Pathogenetic analysis of postlamini:ctomy deformitii:s. Spilll.', 16, 494-502. Schneider, R.e., Chi:rry, G., & Pantek, H. (1954). The syndrome of acute ci:ntral cervical spinal cord injury with special reference to the mechanism involved in hyperextension injuries of the cervical spinto J /":I.'ZlroslIJ'j;, 11, 546-577. Schultz, A., Anderson, G., ()rtengri:n, R., et al. (1982). Loads on the lumbar spine: Validation of a biol1ledwnical anal\'sis by measurements 01" intl'adiscal pressure and myoell'ctric signals. J BOllI.' .Ioillt SlIrg, 6",.1, 713-720. Shono, Y., ivlcAI"ee, P.C., Cunningham, B.\V" et at. (1993), A biomeclwnical analysis of decompression and reconstruction methods in the cervical spine. Emphasis on a carbonfiber-compositc cagt. } BOllI.' .loilZI S/lrg, 7504, 1674-1684. Simmons, E.H. & Bradlev, D.D. (1988). Nturo-myorat flexion deformities (/ the cervical spine. Spille, 756-762. Steel. H.ll. (1968). Anatomical and mechanical conside tions of atlanto-axial articulation. In Proceedings of American Orthopaedic Association. J BOlle Joil/I SlIrg, 5 1481-1482. Sturzenegger, .\1., DiStel"ano, G.. Radano\', B.P., et a!' (199 Presenting symptoms and signs after whiplash injury: T influence of accident lllechanisms. NCllrology, .J.J, 688-6 Sutterlin, C.E., McAfee, P.e., \Vardell, K.E., ct al. (1988) biomechanical evaluation 01" celYlcal spinal stabilizat methods in a bovine model. Static and cyclical loadi Spillt', 13, 793-802. Takebe, K" Vitti, :v1., & Basrnajian, J.\!. (1974). The functi of semispinalis capitis and spli:nitls capitis muscles: electromyographic stud.v. Allat Rec, f79, 477-480. Viano, D.e. &. Gargan, il,'l.F. (1996). Headrest position dur normal driving: Implication to neck injury risk in r crashes. Accidellt Allal fln'I', 28, 663-674. Yoo. I..M .. Kumar"san. S .. Yoganandan. N.. 01 at. (1997). nite elemi:nt anal~'sis of ci:rvical facctectomy. Spille, 964·-969. Wallis, B.J., Lord, S.M., Barnsky, L., et al. (1998J. The p chological profiles of patiellts with whiplash-associa headache. Cephalalgia, IS, 101-105. WertH.', S. (1937), Studies in spontaneous alIas dislocati Acta Ort/lOp Sewld, 23 Suppl. White, A.A. Ill, Johnson, R.~L Panjabi, :\1.M., et al. (197 Biornechanical analvsis of clinical stabilitv in the cerv spill(:. Clill Orl/top, l(N, 85-96. . White, A.A. III & Panjabi. M.\1. (1990). Clillical Bio/llechal oj' tlze Spille. Philadelphia: J.B. Lippincott. Wittenberg, R.l-I., i\loeller, J.. &. \Vhite, A.A. III. (1990). C pressive strength 01" autogenous and alJogenolls b grafts 1"01' thoracolumbar and cervical spine fusion. Spi 15(101. 1073-1077. Yang, K.H., Latour. B.K., & King, A.1. (1992). Computer s ulation of occupant neck rtsponse to airbag deploymen I"rontal impacts. J J3io/llech Ellg, 11.J, 327-331. )'oganandan, N., Kumaresan, S.c., Voo, L., ct a!' (1996a). nite dement modeling of dll: C4-C6 cervical spine u lIed Ellg Phys, 18,569-574. Yoganandan, N., Pintar, EA., \Lliman, D.J., ct al. (1996 I-Iuman head-neck biomechanics under axial tension. l Ellg Pltys, 18,289-294. )'00, J.U., Zou, D.. Edwards, W.T., et al. (1992). Effect of vic,d spine motion on the netlroforaminal dimensions human c('['vical spine. Spinc, 17{ 1131-1136. "~~--------------~...",,.,...,..,..-----~~..... ....-. . - ._ ik ~ Biomechanics of the Shoulder ~: YJ,j i>' f/, Craig 1. Della Va/Ie, Andrew S. Rokito, Maureen Gallagher Birdzel Joseph D. Zuckerma Introduction Kinematics and Anatomy Range of Motion of the Shoulder Complex Sternoclavicular Joint Acromioclavicular Joint Clavicle Glenohumeral Joint and Related Structures Glenoid labrum Joint Capsule Glenohumeral and Coracohumeral ligaments Additional Constraints to Glenohumeral Stability Scapulothoracic Articulation Spinal Contribution to Shoulder Motion Kinetics Muscular Anatomy Integrated Muscular Activity of the Shoulder Complex Forward Elevation External Rotation Internal Rotation EXiension Scapulotr,oracic Motion loads at the Glenohumeral Joint The Biomechanics of Pitching Summary References It ----._------------------------~--------~._-._-~-------_._- """ :;;;Cor Introduction shoulder links the LIpper extremity to the trunk acts in conjunction with the elbow to position hand in space for crfkient function. [t consists of the glenohumeral. acromioclavicular, stcrnoclav· iculat: and scapulolhoracic articulations and the musculature structures that act on them La produce the most dynamic and mobile joint in the bodv (Fig, 12-J). An absence of bony constraints allo\I.'5 a wide range of motion at the expense of stability, which is profor instead by the various ligamentous and musstl1.!cturcs. The biomechanics of the shoulder arc complex, and a complete discllssion necessitates an analysis of the four aforementioned articulations that make up the shoulc!eI- complex. This chapter describes the Acromioclavicular iOin1 @ ~ /. I /' \ , '\ Sternum _ r ($ I \ t?r ~,\ i"-(l-- ,, , \ , Humerus I ~0L-n Scapulothoraclc I i , I I ,I J.I To produce the intricate motions necessary for no n1al functioning of the shoulder complex, the fou aniculations with their associated components ac togcther in a way that produces mobility greate than that afforded by anyone individual aniClda tion, The ability of the shoulder complex to positio the humerus and the remainder of the LIpper cx trcmit~· in space is further augmented by movemen of the spine. A discussion follows of the t\'pcs an ranges of motion for the shoulder complex as whole, with subsequent sections discussing th manner in which motion is achieved at each of th articulat ions, RANGE OF MOTION OF THE SHOULDER COMPLEX 7' Acromion o f . scapula \ Clavicle \ ,)" I Kinematics and Anatomy Stern1i~:~tCular -~ Scapula anatomy of the various aspeclS or the shoulder com plex and shows how their structure allows ror eff cient biotllcchanical function. _ Schematic depiction of the bony structures of the shoulder and their four articulations. The circular insets show front views of the three synovial joints-sternoclavicular, acromioclavicular, and glenohumeral-and a lateral view of the scapulothoracie joint, a bonemuscle-bone articulation. Adapted with permis5ion from De Palma, A.F. (1983). Biomechallin of the shoulder. In Surgery of the Shoulder (3rd 00.. pp.6S-85J. P!I;/adelp!l;d: J.8. Upp;ncorr Shoulder range of motion is traditionally measure in terms of flexion and extension (elevation o movcrncnt of the humenls away from the side of th thorax in the sagittal plane), abduction (elevation i the coronal plane), and intemal·external rotalio (axial rotation of the humerus with the arm held i an aclducted position) (Fig, 12-2), Although durin functional activit.ies these pure motions are rarel seen, we can better understand the complex mo tions of the shoulder by anal~/zing the separate com ponents needed to achieve anyone position, Although forward elevation of 180° is theOl'et cally possible, the average value in men is 167° an in women it is 171°. Extension or posterior eleva lion averages 60· (Boone & Azen. 1979). These va ues are Iimiled by capsular torsion. Abduction i the coronal plane is limited by bony impingemen of the greater tubcr()~ity on the acromion, Forwar elevation in thc plane or the scapula, thereFore, considered to be more Functional because in th plane the inferior portion of the capsule is no twisted and the musculature of the shoulder is op timally aligned for elevation of the arlll (Fig. 12-3 Although shoulder range of motion !10nl1all~' d creases as part of the aging process, physical activ ity can counteract lhis process (Murray ct al 1985). 31 "~ ~'~"~T" "'-_" "_' O' ' ' ' ' =' ' 'r. ....=" "~" ,.,. .,." ,- , ,_: " ' ' '-:' ' '-:' ' ~- ,' ' '., ,~, "'.=.". .- -". . . ,.-. .,.-"_"', ."'_,"', _J"'.;./"', ~. -7"+-r;" ':, :-o"'- :;'?"'~ _P ~"'"' ._--Ii!_'., _,.. •.=.. .. T":::;"': .. Extension Forward flexion ~ it lf , I , I , I I I , ,, :, I I ,, ,, :, --c:+ cs. ..- _........"..'...........-=>1 f I ,, I I !:::xternal and Internal rolation AbduClion Humerus at the side \\ ": '1,1 c B A A, Forward flexion. The humerus is in the sagittal plane. B, Extension. The humerus is in the sagittal plane. C, Abduction. The humerus is in the frontal plane. D, Rotation around the long axis of the humerus. External and internal rotation with STERNOCLAVICULAR JOINT The sternoclavicular joint consists 01" the enlarged medial end of the clavicle and the most superolateral aspect of the manubrium, linking the upper extremity directly to the thorax. In addition, a small facet is present inFeriorly that articulates with the Scapular plane elevation Long axis Long axis c3~~ Elevation in the scapular plane, which is midway between forward flexion and abduction. The humerus is in the plane of the scapula. D the humerus at the side, Internal rotation is shown with arm behind the back, which is a functionally important of this motion. first rib. This is a truc s\TH)vial joint that has a dle-like shape and contains a fibrocartilaginou ticular disc or meniscus that divides it into Cotlllxlrtmcnts. Although the joint itself has little intrinsic st ity, the articular disc in conjunction with ante posterior. C()stoclavicular, and intL't"clavicular ments maintains joint apposition (Fig. 12-4). anterior and posterior sternocla\'icular ligam resist anterior ;:md posterior translatiolls (as we superior displacement), while the costoclm'i lig;:Hllenl, which rUIlS betwcen the undersurfac the medial end 01" the chlviclc and the first rib sists upward as Wi..'l1 as posterior displacemen the clavicle (\'ia its anterior'portion) and is tho to be the major constraint in limiting stcrnoclav br motion. The interclavicular ligament conn the superollledial aspect or the clavicles ,:Hld as in restr;:lining the joint superiorl:v. '1'he post portion 01" the interclavicular ligament also. as with anterior restraint of' the sternoclavicular j Speciflcall)-', the interclavicular ligament tigh with arm depression and is lax when the ann is vated (i\'lorrey & An, 1990). The disc prevents dial disphlcelnent the clavicle, which m<l!' o when C<\IT)-'ing objects at the side, ;:IS well as infe or Sternoclavicular Joint Ant. sternoclavicular Jig. Articular disc ally directed coroid and the anterolaterally d rected trapezoid ligaments, which arc distin structures that serve different biomechanic functions. The smaller coroid ligament acts limit superior-inferior displacement of the clav cle. The quadrilaterally shaped trapezoid is t larger and stronger of the two ligaments and found lateral to the coroid; it resists axial com pression or motion about a horizontal axis. T Post. Protraction-retraction Demonstration of the anatomy of the sternoclavicular joint. Reprinted with permission from DeLee J. & Drez, D. I p~ (/9911). Orthopaedic Sports Medicine. Principles and Practice (p. 464). Phi/adelfJhl~): WB. Saunders Co. Sternum A displacement via articular contact. Although these structures act as important stabilizers, they still al~ Imv for significant motion including up to 50° of axial rotation and 35° of both superior-inferior cleva· lion and anteric)I'-posterior translation (Fig. 12-5). Elevation-depression Clavicle Meniscus l:7== ....... ACROMIOCLAVICULAR JOINT The acromioclavicular joint (Fig. 12·6) lies between the lateral end of the clavicle and the acromion of the scapula (the lateral and anterior extension of the scapular spine) and is subject to the high loads transmitted from the chest musculature to the upper extremity. It too is a synovial joint, but has a planar configuration. A wedge-shaped articular disc, whose function is poorly understood, is found \vithin the joint originating from the superior aspect. Both sides of the articular surface are covered with fibrocartilage and the joint itself slopes inferomedially, causing the lateral end of the clavicle to slightly override the acromion. A weak fibrous capsule encloses the joint and is reinforced superiorly' by' the acromioclavicular ligament. The acromioclavicular ligament acts primarily to restrain both axial rotation and posterior translation of the clavicle. The majority of the joint's vertical stability is provided by the coracoclavicular ligaments that suspend the scapula from the clavicle (Fukuda et aI., 1986). The coracoclavicular ligaments consist of the posteromedi- \ Ant. • B Rotation c Motion at the sternoclavicular joint. A, Top view showin the clavicular protraction and retraction (anteroposterio gliding) in the transverse plane around a longitudinal ax (solid dot) through the costoclavicular ligament, not shown. Motion takes place between the sternum and th meniscus. B. Anterior view showing clavicular elevation and depression (superoinferior gliding) in the frontal pla around a sagittal axis (solid dot) through the costoclavic lar ligament. not shown. Motion occurs between the cla cle and the meniscus. C, Anterior view depicting the clav ular rotation around the longitudinal axis the clavicle of • Coroid ligament Trapezoid li)7ment Coracoclavicular ligamenl Acromioclavicular ligament - Acromion Coracoacromial Intertubercular synovial sheath -1'-----') vasclll~ll' slrlH.:lllrl"S and SL'IYL'S as an altachmcn for Inan\ of Iht.' l1luscles dl41t ~Ict on tilL' sho Tht.' clavicle' also pro\'ides for the normal ap ance 4\[H.! contour or tIJL~ llPP(~J' chest. Elevati IhL' upper extrcmil,\' is aCCOl11lXlnicd b.\· rotati wdl as ck'\-',<1lion or the c1a\'idL.', with approxim -1-0 of c1m·icular dL.'\·ation for evcr~· 10'> of arm \·~ltion, \\·ith the majorit.\, of this motion occu at the stL'I"llocla\'indar joint {Inman, Saunde Ahhot!. 1944 J. GLENOHUMERAL JOINT AND RELATED STRUCTURES Humerus Scapula The coracoclavicular ligament complex consists of the larger and heavier trapezoid ligament, which is oriented laterally, and the smaller coroid ligament, situated more medially. Reproduced with permission from Hollinshead, WH. (/969). Anatomy for Surgeons (Vol. 3), New York: Harper & RO'll, coracoacromial ligament lies on the lateral side of the acromioclavicular joint and runs rrom the most lateral aspect of the coracoid to the medial aspect of the acromion. Although clavicular rotation does occur with arm elevation, Rockwood (1975) found little relative motion between the clavicle and acromion. This has been attributed to synchronous clavicular and scapular rotation with little resullant relative motion at the acromioclavicular joint: the majority of scapulothoracic motion occurs via the stcnlOclavicular joint. Thus, rigid fixation or fusion of the acromioclavicular joint produces little loss or overall shoulder function while sternoclavicular fixation leads to restricted motion or, more commonly. implant failure. Although motion at the acrol'nioclavicular, st chlvicular, and scapulothoracic articulalions is 10 the on:rall function or Ihe shoulder comple cClltral pla.n?r is the glt.'llohullleraljoinl. Tht..' a lar surface of lhe proximal humerus forms a arc and is CO\·cfL.,d with hyalin(' cartilag~ as glenoid fossa. The hUllleral head is rctnwt..'rlc postt..'riorl.\' dirt..'ctcd 3D'" wilh rt..'speci 10 tilL> cond.\'lar plan..: of the distal hUlllerus and has a ward or llledial indination of 45°; this configur gives the humerus an o\'eral! more anterior an eral orienl~\liol\ (Fig. 12-7). The greater and tuberosities lie lateral to Ihe articular surface o proximal hUIl1L'rus thal SCIY('S as tht.' all<:lchlllL'1 for IhL' rotator culT IllUsctdalUrc. The long he Lhe biceps lendon trii\vcrsL.'S till> bicipilal g CLAVICLE The clavicle lies between the two aforementioned articulations, acting as a strut connecting the thorax to the upper extrell1it~'. It is an S-shaped, double-curved bone: the medial two thirds of the body convex anteriorly while the lateral end is concave. It protects the lInderl~,..ing brachial plexus and The two-dimensional orientation of the articular sur the humerus with respect to the bicondyrar axis, Rep ......rth pemllC;SIOfl from Rockwood, C. & Matsen, f (J 990). Shoulder (p. 219). Pillladelphia. ~'V.B. 5.:1lmderc; Co lies between the tuberosities) beneath the tr",nsve,,'se humeral ligament (Fig. 12~8). The proximal humerus articulates with the glenfossa, which itselF is retroverted 7° and superi~ inclined 5° relative to the plane of the scapula 12-9, A & B). This slight superior inclination n,-OV\C!,'5 a significant degree of geometric stability, to resist inferior subluxation or dislocation et al., 1992). The glenoid fossa is shallow and is to contain only approximately one third of the di,wlet"r of humeral head. The bony architecture is by the cartilaginous surface, which is thicker peripherally than it is centrally, acting to slightly but significantly increase the depth of the Greater tuberosity Transverse humeral ligament ----1''';..-1 ---+ Long head of - - - t the biceps Short head of -----,f-tf-the biceps I B A A, The glenoid is retroverted r with respect to the p perpendicular to the scapular plane. B, The glenoid f superiorly approximately 5°. Reprinted ,·vith permission from Simon, S,R, (Ed). (1994). Orthopaedic Basic Science (p. 526-527). Rosemont. IL: AtlQS glenoid as a whole. Although the congruity betw the articular surfaces of the proxin1al humerus glenoid was previously' thought to be somewhat precise, stereophotogrammetric studies (\vhere fine grid is projected onto an object that is then p tographed From two different directions in the text of known reference targets, allowing for a th dimensional reconstruction) have shown articulation to be precise, \vith the deviation f sphericity' of the convex humeral articular sur and concave glenoid articular surface being than 1% (Soslowsky et aI., 1992). Less than 1.5 of translation of the humeral head on the gle surface has been demonstrated in normal subj during a 30 arc of 1110tion (Poppen & Walker, 19 thus, motion at the glenohumeral joint is alm purely rotational. Given the paucity 01' bony straint, stability is instead provided by the caps ligamentous, and muscular structures that round the glenohumeral joint. 0 The long head of the biceps goes in the bicipital groove between the greater and lesser tuberosities. The transverse humeral ligament helps stabilize the biceps tendon in the groove. Reprinted with permission from Neer, C. (1990). Shoulder Reconstruction (p. 29), Philadelphia: WB. Saunders Co. Glenoid Labrum The glenoid labru111 (Fig. 12-10) is a fibrocarti nous rim that acts to deepen the glenoid, provi 50°;(; of the overall depth of the glenohumeral j (\Varner, 1993). It has a triangular configura \'.'hen viewed in cross-section and has firm att Coracoacromial ligament Glenoid --',--'\l;--i. labrum Coracoid process Glenoid --+--~: cavity Coracohumeral ligamenl Tendon of long head of biceps brachii Long head of tTiceps brachll A A, The glenoid labrum is attached to the underlying bony glenoid and is confluent at its area with the long head of the biceps tendon. B. The labrum has a triangular configuration menls inferiorly to the underlying bone, having morc variable and looser atlachments in its superior and anterosuperior portions. The superior portion or the glenoid labrum is connucnt with the tendon of the long head of the biceps and. along with the adjacent slIpraglenoid tubercle. serves as its site of insertion (Moore. 1999). iVlcasurcmenlS of the force needed to dislocate the humeral head under constant compressive pressure have shown that with an intact labrum, the humeral head resists tangential forces of approximately 60% of the compressive load; resection of the labrum reduces the effectiveness of compression·stabilization by 20% (Lippitt ct aI., 1993). Detachment of the superior labrum with anterior-posterior extension (i.e., "SLAP lesion") can occur from traction (repetitive overhead activities or a sudden pull on the arm) or compression (a fall onto an outstretched arm). This lesion can be a cause of severe pain and shoulder instability (hoi, Hsu, & An, 1996) as a result of significant increases in glenohumeral translation whcn compared with the intact shoulder (Pagnani et al., 1995). (Fig. 12-11). when vie'Ned in cross·section and serves to effectively deepen the glenoid, increasing the stability of the glen hllmeral joint. DEmll.- Oblique coronal image showing a tear involving th sertion of the long head of the biceps tendon and superior labrllm (arrow). ~---------- Capsule '/fZ;/-:- The ~:dcnohLll11cral joint capsule has a significant de-of inherent laxity with a stll-face area llull is $: -~~~~t\l"rec ,,,,,,"'-~"-;:;> <"W'~""twice that of the humeral head (Warner, 1993), This ::~jZkJLredtindancyallows for a wide range of mOL ion. Mc,;,;" "idiallv, the capsule attaches both directly onto (an:~;~:;.:terotnferior1y) and beyond the glenoid labrum and ~~-- laterally it reaches [Q the anatomical neck of the ;'~lh~humerlls. Superiorly, it is attached at the base of the coracoid. enveloping the long head of the biceps ~'7F~;~tendon and mnking it an intl'n-articular structure dfB"HFig, 12-1 I), -!:%ihi The capsule nlso has a stabilizing role, tightening ,-._. with various arm positions. In adduction, the capsule is Lattl superiorly and lax inferiorly; with abduction of the upper extremity, this relationship is ,'ev'er<;ect and the inferior capsule tightens. As the is externally rotated, the anterior capsule tightwhile internal rotation induces tightening pasteriOl'lv, The posterior capsule in particular has been to be crucial in maintaining glenohumeral stability, acting as a secondary restraint to anterior dislocation (particularly! in positions of abduction) as well as acting as a primar~y posterior stabilizing structure (Itoi, Hsu, & An, 1996), 'itrtY:' Glenohumeral and Coracohumeral ligaments The three glenohumeral ligaments (SUPCriOI~ middle, and inferior) are discrete extensions of the anterior glenohumeral joint capsule and al'e critical to shoulder stability and function (Fig, 12-12), The superior glenohumeral ligament originates from the anterosupcrior labrum, just anterior to the long head of the biceps, and inserts onto the lesser tuberoSity. It is present in the majority of shoulders bllt only well developed in 50%, The superior glenohumeral ligament acts as the main restraint to inferior translation with the arm in the resting or addllcted position (Warner el aI., 1992), The coracohumeral ligament originates from the lateral side of the base of the coracoid to insert on the anatomical neck of the humerus (Fig, 12-6) (Cooper et aI., 1993). This structure lies anterior to the superior glenohumeral ligament and reinforces the superior aspect of the joint capsule, These ligaments span the rotator interval belween the subscapularis and supraspinatus, Some research incHcates thal these strUClures have a secondary role in preventing 'inferior translation of the shoulder while in the adducted, neutrall.v rotated posilion. The functional significance of the coracohumeral lig rne-nt, however, seems to be related to the overa development of the glenohumeral ligaments in given individual, having a larger role in those with less-developed superior glenohumeral ligame (Warner et aI., 1992), The middle glenohumeral ligament originates i ferior to the superior glenohumeral ligament (at th I o'clock to 3 o'clock position, right shoulder) an inserts fl.1rther laterally on the lesser tuberosit Great variability in the anatomy of this structu has been demonstrated, being absent in as many 30% of shoulders (Curl & Warren, 1996), It ma originate from the anterosuperior portion of th labrum, the supraglenoid lubercle, or the scapul neck, Morphological variants have been descl'ibe including a cord-like variant (clearly distinct fro the anterior band of the inferior glenohumeral lig ment) and a sheet-like variant (blending with th anterior band of the lnferior glenohumeral lig ment). Functionally, the middle glenohumeral lig ment acts as a secondary restraint to inferior tran lations of the glenohumeral joint with the arm the abducted and externally rotated positio (vVarner el aI., 1992). It also sen/cs as a restraint anterior translation, having its maximal effect wi the arm abducted 45°. The inferior glenohumeral ligament originat from the inferior aspect of the labrum and inser on the anatomical neck of the humerus. It has bee shown to have three distinct components (O'Brie et ai., 1990): an anterior band originating from 2 4 o'clock (right shoulder), a posterior compone originating fTom 7 to 9 o'clock (right shoulder), an an axillat), pouch (Fig. 12-12), The inferiOl· glen humeral ligament has the greatest rll11ctional signi icance, acting as the primm)' anterio.' stabilizer the shoulder with the arm in 90' of abduction, A the arm is abducted and externally rotated. the a terior band of the inferior glenohumeral ligame tightens, resisting anterior tninslalion. \·Vith inte nal rotation of the abducted arm, the posterior ban becomes taut and posterior translation is resiste The inferior gleno·humeral ligament complex al serves to resist inferior translation of the glen humeral joint with the arm in the abducted pos tion. Variability as to the size and attachnlent sit of the gle-nohumeral ligaments has been demo strated (Warner et aI., 1992); howe vet; the clinic significance of this has yet to be fully elucidated a though it has been suggested thal absence of t middle glenohumeral ligament may predispose instability (Steinbeck et aI., 1998), A- B A ) AS c D A, Schematic drawing of the shoulder capsule illustrating the location and extent of the inferior glenohumeral ligament complex (IGHLC). A, anterior; P, posterior; B, biceps tendon; SGHL, superior glenohumeral ligament; MGHL, middle glenoo humeral ligament; AS, anterior band; Ap, axillary pouch; PS, posterior band; Pc, posterior capsule. Reprinted with pertnis· sian from O'Brien, SJ, Neves, M.e, Amoczky, S.P, et at. (1990) The anatomy and histology of the inferior glenohumeral ligament complex of the shoulder. Am J Sports Med. 18, 579-584. B. The superior glenohumeral ligament is the primary restraint to inferior translation in the adducted shoulder at neutral rotation. In this position, the middle glen?humeral ligament and the anterior and posterior bands of the inferior gleno· humeral ligament complex remain lax, C, The anterior b is the primary restraint resisting inferior translation of t shoulder at 45<> abduction and neutral rotation, In this tion, the superior glenohumeral ligament, the middle g humeral ligament, and posterior band are lax, D, At 90° duction, the anterior and posterior bands of the inferio glenohumeral ligament cradle the humeral head to pre inferior translation, The posterior band is more significa external rotation, whereas the anterior band plays a gr role in internal rotation. Reprinted wirh permission from W IP, Deng, X.H.. Warren, R,F., et al. (1992). Staric capsuloliga taus restraints ro superior-inferior translations of rhe glenohu joint. Am J Sports Med, 20, 675-678. Experimental Techniques: Ligament Cutting Studies Shoulder Instability A Ligament cutting studies have been instrumental in fur- . thering our knowledge regarding the contribution of a 21-yea:"0Id ma.n fell while skiing onto his right upper extremity, cauSing forceful abduction and external ro" given anatomical structure to overall glenohumeral sta- tation. He noted acute pain in the arm and was unable to bility (Curl & \JVarren, 1996). In this technique, cadaveric specimens are biomechanically tested before and IYlQveit Physical examination revealed loss of the normal contour of the shoulder and painful range of motion. Ra" dlogr~phs showed an anterior dislocation with posterior after selectively cutting sequential structures. A force is then applied in a given arm position, and the transla- superi.()rhumeral head impaction fracture. The patient un- tion that occurs is measured. From this information, the gerwent closed reduction. Postreduction radiographs con" relative contribution that a given structure provides to firlYledreduction of the humeral head with a small bony overall stability can be determined. When a particular ayulsion fracture of the anterior-inferior glenoid rim, pattern of shoulder instability is then identified, the whichrepresents a detachment of the anterior labrum in physician can infer which anatomical structures may be the area of the superior band of the glenohumeral liga- deficient or disrupted so that a rational plan of physical ment insertion (Case Study Figure 12" 1" 1). Structural alteration of bony geometry, ligaments, and therapy or surgical repair can be implemented. • labrum resulted in shoulder instability. The detachment of the anterior labrum and the superior band of the ligament insertion affected primarily the resistance to anterior Additional Constraints to Glenohumeral Stability Sy'l1ovial fluid acts via cohesion and adhesion to further stabilize the glenohumeral joint. Synovial fluid adheres to the articular cartila~e ovcrh;in cr the glenoid and proximal humerus, c;using the b two surfaces to slide along one another: Th~ synovial fluid provides a cohesive force bet\veen the~e two, making it difficult to pull them apart (Simon, 1994). Under normal conditions, the intra-articular pressure \vithin the gknohumeral joint is negative, acting to pull the overlying capsule and glenohumeral ligan1cnts inward. If the integrity of the glcnohun1cral joint capsule is cornpro~1is~d (e.g., venting the capsule) or if a significant effusion exists (normally the glenohumeral joint contains less than 1 cc of fluid), significant increases in translation are observed (Kumar & Balasubramaniam, 1985). Specifically, venting the capsule reduces the force needed for anterior humeral head translation by 55 0ft;, for posterior translation bv 430c), and for inferior translation by 57% (Gibb et ·al., 1991) (Case Study 12-1). translation of the humeral head, resulting in anterior dislocation. !n addition, a concomitant capsule lesion affected the intra"articular negative pressure necessary to pull the humeral head inward. After conservative management. the patient did well for 6 months until he sus" tained a recurrent dislocation. The patient subsequently opted for operative treatment that included repair of the fracture and capsule and a complete period of rehabilita" tion~vithernphasis on joint stability and proprioception. SCAPULOTHORACIC ARTICULATION The scapula is a Flat, triangular bone that lies on the posterolateral aspect of the thorax between the second and seventh ribs. It is angled 30° anterior to the coronal plane of the thOl;X and is rotated slightly toward the midline at its superior end and tilted anteriorly with respect to the sagittal plane Case Study Figure 12-1"1. (Fig. 12-13) (Laumann. 1987). The spine of the scapula gives rise laterally to the acromion process that articulates with the distal clavicle at the acromioclavicular joint. The coracoclavicular ligaments and muscular attachments help to suppon the scapula and stabilize it against the thorax (Fig. 12-6). There is, however, no osseous connection with the axial skeleton. This allows for a wide range or scapular Illotion, including protraction, I"etraction, elevation, depression, and rotation. The scapulothoracic articulation involves gliding of the scapula on lite posterior aspect of the thorax. Interposed between the scapula and the thoracic \vall lie the subscapularis (arising from the costal surface of the body of the scapula) and the serratus anterior, which help to stabilize the scapula against the chest wall and thus prevent "scapular winging" (Fig. 12-14). These two muscles glide along one another to provide greatly enhanced mobility of the shoulder complex as a whole, Elevation of the arm involves rnotion at both the glenohumeral joint and the scapulothoracic articulation. Although the contribution from each varies according to arm position and the specific task bclng performed. the average ratio of glenohumeral to scapulothoracic moL ion is 2: I (Tibone el aI., 1994). Elevation of the arm also induces complex rotatory motion of the scapula, \vith anterior rotation during the first 90° followed by posterior rotation with a total arc of approximately 15° (Morrey & An, 1990). -.J 3 'r Scapular I I Anterior view of the scapulothoracic articulation, a I11llscle·bone articlJlation between the scapula and During scupular motion, the subscaplJlaris muscle. Io attaches broadly to the costa! surface of the scapula on the serratus anterior muscle, which originates or first eight ribs and inserts into the cost<ll surf..'!ce of scapula along the length of its vertebral border. .---------------SPINAL CONTRIBUTION TO SHOULDER MOTION Allhaugh often O\'crlookcd, motion 01" the th and lurnlmr spine contributes to thL' :.\bilit:,· t tion the upper c."tl'(.'l1lit~· in space, {ht.T('b~· l.' ing the on.:r<.111 Illotion and functioll of the sh cOlllpk.·x. f'kxiol1 01" tile spine away frolll an ,il~' altempting ttl reach an object o\'crhcacl t.'n the range of motion <.lllainahlt: (Fig. 12·15), T ponancc or spinal motion in O\·l..'rlH.':'ld act such as throwing ~Illd r<.tcqul.'l sports h<.ts als c1emonst rated. Acromion 0 ~l SerraIus anterior spine I. Scapula Kinetics Scapular orientation on the chest wall. Left. 30~ anterior, Right. 3" upward, Reprinted with permission from Warner JJP: The gross anatomy of the joim surfaces, ligaments, labrum, and capsule. In: Matsen. F.A., FtJ, F.H., Hawkins, R.I (Eds.) The Shoul- der: A Balance of Mobility and Stability. Rosemont, IL: American Academy of Orthopaedic Surgeons, J 993. Numerous llluscll..'s aCI nn the \'arious comp the shoulder complex to provide both m and d\'namic $t~lbilit.\·. Dynamic stabilizati cllrs via sl.~\'cral possible lllcch:'lnisms (Mo An, 1990), including p~lssi\'(' musch.' lensi via a b:,uTic:r effect of the contracted Illuscle prl..'ssi,·L' forces broughl ahout h~' llHISCU!:'U traction, .ioint lllotion that induces lighten or "fmm'-----I activated differently for specific activities. The toid originates from the lateral third of the clav acromion, and scapular spine and inserts on the terolateral aspect of the humerus. The anterior h acts as a strong flexor and internal rotator of humerus, the middle head as an abductor, and posterior head as an extensor and external rot The pectoralis major lies over the anterior c wall and has two heads, a clavicular head orig ing from the side of the clavicle and a sternoco head originating from the sternum, manubri and the upper costal cartilages. The two heads verge at the sternoclavicular joint. The pecto major inserts at the intertubercular groove of humerus between the tuberosities and act adduct and internally rotate the humerus. Se darily, the clavicular head acts as a flexor or fon elevator of the humerus while the sternocostal extends the humerus. The pectoralis minor lies to the pectoralis major, functioning as an impor scapular stabilize!: The pennate subclavius mu lies inferior to the clavicle and Illay assist in cl ular motions. It has a tendinous origin from the teromedial aspect of the first rib and insert lateral bending of the spine enhances the ability to position the upper extremity. Reprinted with permission from Rockwood, C. & Matsen, F (1990), The Shoulder (p. 2' 9). Philadelphia: WB. Saunders Co. Pectoralis major Clavicular Sternal Deltoid Posterior Middle Anterior the passive or ligamentous constraints, or via a redirection of the joint force toward the center of the glenoid. To understand rnusclc function and force transmission. one must consider a given muscle's orientation, size, and activity. Given the multiple articulations present in the shoulder complex, an:y given muscle may" span several different joints, and depending on the position of the upper extremity', its relationship with regard to anyone articulation may change, altering its effect on that joint and the resultant forces or motions produced. MUSCULAR ANATOMY The shoulder musculature can be thought of in layers. The outermost la)-!er consists of the deltoid and pectoralis major muscles (Fig. 12-16). The deltoid forms the normal, rounded contour of the shoulder and is triangular in shape, with anterior, middle, and posterior heads. Each portion of the deltoid is Serra ante Anterior sheath of rectus Pectoralis minor Anterior view showing the superficial muscles (left sh der) and the deep muscles beneath the deltoid and p toralis muscles (right shoulder). the undersurface of the medial clavicle (Morrev & Anterior View An, 1990). BcnL~alh this ouler layer lies the rotator cuff musculature: the supraspinatlls. infraspinatus. subscapularis, and teres minor (Fig.12-17). These four muscles act to abduct and rotate the humerus and act as important glenohumeral stabilizers via both passive muscle tension and dynamic contraction. The supraspinatlls originates from the supraspinatus fossa of the scapula and inserts on the greater tuberosity of the prox.imal humerus. It forms a force couple with the deltoid during abduction of the humerus. The infraspinalllS Biceps and teres minOt- originate from the inferior as- pect of the scapula and insert on the greater (uberosity. These muscles act as external rotators of the humerus. The subscapularis lies on the costal surface of the scapula and inserts on the lesser tuberosity of the proximal humerus. It functions as an important internal rotator of the humerus. The subscapularis. along with tlte mid~ dIe and inferior glenohumcral ligaments, has also been shown to act as an important anterior stabilizer of the glenohumcral joint, particularly with lJ1(.> arm held at 45° of abduction. The teres major muscle (Fig. 12-18), while not part of the rotator culT, also originates from the scapula, but at its inferior angle coursing inferior to the teres minor and then passing anteriorly to insert on the humerus at the intertubercular groove. h functions to assist with arm adduction and internal rotation. The biceps muscle is also involved with motion of the shoulcleI· complex. It is composed of two heads: a short head that originates fTom the tip of the coracoid process of the scapula and a long head that originates I'T om the superior glenoid labrum and supraglenoid tubercle (Fig. 12-8). The tendon of the long head of the biceps lies within the glenohumeral joint and descends between the greater and lesser tuberosities. joining the short head to insert on the bicipital tuberositv of the radius. The biceps fl1nclions to Oex and supinate the forearm and elevates the humerus. The long head of the biceps also acts as a humeral head depressor and, as such, plays a role in maintaining glenohumeral stability (hoi et aI., 1994). Several muscles lie on the back and act directly on the scapula (Fig. 12-18). The outermost layer consists of the trapezium that covers the posterior neck and uppermost portion of the trunk, inserting on the superior aspect of the lateral one third of the clavicle, acromion, and scapular spine. The tl'apez- Subscapularis A Posterior View Supraspinatus " Infraspinatlls B A, Anterior view. The "rotator interval" is a term duced in 1970 to indicate the space between the supraspinatus and subscapularis tendons. The cor humeral ligament lies superficially along its ante where it is readily available for release as indicat long head of the biceps lies deep along its poste and serves as a guide to this interval during surg Reprinted ~'llth permission from Neer. C. (1990). Shou construetiofl (D. 29). Pllild(/e~ohia: !NB. Saunders (0_ rior view_ The two external rotators of the hume ..infraspinatus and teres minor muscles. which are posterior wall of the rotator cuff. Note the medi of the infraspinatus, which is often mistaken at s the border between the infraspinatus and the te Reprinred with permission [rom Rock\'-Ioocl, C. & Mat (/990). The Shoulcier (p. 2) 9). Philadelphia: \>';1 B. Sau ,l L Deltoid Anterior ~-'\-- InfraspInatus Middle '-=--+- Rhomboideus major Posterior Teres minor Teres major ) Latissimus dorsi Posterior view shoWing the superficial muscles (left shoulder) and underlying muscles (right shoulder). ium serves to elevate. retract, and rotate the scapula. The latissimus dorsi covers the inferior portion of the back, inserting on the intertubercular groove of the humerus. It acts to extend. adduct, and internally rotale the humerus. Below these muscles.lie the levator scapulae superiorly, which elevale and inferiorly rotate the scapula, and the rhomboid muscles below, \vhich retract and rotate the scapula. Both of these muscle groups nctLO assist the serratus anterior (which lies laterally on the chest and intercostal muscles inserting onto the me~ dial border of the anterior surface of the scapula) in keeping the scapula fixed to the trunk. INTEGRATED MUSCULAR ACTIVITY OF THE SHOULDER COMPLEX .' .;:. Electromyography allows for the quantification of muscular activity during dynamic conditions. This permits insight into the level of muscular ac- tivity but does not directly indicate forces gene ated. A complete understanding of the latter quires knowledge of lhe moment arm (measur as the distance between the instantaneous ccnt of rotation of thc joint and the distance of mu cular pull) and the physiological cross~section the involvcd muscle ('measured as the muscul volume divided by its length). In the should complex, each motion is associated with mov ment at multiple articulations and the constan changing relationships of the muscular origi and insertions. Given the paucity of osseous stabilil~' at t glenohumeral joint, force generated by one mu cle: (the primary agonist) requires the activati of an antagonistic muscle 50 that a dislocati force does not result (Simon, 1994).. The antag nist usually accomplishes this via an eccentl contraction whereby the mU5cfe is lengthen while actively contracting or via the production a neutralizing force of equal magnitude but in the opposite direction. The relationship between two such muscles is also referred 1O as a force couple (Fig. 12-19). Around the glenohumeral joint. there is a force couple in the coronal plane (between the deltoid and the infedor portion of the rotator curf) and in the transverse plane between the subscapularis muscle anteriorly and the posterior rotator Cliff Illusculature (the infraspinatus and teres minor). Relative Illotion is produced by an imbalance bet\veen the agonist and antagonist that produces torque. The degree of torque and the resultant angular velocity produced is determined by the relative activation of two such muscles or muscle groups. Resultant muscular forces arc determined via an understanding of the cross-sectional area of the activated muscles involved and their orientation at the time of activation. A B The deltoid and the oblique rotator cuff muscles (infraspinatus. subscapularis. and teres minor) combine 10 produce elevation of the upper extremity by means of <1 force couple (two forces equal in magnitude but opposite in direction). With the arm at the side (A), the directional force of the deltoid is upward and outward with respect to the humerus while the force of the oblique rOlator cuff muscles is downward and inward, These two rotational forces can be resolved into their respective vertical and horizontal components. The horizontal force of the deltoid acting below the (enter of rotation of the glenohumeral joint is opposite in direction to the horizontal force of the oblique rotalOrs. which is applied above the center of rolation. These forces acting in opposite directions on either side of the center of rotation produce a powerful force couple. as illustrated by the arm signal (B), The vertical forces offset each other, thereby stabilizing the humeral head on the glenoid and allowing elevation to take place. Adapted with permission from Lucas, 0.8. (1973). Biomechanics of the shoulder joint. Arch 5urg. 107.425. Forward Elevation TIll' most b~\~,dc motion (If thl.· shoukkr compk. \·oln.'s ck\'~lIioll of tIlL' arm ill the scapular p This motion has hel'n studiL'd in depth via both ll'omyograph}' and stcrcopllOlogrammctl-.v. muscles of the shoulder girdk han: subsequ heen grouped according: to relatin.' illlponancL" regard to this rnotion. TIll' first grouping inc the dcILoid <Spccificall~.. lhe anterior and m heads). lrapeziulll (inf..:rior ponion), supraspin and serratus ~lJ)leri()r (Simon, 10(4), The se grouping consists uf' the Iniddlc portion o trapeziulll. the infraspinatus. and thl..' long he the biceps. r\ third group consists or Ilw pos head or the ddlOid. thl.' c1~l\·iclilar Ill,:'~ld of tht.~ toralis major, and lht.' slIpi.:rior ponion or the tr iUIll. The fourth and final group includes I hi.' st head or the pectoralis rnajor. the latissirnlls d and t Itt.' long !It.'ad or the t ricL'ps. Thl.' inter-rt..'!i.lriol1ship bL'l\\·ct.'1l the mus rorces in\·oh'l.'d in shoulder dl.'\'~llion W<lS firsl ied b,\' Inman, who round that the deltoid sllpr~ISpiJlatus \ulrk s}·nergisticall.,· ,,·hile th mainder of till".' rotator cufllllusculalllrc provi hUllleral dcpn:ssing forct.' 10 counter sllbluxati lhl.:.' humeral !lead (Illll1~\I1. Saunders. & Ab 19-1-1). Thus. the vertic'l1h· orien'ed pull o[ th (oid is offset h~' a net inferior rorce crcat(:d b~' th fraspinatlls, suhscapularis, anc! teres millOI'. EIt.:.'ctrom.,'ographic studies have shown tlwt lhe supraspilwluS and deltoid ~\rc acti\'L' throug the nlllgt..' of ann clev"ltioll. ThL' supraspinatus. ever, is felt to h~l\'c a larger role in initiating a tion. As the arm is progrcs~i\'d}" dL'\'~\ted from side, the 1ll0!llCI\t arm of till".' deltoid improvl. suiting in a larger force in I\.'lation to thc supras tus (Fig. 12-20), The p<.:rccllwge 01" shearing or cal force created by the deltoid likewise ck'crt with increasing arnollllts or ahduction. The an pull or the supraspinalUs'is more l'OIlSI~lJll a pro:dmalcl.,· 7,=,°, acting not onl.\· to elcvate or a the arm but also to compress the hurneral within the glenoid. Thl.' remaining roW tor cull cles pull at approxil1latd~' -15°, which is din;(:ll" {'edod}; (slightl,'· highcr in the ter('s minor at 55 sulting in forces lh~lt l.·quall}' compress ;:111(1 de the humeral head to maintain glcnohulllcral s ity (Fig. 12-21). Sl~\('ctive anesthetic block or tht' n:dllary (and resulting dcILoid paralysb) dcnlOnsll·atl.' forward elevation is possihle alb~it signinc weakencd. Lik.:,\·isc. a supr~lscaplilar nerve Subscapularis Supraspinatus \ ff,o':::;7'1:_u..p",ra::-s-,-p_in_a-tIU.S -~------­ I As the arm is abducted to 90", the direction of pull of the deltoid approximates that of the supraspinatus. Therefore, patients with a large tear of the rotator cuff often can actively maintain the arm abducted to 90° but may not be capable of actively abducting to 90", Reprinted with permission from Simon, S.R. (Ed). (/994) Orthopaedic Basic Supraspinatus Science (p. 527). Rosemont, IL: AAOS. I --I-:~~~~~~~ 9 and the resultant supraspinatlls paralysis induced has a similar eHeet. However, a block of both nerves results in a loss of arm elevation (Colachis, & Strohm, 1971; Howell ct a\., 1986) (Case Study 12-2). \Vhen pure abduction is compared with pure forward elevation, the same basic relationships are seen with the rotator cliff acting to stabilize the glenohumeral joint while the deltoid provides the necessary torque. Forward flexion results in activation of the anterior and middle deltoid (73 and 62% activity, respectively') \vith stability provided mainly by the supraspinatus. infraspinatus, and latissimus dorsi, the lalter being particularly active (25°;0 activation) with forward flexion beyond 90°. Pure abduction requires similar muscular activity; however, the subscapularis shows increased activation as it acts as the prime stabilizer via eccentric contraction. Infraspinatus - Teres minor External Rotation The primary external rotator of the humerus is the infraspinatus. with significant contributions made by the posterior head of the deltoid and the teres minOI: With any given amount of shoulder abduction, electromyography reveals the prime external rotator to be the infraspinatus. The subscapularis is similarly active but serves an antagonistic role as the main stabilizer preventing anterior displacement of The angle of pull of the subscapularis (top) is approximately 45°. The angle of pull of the infraspinatus (bottom) is also approximately 45°, and the teres minor (bottom) is approximately 55°. These vectors result in nearly equal glenohumeral joint compression and humeral head depression. The supraspinatus (center) is essentially horizontal in its orientation, resulting in compression of the glenohumeral joint. • ----_._.__ ._----_._~~----~-_._._-----~- ----._- Subacromial Impingement Syndrome and Rotator Cuff Tear ·. A 53-year-old right-hand-dominant woman who presents right shoulder pain of 6 months' duration, The patient noted increasing pain at night and with overhead activities of daily living such as putting on a shirt and brushing her hair. The pain is unresponsive to pain killers. Physical examination revealed diffuse tenderness over the shoulder and deltoid with.painful forward elevation above 60:3 and internal fOla- still exert Its effects by Wc1Y of the "spans" of the bridge. Thus. patients may have a "functional" rotator Cliff tear in \-vhich they are s!ill able to periorm overhead activities. Ii the tearing force presem at the t\:'IO suppons is great enough. howe'/er, worsening of the tear will occur. tion limited to her ipsilateral greater trochanter. She had a positive Neer sign (painful forward ele'l<ltion between 60 and 120°) and a positive Hawkins sign (pain with passive abduction and. internal rotation). A subacromial impingement test was positive with near complete relief of pain and restoration of fOl\lvard elevation to 1500 but persistent \;veakness on resisted shoulder abduction. A diagnosis of subacromial impingement syndrome is made. The patient was initially managed with conservative treatment. At 5 weeks follow-up. the patient had persistent pain and an MRI of her shoulder revealed a full thickness tear or the supraspinatus tendon. The patient opted for operative management (Case Study Fig, 12-1-1). Biomechanically, rotator cuff tears have been likened to a suspens.ion bridge, whereby the free margin of the tear corre· sponds to the cable and the residual atlachrnents correspond to the supports at each end of the cable's span. This coniiguration allows the muscle that has been torn at its insertion to the humeral head with external rotation. As the amount of shoulder abduction is increased, the pos~ terior deltoid increases in cff-!cicncy as an accessory external rotator the humerus secondary to improvement of its moment ann. or Internal Rotation Internal rotation of the shoulder is accomplished by the subscapularis. sternal head of the pectoralis major, lHtissinlus dorsi, and teres major. The subscapularis is active during all phases of' internal rotation, with decreased relativc activity seen with cxtremes of abduction. 1n the samc \Va.v. activity of the sternal head of the pectoralis major Hnd the latissimus dorsi decreases with abduction. However. the posterior and middle heads of the dehoid compensate with increased eccentric activity during internal rotation while the arm is abducted. Extension Extension of the upper extremity is accompl by the postcrior and middk heads of the de The supraspinatus and subscapularis are also tillually active throughollt arm extension, res forces via eccentric activil~« that would tend to c anterior dislocation. Scapulothoracic Motion at the scapulothoracic articulation a maintenance or deltoid tcnsion, allowing maintain optimal power regardless of arm lioll. \-Vith forward elevation at the arm, scapula rotates, increasing stability at the g humeral joint and decreasing the tendency 1'0 pingement or the rOlator cufe bcncmh acromion (Fig. 12-22). A rotalional rorc~ CO (two equal. noncollinear, parallel but opposite Motion forces) bet\veen the upper trapezium, levator and upper serratus anterior with concontraction of the lo\ver trapeziurn and cP'T"llllS anterior leads to scapular rotation that is for full forward elevation (Fig. 12-23) 1994). .u,","""o> AT THE GLENOHUMERAL JOINT glenohumeral joint is considered to be a load-bearing joint. Although calculations the exact forces acting on it are challenging thc large number of involved muscular and possible positions attainablc, sevsin1plifying assumptions allow an estimaof the magnitude of these forces. A freediagram of a person holding the upper held at 90' of abduction can be utias an example; it is assumed that only! the ue"cm, muscle is active and that it acts at a distance of 3 cm from the center of rotation of the humeral head. Three forces are then considere the deltoid muscle force (D), the weight of th arm (equivalent to 0.05 body weight [BW] actin at the center of gravity of the limb, 3 cm), an the joint-reactive force at the glenohuIl1eral joi (1). The joint reactive force and the force of th deltoid, being nearly! parallel, are considered be a force couple and are of equal but opposi magnitude, The force on the glenohumeral joi \vhen holding the arm at 90° of abduction can b estimated to be one-half of body weight (Calcul tion Box 12-1, Case A). 11' a weight (W) of 2 kg added (equivalent to 0,025 body weight of an SOmale) to the hand of the outstretched extremi held at 90° of abduction, a sirnilar calculation ca be made (Calculation Box 12-1, Case B). Experimentally, loads at the glenohumeral joi and the forces necessary! for arm elevation hav been detcrmined. These forces have been found be greatcst at 90° of elevation, with the deltoid for equivalent to 8.2 times the weight of the arm an Forward elevation or abduction 0-120 o of the arm requires synchronous rotation of the scapula. Reprinted "'lith permission {rom Simon, S.R. (Ed). (1994). Orthopaedic Basic Science (p. 1 527-535). Rosemont, IL: AAOS. - - - - - - - Rotation of the scapula is produced by the synergistic contractions of the lower portion of the serratus anterior and the lower trapezius, with the upper trapezius, levator scapulae, and upper serratus anterior. Reprinred with permission from Simon, 5.R. (Ed). (1994). Orthopaedic Basic Science (p. 527-535). Rosemont. IL: AAOS the joint reactive force equivalent to 10.2 times the weight of the arm (Inman, Saunders, & Abbott, 1944). More recent investigations of these same forces utilizing the assumption that the muscular force is proportional to its area multiplied by its ac~ tivity as determined by electromyography have yielded similar values. with a maximal joint reactive force of 89% of body weight seen at 90 0 of elevation in the scapular plane (Poppen, & Walke.; 1978). THE BIOMECHANICS OF PITCHING Pitching has been divided into five stages: wind up, earl.y cocking. late cocking, acceleration, and follow-through (Tibone et 'II., 1994). The deltoid has been found to be responsible for elevation and abduction of the humerus in the carly phases. followed by increased activation of the ratmar cuff musculature in the late cocking phase aCling b rotate the humerus and prevent anterior su tion of the glenohumeral joint (Eames & 1978). Specifically, the supraspinatus acts in t cocking phase to draw the hUllleral head low glenoid, the infraspinatus and teres minor p humeral head posteriorly, and the subscap bOlh prevents excessive external rotation humerus ancl contracts eccentrically to stress on the anterior shoulder (Tibone et 'II., The importance of scapular (and thus glenoi bilization has also been recognized, and the se anterior has been shown to nre actively in t cocking phase; this provides a stable platfo humeral motion. Thus. coordinated. sequenti vation of the shoulder musculature is needed vent anterior subluxation of the glenohumera and the overuse tendinitis that can ensue. ~~,;;~ -/j ,0;"0" .Calculation of Reaction Forces Estimates of the reaGlon force on the glenohumeral joint are obtained with the use of simplifying assump:, lions (Poppen & Walker, 1978). ;;; Case A. In this example, the arm is in 90° of abduction, ;~(:and it is assumed Ii that only the deltoid muscle is active. The force produced through the tendon of the deltoid muscle (0) :,~,acts at a distance of 3 em from the center of rotation of the ,,'joint (indicated by the hollow circle). The force produced by ~, the weight of the arm is estimated to be 0.05 times body \.veight (BvV) and acts at a distance of 30 cm from the center of rotalion. The reaClion force on the glenohumeral joint 0) may be calculated with the use of the equilibrium equation o is approximately one-half body weight. Because 0 are of equal magnitude; thus, the joint reaction force is also approximately one-half body weigll!. Case B. Similar calculations can be made to determine the value for 0 when a weight equal to 0,025 times body weight is held in the hand with the arm in 90" of abduction. ~M = 0 (30 ern x .05 BW) + (60 em " .025 BW) (D x 3 em) ~ 0 D = (30 em " .05 BW) + (60 em " .025 BW) 30m that states that for a body to be in moment equilibrium, the 5;Jm of the moments must equal zero. In this example, the and J are almost paraUel but oppOSite. they form a force couple and D = 1 BW moments acting clockwise are considered to be positive and counterclockwise moments are considered to be negative. ':: ~M = Once again. 0 and J are essentially equal and opposite. forming a force couple. Thus. the joint reaction force is ap- 0 (30 em x .05BW) - (D " 3 em) = 0 proximately equal to body weight. D = 30 em x .05 BW 3cm D = 0.5 BW 3cm 3cm I Force 0 ......... ----i""X'Force J I I II I ! I, ,.ossw ;• , 30cm ): Calculation Box Figure 12-1-1. &.............. . . , ;. _ ·········~)~~=----=t;;;;_;;:;---r ! 30cm .~,----::-------> 60 em Calculation Box Figure 12-1-2. __ .-._ _. Sumrnarv F The shoulder consists of the glenohumeral, ncromioclavicular. stcrnoclaviculat: and scapulothoracic articulations and the muscular structures that act on them to produce the most mobile joint in the body. 2 The sternoclavicular joint, which connects the medial end of the clavicle to the manubrium. links the uppe.· ex trernity to the thorax. The aniculnr disc within the joint and the ligaments that surround provide stability while allowing for significant rota tion of the clavicle. lillie relative motion is seen be tween the clavicle and acromion at the acromio clavicular joint. 3 The glenohumeral joint. while known to be precise articulation, is inherently unstable becaus the glenoid fossa is shallow and is abl.e to contai only approximately onc third of the diameter of th humeral head. Stability is instead provided by th capsular, musculal~ Llnd ligamentous structures that surround it. 4 The three glenohumeral ligaments (superior, middle. and inferior) are discrete extensions of Ihe anterior glenohumeral joint capsule and are critical to shoulder slability and function. S" The inferior glenohumeral ligament has the most functional significance (particularly the anterior bane!), acting as the primary anterior stabilizer of the shoulder when the arm is 'lb· ducted 90'. ·6. The inlegrity of the shoulder capsule and the negative intra-articular force it maintains also plays an integral pan in mainlaining shoulder stability. .)~ Elevation of the arm involves motion at both the glenohumeral joint and the scaplliothoracic articulation. ?;S) Movements of the spine assist the shoulder in positioning the UppCt' extrcmity in space. {g) The muscles around the shouldcr contribute lO st~~'bility via a barricr effect by producing compressive forces on the glenohumeral joint and byeccentric contraction. ~Q';The glenohumeral joint is a major load-bearing joint with forces equivalent to one-half body weight produced when holding the arm in an outstretched position. REFERENCES :\gur, A.M.R .. lee. \1., &. Grant. J.e. (1991). :lIla ... o":lwllomy (9th NI.). Bahimorl..': Williams &. Wilkins. Baml..'s, D.A. &. Tullos, H.S. (1978). An analysis of 100 symptom'llie b<tsl..'ball playt.'fs...\11/ J Sports Mel!. 6. 62-67. Boonc. D.C. &. :\zcn. S.P. (1979). Nornwl range of Illotion of joints in male subjl.'cts. 1 Bow: loiHt S/frg. 61. 756-759. Colaehis, S.C. Jr. &. Strohm, B.R. (1971). Effl..'ct of suprascapul;lr and axillary nervc blocks on muscle farer: in upper I..'.xtrl..'milv. Arch Phvs :\Jed RciUlbil. .12, 21-29. Cooper, D.E., O'Briel~, 5.1., Amol:Zk:". S.P., 1..'1 ;;d. (1993). The structurc and function of the coraCOllllllll.'ral ligamcnt: An anatomic and mil.:roscopic study. J Shoulder Elboll· SlIrg. 2, 70-77. Curl. L.A. &. Warrcn, R.F. (J 996). Glenohumeral joinl stabil· ity: ScI",'ctin' culting studies on the slatic: c<\psubr re· siminis. Clill Onhop, 330, 54-65. Delee J. & Ora, D. (199.:1). Onhopac:l!ic Spons Mc:dicillt·. Prillciplt:s awl 1)raelicc (p. 464). Phibdelphin: W.B. Saunders. Dc P.dlll':\. A.F. (1983). BiollH.'chanics of the shoulder. In SI/rgery tlte Shollider (3rd cd .. pp.65-85). Philacldphia: J,B. Lippincolt. or Fukllda, K.. Cr;lig. E.\·.. All. K.. t·l ;d. 11{J~6i. Bihmt· :-llId,- e,l Illl· ligallll·lll('ll:' :-~:-tt·1ll (If thl' :lcnHnicH jllill!..1 HfJlil' .If/ill; Sltr::. MS. ·t3.t-~.tO. Gihh. T.IL Sidle,. J.:\ .. lbn·~lll;tll. D.T.. d :11. (1991) ft...:! (d- l"~lp:-lIlar \l·lIling oil gk·lIohllllkr:d 1:,'\iiY. !/f(ip. 268. 110-11i. l!ollillShl·"d. \\'.11. (IY6t)). ·\lIli//llllf /n,. Sllr.::OJl/\ (VoL York: Harp..·.. ...\: Ih,\\. lIe,\,-dl. 5.:\1.. IIlHdlcr:-kg. :\ ..\\.. S..· ga, D.lI .. l:I :11 Cbrifil:;ltion 01 till.· nd .., Io! Ih~· :-upr:l:-pill:IlU:- n :-houldl·r 11llKliclll. 1 H"'I/, 11'iHI Slfl~. 00:\. 3YS-·W [nm:\ll. V.T.. S;\Illllk·r:-. J.I3 .. & :\blwll. L.C. (19.t ..+.!. lifl!\S nn Ihe llln;.:lillll III till· :-houldn jlllnl . .1 !J SllJ·.~. 26..\. 1-30. Iloi. E., I-bll. I-I.e.. l\: All. h.\. (199{\1. BiOlllt'ch:lnica g;llioll of lhl: gkllohllllll:r;d join!.) SltlJ//lcla Uhow 40i- ..L~~. Iloi. E.. \lol/.kin. ~.E . .\IMr..·.\. B.F.. d :tl. {IY9-ii. SI illlKlioll Ilf tht· long hC:ld of Iht· hic..·p:- ill till· han po:-,itilln. J Sltould..r FlblOw Sur;.!.. 3, 135-1-I:!. hoi. E. .\l(llzkin. \:.E. . .\1{)lTt·'. 13.17.. l'l ~tl. (1992). inclill:ltiol\ ~ll\d illk·ri(lr :-1;lbilil, 1'\ lilt· ... huuldCl". tit'/" r:1l}('1~' Sun.;. I. 131-1.,9. "-l1m;lr. \'.1'. & lbb:-uhr;ltll;tlli:llll, P. (1')S51. Tilt· rC n\I,:,ph~·rjt· pr~· . . :-t1r~' ill :-1:lhili/ing lllv :-h(,tlllkr: i\ ll\~·llt:d ~tltd\. J nUll<' Jpiur Sill.!.:. (,iii. 71'L·711. L'llilll'11HI. t'. i 1'J~7). Kin~·~i()I(I~'.\' (Ii I h\.· sil(llillkr jo K(,ih~·1. ~·t a!. i. E.d~.I. ,I,,;/inlildl'l' N'Pi!;Cl'iJl,'IIi. Sprinl,'n· \"nl:q!. I 'J~ 7. Lippitt. S,B .. \·"JHkrho()ft. .J.F. GICI1(lhllI11~'!":\1 :;t~lhijit' q\l'lnlil'lli\~· :ll1;I!.,:;i:, . .1 J[;\I'1'i~, S.L. ~·t ~i ('(JJICI\'il.\·l:olllpr,, SJIO/lldrJ' IJ;'II\\' Sur::, }, 2 I"rllill [.tll·,t:'. I).B, (1973). Biolll(·Ch:ltli\.'s 01 tht· :;hntlltlt·r in Sw·:,;. 107. ~2~ . .\bbC:l. E. Fl!. F.. \.\: 1[;1\\ kill:-. R. 1Fd ..... \ \ 19921. rh,· :1 J3a!IiIiC" iI! _\//Ohilizy rll/{; Stahility. RO:-Clllill1l. I I\';II!. Colorado: \\'lIl"bhop Suppllrkd b., dl~· :\(adt'my (If Orth(Jp~lt·dic Sllr~"·{)I\:-. Ih..• X<llional of r\nhrili:- alld .\llI:-nllo~k~·k!:t1 Skill I)i~t·a:-;..,:-. t i t-:Ill ShouldtT ;1Ilt! ElbO\\ Suq:!t·OII:-. Ih ..· On!lopa :-t·~lr\.>h and E.hIGII il'" F('lllltbi iOll.] .\lilffcy. B.F. & :\n. I\..\:. i 1990), Biolllt·dl:lllk:o, (If the III c.:\. ROi..·k\\(I(,t! \'-. F.:\ . .\1~lt:-t·1l III {Elk). The lpp. 10~-2.t:;L rhibddphi~l: \\".B. S;illl1(lt-r:-. .\I(lor..·. h.L. (19t)\ji. Cliuh:{:lly O,-it'lth'd ..Ii/ili/IJllY 1 l)hibtklphi"l: LippilKllti \\'illl,tlll:- & \\"ilkin:-. .\hUTt·'·. B.F. l\: :\11. "-.X. (1990). BiCll11t'I.:h:lllk:-. of l tit-I'. In C:\. Ritt"kW(IOd l\: 17.:\. '\\;il:-t·1l 111 {Ed~.). rI d,·,: PhiliHk·lphia: W.B. S:llln.kr.... .\lurr:I\·. ;\-I.P.. GMt.. D.R .. Cardlll·r. C. .\I.. ('1 ,d. l19~5 tin' motion ;llld Illll:-r.:k :-lrt·ngth of normal \\Olllt'll in Iwo ;l~e ~rollp:-,. Clill Onhflp. /92, 193. :\cl..'r. C. (1990i. Shoulder UC·(·flll:'In/oiIlJl (p. ]9). phi;t: W.B. Saundns Co. O'Brit·n. Sol .. \ne:- . .\1.C. "l1Hli.:l.k~. 5.1'.. d al. (1 :lnalomy and hi:'lology 01 Ih ..· inkriof gknohullw lIh:nl cOlllpkx of I lit· shoulder. :\111 J Sports 579-3$4. Pagn:llii. ,\-l.J .. Dt'llg, X.Il .. Wan\'Il. R.F.. l"l ;11. (1995), k:-;iolls of the superior 1l<ll"Ii(lli of lilt' glt'noid b glenohullleral tr'lmlali(JIl:- . .I1J<"h,·.Il,1illl SIlI".:;. 77:\. Popp(,ll, N.K. & Walh'r. P.S. (1978). rorcl.'s al Ih(' giL'nohllnlt:r,1! join! in ,lbdllClioll. Clill OnJ/lJp. 135. 165-[ 70. poppell, N.K. & W~lkl.'r, IJ.S. (1976). NOl"m~1 .111<1 abnormal motion of thl.' shoulder. J BOlle Joi/lt SlI(!!.. 58..\, 195-20 I. Rochmod. c..-\. h. (1975). Disloc:.uions about lht: shoulder. In C.A. Rockwood Jr. ..\: D.I J _ Grl.·cn (Eds.), Fmc/m't's (1st ,'j l.'d .. Vol I, PI'. 624-815). Philadelphia: J.B. Lippincott. C. ..\: Malst:ll. F. (1990). The ')·!lo/ddt·" (p. 219J. Philaddphia: W.O. Saunders Co. S.R. (Ed). (1994). Orthopaedic Basic .'idCllce (p. 327). Rosemont, IL: AAOS. L.J., Flaww, E.L., Bigliani. L.U., C[ HI. (1992). Articular geonll:lr~' IhL' glenohumeral joillt. Clill Ol"lhop, or 28;, 181-190. Sh.:inbt:ek, J., Lilje:lIq\'isl. U., 6: Jcrosh, J. (199-0_ The contribution to <In[aior shoulder stnbility. J Sho/!lder bOIl'SlIrg, 7(2J.122-6. Tibont:. J .. P;Hd, R., Jobe:, F.W.. t:l a!. (199-0. The Shoul fUIKtional analOmy. hi(1/1h:<:hanics and kinesiology. In De:Lt:1: 6: D. Dra (Eds.), OrillOpa.:dic SPOrtS ,\Iedic IJhilnddphia: W.B. Sallnde:l"s Co. W:ll'I1C.... J.P, (1993). Thl; gross all~lIomy of the joint surfa lig<tll1<.'nts, labrum :lIld c.::lpsuk. In F.A. Matsen, F.t!. F R.J. H.l\\·kins (Eels.), Thl.' Should{'/": A !3alaJ/n: of .\fob and Flllictioll (PI'. 7-17). Rosemont, It: AAOS. Wnrncr, J.P., Dcng, X.I-! .. \Varren, R.E et al. (1992). SU l::lflsllloligamenlOlls rt.'siraints to superior-inferior tr:lf1 lions of the g!t:!lohulllcral joint. Alii } SPOI'1S ,\Jed, 675-678. an.atolll)' or Ihl: gle:nohlimcrallig~lIne:nlOlis compkx and its :/;' ~ L • O>-~ '" "'"!'.,-_-.!'.!','!'.A__"'zg!,!,,,.,,!,!,,"""'_.","_!,,,~".B'" "'_........"_...,;"t.~".,,__,,,",_"_,·c"-, ___·,"'.", "_--",,---~. ·"",~c_,,,:-,-,,,,~,~,,_,,,,;;o..,,,,,,~,?,,_,,~,,,_--- __- __,,,,,,"7.2'}"'...!,,",,J.7!!J~'!;?3?WC'--"; _ ..!_".xa"'..", ;. ""'.."'""_, Biomechanic of the Elbow Laith M. Jazrawl~ Andrew 5. Rokito, Maureen Gallagher Bird Joseph D. Zuckerm Introduction Anatomy Kinematics Carrying Angle Elbow Stability Kinetics Electromyography Elbow Joint Forces Artieular Surface Forces Calculation of Joint Reaction Forces at the Elbow Summary References . , • f ,,Introduction ~, IX ~/ t Wi, t, ! ·i The elbow is a complex joint that functions as a fulcrUIll for the forearm lever system that is responsi- ~•. ",.' ~ j I I . hie for positioning the hand in space. A detailed lin-Hi;:Xderstnnding of the biomechanics of elbow function J,;': \ I ,is essential for the clinician to e1Tcctivcly treat 'pathological conditions affecting the elbow joint. ~"'" ~:'.•,.,· .;;'.A /'l ato III V Wi I't: The elbow joint complex allows types of molion: nexion-extension and pronation-supination. ~ ,•.... '.' .,/ 1\\10 i, The humcroulnar and hUlllcroradial articulations ~ allow elbow Oexion and extension and arc classified i as ginglymoid or hinged joints. The proximal ra- ~,: dioulnar articulation allows forearm pronation and i supination and is classified as a u'ochoid joint. The elbow joint complex. when considered in its en~ tircty. is therefore a tl'ochleoginglymoid joint. The trochlea and capitellum of the distal hUl11cnls arc internally rotated 3 to 8' (Fig. 13-IC) and in 94 to 98' of valgus with respect to the longitudinal axis of the humerus (Fig, 13-IA). The distal humerus is anteriorly angulated 30° along the long axis of the humenls (Fig. 13-1 B). The a.-licular surface of the ulna is oriented in approximately 4 to 7° of valgus angulation with respect to the longitudinal axis of its shaft (Fig. 13-2;\). The distal humerus is divided into medial and lateml columns that terlllinate distally with the trochlea connecting the two columns (Fig. 13-3). The medial column diverges [Tom the humeral shaft at a 45° angle and ends approximatel,Y 1 cm proximal to the distal end of the trochlea. The dislal one third of the medial column is composed of cancellous bone, is ovoid in shape, and represents the medial epicondyle. The lateral column of the distal humerus diverges at a 20° angle from the humerus at the same level as the medial column and ends \'.lith the capitelIUlll. The trochlea is in the shape of a spool and is comprised of medial and lateral lips with an intervening sulcus. This sulcus articulates with the semilunar notch of the proximal ulna. The articular surface of the trochlea is covered by hyaline cartilage in an arc of 330°. The capitellum, comprising al0105t a perfect hemisphere. is covered by hyaline cartilage forming an arc of approximately 180°. Tbe mticular surface of the ulna is rotated 30' poste· Jiorly with respect to ilS long ax.is. This matches the 30° anteIior angulation of the distal humerus. which helps pl"Ovide stability to the elbow joint in full extension I I rf i, i ~i ~. K ~~ t' ~ ~ i .,. , ~. ~ I :: $ ii; ( ~.. < I. l ~ I ~ r: t ~ I t r 30' B A c Angular orientation of the distal humerus in the anteroposterior (A), lateral (B). and axial (C) projections. 4' ~. I A B Angular orientation of the proximal ulna in the anteroposterior (A) and later