Veterinary Surgery
31:201–211, 2008
Proximodistal Alignment of the Canine Patella:
Radiographic Evaluation and Association with Medial and
Lateral Patellar Luxation
AYMAN A. MOSTAFA, BVSc, MVSc, DOMINIQUE J. GRIFFON, DMV, PhD, Diplomate ACVS & ECVS, MICHAEL W.
THOMAS, DVM, Diplomate ACVR, and PETER D. CONSTABLE, BVSc, PhD, Diplomate ACVIM
Objectives—To evaluate the contribution of proximodistal alignment of the patella to patellar
luxation, and to evaluate the structures contributing to proximodistal alignment of the patella
relative to the femoral trochlea.
Study Design—Retrospective study using a convenience sample.
Animals—Medium to giant breed dogs (n ¼ 106).
Methods—Medical records and stifle radiographs of 106 dogs were reviewed. Radiographic measurements evaluated the proximodistal alignment of the patella with respect to the femoral trochlea,
distal aspect of the femur, and proximal aspect of the tibia. Measurements were compared between
dogs with clinically normal stifles (controls; n ¼ 51 dogs, 66 stifles), and dogs with a clinical diagnosis of medial patellar luxation (MPL, n ¼ 46 dogs, 65 stifles) or lateral patellar luxation (LPL,
n ¼ 9 dogs, 11 stifles) using ANOVA.
Results—In dogs with MPL, the ratio of patellar ligament length (PLL) to patellar length (PL) was
increased, as was the ratio of the distance from the proximal aspect of the patella to the femoral
condyle (A) to PL (Po.0001). Dogs with LPL had a decreased A:PL (P ¼ .003) and an increased
ratio of the proximal tibial length (PTL) to distal tibial width (DTW; P ¼ .009).
Conclusions—MPL is associated with a relatively long patellar ligament and patella alta in medium
to giant breed dogs. LPL is associated with a relatively long proximal tibia and patella baja. Values
for PLL:PL42.06 and A:PL42.03 are suggestive of the presence of patella alta, whereas a value for
A:PLo1.92 is suggestive of patella baja.
Clinical Relevance—Measurements of both PLL:PL and A:PL are recommended in dogs with
patellar luxation, and surgical correction should be considered in those with abnormal values.
r Copyright 2008 by The American College of Veterinary Surgeons
18–29% of dogs, and more likely in dogs weighing
420 kg.3,4 Reluxation accounts for 30–48% of these
postoperative complications3,5 and 65–86% of the major
complications after surgical repair of patellar luxation in
dogs.3,4 The relatively high frequency of patellar reluxation may be related to inadequate evaluation of underlying skeletal deformities leading to incomplete surgical
correction.4 Although proximodistal transposition of the
tibial crest has been proposed in dogs, its indications and
INTRODUCTION
M
EDIAL PATELLAR luxation (MPL) is a common orthopedic disease in dogs with 95% of
affected dogs having related structural abnormalities.1,2
Lateral patellar luxation (LPL) occurs less frequently
than MPL.2 Recurrence of patellar luxation is one of the
most common complications associated with surgical
management of MPL or LPL. Complications develop in
From the Department of Veterinary Clinical Medicine, College of Veterinary Medicine, University of Illinois at Urbana-Champaign, IL.
Address reprint requests to Ayman A. Mostafa, BVSc, MVSc, Department of Veterinary Clinical Medicine, University of Illinois at
Urbana-Champaign, 1008 W Hazelwood Dr, Urbana IL 61802. E-mail: mostafa@uiuc.edu.
Dr. Constable’s current address is Department of Veterinary Clinical Sciences, Purdue University, West Lafayette IN.
Submitted August 2007; Accepted December 2007
r Copyright 2008 by The American College of Veterinary Surgeons
0161-3499/08
doi:10.1111/j.1532-950X.2008.00367.x
201
202
PROXIMODISTAL ALIGNMENT OF THE CANINE PATELLA
effects remain undetermined.6 Surgical treatment of patellar luxation traditionally includes desmotomy, contralateral imbrication of soft tissue, trochleoplasty, and
transposition of the tibial crest. When combined, these
techniques restore the quadriceps alignment and adequately deepen the trochlear groove of the femur7; however, these procedures do not address proximodistal
malalignment of the patella relative to the femoral trochlea, which may contribute to postoperative recurrence of
patellar luxation.
Proximodistal malalignment of the patella includes
patella alta and baja, which refer to a patella located too
proximally (high riding patella) or distally (low riding
patella) within the femoral trochlea, respectively. Among
the radiographic techniques described to evaluate the
proximodistal alignment of the patella in humans, Insall’s
index provides the most applicable and convenient method of diagnosing patella alta in humans. Insall’s index is
obtained by dividing the greatest diagonal length of the
patella by the length of the patellar ligament to calculate
a ratio, which is normally equal to 1.0. An Insall’s index
value o0.80 is consistent with patella alta, whereas a
ratio 41.20 indicates the presence of patella baja.8–10
Insall’s index has been modified in humans to overcome the artifact created by long distal patellar facets
that artificially increase the ratio and may mask the presence of patella alta. Insall’s modified index is obtained
by dividing the distance between the inferior aspect of
the articular surface of the patella and the insertion of the
patellar ligament by the length of the patellar articular
surface. In spite of this modification, the clinical utility of
Insall’s index and Insall’s modified index remains limited
by the difficulty in precisely determining the point of insertion of the patellar ligament on the tibial tuberosity.11
An alternative technique has been developed in humans
using anteroposterior radiographs to address this issue.9
Using this technique, patella alta should be suspected if
the distance from the inferior pole of the patella to a line
drawn across the distal aspect of the femoral condyles is
420 mm.9
Evaluations of the proximodistal alignment of the
canine patella have been limited to the application of
Insall’s index to large breed dogs.6,12 The ratio of patellar
ligament length (PLL) to patellar length (PL) was
increased in 30 large breed dogs with MPL, with
ratios 41.97 being consistent with patella alta.12 Based
on the association between increased PLL:PL and MPL,
patella alta was suspected to play a role in the development of MPL in large breed dogs.6
Most studies in humans and animals have focused on
the diagnosis of patella alta and baja using techniques
relying on the length of the patellar ligament, but have
ignored the potential effects of distal femoral or proximal
tibial abnormalities on the position of the patella in
relationship with the femoral trochlea. Although several
authors have described anomalies of the distal portion of
the femur13 and proximal portion of the tibia in dogs,14–17
the relationship of these anomalies to the extensor mechanism has not been characterized. Incomplete knowledge
of the anatomic structures contributing to the proximodistal alignment of the patella may limit the success of
surgical management of patellar luxation.
The impetus for this study was our observation that
measurements of PLL and PL in clinical cases were influenced by the presence of enthesiophytes on the distal
patella and proximal tibia. In addition, reliance on
PLL:PL ignores the potential effect that distal femoral
or proximal tibial abnormalities could exert on the position of the patella in relationship to the femoral trochlea. We hypothesized that a proximal or distal position of
the patella contributes to medial or LPL in medium,
large, and giant breed dogs. We also hypothesized that
structures other than the patellar ligament contribute to a
proximodistal malalignment of the patella. Our 1st objective was to evaluate the reproducibility of radiographic
measurements characterizing the proximodistal alignment of the patella in medium to giant-breed dogs. The
2nd objective was to evaluate the potential contribution
of patellar position (alta or baja) to the side of patellar
luxation (medial or lateral) in medium to giant-breed
dogs and our 3rd objective was to evaluate structures
contributing to the proximodistal alignment of the patella
relative to the femoral trochlea.
MATERIALS AND METHODS
Dogs
Medical records and stifle radiographs of medium (minimum body weight, 11 kg) to giant-breed dogs, 5 months of
age admitted between March 1997 and February 2006, were
retrieved. The control group consisted of dogs with no history
of stifle disease and normal stifle radiographs. Dogs with displaced diaphyseal femoral fractures were excluded from study.
The case definition for a diagnosis of MPL or LPL did not
require a radiograph and was based on a history of intermittent pelvic limb lameness and the presence of a grade I–IV
patellar luxation on orthopedic examination of the affected
pelvic limb.18,19 Additional inclusion criteria for dogs with
patellar luxation included no history of trauma, no detectable
evidence or knowledge of previous stifle surgery, and no
evidence of pathologic changes in the stifle joint other than
patellar luxation with or without cranial cruciate ligament
(CCL) rupture and degenerative joint disease (DJD).
All radiographs included in the study were approved in
terms of quality and positioning by a board certified radiologist. Positioning was judged as satisfactory if both femoral
condyles were superimposed on lateral projections and if both
fabellae symmetrically bisected the femoral cortices in cranial
views and the patella was centered over the trochlear groove
MOSTAFA ET AL
203
Table 1. Definitions of the Abbreviations of the Radiographic
Measurements
PL
PLL
A
FCL
FW
PTL
DTW
PTW
FTL
J
Patellar length
Patellar ligament length
Vertical distance between the proximal pole of the
patella and transcondylar axis of the distal femur
Femoral condylar length
Femoral width
Proximal tibial length
Distal tibial width
Proximal tibial width
Femoral trochlear length
Distance between proximal extent of femoral
trochlear ridges and tibial tuberosity
(in the control group). Dogs were included in this study if the
lateral projection of the stifle extended to the diaphysis of the
femur and tibia, allowing all the measurements required.
Radiographic Measurements (Table 1)
All radiographs were evaluated to study the proximodistal
alignment of the patella, the conformation of the distal aspect
of the femur and proximal aspect of the tibia, the presence and
severity of joint effusion, and the presence and degree of
DJD.18 Radiographic signs of DJD were graded as 0 (no enthesiophytes), 1 (1 mm enthesiophytes), 2 (2–3 mm enthesiophytes), and 3 (43 mm enthesiophytes), based on a
previous scoring system.20–22 Radiographic measurements
were obtained from 3 populations of medium to giant-breed
dogs: normal dogs, dogs with MPL, and dogs with LPL.
The reproducibility of radiographic measurements was
evaluated by calculating intra-observer repeatability using the
coefficient of variation (CV) for each variable and by determining the effect of stifle flexion on measurements. CV was
based on evaluations performed by the same investigator
(AM) on 3 separate days of 20 randomly (using a simple random sample) selected dogs with patellar luxation.
Fig 1. Mediolateral radiograph of a clinically normal canine
stifle illustrating the measurements made to determine patellar
length (PL) and patellar ligament length (PLL).
Evaluation of the Position of the Patella in Relation to the
Transcondylar Axis of the Femur (A:PL)
The length of a vertical line (A) connecting the proximal
pole of the patella and the transcondylar axis of the distal
femur, measured from a caudocranial or craniocaudal radiograph (Fig. 2), was indexed to the PL to provide A:PL. The
value for PL was measured on the corresponding mediolateral
view of the stifle because superimposition over the distal aspect of the patella prevented adequate observation of the distal pole of the patella on caudocranial and craniocaudal
projections.
Evaluation of the Distal Aspect of the Femur
Evaluation of the Patellar Ligament in Relationship to the
Patella (PLL:PL)
The PLL was indexed to the greatest PL on a mediolateral
radiograph (Fig 1) and the modified PLL:PL calculated. The
PLL represented the distance from the point of origin of the
ligament on the distal aspect of the patella to its insertion on
the proximal extent of the tibial crest (tibial tuberosity).14,23
Measurement of PL in dogs with DJD was problematic
because of the presence of enthesiophytes at the distal aspect
the patella. Enthesiophytes were excluded from the measurement of PL in these dogs by extrapolating the normal contour
of the patella. When contralateral radiographs were available,
this extrapolation was based on superimposition of the normal
patella.
The femoral condylar length (FCL) was the length of a
vertical line extending between the distal aspect of the femoral
condyles and the proximal extent of the femoral trochlear
ridges (X; Fig 3). This value was normalized by the femoral
width (FCL:FW) to mitigate the effects of size and magnification. The femoral trochlea length (FTL, Fig. 4) was the
distance between a line across the proximal extent of the femoral trochlear ridges (X) and the cranial distal femoral physis
or the extensor fossa (Z). The (FW, Fig 3) was defined as the
width of the femur measured at a distance equal to FCL from
the proximal aspect of the trochlea (X). The greatest diagonal
PL was measured as described previously. The lengths of the
femoral condyle (FCL) and trochlea (FTL) were expressed
relative to the length of the patella (FCL:PL, FTL:PL) and
compared between groups.
204
PROXIMODISTAL ALIGNMENT OF THE CANINE PATELLA
Fig 2. Caudocranial view of a clinically normal canine stifle
illustrating the measurement of the vertical line (A) connecting
the proximal pole of the patella to the transcondylar axis.
Fig 4. Evaluation of the relationship between the patellar
mechanism, distal femur and proximal tibia on a mediolateral
radiograph of a clinically normal canine stifle. O, insertion of the
patellar ligament on the tibial tuberosity; X, proximal extent of
the femoral trochlea; Z, origin of the common digital extensor;
FTL, femoral trochlear length; J, distance between O and X.
Evaluation of the Proximal Aspect of the Tibia
The proximal tibial length (PTL) was measured as the distance between the cranial point of the medial articular surface
of the tibial plateau (I) and the tibial tuberosity (O; Fig 5).
The proximal tibial width (PTW) was measured as the distance between the point of insertion of the patellar ligament
on the tibial tuberosity (O) and the caudal point of the
medial articular surface of the tibial plateau (point of
insertion of caudal cruciate ligament; S). The distal tibial
width (DTW) was measured at a distance equal to twice
the proximal tibial width (PTW) from the cranial aspect
of the medial tibial plateau (I).24 The ratios of PTL to
DTW (PTL:DTW) and PTW to DTW (PTW:DTW) were
calculated.
Relationship Between the Patellar Mechanism, Distal
Aspect of the Femur, and Proximal Aspect of the Tibia
Fig 3. Evaluation of the distal femur on a mediolateral radiograph of a clinically normal canine stifle. X, proximal extent of the femoral trochlea; FCL, femoral condylar length;
PL, patellar length; FW, femoral width, which is measured at a
distance equal to FCL from X.
This relationship was evaluated with measurements that
spanned the distal femur and proximal tibia. The measurement ‘‘J’’ was defined as the distance between a line extending
from the proximal extent of the femoral trochlear ridges (point
‘‘X’’) to the proximal aspect of the tibial crest ‘‘O’’ (Fig 4). The
following femorotibial ratios were evaluated: the proximal
tibial width to the FCL (PTW:FCL), the sum of PL and PLL
to the sum of FTL and PTL ({PLL þ PL}:{PTL þ FTL}), and
the sum of PL and PLL to the distance between the proximal
trochlea and the insertion of the patellar ligament
({PLL þ PL}/J).
205
MOSTAFA ET AL
Fig 5. Evaluation of the proximal tibia on a mediolateral
radiograph of a clinically normal canine stifle. I, cranial extent
of the medial tibial plateau; S, caudal extent of the medial
tibial plateau; O, insertion of the patellar ligament on the tibial
tuberosity; PTL, proximal tibial length; PTW, proximal tibial
width; DTW, distal tibial width.
Angle of Flexion of the Stifle
The angle of the stifle joint was defined as the angle formed
between the long axis of the distal femur and proximal aspect
of the tibia (Fig 6). This angle was determined by drawing a
segment (FW) between the femoral cortices at a distance equal
to the length of the femoral condyle from the proximal extent
of the trochlea ‘‘X’’. A segment ‘‘D’’ parallel to FW was
drawn 20 mm proximal to FW. A line was drawn from joining
the centers of the 2 segments (FW and D) to define the long
axis of the distal femur. The long axis of the proximal tibia
was drawn by connecting the cranial point of the medial
articular surface of the tibial plateau to the midpoint of the
segment used to measure DTW.24 The stifle joint angle was
measured at the intersection of these 2 axes.
Statistical Analysis
Data are reported as mean SD. Significance was set
Po.05. Intra-observer variability for each radiographic measurement was evaluated by calculating the CV for each variable obtained by 3 measurements performed on radiographs
of 20 randomly selected dogs with patellar luxation. A 95%
confidence interval (CI) for normal values was calculated for
selected measurements from control dogs. Group means were
compared with control using analysis of variance (PROC
MIXED, SAS 9.2, Cary, NC), with stifle nested within dog
because analysis was based on 1 or 2 stifles for each dog.
Fig 6. Measurement of stifle joint angle on a mediolateral
radiograph of a clinically normal canine stifle. I, cranial extent
of the medial tibial plateau; X, proximal extent of the femoral
trochlea; DTW, distal tibial width; FCL, femoral condylar
length; FW, femoral width; D, a segment parallel to FW and
drawn 20 mm proximal to FW.
Nonnormally distributed variables were log transformed before ANOVA was performed. The effect of stifle angle on
femorotibial measurements was examined by regressing each
femorotibial measurement (as well as the square of the femorotibial measurement to explore for the presence of a curvilinear relationship) against stifle angle using multivariable
regression (PROC REG, SAS 9.2) with MPL and LPL as
covariates. The relationship between DJD score and PLL:PL,
A:PL, FCL:FW, and FCL:PL for all radiographs was explored by calculating the Spearman’s rank correlation coefficient (Spearman’s rho; PROC CORR, SAS 9.2).
RESULTS
Body weight and age did not differ significantly between the 3 groups of dogs studied (control, MPL, LPL).
Mean ( SD) body weight was 28.1 9.8 kg and age
3.1 2.8 years (range, 5.5 months to 13.5 years).
206
PROXIMODISTAL ALIGNMENT OF THE CANINE PATELLA
Dogs with Clinically Normal Stifles (Controls)
Fifty-one dogs (66 stifles) met the criteria for inclusion
in the control group. Thirty-six (71%) dogs were admitted for fractures, affecting the pelvis, femoral head, neck,
greater trochanter, proximal femoral diaphysis (nondisplaced), or distal tibia with no evidence of stifle disease
(stifle measurements were performed on the fractured
limb in 18 dogs, the normal limb in 11 dogs, and both
fractured and normal limbs in 7 dogs). Six dogs (12%)
were clinically and radiographically normal with no evidence of orthopedic disease, and 12% also had a CCL
tear on the opposite stifle with no evidence of stifle disease on the measured side. Two dogs (4%) had radiographic evidence of bilateral coxoarthrosis but normal
stifle joints, and 1 dog had bilateral hip dysplasia (total
hip replacement had been performed on both hips).
Breeds in the control group were: 26 (51%) Labrador
Retrievers; 4 (8%) Golden Retrievers; 2 (4%) each of
Doberman, Australian Cattle, German Shepherd, Border
Collie, and Australian Shepherd; and 1 (2%) each of
Rottweiler, Great Dane, Boxer, Bull Mastiff, Saint Bernard, Great Pyrenees, and Irish Wolfhound; and 4 dogs
were mixed breeds.
Dogs with Patellar Luxation (Diseased Group)
Patellar luxation was diagnosed in 55 dogs (76 stifles);
84% (46 dogs) had MPL and 16% (9 dogs) had LPL.
Twelve of 65 (19%) stifles with MPL had grade I luxation, 31 (48%) had grade II, 13 (20%) had grade III and 9
(14%) had grade IV MPL. Among 11 stifles with LPL, 2
(18%) had grade I, 3 (27%) grade II, 2 (18%) grade III
and 4 (36%) had grade IV LPL. Among the 46 dogs (65
stifles) with MPL, 27 (59%) were affected unilaterally and
19 (41%) had bilateral luxations. Seven of 9 (78%) LPL
dogs had unilateral luxation whereas both limbs were
affected in 2 (22%) dogs.
Radiographic evidence of DJD was identified in 25 of
76 (33%) stifles with patellar luxation. Degenerative
changes were scored as: 1 in 8 stifles (10.5%), 2 in 12
(15.7%), and 3 in 5 (6.6%). Orthopedic conditions associated with patellar luxation included hip dysplasia
(n ¼ 16 dogs, 29%), CCL disease (n ¼ 10, 18%), hip dysplasia and CCL tear (n ¼ 4, 7%), healed femoral neck
fracture (n ¼ 1, 2%), slipped capital femoral physis
(n ¼ 1, 2%), 4th lumbar vertebrae mass (n ¼ 1, 2%),
and panosteitis (n ¼ 1, 2%). Thirteen of 46 (28%) dogs
with MPL had CCL disease whereas only 1 of 9 (11%)
dogs with LPL had CCL disease. Patellar luxation was
the only orthopedic disease in 21 (38%) dogs.
Dogs with patellar luxation included 16 mixed breeds
(29%), 10 (18%) Labrador Retrievers, 3 (5%) each of
Cocker spaniel and Akita, 2 (4%) each of German Shep-
herd, Bulldog, Husky, Cavalier King Charles Spaniel,
and Newfoundland, and 1 (2%) each of Great Dane,
Boston Bull Terrier, Golden Retriever, Flat Coated Retriever, Bull Mastiff, Shar Pei, Soft Coated Wheaten
Terrier, Keeshond, Basset Hound, Australian Cattle dog,
Bouvier des Flanders, Border Collie, and Weimaraner.
Reproducibility of Radiographic Measurements
The CV of radiographic measurements varied between
0.96% and 2.11% (Table 2). Stifle angle did not differ
between groups (Table 2), varying from 43–1341. The
median (range) of stifle angle was 861 (43–1301) in the
control group, 931 (60–1341) in the MPL group and 941
(57–1221) in the LPL group. PLL varied with stifle angle
in a curvilinear manner: PLL ¼ 3.00 þ 0.056(angle)0.00040(angle)2 (R2 ¼ 0.18, Po.0001) with a maximum value for PLL occurring at a stifle angle of 701.
The value for PLL:PL also varied with stifle angle in a
curvilinear manner: {PLL:PL} ¼ 1.07 þ 0.22(MPL) þ
0.023(angle)0.00014(angle)2 (R2 ¼ 0.32, Po.0001; Fig.
7) with a maximum value for PLL:PL occurring at 901.
The PLL to PL ratio was constant over a stifle angle
range of 70–1101. The indice {PLL þ PL}:{PTL þ FTL}
was weakly associated with stifle angle in a curvilinear
manner (R2 ¼ 0.072, P ¼ .008). Indices that were curvilinearly affected by stifle angle included PLL (Po.0001)
and measurements crossing the joint: J (r ¼ 0.60,
Po.0001), {PLL þ PL}:J (r ¼ þ 0.56, Po.0001). Radiographic evaluation indicated that DJD score was positively correlated with PLL:PL (rs ¼ þ 0.22, P ¼ .011) but
was not correlated with A:PL (P ¼ .11), FCL:FW
(P ¼ .46), or FCL:PL (P ¼ .19).
Proximodistal Alignment of the Patella
The 95% CI of the modified PLL:PL was 1.97–2.06
for controls, 2.18–2.29 for dogs with MPL, and 1.77–2.02
for dogs with LPL. The value for PLL:PL was increased
in dogs with MPL (Po.0001) but did not differ in dogs
with LPL compared with controls (Table 2).
The 95% CI of A:PL was 1.92–2.03 for control, 2.18–
2.37 in dogs with MPL, and 1.49–1.95 for dogs in the
LPL group. The value for A:PL was greater in the MPL
group (Fig. 8, Po.0001) and lower in dogs with LPL
(Fig. 9, P ¼ .003) compared with the control group. A:PL
was evaluated from craniocaudal views in 44 of 66 (67%)
normal stifles, in 8 of 11 (73%) stifles with LPL and 49 of
65 (75%) stifles with MPL.
The height of the femoral condyle (FCL:PL) was increased relative to the length of the patella in both diseased groups, compared with controls (Table 2). The
length of the femoral trochlea normalized by the size of
the corresponding patella (FTL:PL) was only increased
207
MOSTAFA ET AL
Table 2. Mean Value for the Coefficient of Variation (CV) and Mean ( SD) Values for Radiographic Measurements Obtained on the Distal Femur,
Proximal Tibia, and Stifle Joint, for Dogs with Clinically Normal Stifle Joints (Control, n ¼ 51 Dogs, 66 Stifles), Dogs with Medial Patellar Luxation
(MPL, n ¼ 46, 65 Stifles), and Dogs with Lateral Patellar Luxation (LPL, n ¼ 9, 11 Stifles)
Measurement
PL (patella length) (mm)
PLL (patellar ligament length) (mm)
PLL:PL (patellar ligament length to patellar length)
A (vertical distance between the proximal pole of the patella and
transcondylar axis of the distal femur) (mm)
A:PL
Distal femoral measurements
FCL:FW (femoral condylar length to femoral width)
FCL:PL
Proximal tibial measurements
PTL:DTW (proximal tibial length to distal tibial width)
PTW:DTW (proximal tibial width to DTW)
Femorotibial measurements
PTW:FCL
{PLL þ PL}:{PTL þ FTL}
{PLL þ PL}:J ({PLL þ PL}to distance between proximal
extent of femoral trochlear ridges and tibial tuberosity)
Stifle angle
CV (%)
Control
MPL
LPL
1.43
0.92
2.11
0.39
23.7 47.4 2.02 46.8 3.4
6.7
0.20
8.9
19.7 4.4
44.8 11.1
2.23 0.23
45.9 15.5
21.1 40.4 1.90 36.2 1.7
2.4
0.14
7.6
1.77
1.98 0.23
2.28 0.39
1.72 0.34
1.48
2.08
2.52 0.23
1.96 0.13
2.59 0.24
2.08 0.15
2.82 0.34
2.07 0.14
2.64
1.60
1.61 0.16
2.64 0.27
1.58 0.18
2.63 0.31
1.88 0.32
3.06 0.43
1.43
1.43
1.04
1.01 0.04
1.12 0.06
0.91 0.08
1.04 0.07
1.14 0.10
0.93 0.07
1.00 0.05
1.01 0.10
0.83 0.07
0.96
88.3 20.0
94.6 15.7
92.6 20.6
Po.05 Compared with Control.
in the MPL group. Also, in dogs with MPL the proximal
tibia was approximately wider (95% CI of PTW:FCL,
1.028–1.061, P ¼ .034) than in dogs without patellar
luxation (95% CI of PTW:FCL, 1.005–1.026). In dogs
with LPL, the proximal tibia was proportionally longer
(95% CI of PTL:DTW, 1.65–2.11, P ¼ .009) and wider
(95% CI of PTW:DTW, 2.76–3.37, P ¼ .008) than in
dogs without patellar luxation (95% CI of PTL:DTW,
1.57–1.65; 95% CI of PTW:DTW, 2.58–2.71). Compared
with the control group, the overall length of the patellar
mechanism (PLL þ PL) of dogs with LPL was propor-
Fig 7. Curvilinear relationship between the value for
PLL:PL and stifle angle for dogs with clinically normal stifle
joints (n ¼ 51 dogs, 66 stifles), dogs with medial patellar luxation (n ¼ 46, 65 stifles), and dogs with lateral patellar luxation
(n ¼ 9, 11 stifles). The filled circles are the measured PLL:PL,
the open circles are the fitted polynomial regression line. PL,
patellar length; PLL, patellar ligament length.
tionally shorter than the distal femoral condyle and
proximal tibia, based on 2 ratios: ({PLL þ PL}:
{PTL þ FTL}, P ¼ .0004; {PLL þ PL}: J, P ¼ .023).
DISCUSSION
Although patellar luxation remains more common in
small breeds, this study focused on larger dogs because of
the increasing incidence of MPL and greater risk of postoperative complications in large dogs.4,5,25 In addition,
proximodistal malalignment of the patella (i.e. patella
alta) has previously been proposed as a predisposing factor to postoperative recurrence of patellar luxation.6 The
main findings of our study reported were: (1) the intraobserver variability of radiographic measurements was
o2.5%; (2) measurements influenced by stifle angle included those that crossed the joint as well as PLL; (3) the
normal range (95% CI) of the modified PLL:PL and
A:PL were 1.97–2.06, and 1.92–2.03, respectively. Therefore, values above or below these CI ranges might be
considered to represent a pathologic condition (i.e. patella alta or baja, respectively); (4) MPL was associated
with patella alta, whereas LPL was associated with patella baja; (5) dogs with MPL had a relatively long patellar ligament compared with controls; and (6) dogs with
LPL had a relatively longer proximal tibia than controls.
The conventional Insall’s index26 relies on observation
of the patella and patellar ligament on lateral radiographs. The point of insertion of the patellar ligament has
previously been associated with a small indentation on
the tibial tubercle which can be used as a landmark to
208
PROXIMODISTAL ALIGNMENT OF THE CANINE PATELLA
Fig 9. Mediolateral and caudocranial radiographs of the left
stifle joint of a dog with a grade IV lateral patellar luxation
and patella baja. Note the distal and lateral displacement of the
patella (arrows). PLL:PL ¼ 1.82 and A:PL ¼ 1.23. PL, patellar length; PLL, patellar ligament length.
Fig 8. Ventrodorsal view of the pelvis and mediolateral radiographs of the stifles of a dog clinically diagnosed with bilateral medial patellar luxation (grade IV on the left and grade
II on the right) and patella alta. Note the marked proximal and
medial displacement of the left patella (arrows). PLL:PL ¼ 3.1
and A:PL ¼ 3.1 for the left stifle; PLL:PL ¼ 2.3 and
A:PL ¼ 2.6 for the right stifle. PL, patellar length; PLL, patellar ligament length.
measure the PLL in dogs.6 However, identification of this
landmark as well as the patellar ligament itself is difficult
in dogs with joint effusion or DJD. All dogs with patellar
luxation in our study (subjectively) appeared to have a
similar degree of joint effusion based on radiographs and
the only dogs without joint effusion were those without
stifle disease (control group). These 2 populations differ
by more than 1 variable, preventing an evaluation of the
effect of joint effusion on radiographic measurements. To
palliate this potential issue, all measurements in our study
were specifically designed to rely on bony landmarks
rather than observation of soft tissue structures. For example, the modified PLL:PL and a ratio (A:PL) derived
from the human literature were selected to evaluate the
proximodistal alignment of the patella.
Values for PLL:PL were previously reported to be
independent of stifle angle.12 However, our findings
indicated that the value of modified PLL:PL varied in a
curvilinear manner with stifle angle (Fig 7), with the
maximal value occurring at an angle of 901. The discrepancy between our findings and the previous report
may reflect the larger number of observations in our
study (106 versus 13 large breed dogs), which increased
statistical power. The large range of flexion in our study
(43–1341 versus 75–1481 in the previous study) may also
have contributed to the difference between the 2 studies,
in that the value for modified PLL:PL remained constant
over a stifle angle range of 70–1101. Although the influence of stifle flexion on the value for modified PLL:PL
may affect the assessment of individual patients, the angle
of flexion did not differ between our 3 groups, thereby
permitting comparisons to be made between groups.
The vertical distance between the proximal extent of
the patella and transcondylar axis of the distal femur divided by the length of the patella (A:PL) did not appear
to be affected by stifle angle in our study; however, our
ability to evaluate the influence of stifle position on the
value for A:PL was limited by our inability to determine
the degree of flexion of the stifle on caudocranial or
craniocaudal radiographs. Another limitation is the potential effect of magnification between caudocranial and
craniocaudal radiographic views of the stifle joint. Nevertheless, the index A:PL appears clinically valuable when
observation of the patella and the patellar ligament is
MOSTAFA ET AL
impaired on mediolateral radiographs because of superimposition of the luxated patella over the distal femur. In
these cases, the PL can be easily determined on the
caudocranial or craniocaudal view.
Values for the 2 indices used to evaluate proximodistal
alignment of the patella (modified PLL:PL, A:PL) were
greater in dogs with MPL than in controls. Based on the
95% CI we calculated, PLL:PL values 42.06 are consistent with patella alta. This value is comparable to the
ratio (1.97) previously proposed.6 Our findings therefore
confirm the previously reported association between
MPL, patella alta, and an increased length of the patellar ligament.6 The ligament may be inherently long in
these dogs or may have stretched secondary to patellar
luxation. The positive correlation between radiographic
scores of DJD and PLL:PL supports the second mechanism. Proximal displacement of the patella in dogs with
patella alta may create a patellofemoral articulation that
extends proximal to the femoral trochlear groove during
extension of the stifle joint.6 The potential loss of the
buttressing effect of the proximal trochlear ridges may be
exacerbated in dogs with MPL by the presence of a relatively wide proximal tibia that may reduce the retropatellar pressure (femoropatellar contact force). This loss of
patellar pressure with the femoral condyle would facilitate medial luxation of the patella, especially in dogs with
shallow trochlea, medial tibial torsion, coxa vara, or lateral torsion of the distal aspect of the femur.27,28 The only
other clinically relevant difference between control and
MPL groups was the presence of longer femoral condyles
in dogs with MPL. This finding may be because of
secondary degenerative changes affecting the proximal
aspect of the femoral trochlear ridges. Our inability to
establish a correlation between DJD and the length of
femoral condyles may be explained by the fact that secondary changes were scored based solely based on the
presence and size of patellar enthesiophytes.21
Although the scope of our study is limited by the small
number of dogs with LPL, the PLL did not appear to be
abnormal in this group. Although values for modified
PLL:PL did not differ between controls and dogs with
LPL, values for A:PL were decreased in dogs with LPL.
This change is unlikely to reflect a difference in position
and radiographic positioning. Indeed, the percentage of
dogs radiographed in a craniocaudal position was greater
in the LPL (73%) than in the control (67%) group. Under these conditions, the distance between the patella and
the femoral condylar axis (A) would have been more
magnified, thereby increasing A:PL in dogs with LPL.
The proximal tibia in dogs with LPL was also relatively
longer than that in controls, potentially contributing to
the distal position of the patella in spite of a normal
patellar ligament. Based on these findings, a thorough
evaluation of the proximal aspect of the tibia is recom-
209
mended in dogs with LPL and based on our 95% CI,
dogs with PTL:DTW values 41.65 should be considered
at risk for developing patella baja. This finding explains
the differences in measurements evaluating the relationship between the patellar mechanism ({PLL þ PL}:
{PTL þ FTL}; {PLL þ PL}: J) between control dogs
and dogs with LPL. The medial ridge of the trochlear
groove is normally thicker than the lateral ridge in its
midportion, predisposing a low riding patella to luxate
laterally rather than medially,29 especially when combined with lateral tibial torsion or coxa valga.19
Surgical treatment of patella alta with recurrent patellar luxation has been reported to restore normal patellar position in humans.30 The tibial tubercle is
transposed both distally and medially to correct patella
alta, and elevated anteriorly to reduce patellofemoral
joint reaction forces, thereby allowing normalization of
the patellar index.31–36 Our results suggest that positioning the tibial tuberosity distally during routine tibial tuberosity transposition in dogs may be indicated in order
to move the tibial tuberosity distally and correct patella
alta, when indicated. Cranial transposition is unlikely to
help because the proximal tibia appeared relatively wider
than the femoral condyle in our dogs with MPL. The
caudalization previously reported after tibial transposition in dogs may, in fact, be beneficial.7,37 The optimal
method for surgical correction of patella baja remains
unclear. In humans, patella baja is corrected by proximal
transfer of the tibial tuberosity38,39 or Z-plasty lengthening of the patellar ligament.38,40 Proximal transposition
of the tibial crest in dogs is limited by the bone stock
available in that region and would palliate, but not address, any underlying malformation of the proximal tibia.
Thorough evaluation of the morphology of the tibia in
a larger population of dogs with LPL appears warranted
to further define the treatment options for patella baja in
affected dogs. We believe that the measurements we describe may better estimate the indication and extent of a
proximodistal transplantation of the tibial tuberosity during surgical correction of patellar luxation in dogs. This
may, in the future, reduce the frequency of recurrence after
surgical treatment of this condition. Although our results
support an association between proximodistal malalignment of the patella (patella alta versus baja) and the side of
patellar luxation (medial versus lateral), the retrospective
nature of our study precludes any conclusion about the
primary versus secondary nature of these changes. Early
evaluation of immature dogs with patellar luxation may
help further elucidate the contribution of proximodistal
malalignment of the patella as a cause of patellar luxation.
Preoperative measurement of modified PLL:PL and
A:PL is recommended in dogs with patellar luxation and
surgical correction should be considered in those with
abnormal values. Medium to giant-breed dogs with
210
PROXIMODISTAL ALIGNMENT OF THE CANINE PATELLA
PLL:PL values 42.06 or A:PL values 42.03 are considered to have patella alta, whereas dogs with A:PL values
o1.92 are considered to have patella baja. MPL is associated with patella alta, whereas LPL is associated with
patella baja. Patella alta is associated with an increased
PLL, whereas patella baja is associated with an increased
PTL. A thorough evaluation of the proximal tibia is
recommended in dogs with LPL because dogs with
PTL:DTW values 41.65 are considered at high risk for
developing patella baja.
ACKNOWLEDGMENTS
We thank the medical records and radiology services of the
Veterinary Teaching Hospital for their assistance in this
study.
REFERENCES
1. LaFond E, Breur GJ, Austin CC: Breed susceptibility for
developmental orthopedic diseases in dogs. J Am Anim
Hosp Assoc 38:467–477, 2002
2. Roush JK: Canine patellar luxation. Vet Clin North America
Small Anim Pract 23:855–868, 1993
3. Gibbons SE, Macias C, Tonzing MA, et al: Patellar luxation
in 70 large breed dogs. J Small Anim Pract 47:3–9, 2006
4. Arthurs GI, Langley-Hobbs SJ: Complications associated
with corrective surgery for patellar luxation in 109 dogs.
Vet Surg 35:559–566, 2006
5. Piermattei DL, Flo GL: The stifle joint, in Piermattei DL, Flo
GL (eds): Handbook of Small Animal Orthopedics and
Fracture Repair (ed 3). Philadelphia, PA, Saunders, 1997,
pp 516–534
6. Johnson AL, Broaddus KD, Hauptman JG, et al: Vertical
patellar position in large-breed dogs with clinically normal
stifles and large-breed dogs with medial patellar luxation.
Vet Surg 35:78–81, 2006
7. Towle HA, Griffon DJ, Thomas MW, et al: Pre- and postoperative radiographic and computed tomographic evaluation of dogs with medial patellar luxation. Vet Surg
34:265–272, 2005
8. Insall J, Salvati E: Patellar position in the normal knee joint.
Diagnostic Radiol 101:101–104, 1971
9. Carson WG, James SL, Larson LL, et al: Patellofemoral disorders. Physical and radiographic evaluation. Clin Orthop
Rel Res 185:178–186, 1984
10. Simmons E, Cameron JC: Patella alta and recurrent dislocation of the patella. Clin Orthop Rel Res 274:265–269, 1992
11. David- Chaussé J, Vignes L: Critical study of patella alta. II.
Patella alta and femoropatellar pathology. Rev Rhum Mal
Osteoartic 49:507–513, 1982
12. Johnson AL, Probst CW, DeCamp CE, et al: Vertical position of the patella in the stifle joint of clinically normal
large-breed dogs. Am J Vet Res 63:42–46, 2002
13. Riser WH, Parkes LJ, Rhodes WH, et al: Genu valgum: a
stifle deformity of giant dogs. Vet Radiol Ultrasound
10:28–37, 1969
14. Osmond CS, Marcellin-Little DJ, Harrysson OLA, et al:
Morphometric assessment of the proximal portion of the
tibia in dogs with and without cranial cruciate ligament
rupture. Vet Radiol Ultrasound 47:136–141, 2006
15. Read RA, Robins GM: Deformity of the proximal tibia in
dogs. Vet Rec 111:295–298, 1982
16. Selmi A, Padilha FJ: Rupture of the cranial cruciate ligament
associated with deformity of the proximal tibia in five dogs.
J Small Anim Pract 42:390–393, 2001
17. Macias C, McKee WM, May C: Caudal proximal tibial deformity and cranial cruciate ligament rupture in smallbreed dogs. J Small Anim Pract 43:433–438, 2002
18. Slatter D: Patellar luxation, in Slatter D (eds): Textbook of
Small Animal Surgery (ed 3). Philadelphia, PA, Saunders,
2003, pp 2122–2133
19. DeAngelis M: Patellar luxation in dogs. Vet Clin North Am
1:403–415, 1971
20. Budsberg SC, Johnson SA, Schwarz PD, et al: Efficacy of
etodolac for the treatment of osteoarthritis of the hip joint
in dogs. J Am Vet Med Assoc 214:206–210, 1999
21. Havig ME, Dyce J, Kowaleski MP, et al: Relationship
of tibial plateau slope to limb function in dogs treated
with a lateral suture technique for stabilization of cranial
cruciate ligament deficient stifles. Vet Surg 36:245–251,
2007
22. Hurley CR, Hammer DL, Shott S: Progression of radiographic evidence of osteoarthritis following tibial plateau
leveling osteotomy in dogs with cranial cruciate ligament
rupture. J Am Vet Med Assoc 230:1674–1679, 2007
23. Evans HE: Ligaments and joints of the pelvic limb, In Evans
HE, ed., Miller’s Anatomy of the Dog (ed 3). Philadelphia,
PA: WB Saunders, 1993, pp 244–257
24. Abel SB, Hammer DL, Shott S: Use of the proximal portion
of the tibia for measurement of the tibial plateau angle in
dogs. Am J Vet Res 64:1117–1123, 2003
25. Remedios AM, Basher AW, Runyon CL: Medial patellar
luxation in 16 large dogs: a retrospective study. Vet Surg
21:5–9, 1992
26. Brattstrom H: Patella alta in non-dislocating knee joints.
Acta Orthop Scand 41:578–588, 1970
27. Hulse DA: Pathophysiology and management of medial patellar luxation in the dogs. Vet Med Small Anim Clin
76:43–51, 1981
28. Kaiser S, Cornely D, Golder W, et al: Magnetic resonance
measurements of the deviation of the angle of force generated by contraction of the quadriceps muscle in dogs
with congenital patellar luxation. Vet Surg 30:552–558,
2001
29. Carpenter DH, Cooper RC: Mini review of canine stifle joint
anatomy. Anat Histol Embryol 29:321–329, 2000
30. Al-Sayyad MJ, Cameron JC: Functional outcome after tibial
tubercle transfer for the painful patella alta. Clin Orthop
Rel Res 396:152–162, 2002
31. Hughston JC, Georgia C: Subluxation of the patella. J Bone J
Surg 50A:1003–1026, 1968
32. Hauser EDW: Total tendon transplant for slipping patella: a
new operation for recurrent dislocation of the patella. Surg
Gynecol Obstet 66:199–214, 1938
MOSTAFA ET AL
33. Jones JB, Francis KC, Mahoney JR: Recurrent dislocating
patella: a long- term follow-up study. Clin Orthop 20:230–
240, 1961
34. Nicholas JA, Freiberger RH, Killoran PJ: Double contrast
arthrography of the knee. Its value in the management of
two hundred and twenty-five knee derangements. J Bone J
Surg 52A:203–220, 1970
35. Caton J, Mironneau A, Walch G, et al: Idiopathic high
patella in adolescents. A propos of 61 surgical cases. Rev
Chir Orthop Reparatrice Appar Mot 76:253–260, 1990
36. Palmer SH, Servant CTJ, Maguire J, et al: Surgical reconstruction of severe patellofemoral maltracking. Clin Orthop Rel Res 419:144–148, 2004
211
37. L’Eplattenier H, Montavon P: Patellar luxation in dogs and
cats: management and prevention. Compend Contin Educ
Pract Vet 24:292–298, 2002
38. Grelsamer RP: Patella baja after total knee arthroplasty. Is it
really patella baja? J Arthroplasty 17:66–69, 2002
39. Paulos LE, Wnorowski DC, Greenwald AE: Infrapatellar
contracture syndrome: diagnosis, treatment, and long-term
follow up. Am J Sports Med 22:440–449, 1994
40. Dejour D, Levigne C, Dejour H: Postoperative low patella:
treatment by lengthening of the patellar tendon. Rev
Chir Orthop Reparatrice Appar Mot 81:286–295,
1995