DOI = 10.1177/0363546503261360
Winner of the 1999 Aircast Award
Knee Laxity Does Not Vary With the
Menstrual Cycle, Before or After Exercise
Michael J. Belanger,* MD, Douglas C. Moore,†‡ MS, Joseph J. Crisco III,‡ PhD,
‡
‡
‡
Paul D. Fadale, MD, Michael J. Hulstyn, MD, and Michael G. Ehrlich, MD
From the *Department of Orthopaedics, Harvard Medical School/Memorial Hospital of Rhode
‡
Island, Pawtucket, Rhode Island, and Bioengineering Laboratory, Department of
Orthopaedics, Brown Medical School/Rhode Island Hospital, Providence, Rhode Island
Background: An intriguing explanation for the disproportionately high rate of anterior cruciate ligament injury in female athletes
is that the structural properties of the anterior cruciate ligament are affected by the menstrual hormones. Whether this actually
occurs, however, is the subject of ongoing debate.
Hypotheses: (1) Anterior cruciate ligament laxity is different in the follicular, ovulatory, and luteal phases of the menstrual cycle,
and (2) exercise exacerbates the difference in anterior cruciate ligament laxity in the 3 phases.
Methods: Over the course of 10 weeks, repeated knee laxity measurements were taken on 27 high-level female athletes, before
and after exercise. Point in the menstrual cycle was determined with charts of waking temperature and menstruation. The independent effects of menstrual phase and exercise were evaluated using generalized estimating equations.
Results: Data from 18 participants were included in the final analysis. There were no significant differences in anterior cruciate
ligament laxity in any of the 3 menstrual phases, before or after exercise.
Conclusions: Anterior cruciate ligament laxity is not significantly different during the follicular, ovulatory, and luteal phases of
the menstrual cycle, and bicycling exercise does not exacerbate or create any differences in anterior cruciate ligament laxity.
Keywords: anterior cruciate ligament (ACL); menstrual cycle; knee laxity; exercise; KT-2000 arthrometer
It is widely accepted that female athletes sustain disproportionately more ACL injuries than do male athletes who
compete in similar sports.2,10,15,20,31,44 A major stabilizer of
the knee, the ACL is frequently injured without any precipitating collision or traumatic event. Depending on the
study, and the sport being investigated, the ACL injury
rate in female athletes has been reported to be anywhere
from 2 to 10 times the rate in male athletes.2,20 The
increased rate of ACL injury in female athletes is thought
to reflect real, gender-related differences in anatomy, physiology, training, and conditioning, as opposed to simple differences in the level of participation in sport.
There are several factors that have been posited as
likely contributors to the disparity in ACL injury rates in
male and female athletes.2,16,19,20,40 They are often separated into 1 of 2 general categories, depending on whether
they directly affect the internal structures in the knee
joint. Those that influence the anatomy and physiology of
the knee directly are called intrinsic factors and include
such things as generalized ligamentous laxity, ACL size,
femoral notch dimensions, limb alignment, and ligamentous physiology, including the response to circulating hormones. Extrinsic factors, which are more remote but nevertheless influence the development of loads in the joint,
include the level of strength and conditioning, body
mechanics, neuromuscular performance, and footwear.
Currently, there is evidence associating each of the intrinsic and extrinsic factors with ACL injury in female athletes. However, to this point there has been no clear-cut
demonstration of cause and effect.
Of the intrinsic factors believed to influence ACL injury
risk in women, one of the more intriguing is the possibility
that the tissue-level material properties of the ACL are
affected by the normal fluctuations of the hormones associated with the menstrual cycle.24,25,28,37,38,41 Clinically, ACL
injuries in female athletes have been reported to occur
more frequently than expected during certain phases of
the menstrual cycle. For example, Wojtys et al found that
ACL injury risk is increased during the ovulatory (middle)
†
Address correspondence to Douglas C. Moore, MS, Bioengineering
Laboratory, CORO West, Suite 404, 1 Hoppin Street, Providence, RI
02903 (e-mail: douglas_moore@brown.edu).
This work was presented at the 1999 annual meeting of the American
Orthopaedic Society of Sports Medicine where it received the Aircast
Award for Basic Science Research
No author or related institution has received financial benefit from
research in this study.
The American Journal of Sports Medicine, Vol. 32, No. 5
DOI: 10.1177/0363546503261360
© 2004 American Orthopaedic Society for Sports Medicine
1150
Vol. 32, No. 5, 2004
phase of the cycle, in which there is a surge in estrogen
production, and less frequently in the luteal phase,41 and
other investigators have found that injury risk is highest
during the follicular phase.28,38 In laboratory studies, estrogen (and progesterone) receptors have been found in ACL
fibroblasts and in the walls of the blood vessels in the
ACL,24 and tissue culture studies have demonstrated that
estrogen suppresses ACL fibroblast function.25 There is
also evidence that chronic, pregnancy-level estrogen
administration reduces ACL failure load in rabbits.37
Taken together, these findings suggest that the menstrual
hormones, and estrogen in particular, may have the potential to directly affect the strength of the ACL.
Despite the interesting laboratory and basic science
studies suggesting that the menstrual hormones are capable of influencing the structural properties of the ACL,
most of the existing clinical data indicate that this does not
occur in vivo.3,9,22 However, there has never been an investigation of the combined effects of the menstrual cycle and
exercise on the ACL. Because ACL injuries typically occur
during athletic events, it is possible that the associated
activity could exacerbate or magnify subtle differences in
ACL structural properties that are otherwise undetectable
prior to activity. Accordingly, this study was performed to
bridge this gap. Serial measurements of anterior knee laxity, taken before and after a defined exercise protocol, were
correlated with phase in the menstrual cycle (follicular,
ovulatory, or luteal). We hypothesized that (1) knee laxity
would increase and/or decrease as a function of phase in
the menstrual cycle, and (2) exercise would elicit or exacerbate any menstrual cycle–associated changes in knee laxity.
METHODS
Menstrual cycle and exercise-associated changes in ACL
laxity were investigated by correlating anterior tibial
translation with time point in the menstrual cycle. To do
so, knee laxity testing (arthrometry) was performed on
female collegiate athletes twice weekly for 10 weeks, before
and after completing a brief exercise routine. All protocols
were approved by the institutional review boards at Rhode
Island Hospital and Brown University, both of which follow
the National Institutes of Health guidelines on the use of
human subjects in research.
Subjects
Twenty-seven female volunteers were recruited through
the athletic department at Brown University, Providence,
Rhode Island. All were collegiate (26) or high-level recreational (1) athletes. At the time of recruitment, a menstrual history was obtained from each volunteer, and each
was given a standard knee exam, which included anterior
and posterior drawer, Lachman, pivot shift varus and valgus stability, McMurray, and range-of-motion testing. To be
considered for inclusion in the study, the volunteers had to
be in good health, have a normal knee exam, and be actively training for their sports. Women who had a history
of ACL injury, reported irregular menstrual cycles or
Knee Laxity Does Not Vary With the Menstrual Cycle 1151
amenorrhea, or had an abnormal knee exam were excluded
from consideration.
Menstrual Record
Menstrual cycle length and day of ovulation were determined from the charts of waking temperature and menstruation. Once enrolled, the volunteers were asked to
record their temperatures each morning before rising from
bed for the full 10-week duration of the study. They were
also asked to record the days they started and finished
their menstrual periods. To facilitate record keeping, each
volunteer was issued a set of disposable oral thermometers
(TempaDot, PyMaH Corporation, Flemington, NJ; accuracy ±0.2°F) and standard basal body temperature charts
commonly used to monitor fertility.
The dates of knee arthrometry were converted to day in
cycle, and the individual cycle lengths were normalized to
28 days to facilitate comparison. Normalization of cycle
length was done via simple proportional scaling (for example, if a subject’s cycle was 31 days long, each day would be
multiplied by 31/28 and rounded to the nearest whole day).
The normalized 28-day cycle was then divided into 3
phases—follicular (days 1-9), ovulatory (days 10-14), and
luteal (days 15-28)—to facilitate comparison of knee laxity
when both estrogen and progesterone are low (follicular
phase), estrogen is high (ovulatory phase), and estrogen
and progesterone are both high (luteal phase).9,22,32,41,42
Knee Arthrometry
Both knees of each volunteer were tested 2 times per week
for 10 weeks, before and after exercise. Arthrometry was performed with the use of a KT-2000 arthrometer (Medmetric
Corp, San Diego, Calif). During testing, the force and displacement outputs from the KT-2000 arthrometer were
recorded with a computer-based digital data acquisition
system. The recorded data were plotted after each test to
confirm the integrity of the data acquisition and the consistency of the consecutive KT-2000 testing cycles.
All knee laxity tests were performed by a single examiner (MJB) using the standard protocols outlined in the
KT-2000 arthrometer user’s guide. The volunteers were
positioned supine, with the knees flexed to approximately
25° and the lower extremities supported by the KT thigh
and foot support platforms. The KT-2000 was affixed to the
lower leg with the arthrometer’s joint line arrow at the
level of the joint line of the knee, and the soft tissues at the
knee were conditioned by repeatedly cycling (pushing and
pulling) the KT-2000’s force handle. Once the soft tissues
were conditioned, the testing reference position was established and the KT-2000 was zeroed. With the KT-2000
zeroed, the comfort and relaxation of the volunteers were
verified, and 5 consecutive “pull-push” testing cycles (pull
anteriorly to 134 N [30 lb], push posteriorly to 89 N [20 lb],
release) were performed to yield 3 smooth, consistent tests
for analysis. Testing was performed at approximately the
same time each day (between 7 AM and 10 AM), and the
right knee was always tested first. During testing, the
examiner was blinded to the volunteer’s menstrual phase.
1152 Belanger et al
The American Journal of Sports Medicine
Exercise Protocol
Following the first (preexercise) arthrometry session each
day, the volunteers rode a stationary bicycle (Lifecycle, Life
Fitness, Franklin Park, Ill) for 20 minutes, and then they
were retested. During the ride, pedaling speed was maintained at 80 revolutions per minute against a resistance
set to yield 685 W of energy output. The stationary bicycle
was approximately 30 ft from the arthrometry station, and
less than 5 minutes elapsed between the end of the ride
and the second (postexercise) series of knee arthrometry
tests. The protocol used for postexercise testing was the
same as that used for the preexercise testing.
Data Reduction and Analysis
The load and displacement data acquired with the KT2000 arthrometer during each arthrometry session (preexercise and postexercise) were reduced to yield anterior tibial translations and compliance index values for comparison. To start, the 5 load-displacement plots from each session were compared by a blinded reviewer, and 3 smooth,
consistent curves from each session were retained for
analysis (2 tests from each session were deleted to eliminate outliers and other spurious data points). Anterior tibial translations at 134 N of pull (30 lb) were then determined for each of the 3 retained tests, and corresponding
compliance indexes were calculated by subtracting the
translation at 89 N (20 lb) from the translation at 134 N.
Each set of values was then averaged, yielding single values for anterior tibial translation at 134 N and compliance
index for each arthrometry session.
The experimental design in this study was a partially
balanced repeated-measures complete block, with a factorial structure. The factors of interest were phase in the
menstrual cycle (follicular, ovulatory, or luteal), exercise
(preexercise and postexercise), and leg (left or right).
Outcome measures were anterior tibial translation at 134
N and compliance index. The independent effects of menstrual phase, exercise, and leg were evaluated using generalized estimating equations (GEE), which automatically
corrected for the correlations within a given volunteer’s
repeated measurements (SAS, version 8.0, SAS Institute,
Inc, Cary, NC). Significance was accepted when P < .05.
Confidence intervals were calculated to provide additional
insight.
Two strategies were used to estimate the measurement
reliability. The specific reliability of our arthrometry technique was assessed with a small, post hoc repeatability
study in which a single examiner (MJB) performed 4 KT2000 arthrometer examinations on each leg of 4 male volunteers. The testing and data analysis followed the techniques outlined above, with the exception that all of the
exams were performed at 1 sitting, and testing of the right
and left knees was alternated. Variance components analysis was used to rank the sources of variability in the measurements (ie, volunteer, leg, exam, replicate), and intraclass correlation coefficients were used to assess the
repeatability of the measurements (SAS, version 8.0). The
reliability of the measurements made during the study
Figure 1. Typical graph of anterior tibial translations (laxity) at
134 N anterior pull, as a function of day in cycle and exercise
(participant RSR, right leg). There was no obvious menstrual
cycle–related variation in knee laxity for any subject.
was assessed with confidence intervals (mentioned above)
and by calculating intraclass correlation coefficients using
the repeated preexercise knee laxity data from 5 randomly
selected days during the cycle (SAS, version 8.0).
RESULTS
Data from 18 of the 27 volunteers originally enrolled in the
study were included in the final statistical analysis. Data
from 7 subjects were dropped due to inadequate attendance or failure to provide complete basal body temperature charts. Data from 2 additional subjects were dropped
because they failed to menstruate over the course of the
study. The average age, height, and weight of the 18 volunteers who completed the study were 20.4 ± 3.3 years,
67.2 ± 3.6 in, and 151 ± 29 lbs, respectively. The length of
their menstrual cycles averaged 28.9 ± 4.1 days (range, 2238 days), with an ovulatory temperature spike at 16.0 ± 2.6
days. The median number of days the participants were
tested was 14.
In general, for a given subject the test days were randomly distributed over the course of the menstrual cycle,
and there was a fair amount of day-to-day variability in
anterior tibial translation (Figure 1). Because testing was
performed on set calendar dates, the days in the cycle on
which measurements were made varied from subject to
subject. However, when the data from all of the subjects
were pooled, there were multiple data points for each day
in the cycle (Figure 2). There was no obvious pattern in the
day-to-day variability in anterior tibial translation in any
of the individual subjects, nor was there any obvious cyclical association between day in menstrual cycle and knee
laxity, before or after exercise, for the group as a whole
(Figure 2).
When the data from all participants were summarized
(before and after exercise), the range of measured anterior
tibial translation at 134 N was 0.28 mm to 11.24 mm. Most
Vol. 32, No. 5, 2004
Knee Laxity Does Not Vary With the Menstrual Cycle 1153
TABLE 1
Measured Anterior Tibial
Translation at 134 N (30 lb)
Before Exercise
Figure 2. Anterior tibial translations (laxity) at 134 N anterior
pull for both legs of all subjects, before and after exercise.
Note that although testing was conducted twice each week,
there were multiple data points for each day in the normalized cycle.
Menstrual
Phase (days) Left Knee
Right Knee
Left Knee
Right Knee
5.05 ± 1.26
4.02 ± 1.59
5.57 ± 1.70
3.93 ± 1.84
5.62 ± 1.63
3.77 ± 1.74
5.85 ± 1.54
4.10 ± 1.56
5.41 ± 1.46
3.85 ± 1.53
5.63 ± 1.30
4.01 ± 1.52
Follicular
(1-9)
Ovulatory
(10-14)
Luteal
(15-28)
TABLE 2
Summary of Main Effects for Anterior
Tibial Translation at 134 N (30 lb)
Factor
Level (days)
Stage
Follicular (1-9)
Ovulatory (10-14)
Luteal (15-28)
Left
Right
Preexercise
Postexercise
Lega
Exercise
Figure 3. Anterior tibial translation (laxity) as a function of leg
and phase in menstrual cycle, before (left side) and after
(right side) exercise (mean ± SD, 134 N pull). There were no
statistically significant differences in knee laxity between any
of the 3 menstrual phases.
(>95%) of the translations fell between 1 mm and 9 mm.
Before exercise, the mean anterior tibial translation was
4.53 ± 1.91 mm (for both legs and all participants),
whereas after exercise it averaged 4.77 ± 1.91 mm. There
were only small (<2 mm) differences in the measured
anterior tibial translations as a function of menstrual
phase and leg (Table 1, Figure 3).
Statistical analysis of the raw anterior tibial translation
data revealed no significant differences in knee laxity as a
function of phase in the menstrual cycle (Table 2). The
method we used (GEE) estimated anterior tibial translations during the follicular, ovulatory, and luteal phases of
the menstrual cycle of 4.6 mm, 4.8 mm, and 4.7 mm,
respectively, with confidence intervals for each measurement of approximately ±0.6 mm. Similarly, our bicycling
exercise protocol had no significant effect on knee laxity
(Table 2). There was, however, a statistically significant
(P > .05) difference in anterior tibia translation in the right
and left legs, with the left leg being more lax than the right
by approximately 38% (1.5 mm). The compliance index
results were similar to those of anterior tibial translation
(Table 3): no significant differences due to menstrual phase
or exercise but a significantly higher compliance index in
the left leg than in the right leg (P < .05).
After Exercise
a
Laxity
Estimate (mm)
95% Confidence
Interval
4.6
4.8
4.7
5.5
4.0
4.6
4.8
4.0-5.2
4.2-5.4
4.1-5.3
4.9-6.0
3.3-4.6
4.0-5.2
4.2-5.4
P < .05.
TABLE 3
Summary of Main Effects for Compliance Index
Factor
Level (days)
Stage
Follicular (1-9)
Ovulatory (10-14)
Luteal (15-28)
Left
Right
Preexercise
Postexercise
Lega
Exercise
a
Laxity
Estimate (mm)
95% Confidence
Interval
1.3
1.3
1.3
1.7
0.9
1.3
1.3
1.1-1.4
1.1-1.5
1.1-1.5
1.5-1.9
0.7-1.1
1.1-1.4
1.2-1.5
P < .05.
The results of our repeatability study indicated that
most of the variability in our anterior tibial translation
measurements was due to differences in subjects and legs
(together, 77.6%), although a fair amount of variability in
the measurements (18.6%) could also be attributed to differences in how the volunteers were positioned or how the
KT-2000 arthrometer was affixed for each evaluation.
Because the overall translations were small, the magnitude of the effect was also small (<0.2 mm/evaluation). The
intraclass correlation coefficient calculated using the
means of the 3 replicates (for anterior tibial translation at
134 N) in each of the 4 evaluations in the repeatability
study was .93. The intraclass correlation coefficient calcu-
1154 Belanger et al
lated using the preexercise knee laxity data (again, for
anterior tibial translation at 134 N) from 5 randomly
selected days during the larger study was .61.
DISCUSSION
This study was performed to determine whether knee laxity changes as a function of the normal cyclical fluctuations in the hormones associated with the menstrual cycle
and whether these changes might be exacerbated by exercise. To do so, repeated KT-2000 arthrometer measurements were performed on 18 high-level female athletes,
before and after periods of stationary bicycling, and correlated with phase in the menstrual cycle. Our results suggest knee laxity does not vary significantly with changes in
the menstrual cycle. In particular, we found no differences
in knee laxity or compliance index in the follicular, ovulatory, and luteal phases of the menstrual cycle, nor did we
see any significant increases after exercise in any of the 3
menstrual phases.
We used GEE to analyze our data because, although
GEE rely on the same basic assumptions as standard
analysis of variance (ANOVA) and repeated-measures
ANOVA, they provide greater ability to deal with missing
values and to appropriately adjust the results where observations are not independent (which is expected when there
are repeated measures). The experimental design in our
study used repeated measures with a factorial structure.
Being repeated measures, the analysis had to account for
the correlations between the repeated measures to provide
correct degrees of freedom for comparisons as well as provide the correct standard errors. Failure to account for the
repeated measures results in overly optimistic (low) standard errors and a high probability of a type II error (rejection of the null hypothesis when it is in fact true). GEE
adjust standard errors, allow missing values, and provide
the flexibility to assume different correlation structures
among the repeated measures. To confirm our findings
with GEE, the analysis was rerun using a standard repeated-measures ANOVA framework. The data were
simplified so that multiple observations for a single phase
were combined (eg, used the mean from multiple days in
stage 1). This was necessary to avoid dropping people,
because some people have more observations than others.
The results and conclusions were again the same: there
was a difference between left and right legs but no significant effect of menstrual phase or between preexercise and
postexercise.
Because we found no statistically significant differences
in preexercise anterior tibial translation in the 3 phases of
the cycle, it is possible that (1) there was a relatively large
change in laxity that we simply missed, (2) there was a
small change in laxity that we were unable to detect, or (3)
there was no change in laxity. Of the 3, we feel the first is
the least likely, given the relative accuracy of our techniques. The 95% confidence interval for our measurement
of anterior tibial translation at 134 N was 1.2 mm. A post
hoc power calculation found that if the true difference had
been roughly 20% (or approximately 1.0 mm), we would
The American Journal of Sports Medicine
have had 80% power to detect it. We would have had
greater than 95% power to detect a 30% (approximately
1.5 mm) difference. However, based on our analysis, it
appears that anterior tibial translations for each of the 3
phases were within 0.1 mm to 0.2 mm of each other. If, in
fact, these small differences were real, we would not have
been able to detect them. We would argue, however, that
they would have very little impact clinically because they
would be dwarfed by the normal increases in knee laxity
seen with vigorous exercise, which have been reported by
various groups to be anywhere from a low of 0.5 mm for
running and basketball36,39 to a high of 2.0 mm to 2.2 mm,
also for basketball.34
It is possible that our results were skewed by the relatively high drop-out rate (data from 9 of the original 27 volunteers were ultimately dropped from the analysis). We
believe this is unlikely as there is no indication that the
dropouts were in any way related to the hypothesis of the
study—2 were oligomenorrheic, and the remainder simply
failed to show up for testing. Because the study was observational as opposed to interventional, drop-out rate would
not be influenced by treatment or outcome. In the end,
although there is no way to know for sure whether the laxities of the dropouts would have been the same or different
than the study participants, there is no reason that they
should not have been representative.
Similarly, our scheme for normalizing the menstrual
cycles of all participants to 28 days via proportional scaling could have introduced some error. Accordingly, we performed 2 sensitivity analyses to see if this uncertainty
affected the results. First, we analyzed the sensitivity of
our results to the cutoff points for each phase—follicular
(days 0-9), ovulatory (days 10-14), and luteal (days 15-28).
To do so, we reran our analysis, restricting the input to
data points that were at least 2 days from the boundaries
of the follicular and luteal phases and 1 day from the
boundary of the ovulatory phase (it would be better to use
2 days for all phases, but that would have left only 1 day
for phase 2), as well as a stratified analysis that dropped
ovulatory phase data (because it is the shortest). This
restricted the analysis to 628 data points instead of 940
(roughly one third removed) but produced nearly identical
results and conclusions; the means for each phase changed
by less than .1 mm, and the P values and confidence intervals were very similar. Our second sensitivity analysis
involved the use of just those women with cycle lengths between 26 and 30 days (9 of the 18 women). Again, this analysis provided nearly identical results for mean laxity by phase
as the models that included all of the women; the mean
laxities were within 0.1 mm of one another, and the range
of the laxities across the phases was the same (0.2 mm).
There have been 3 recently published studies of the
effect of menstrual-related hormones on knee laxity.3,9,22
Our results are generally consistent with those of Arnold
et al3 and Karageanes et al.22 Arnold et al measured serum
relaxin levels and knee laxity in 57 female athletes (and 5
men) each week for 4 weeks. Although they found that
relaxin levels varied over the 4-week course of their study
(P = .035), they did not detect any changes in knee laxity
Vol. 32, No. 5, 2004
(P = .901).3 Similarly, Karageanes et al performed repeated
KT-1000 arthromter measurements in 26 female high
school athletes over 8 weeks and found no significant
changes in knee laxity over the course of the menstrual
cycle.22 On the other hand, our findings conflict with those
of Deie et al, who reported data from repeated (2 or 3 times
per week) KT-2000 arthromter knee laxity measurements
on 16 women (and 8 men) over the course of 4 weeks.9 They
found significant differences in anterior tibial translation
in the follicular and luteal phases of the cycle at 134 N
anterior pull and in all 3 phases at 89 N anterior pull. This
is surprising given that the differences they detected were
relatively small (approximately 0.5 mm), whereas the
standard deviations of their measurements were comparatively large (approximately 0.7-1.0 mm).
We evaluated knee laxity after exercise, in addition to
evaluating it before exercise, to determine whether physical activity might exacerbate or magnify any potential
hormone-related changes in ACL structural properties.
Because ACL injuries in female athletes typically occur
during active participation in sporting events, we speculated that there might be a synergistic effect of exercise
and the menstrual hormones. However, we found no significant increase in anterior knee laxity caused by exercise, nor did we see any phase-specific exercise-induced
changes in knee laxity attributable to the menstrual cycle.
This was unexpected, as other studies have found significant but small (0.5-0.6 mm) increases in knee laxity after
basketball39 and a muscular-fatiguing exercise protocol.36
It is possible that the bicycling exercise protocol we
selected was too mild to elicit a response, as the peak
strains developed in the ACL during bicycling are relatively low compared with other activities and knee rehabilitation activities.11 At the outset of the study, we made the
choice to use a bicycling exercise protocol because we
wanted an exercise regimen that could be standardized,
which would have been difficult with other activities, such
as running. In retrospect, we may have been better off with
a more aggressive though less standardized protocol.
Assessing the change in ligament structural properties
in vivo is a challenge. Although ACL strains can be measured directly using arthroscopically implanted displacement transducers,5,6 they are impractical for studies like
ours that require repeated measurements over several
weeks. Accordingly, we chose to use knee laxity as an indirect measure of ACL stiffness. The ACL is the dominant
structural element in the knee limiting anterior tibial
translation, accounting for the majority of the resistance to
anterior tibial translation (stiffness) at low loads8,13,27 and
86% of the load developed at 5 mm of anterior displacement.8 We used the KT-2000 arthromter to measure knee
laxity because it is commonly used for knee ligament
arthrometry clinically,4 and it has been widely used for
clinical and basic science ACL-related research.12 The
average error of the displacement-measuring portion of the
device has been reported to be 0.13 mm (±0.12 mm).23 Used
by a single observer for repeated measures on a single
knee, KT-2000 measurements can be repeatable and reliable, with intraclass correlation coefficients as high as .932
Knee Laxity Does Not Vary With the Menstrual Cycle 1155
(at 134 N)29 and 90% and 95% confidence intervals on the
order of ±1.5 mm43 and ±1.65 mm,29 respectively.
Previous investigations have found that the reliability of
laxity-measuring devices such as the KT-2000 arthrometer
decreases when the results of multiple observers are combined.29 In this study, a single operator (MJB) performed
all of the knee laxity measurements, thus eliminating the
variability and bias associated with multiple operators. In
our tester’s hands, repeated measurements of a single knee
in our repeatability study were very reliable (95% confidence interval of ±0.6 mm), although he did appear to test
the right leg differently than the left (approximately 38%
of the measurement variability was due to leg-to-leg differences). Side-to-side differences in healthy volunteers
have been reported by other investigators.29,43 Accordingly,
in analyzing our data we accounted for this source of variability by using a model that evaluated the displacements
of each leg separately and then compared them with one
another.
Possible correlations between increased ligamentous
laxity and the risk of injury have been explored by several
investigators. In an early study, Nicholas evaluated generalized ligament laxity and knee ligament tears in 139 professional football players over the course of 5 years.30
Overall, he found 37 knee ligament ruptures, 28 in the 39
players with 3 or more indices of looseness and 9 in the 100
players with 2 or fewer indices of looseness, leading him to
conclude that increased looseness increased the likelihood
of knee ligament rupture. More recently, in a study of 675
male soldiers in the Spanish Air Force, Acasuso Diaz et al
found a significant increase in musculoligamentous lesions
(though not ACL tears specifically) associated with
increased laxity.1 Other authors have found no correlation
between laxity and injury, however. In 1975, Godshall
reported that he found no correlation between loose jointedness and knee ligament injuries in an 8-year study of
high school athletes using protocols similar to those used
by Nicholas.14 Subsequently, Kalenak and Morehouse
found similar numbers of knee injuries in 401 collegiate
football players classified as tight (24/43, 55.8%) or loose
(19/43, 44.2%) jointed via objective biomechanical testing.21 The relationship between knee laxity and knee
injury remains controversial.
There are less data on the relationship between laxity
and other ligamentous structural properties, and most of
what exists pertains to the properties of the ACL in reconstructed knees. In a 1994 study on the healing of ACL
reconstructions in dogs, Beynnon et al found significant
linear correlations between anteroposterior knee laxity
and graft stiffness, and anteroposterior knee laxity and
ultimate strength,7 leading the authors to conclude that
increased anterior tibial translation may indicate that
ACL grafts have weakened or reduced structural properties. However, in a review of data from several studies,
Grood et al found only sporadic correlations between
anteroposterior knee translation and graft structural properties.17 Furthermore, a review of the data of Hart et al on
the effects of pregnancy on the cellular activity and tissue
mechanical properties in the rabbit medial collateral liga-
1156 Belanger et al
ment (MCL) reveals no correlation between MCL laxity
and stiffness, failure load, or failure stress; they found that
laxity decreased just prior to parturition, whereas stiffness, failure load, and failure stress were unaffected.18
Although we found no menstrual-related or exerciseexacerbated changes in knee laxity in any of the 3 menstrual phases, it is still possible that the menstrual cycle
influences injury risk. In a recent study of 65 women with
ACL injuries, the injuries appeared to occur more frequently in the ovulatory phase of the cycle, as documented
by urine hormone metabolite analysis.41 Rather than affect
the ACL directly, the menstrual hormones may affect ACL
injury risk by altering neuromuscular performance.
Although reaction time does not appear to be influenced by
the menstrual cycle,26,33 there is evidence that skeletal
muscle function may be.35
In summary, our repeated KT-2000 arthromter testing of
female athletes over the course of 10 weeks revealed no
significant differences in knee laxity in any of the 3 phases
of the menstrual cycle, nor were there any exerciseinduced exacerbations in laxity in any of the 3 phases with
our bicycling exercise protocol. Viewed in light of the fact
that the ACL dominates the resistance of the tibia to
anterior translation (laxity),8,13,27 our results, and those of
other groups that have done similar studies,3,22 suggest
that the menstrual hormones do not affect the ACL. If they
do, the effect is certainly smaller in magnitude than the
changes associated with exercise.
ACKNOWLEDGMENT
The authors thank Gail Connolly for facilitating volunteer
recruitment, Daniel P Labrador and Robert D McGovern
for assistance with testing and data analysis, and Daniel
Gottlieb for consultation on the statistical analysis.
Funding for this work was provided by the RIH Orthopaedic Foundation, Inc, and University Orthopedics, Inc.
REFERENCES
1. Acasuso Diaz M, Collantes Estevez E, Sanchez Guijo P. Joint hyperlaxity and musculoligamentous lesions: study of a population of
homogeneous age, sex and physical exertion. Br J Rheumatol.
1993;32:120-122.
2. Arendt E, Dick R. Knee injury patterns among men and women in collegiate basketball and soccer: NCAA data and review of literature. Am
J Sports Med. 1995;23:694-701.
3. Arnold C, Van Bell C, Rogers V, et al. The relationship between serum
relaxin and knee joint laxity in female athletes. Orthopedics.
2002;25:669-673.
4. Bach BR. Knee laxity testing devices. In Scott WN, ed. The Knee. St.
Louis, Mo: Mosby Year Book; 1993:673-700.
5. Beynnon B, Howe JG, Pope MH, et al. The measurement of anterior
cruciate ligament strain in vivo. Int Orthop. 1992;16:1-12.
6. Beynnon BD, Fleming BC. Anterior cruciate ligament strain in-vivo: a
review of previous work. J Biomech. 1998;31:519-525.
7. Beynnon BD, Johnson RJ, Toyama H, et al. The relationship between
anterior-posterior knee laxity and the structural properties of the
patellar tendon graft: a study in canines. Am J Sports Med.
1994;22:812-820.
The American Journal of Sports Medicine
8. Butler DL, Noyes FR, Grood ES. Ligamentous restraints to anteriorposterior drawer in the human knee: a biomechanical study. J Bone
Joint Surg Am. 1980;62:259-270.
9. Deie M, Sakamaki Y, Sumen Y, et al. Anterior knee laxity in young
women varies with their menstrual cycle. Int Orthop. 2002;26:154156.
10. Ferretti A, Papandrea P, Conteduca F, et al. Knee ligament injuries in
volleyball players. Am J Sports Med. 1992;20:203-207.
11. Fleming BC, Beynnon BD, Renstrom PA, et al. The strain behavior of
the anterior cruciate ligament during bicycling: an in vivo study. Am J
Sports Med. 1998;26:109-118.
12. Freedman KB, D’Amato MJ, Nedeff DD, et al. Arthroscopic anterior
cruciate ligament reconstruction: a metaanalysis comparing patellar
tendon and hamstring tendon autografts. Am J Sports Med.
2003;31:2-11.
13. Fukubayashi T, Torzilli PA, Sherman MF, et al. An in vitro biomechanical evaluation of anterior-posterior motion of the knee: tibial displacement, rotation, and torque. J Bone Joint Surg Am. 1982;64:258264.
14. Godshall RW. The predictability of athletic injuries: an eight-year
study. J Sports Med. 1975;3:50-54.
15. Gray J, Taunton JE, McKenzie DC, et al. A survey of injuries to the
anterior cruciate ligament of the knee in female basketball players. Int
J Sports Med. 1985;6:314-316.
16. Griffin LY, Agel J, Albohm MJ, et al. Noncontact anterior cruciate ligament injuries: risk factors and prevention strategies. J Am Acad
Orthop Surg. 2000;8:141-150.
17. Grood ES, Walz-Hasselfeld KA, Holden JP, et al. The correlation
between anterior-posterior translation and cross-sectional area of
anterior cruciate ligament reconstructions. J Orthop Res.
1992;10:878-885.
18. Hart DA, Reno C, Frank CB, et al. Pregnancy affects cellular activity,
but not tissue mechanical properties, in the healing rabbit medial collateral ligament. J Orthop Res. 2000;18:462-471.
19. Huston LJ, Greenfield ML, Wojtys EM. Anterior cruciate ligament
injuries in the female athlete: potential risk factors. Clin Orthop.
2000:50-63.
20. Hutchinson MR, Ireland ML. Knee injuries in female athletes. Sports
Med. 1995;19:288-302.
21. Kalenak A, Morehouse CA. Knee stability and knee ligament injuries.
JAMA. 1975;234:1143-1145.
22. Karageanes SJ, Blackburn K, Vangelos ZA. The association of the
menstrual cycle with the laxity of the anterior cruciate ligament in
adolescent female athletes. Clin J Sport Med. 2000;10:162-168.
23. Kowalk DL, Wojtys EM, Disher J, et al. Quantitative analysis of the
measuring capabilities of the KT-1000 knee ligament arthrometer. Am
J Sports Med. 1993;21:744-747.
24. Liu SH, al-Shaikh R, Panossian V, et al. Primary immunolocalization of
estrogen and progesterone target cells in the human anterior cruciate
ligament. J Orthop Res. 1996;14:526-533.
25. Liu SH, Al-Shaikh RA, Panossian V, et al. Estrogen affects the cellular
metabolism of the anterior cruciate ligament: a potential explanation
for female athletic injury. Am J Sports Med. 1997;25:704-709.
26. Loucks J, Thompson H. Effect of menstruation on reaction time. Res
Q. 1968;39:407-408.
27. Markolf KL, Mensch JS, Amstutz HC. Stiffness and laxity of the
knee—the contributions of the supporting structure: a quantitative in
vitro study. J Bone Joint Surg Am. 1976;58:583-594.
28. Myklebust G, Engebretsen L, Hoff Brækken I, et al. Prevention of
anterior cruciate ligament injuries in female team handball players: a
prospective intervention study over three seasons. Clin J Sport Med.
2003;13:71-78.
29. Myrer JW, Schulthies SS, Fellingham GW. Relative and absolute reliability of the KT-2000 arthrometer for uninjured knees: testing at 67,
89, 134, and 178 N and manual maximum forces. Am J Sports Med.
1996;24:104-108.
30. Nicholas JA. Injuries to knee ligaments: relationship to looseness and
tightness in football players. JAMA. 1970;212:2236-2239.
Vol. 32, No. 5, 2004
31. Nielsen AB, Yde J. Epidemiology of acute knee injuries: a prospective
hospital investigation. J Trauma. 1991;31:1644-1648.
32. Pennington GW, Naik S. Hormone Analysis: Methodology and Clinical
Interpretation. Boca Raton, Fla: CRC Press; 1981.
33. Pierson WR, Lockhart A. Effect of menstruation on simple reaction
time and movement time. Br Med J. 1963;1:796-797.
34. Sakai H, Tanaka S, Kurosawa H, et al. The effect of exercise on anterior knee laxity in female basketball players. Int J Sports Med.
1992;13:552-554.
35. Sarwar R, Niclos BB, Rutherford OM. Changes in muscle strength,
relaxation rate and fatiguability during the human menstrual cycle. J
Physiol. 1996;493(pt 1):267-272.
36. Skinner HB, Wyatt MP, Stone ML, et al. Exercise-related knee joint
laxity. Am J Sports Med. 1986;14:30-34.
37. Slauterbeck J, Clevenger C, Lundberg W, et al. Estrogen level alters
the failure load of the rabbit anterior cruciate ligament. J Orthop Res.
1999;17:405-408.
Knee Laxity Does Not Vary With the Menstrual Cycle 1157
38. Slauterbeck JR, Fuzie SF, Smith MP, et al. The menstrual cycle, sex
hormones, and anterior cruciate ligament injury. J Athletic Train.
2002;37:275-280.
39. Steiner ME, Grana WA, Chillag K, et al. The effect of exercise on
anterior-posterior knee laxity. Am J Sports Med. 1986;14:24-29.
40. Traina SM, Bromberg DF. ACL injury patterns in women. Orthopedics.
1997;20:545-549, Quiz 550-551.
41. Wojtys EM, Huston LJ, Boynton MD, et al. The effect of the menstrual cycle on anterior cruciate ligament injuries in women as determined
by hormone levels. Am J Sports Med. 2002;30:182-188.
42. Wojtys EM, Huston LJ, Lindenfeld TN, et al. Association between the
menstrual cycle and anterior cruciate ligament injuries in female athletes. Am J Sports Med. 1998;26:614-619.
43. Wroble RR, Van Ginkel LA, Grood ES, et al. Repeatability of the KT1000 arthrometer in a normal population. Am J Sports Med.
1990;18:396-399.
44. Zelisko JA, Noble HB, Porter M. A comparison of men’s and women’s
professional basketball injuries. Am J Sports Med. 1982;10:297-299.