Clinical Orthopaedics and Related Research®

, Volume 473, Issue 8, pp 2548–2558

Are There Sex Differences in Knee Cartilage Composition and Walking Mechanics in Healthy and Osteoarthritis Populations?

  • Deepak Kumar
  • Richard B. Souza
  • Karupppasamy Subburaj
  • Toran D. MacLeod
  • Justin Singh
  • Nathaniel E. Calixto
  • Lorenzo Nardo
  • Thomas M. Link
  • Xiaojuan Li
  • Nancy E. Lane
  • Sharmila Majumdar
Symposium: Sex Differences in Musculoskeletal Disease and Science

Abstract

Background

Women are at a greater risk for knee osteoarthritis (OA), but reasons for this greater risk in women are not well understood. It may be possible that differences in cartilage composition and walking mechanics are related to greater OA risk in women.

Questions/purposes

(1) Do women have higher knee cartilage and meniscus T than men in young healthy, middle-aged non-OA and OA populations? (2) Do women exhibit greater static and dynamic (during walking) knee loading than men in young healthy, middle-aged non-OA and OA populations?

Methods

Data were collected from three cohorts: (1) young active (< 35 years) (20 men, 13 women); (2) middle-aged (≥ 35 years) without OA (Kellgren-Lawrence [KL] grade < 2) (43 men, 65 women); and (3) middle-aged with OA (KL > 1) (18 men, 25 women). T and T2 relaxation times for cartilage in the medial knee, lateral knee, and patellofemoral compartments and medial and lateral menisci were quantified with 3.0-T MRI. A subset of the participants underwent three-dimensional motion capture during walking for calculation of peak knee flexion and adduction moments, flexion and adduction impulses, and peak adduction angle. Differences in MR, radiograph, and gait parameters between men and women were compared in the three groups separately using multivariate analysis of variance.

Results

Women had higher lateral articular cartilage T (men = 40.5 [95% confidence interval {CI}, 38.8–42.3] ms; women = 43.3 [95% CI, 41.9–44.7] ms; p = 0.017) and patellofemoral T (men = 44.4 [95% CI, 42.6–46.3] ms; women = 48.4 [95% CI, 46.9–50.0] ms; p = 0.002) in the OA group; and higher lateral meniscus T in the young group (men = 15.3 [95% CI, 14.7–16.0] ms; women = 16.4 [95% CI, 15.6–17.2] ms; p = 0.045). The peak adduction moment in the second half of stance was lower in women in the middle-aged (men = 2.05 [95% CI, 1.76–2.34] %BW*Ht; women = 1.66 [95% CI, 1.44–1.89] %BW*Ht; p = 0.037) and OA (men = 2.34 [95% CI, 1.76–2.91] %BW*Ht; women = 1.42 [95% CI, 0.89–1.94] %BW*Ht; p = 0.022) groups. Static varus from radiographs was lower in women in the middle-aged (men = 178° [95% CI, 177°–179°]; women = 180° [95% CI, 179°–181°]; p = 0.002) and OA (men = 176° [95% CI, 175°–178°]; women = 180° [95% CI, 179°–181°]; p < 0.001) groups. Women had lower varus during walking in all three groups (young: men = 4° [95% CI, 3°–6°]; women = 2° [95% CI, 0°–3°]; p = 0.013; middle-aged: men = 2° [95% CI, 1°–3°]; women = 0° [95% CI, −1° to 1°]; p = 0.015; OA: men = 4° [95% CI, 2°–6°]; women = 0° [95% CI, −2° to 2°]; p = 0.011). Women had a higher knee flexion moment (men = 4.24 [95% CI, 3.58–4.91] %BW*Ht; women 5.40 [95% CI, 4.58–6.21] %BW*Ht; p = 0.032) in the young group.

Conclusions

These data demonstrate differences in cartilage composition and gait mechanics between men and women in young healthy, middle-aged healthy, and OA cohorts. Considering the cross-sectional nature of the study, longitudinal research is needed to investigate if these differences in cartilage composition and walking mechanics are associated with a greater risk of lateral tibiofemoral or patellofemoral OA in women. Future studies should also investigate the relative risk of lateral versus medial patellofemoral cartilage degeneration risk in women compared with men.

Level of Evidence

Level III, retrospective study.

Introduction

Knee osteoarthritis (OA) is a leading cause of disability in noninstitutionalized adults resulting in significant knee pain and difficulties with activities of daily living [14, 21, 24]. In the United States, approximately 5% of American adults older than 25 years of age and 12% older than 65 years of age are affected by knee OA [15, 32]. However, women are at a significantly greater risk of knee OA than men, particularly after 50 years of age [3, 50]. A systematic review and meta-analysis on published literature on risk factors for onset of knee OA reported a pooled odds ratio of 1.84 (95% confidence interval [CI], 1.32–2.55) for women compared with men from nine studies that reported effect sizes for sex [3]. However, the reasons for greater risk of knee OA in women are not well understood.

It may be possible that there are differences in cartilage composition between men and women that predispose women to a greater risk of cartilage degeneration, but these differences cannot be found on plain radiographs or anatomic MRI [37, 42]. Quantitative MRI techniques, however, including T relaxation time mapping, enable the assessment of cartilage composition noninvasively and are useful to visualize the loss of collagen matrix that occurs early in the OA disease process [1, 33, 44, 45]. An increase in T relaxation times indicates loss of proteoglycans and possibly a concomitant increase in hydration [1, 33, 34, 38]. T relaxation times have been shown to detect early knee cartilage degeneration in multiple studies [33, 49, 53, 55].

Another factor that has been shown to be related to the risk of knee OA onset and progression is frontal plane alignment (varus or valgus) [20, 46] and loading at the knee during walking [2]. Varus alignment is associated with a greater risk of medial knee OA and valgus alignment is associated with a greater risk of lateral knee OA [20, 46]. During walking, the loading in the frontal plane and sagittal plane has been shown to be related to greater risk of onset and progression of knee OA [9, 13, 31, 39]. However, it is not known if there are differences in static and dynamic measures of knee loading that are present between men and women in healthy and OA populations that could be related to greater prevalence of knee OA in women.

The primary objectives of this study were to (1) evaluate whether women have higher knee articular and meniscus cartilage T relaxation times when compared with men in young healthy, middle-aged non-OA, and OA populations; and (2) to assess if women exhibit greater static and dynamic (during walking) knee loading compared with men in young healthy, middle-aged non-OA and OA populations, which could explain the sex-related difference in the prevalence of OA. The study hypotheses were that articular cartilage relaxation times, static varus, and gait mechanics will be different between men and women in all three cohorts.

Materials and Methods

Study Design and Setting

Secondary data analyses from an observational cross-sectional study were conducted at the University of California-San Francisco, which is a tertiary care medical and research institution.

Participants

Data reported in this study are secondary analyses from participants recruited from the community for two studies. The first cohort was of “young healthy” volunteers (n = 33; men:women = 20:13; age = 28 ± 4 years; body mass index [BMI], 23 ± 2 kg/m2 between the ages of 20 and 35 years who were physically active (at least 150 min/week of moderate to vigorous physical activity) without any history of knee pain or any other lower extremity injuries that would confound the evaluation of walking patterns. For the second study, participants were recruited for an observational longitudinal study on knee OA. The inclusion criteria for patients with “OA” (n = 42; men:women = 18:25; age = 58 ± 10 years; BMI, 25 ± 4 kg/m2) were age > 35 years, knee symptoms consistent with OA (pain, aching, or stiffness on most days per month during the past year or use of medication for knee pain on most days per month during the past year), and definite radiographic evidence of knee OA (Kellgren-Lawrence [KL] > 1). The inclusion criteria for “middle-aged non-OA” participants (n = 108; men:women = 43:65; age = 50 ± 9 years; BMI = 24 ± 4 kg/m2) were age > 35 years, no knee pain or stiffness in either knee or use of medications for knee pain in the last year, and no radiographic evidence of OA (KL ≤ 1) on either knee. The exclusion criteria for all subjects were (1) concurrent use of an investigational drug; (2) history of intraarticular fracture or surgical intervention in the study knee; (3) conditions other than OA that limit lower extremity function and mobility and/or would confound the evaluation of function; and (4) contraindications to MRI. All subjects signed a written informed consent form before participation in the study, and all protocols were approved by the University of California, San Francisco Committee on Human Research.

Radiographs

To determine the presence and severity of OA, all middle-aged and OA participants underwent bilateral weightbearing and fixed-flexion posteroanterior knee radiographs with the aid of a Synaflexer device (Synarc, Newark, CA, USA) [8]. A radiologist with more than 20 years of experience in musculoskeletal imaging (TML) performed the KL scoring of the tibiofemoral compartment from these radiographs [29]. Lower extremity alignment was assessed by a reader (DK) blinded to subject information using a standing, AP radiograph in which the hips, knees, and ankles were visible. Alignment was determined by the angle (varus < 180°, valgus > 180°) of the mechanical axes of the femur and tibia [26]. These data were available from all participants in the middle-aged non-OA and OA groups and a subset of participants from the young healthy group (n = 16; 10 men, six women).

MRI Acquisition

All imaging was performed with 3.0-T GE (General Electric Healthcare, Waukesha, WI, USA) MR scanners (Signa HDx for young subjects, MR 750w for remaining subjects) using an eight-channel phased array transmit/receive knee coil (Invivo, Orlando, FL, USA). Bilateral knees were imaged for the young group and data from the right knee were used for the analyses in this study. The knee was selected at random by the participants for the middle-aged non-OA group. For the OA group, the knee with worse radiographic severity, or worse symptoms if similar radiographic severity, was imaged. Imaging sequences for segmentation of cartilage regions of interest and quantification of MR relaxation times were acquired (Table 1).
Table 1

MR acquisition parameters for the two studies

Sequence

Parameters

Purpose

For young healthy group

 Sagittal 3-D fat-saturated high-resolution spoiled gradient echo

TR/TE = 15/6.7 ms, flip angle = 18, FOV = 14 cm, matrix = 512 × 512, slice thickness = 1 mm, bandwidth = 31.25 kHz, NEX = 1, acquisition time = 8 minutes

Cartilage segmentation

 T quantification sequence

TSL = 0/10/40/80 ms, prep TE = 0/13.7/27.3/54.7 ms, FOV = 14 cm, matrix = 256 × 128, views per segment (VPS) = 64, time of recovery = 1.2 seconds, slice thickness = 4 mm, number of slices = 26, acquisition time = 9 minutes 30 seconds

Articular and meniscal cartilage T relaxation times

For middle-aged control and OA groups

 Sagittal 3-D fat-saturated fast spin echo Cube

TR/TE = 1500/26.69 ms, FOV = 16 cm, matrix = 384 × 384, slice thickness = 0.5 mm, echo train length = 32, bandwidth = 37.5 kHz, NEX = 0.5, acquisition time = 10.5 minutes

Cartilage segmentation

 T quantification sequence

TR/TE = 9/2.6 ms, time of recovery = 1500 ms, FOV = 14 cm, matrix = 256 × 128, slice thickness = 4 mm, bandwidth = 62.5 kHz, time of spin-lock (TSL) = 0/2/4/8/12/20/40/80 ms, frequency of spin-lock = 500 Hz, acquisition time = 11 minutes

Articular and meniscus cartilage T relaxation times

3-D = three-dimensional; FOV = field of view; NEX = number of excitations; OA = osteoarthritis.

Quantitative MRI Analyses

Sagittal high-resolution images (spoiled gradient echo or Cube) were rigidly registered to the first image of the T sequence (TSL = 0 ms) and used for cartilage segmentation. Medial femoral condyle (MF), medial tibia (MT), lateral femoral condyle (LF), lateral tibia (LT), patella (P), Trochlea (TrF), medial meniscus, and lateral meniscus cartilage compartments were segmented semiautomatically (automated edge detection and manual correction) on multiple slices using in-house software program developed with Matlab (Mathworks, Natick, MA, USA) based on edge detection and Bezier splines [6]. T relaxation time maps were constructed by two-parameter fitting of the image intensity (voxel-by-voxel) for four T-weighted images using a Levenberg-Marquardt monoexponential fitting algorithm developed in-house: S(TSL) = S0 exp(−TSL/T), where TSL is the spin lock time and S0 is the signal intensity when TSL = 0 ms; S(TE) = S0 exp(−TE/T2). The TSLs used for the fitting were matched for the two studies as closely as possible, ie, T (TSL 0, 10, 40, and 80 ms for young group; TSL 0, 12, 40, and 80 ms for remaining subjects). However, for meniscus, only the first three of the mentioned four TSLs for T-weighted images were used [44]. T-weighted images with the longest TSL had a very low signal-to-noise ratio (< 5) for meniscus resulting from short T in meniscus, respectively, and therefore were not used during map reconstruction. The cartilage regions of interest were overlaid onto the T maps. The cartilage splines were adjusted manually to avoid synovial fluid or surrounding anatomy. To eliminate artifacts resulting from partial volume effects with synovial fluid, voxels with relaxation time ≥ 100 ms for cartilage and ≥ 40 ms for meniscus in the T relaxation time maps were excluded before quantification. These techniques have been shown to be reproducible [4, 5, 35]. Mean T values were calculated for three knee articular cartilage compartments: medial knee (mean of MF and MT), lateral knee (mean of LF and LT), and patellofemoral (mean of P and TrF) and for two knee meniscus cartilage compartments: medial meniscus and lateral meniscus.

Walking Gait Analyses

A subset of the participants underwent walking gait analyses (n = 25 of 33 for young healthy group, n = 78 of 109 for middle-aged non-OA group, and n = 29 of 42 for OA group). The subset consisted of participants who agreed to participate in the motion analysis session. Subjects walked at their self-selected speed while three-dimensional kinematic data were collected at 250 Hz using a passive 10-camera system (VICON, Oxford Metrics, UK), and kinetic data were collected at 1000 Hz from two embedded force platforms (AMTI, Watertown, MA, USA). Nine and a half millimeters spherical retroreflective markers were placed on bony landmarks of bilateral lower extremities for identification of joint centers and rigid clusters placed bilaterally on the lateral surface of the subject’s thighs, legs, and heel shoe counters were used to track segment motions [48]. A trial was considered acceptable when there was clean foot strike on any of the force platforms and the speed was within ± 5% of the first good trial. At least five good trials were collected from both lower extremities in young subjects and data from the right lower extremity were used in these analyses. For the remaining subjects, data from the extremity that was scanned were used. Kinematic and kinetics were calculated using Visual3D (C-motion, Georgetown, MD, USA). All net joint moments are expressed as external moments (Nm) and normalized to body weight (BW) and height (Ht) (%BW*Ht). Variables were calculated for the stance phase when the foot was in contact with the ground and included peak external knee flexion moment, first peak knee adduction moment, second peak knee adduction moment, and peak adduction angle. The average of five trials was calculated for each subject. The first peak of adduction moment and peak varus have been related to medial compartment loading [31]. Additionally, the first peak knee adduction moment has been related to increased risk of medial OA progression [39]. Walking with greater toe-out was found to be related to lower second peak knee adduction moment and a reduced risk of OA progression [7]. Finally, it has been shown that the adduction moment by itself is not sufficient to describe knee loading during walking and the knee flexion moment also contributes to knee loading [52].

Statistical Analyses

Mean and 95% CIs were calculated for all outcome measures. Age and BMI were compared between the sexes in each group using independent sample t-tests. Because the outcomes were at the ratio level of measurement, were from independent samples, and had homogenous variance, multivariate analysis of variance was used to compare the MRI and gait outcome measures between men and women separately in the three groups. The analyses were repeated with radiographic alignment as a covariate to ascertain the effect of alignment on sex differences in the middle-aged and OA groups. The number of subjects with radiographic alignment data was very low for the young group. Distribution of KL was compared between sexes in the middle-aged non-OA and OA groups using chi-square tests. Significance was set at p < 0.05. Effect size was calculated using the partial eta squared (ηp2) statistic. All analyses were performed using IBM SPSS Version 22 (IBM Corporation, Armonk, NY, USA).

Demographics

There were no differences in age or BMI (Table 2) between any groups, except women in the young healthy group had lower BMI than the men when all the participants were considered.
Table 2

Age and BMI for the participants in the three groups

Variable

Young healthy

Middle-aged control

Osteoarthritis

Men

Women

p value

Men

Women

p value

Men

Women

p value

For T and T2 comparisons

 Number

20

13

 

43

65

 

18

25

 

 Age* (years)

28 (26–29)

29 (27–31)

0.285

49 (46–51)

51 (48–53)

0.287

57 (52.7–61.8)

58 (55–62)

0.702

 BMI* (kg/m2)

23 (22–24)

22 (20–23)

0.024

25 (24–26)

24 (24–24)

0.081

26 (24–27)

25 (25–26)

0.087

 KL

   

KL0 = 26, KL1 = 17

KL0 = 30, KL1 = 35

0.168

KL2 = 9, KL3 = 8, KL4 = 1

KL2 = 10, KL3 = 12, KL4 = 3

0.693

For gait mechanics comparisons

 Number

15

10

 

29

49

 

13

16

 

 Age* (years)

27 (25–30)

29 (27–32)

0.297

49 (45–52)

51 (48–53)

0.408

58 (52–63)

62 (57–67)

0.252

 BMI* (kg/m2)

23 (22–24)

22 (21–23)

0.177

25 (24–26)

24 (23–25)

0.083

26 (24–28)

24 (23–28)

0.142

* Values are mean with 95% confidence intervals in parentheses; p values from independent sample t-tests; p value from chi square tests; BMI = body mass index; KL = Kellgren-Lawrence.

Results

Sex Differences in Articular and Meniscus Cartilage Composition

Women had higher articular cartilage T than men in the lateral (p = 0.017, ηp2 = 0.137 for T) and patellofemoral (p = 0.002, ηp2 = 0.227 for T) compartments in the OA group, but not in the other two groups (Fig. 1). There were no significant differences in medial knee cartilage T relaxation times between men and women in any of the groups (p > 0.05) (Fig. 1). After adjusting for radiographic alignment, the differences in the lateral T were no longer statistically significant in the OA group (p = 0.036). However, after adjusting for radiographic alignment, the differences in patellofemoral T remained significant (p = 0.002), and the differences in medial T were observed to be significant (p = 0.036) with women having higher T.
Fig. 1A–C

T relaxation times for men (black bars) and women (gray bars) are shown for the medial (A), lateral (B), and patellofemoral (C) articular cartilage compartments in the three groups. Error bars indicate 95% CIs. *Significant differences between men and women at p < 0.05.

For the meniscus, lateral meniscus T was higher in women in the young healthy group (p = 0.045, ηp2 = 0.140) (Fig. 2). The medial meniscus T was not statistically different between men and women in any group (Fig. 2).
Fig. 2A–B

T relaxation times for men (black bars) and women (gray bars) are shown for the medial (A) and lateral (B) menisci in the three groups. Error bars indicate 95% CIs. *Significant differences between men and women at p < 0.05.

Sex Differences in Static and Dynamic Loading Parameters

In the middle-aged and OA groups, women had a lower second peak knee adduction moment (p = 0.037, ηp2 = 0.056 for middle-aged group; p = 0.022, ηp2 = 0.179 for OA group) (Fig. 3). Women also had lower peak varus angle during walking in all three groups (p = 0.013, ηp2 = 0.241 for young healthy; p = 0.015, ηp2 = 0.075 for middle-aged group; p = 0.011, ηp2 = 0.216 for OA group) as well as lower varus angle as measured from radiographs in the middle-aged (p = 0.002, ηp2 = 0.090) and OA (p < 0.001, ηp2 = 0.278) groups (Fig. 3). The differences in second peak adduction moment and peak varus during walking in the middle-aged and OA groups were not significant after adjusting for radiographic alignment (p > 0.05). In the sagittal plane, women had a higher knee flexion moment (p = 0.032, ηp2 = 0.184) in the young group. The differences in walking speed between men and women were not significant in any of the groups (men = 1.4 ± 0.1 m/s, women = 1.4 ± 0.2 m/s in young healthy; men = 1.6 ± 0.2 m/s, women = 1.5 ± 0.2 m/s in middle-aged group; men = 1.5 ± 0.2 m/s, women = 1.5 ± 0.3 m/s in OA group) (Fig. 3).
Fig. 3A–E

First and second peak adduction moment (AB), flexion moment (C), peak varus during walking (D), and static varus from radiograph (E) are shown for men (black bars) and women (gray bars) in the three groups. Error bars indicate 95% CIs. *Significant differences between men and women at p < 0.05.

Discussion

Women have a significantly greater risk of knee OA than men, especially with increasing age. However, the reasons for greater risk of knee OA in women are not well understood. Walking mechanics are implicated in the pathogenesis of cartilage degeneration in OA. Hence, it may be possible that there are differences in cartilage composition and walking mechanics between men and women that predispose women to a greater risk of OA. The objective of this study was to evaluate the differences in knee cartilage MR relaxation times, and static and dynamic measures of knee loading, between men and women in young healthy, middle-aged healthy, and OA populations. Our results show that women have higher MR relaxation times in the lateral and patellofemoral compartments in the OA group and women have lower second peak adduction moment in the middle-aged and OA groups. Women also had lower static and dynamic varus in the middle-aged and OA groups and lower varus during walking in all groups.

The study has several limitations that need to be considered while interpreting the findings. The sample size in the young healthy and OA groups was relatively small, especially for biomechanical parameters; and the number of men and women was not balanced in each group. Hence, these findings would need to be replicated in larger populations. Because these results are from secondary analyses of existing data, we used all existing data to maximize the sample size. Additionally, the cross-sectional nature of the study allowed for the detection of an association, but we cannot infer causality from this study or make interpretations about associations of these outcomes with risk of knee OA. Longitudinal studies are needed to investigate if there are sex differences in cartilage composition and walking mechanics in these populations over time. Earlier studies have reported radiographic joint space width. We report KL scores and quantitative MR imaging data. Quantitative MR data provide evidence of early cartilage degeneration that may not be visualized using radiographic joint space width measurements. The use of different imaging sequence in the young group versus the middle-aged and OA groups did not allow comparison of MR parameters among the three cohorts. However, the goal of this study was to evaluate the differences between men and women in the three groups. Also, we did not make adjustments for multiple comparisons; therefore, p values close to 0.05 should be interpreted with caution.

We observed that women in the OA group had higher articular cartilage T in the lateral compartment compared with men. These differences had medium to large effect sizes in our cohort. Higher T indicates worse cartilage composition with lower proteoglycan content. Wise et al. [54] reported that women have a greater prevalence of joint space narrowing in the lateral compartment than men in a sample of 5202 knees of people with and without knee OA from the Multi-center Osteoarthritis Study (MOST). Our MRI findings partially support this observation with higher T in the lateral compartment in the OA group but not in the middle-aged group that did not have radiographic OA. Use of quantitative MR imaging allows us to observe the cartilage degeneration while controlling for the presence of radiographic OA. This is important considering the fact that interventions to prevent the onset of OA are likely to be more effective before radiographic changes have occurred. Furthermore, it has been shown that MR relaxation measures of cartilage predict future progression of OA [36, 43]. Wise et al. [54] also reported that women have greater prevalence of valgus alignment, which was also seen in our cohorts. After adjusting for lower extremity alignment, Wise et al. found that the prevalence of lateral joint space narrowing between women and men was no longer significant, and women had a greater prevalence of medial joint space narrowing. We observed similar results in that the differences in lateral articular cartilage T were no longer statistically significant after adjusting for radiographic alignment, whereas the differences in medial T became significant. These results indicate that greater degeneration of the lateral compartment in women may be partially related to more valgus alignment of the lower extremity in women, whereas the differences in the medial compartment may be related to other factors. However, longitudinal studies are needed to confirm these observations.

We also observed greater T in the patellofemoral articular cartilage in women compared with men in the OA group. These differences were significant even after adjusting for radiographic alignment. Wise et al. [54] did not report data on patellofemoral compartment in their study. However, using MOST data, Glass et al. [22] recently reported that women have a greater prevalence of patellofemoral OA than men. Furthermore, in knees with patellofemoral OA, women reported greater severity of pain for all KL grades than men. The authors speculated that the mechanisms underlying tibiofemoral OA may be different from those underlying patellofemoral OA, as has been suggested in other studies [12, 25]. Our data support these observations because lower extremity alignment affected the differences in tibiofemoral cartilage T but not the differences in patellofemoral cartilage T. One anatomic factor that has been shown to be related to patellofemoral OA that we did not evaluate in our study is trochlear dysplasia [28]. Furthermore, it may be possible that the interaction of sex and alignment may be different for medial and lateral patellofemoral compartments, although recent studies have not reported it [23]. In this study, we did not evaluate medial and lateral patellofemoral compartments separately. Hence, future work needs to consider trochlear dysplasia and other factors that may be related to higher patellofemoral T in women and differences in medial and lateral patellofemoral compartments.

We did not find a significant difference between sexes in articular cartilage composition in the young and middle-aged groups. A vast majority of earlier studies on sex differences in knee cartilage have focused on cartilage thickness and volume and have reported lower cartilage volume and thickness in women compared with men in healthy young and older individuals without knee OA [11, 16, 18, 19, 41]. However, it is not known if thicker cartilage by itself is protective against knee OA [17]. One study specifically studied sex differences in articular cartilage T2 in young healthy adults (20–29 years) and did not observe any significant difference [40]. Our results are in agreement with theirs because we did not observe significant differences in cartilage composition between men and women in the young healthy and middle-aged non-OA groups. Further studies are needed on sex differences in articular cartilage composition from larger longitudinal cohorts like the Osteoarthritis Initiative.

For the meniscus, we did not observe statistically significant differences between men and women in any group except for higher lateral meniscus T was higher in women in the young healthy group. There are two published reports of sex differences in meniscus composition from the one research group [10, 51]. They reported higher T2 for the posterior horn of the medial meniscus in women compared with men irrespective of the age. In our cohort, the difference for medial posterior horn was also not significant for any group (results not shown). The authors in the previous studies speculated that the difference in meniscus composition was related to walking mechanics. However, we observed less medial loading in women compared with men as would be expected with the greater valgus alignment. It may be possible that women have greater loading over the lateral compartment during walking, although it is not possible to confirm this from the data in this study.

We observed differences in gait biomechanics characterized by lower second peak knee adduction moment and lower varus during walking in women compared with men. Another study reported lower knee adduction moment in women compared with men with knee OA [47], similar to our findings. These data suggest that women may experience lower loading over the medial compartment than men. Further evidence comes from the well-established observation of greater valgus alignment in women [54] as was also seen in our cohorts. However, the differences in the first peak knee adduction moment were not significant. A reduction in the second peak knee adduction moment has been associated with a reduced risk of OA progression, and there have been studies on gait retraining interventions to reduce the second peak knee adduction moment [7, 27]. We also did not observe significant differences in medial cartilage or meniscus T in any of the groups. Hence, further research is needed to assess if a lower second peak knee adduction moment but not first peak knee adduction moment leads to lower medial loading in women compared with men.

The differences in adduction moment and peak varus during walking were not significant after adjusting for the radiographic alignment suggesting that lower extremity alignment may be a key factor related to differences in gait mechanics between men and women. It is not possible to directly estimate lateral compartment loading from these data. However, it may be possible that observed differences in lateral T may be related to mechanical factors. Previous studies in gait mechanics that explored sex differences report higher knee flexion moment in preswing in healthy young women compared with men [30]. In this earlier study, women were observed to have higher knee flexion moment in early stance as well but the difference was not significant. We did observe a greater peak flexion moment in women in early stance compared with men. The difference could be the result of populations or the difference in joint moment normalization (body mass and height versus body weight and height). Longitudinal studies are needed to investigate the effect of differences in gait mechanics between sexes on the risk of OA incidence and progression.

In conclusion, we observed that women with radiographic knee OA had higher lateral and patellofemoral compartment MR relaxation times compared with men indicating deterioration in cartilage composition. We also observed significantly higher varus and lower second peak adduction moment in women compared with men. These data demonstrate differences in cartilage composition and gait mechanics between men and women in young healthy, middle-aged healthy, and OA cohorts. Future longitudinal studies are needed to investigate if these differences in cartilage composition and walking mechanics are associated with a greater risk of lateral tibiofemoral or patellofemoral OA in women. Future studies should also investigate the relative risk of lateral versus medial patellofemoral cartilage degeneration risk in women compared with men.

Acknowledgments

We thank Anthony Luke for providing equipment and facilities for motion capture and Joseph Schooler, Nishant Neel, and Kelly Bauer for data processing.

Copyright information

© The Association of Bone and Joint Surgeons® 2015

Authors and Affiliations

  • Deepak Kumar
    • 1
    • 2
  • Richard B. Souza
    • 1
    • 3
  • Karupppasamy Subburaj
    • 1
    • 4
  • Toran D. MacLeod
    • 1
    • 5
  • Justin Singh
    • 1
  • Nathaniel E. Calixto
    • 1
  • Lorenzo Nardo
    • 1
  • Thomas M. Link
    • 1
  • Xiaojuan Li
    • 1
  • Nancy E. Lane
    • 6
  • Sharmila Majumdar
    • 1
  1. 1.Musculoskeletal Quantitative Imaging Research Group, Department of Radiology and Biomedical ImagingUniversity of California-San FranciscoSan FranciscoUSA
  2. 2.Division of Physical Therapy, Department of Health ProfessionsMedical University of South CarolinaCharlestonUSA
  3. 3.Department of Physical Therapy and Rehabilitation ScienceUniversity of California-San FranciscoSan FranciscoUSA
  4. 4.Engineering Product DevelopmentSingapore University of Technology and DesignSingaporeSingapore
  5. 5.Department of Physical TherapyCalifornia State UniversitySacramentoUSA
  6. 6.Center for Musculoskeletal HealthUniversity of California at Davis Medical CenterSacramentoUSA

Personalised recommendations