Sex Differences in the Correlation Between Restricted Ankle Dorsiflexion and Knee Joint Biomechanics During Gait-Focus on the Knee Adduction Moment-Susumu the Correlation Between Restricted Ankle Dorsiflexion and Knee Joint Biomechanics During Gait -Focus on the Knee Adduction Moment-.

Potential differences between sexes in knee joint biomechanics due to restricted ankle dorsiflexion (DF) may influence the increased prevalence of osteoarthritis (OA) in women. The purpose of this study is to investigate sex differences in the knee adduction moment (KAM) during gait under unrestricted and restricted ankle DF conditions. Forty-nine healthy volunteers (24 men: 21.7 ± 2.5 years old; 25 women: 21.3 ± 1.2 years old) were fitted with a custom-made ankle brace to restrict ankle DF. The ankle DF angle, knee extension angle, and KAM (1 st and 2 nd halves of the stance phase) were measured during the stance phase while walking. The decrease in the ankle DF angle correlated with an increase in the knee extension angle during the 1 st half of the stance phase in both sexes (r=0.52 and 0.61, both p<0.01, respectively). The ankle DF decrease correlated with an increase in the 1 st and 2 nd KAM only in women (r=-0.50 and p<0.05; r=-0.53 and p<0.01, respectively).Multiple regression analysis revealed that decreases in the ankle DF and knee flexion angle accounted for 24% of the increase in the 2 nd KAM in women. Restricted ankle DF was related to an increase in the 1 st and 2 nd peak KAM in women, but not in men. In particular, the increase in the 2 nd KAM in women related to a greater knee extension angle (hyperextension) and a smaller ankle DF angle. The present findings support the use of range of motion exercises to increase ankle DF to reduce KAM and prevent knee OA, especially in women.


Introduction
Knee osteoarthritis (OA) is a highly prevalent joint disease among middle-aged and elderly people [1,2]. The prevalence of knee OA is higher in women than in men [1]. Mechanical stress (medial compartment contact force) to the knee joint relates to this prevalence [3,4], and the knee adduction moment (KAM) during gait is a surrogate value of the medial compression force [5,6]. Based on longitudinal studies [7][8][9], higher KAM is a risk factor for medial knee OA. Several studies report that decreasing the KAM during the gait is an effective strategy for preventing knee OA [10][11][12][13].
On the other hand, a previous biomechanical gait study showed that restricting the ankle dorsiflexion (DF) angle to ~8° during the terminal stance increases the KAM and the knee extension angle (knee genu recurvatum) in healthy adults [14], and suggested that improving and/or maintaining the ankle DF angle could prevent medial knee OA. Although some studies report that women have greater knee joint laxity (hyperextension) compared with men, potential sex differences in the biomechanical effects due to a restricted ankle DF were not investigated.
The ankle DF angle is often restricted as a result of an ankle sprain [15], ankle fracture [16], or Achilles tendon injuries [17]. In addition, aging may reduce ankle DF, especially in women [18,19].
Although sex differences may affect the gait mechanism when the ankle DF is restricted, no studies have yet investigated the potential sex-related biomechanical effects of reduced ankle DF during gait. The present study investigated whether the correlation between restricted ankle DF and knee joint biomechanics during the gait differs between the sexes, focusing on the KAM. We hypothesized that a reduced ankle DF would correlate with an increase in the KAM and a decreased knee flexion angle (increased knee extension) during the terminal stance phase, which usually requires ankle DF, to a greater extent in women than in men.

Participants
Forty-nine healthy volunteers (24 men and 25 women) participated in the study. Individuals from Nagoya University responded to flyers posted on campus and verbal requests for subjects. Subjects were excluded from participation if they had a current lower extremity injury, knee OA, history of ankle injury, or any medical condition that would affect their ability to perform the study tasks. The Ethics Committee of Nagoya University approved the study, and, after the nature of the study was explained to the participants, all subjects provided written informed consent to participate in the study.

Brace to Restrict Ankle DF
A custom-made ankle brace for the right ankle (Matsumoto P & O Co., Ltd., Komaki, Japan) was used to restrict ankle DF using a dial lock joint ( Figure 1) while allowing for unrestricted plantar flexion.
The brace was made of Torelac clear material and the lock was adjustable to restrict motion at various degrees of ankle DF. The restricted condition was set to 0° of DF or natural alignment, based on a previous study [14] demonstrating a significantly increased KAM under this condition compared with the unrestricted condition.
Ankle DF was examined in the unrestricted condition and restricted condition (restricted to 0°) in a non-weight-bearing state. Subjects wore the ankle brace during the unrestricted condition, but the sagittal plane ankle range of motion was not restricted.

Instrumentation
Three-dimensional kinematic data were obtained using a 7-camera motion analysis system (Vicon Nexus; Vicon Motion System, Oxford, UK). Kinematic data were sampled at 120 Hz and digitally recorded. Ground reaction forces were collected at 120 Hz using custom-made two-piece (left-right) force plates (Kistler Z11942; 2×0.8 m, Z12091; 2×0.4 m; Kistler, Amhurst, NY) embedded in the floor. The motion capture and force plates were temporally synchronized.

Procedures
Data were collected at the gait analysis laboratory at Rosai Rehabilitation Engineering Center. Thirty-five reflective markers  Note: The brace limits ankle dorsiflexion using a dial lock joint to allow for plantar flexion.
Each subject was evaluated in two brace conditions (unrestricted and restricted) in random order. Participants performed the tasks barefoot with only the right foot secured in the brace. Cadence was controlled by a metronome set at the standard pace for Japanese men and women according to a previous report (113 steps/minute and 115 steps/minute, respectively) [23]. Subjects were asked to walk along a 10-m walkway, and three trials under each brace condition were performed. Practice trials allowed subjects to become familiar with the brace.

Data Analysis
The Vicon Plug-in Gait model was used to quantify the ankle sagittal plane angle, and knee sagittal and frontal plane angles.
Kinematic data were filtered using the Woltring filter with a predicted mean squared error value of 20. The Plug-in Gait model was used to define rigid body segments and joint angles between these segments. The hip joint center was defined using the Davis method [24]. The knee and ankle joint centers were defined as the midpoint between the lateral and medial femoral epicondyle and medial and lateral malleolus, respectively. Euler angles were The right foot data of the stance phase were used in the final analysis. All data were normalized to 100% of the stance phase. The biomechanical values used were the peaks of the 1 st half (0%-50%) and 2 nd half of the stance phases (50%-100%). The peak ankle DF angles, peak knee flexion angle during the 1 st half, minimum peak of the knee flexion angle during the 2 nd half, peak knee adduction angles, peak knee flexion moment during the 1st half, negative peak of the knee flexion moment, and peak knee adduction moments were measured for the braced (right) limb. For each subject, dependent variables of interest were identified from each trial and averaged over three trials.
Intratester reliability of the kinematics and kinetics were assessed using the intraclass correlation coefficient in the same measurement setting as described in a previous study [14]. The KAM, knee adduction angle, knee flexion angle, and ankle DF angle had an intraclass correlation coefficient ranging from 0.91 to 0.99, and that of the knee flexion moment was 0.59.

Statistical Analyses
Prior to the analysis, normal distributions were confirmed by Shapiro-Wilk analysis. Differences between the biomechanical data under the unrestricted and restricted conditions were assessed using a paired t-test (normal distribution) or Wilcoxon signedrank test (non-normal distribution). Differences in values between men and women were assessed using an unpaired t-test (normal distribution) or Mann-Whitney U test (non-normal distribution).  (Table 1). Note: ES, effect size; * Non-normal distribution; ** Mann-Whitney U test Table 2 shows the correlations among the changes in the ankle In the 2 nd half of the stance phase, however, the correlation was  Unrestricted and restricted ankle range of motion differences in the correlation coefficients of biomechanical values during the 1 st (0%-50%) and 2 nd halves (50%-100%) of the stance phase.

Discussion
The main findings of the present study focusing on the KAM were that restricted ankle DF during the 1 st and 2 nd halves of the stance phase significantly correlated with an increase in the KAM only in women. In a healthy population, KAM is reported to have two peaks: one at mid-stance and the other at the terminal stance phase [25,26]. with the maximum ankle DF required through the gait cycle [27].
Therefore, the effect of reduced ankle DF on the KAM is thought to be larger during terminal stance (the 2 nd half of the stance phase).
In women, the increased KAM in the 2 nd half of the stance phase related not only to decreased ankle DF, but also to a decreased knee flexion angle (increased knee extension angle) ( Table 2).
Because we did not investigate how restricting ankle DF led to  Table 1) indicating hyperextension of the knee. Joint laxity in women [20][21][22] which predisposes women to knee hyperextension, might be the influencing factor of the finding of the increased KAM.
Additionally, hyperextension can lead to low proprioception of the knee joint [28], which is assumed to result in poor control of knee joint function and biomechanics.
From these findings, restricting the ankle DF angle in women could affect the knee joint angle, leading to a higher KAM compared with that in men. Previous studies reported that an increase in the KAM is related to progression of medial knee OA [7][8][9], and the prevalence rate of knee OA is higher in women than in men [1]. Women are particularly predisposed to age-related ankle DF restriction [18]. The present finding might support the use of range of motion exercises to increase ankle DF to reduce the KAM and prevent knee OA, especially in women.
The present study revealed a relationship between decreased (i.e., 2.0° of knee extension). We assume that the lack of a significant correlation was due to a ceiling effect for knee hyperextension.
Restricted ankle DF due to ankle and subtalar arthrodesis [31] and total ankle replacement [32] result in biomechanical changes to the gait, with the peak ankle DF angle during gait changing from 4.9° and 7°, respectively, to 6.4° and 6° after the operations.
The postoperative changes in the ankle DF angle were similar to restriction of the ankle DF angle in the present study. The total ankle replacement decreased the ankle pain, thereby improving walking speed and cadence [32]. The amount of ankle DF obtained for the terminal stance postoperatively was not adequate. The impact of the postoperative decrease in ankle DF and its influence on the knee joint, especially in women, should be considered. The studies did not evaluate other components of the knee joint biomechanics, such as KAM.
The present study has several limitations. The brace was one size and only fitted to the right ankle. The amount of reduction in the ankle DF angle could not be controlled for each subject. Therefore, we evaluated the correlation between a reduced ankle DF and other biomechanical values. We could not determine the amount of the decrease in the ankle DF angle that results in increased KAM for the 1 st and 2 nd halves of the stance phase. In addition, because the ankle DF restriction was immediate and forced, we could not assess the long-term effects of restricted ankle DF. Finally, a third limitation is that the subjects were all healthy and young, and the effects of restricted ankle DF should also be evaluated in middleaged subjects with a greater prevalence of knee OA.

Conclusion
Restricted ankle DF induced using a custom-made brace correlated with the increase in the KAM during the 1 st and 2 nd halves of the stance phase in women, but not in men. The increase in the 2 nd KAM in women significantly correlated with a decrease in the ankle DF angle and an increase in the knee extension angle.
Compared with men, women are more predisposed to an increased KAM due to decreased ankle DF.