2.1 Study design and participants

We performed a pilot cross-sectional study at a single facility, including 21 participants (mean age: 75.6 years; age range: 64–85 years), between October 2022 and September 2023. All participants, diagnosed with unilateral or bilateral KOA by an orthopedic surgeon, were attending an orthopedic clinic, receiving outpatient rehabilitation, and able to walk independently. Exclusion criteria included passive knee joint flexion ROM below <67°(the minimum knee flexion required for normal-level walking),⁹ presence of other conditions affecting gait (e.g., post-traumatic disorders and rheumatoid arthritis), history of intra-articular corticosteroid injections within one month of measurements, and cognitive or higher brain dysfunction affecting understanding of measurement content or performance of tasks.

For eligible participants, we measured height, weight, and body mass index (BMI). We also collected data on the duration of outpatient rehabilitation, use of other treatments (e.g., analgesics, nonsteroidal anti-inflammatory drugs, and hyaluronic acid injections) within one month of measurements, Kellgren–Lawrence (KL) grade, and femorotibial angle (FTA) from the medical records. An orthopedic surgeon assessed posterior osteophytes from lateral knee radiographs to determine their impact on knee joint ROM.

2.2 Assessment of SKG

Gait was recorded using a high-speed camera (ZV-1, Sony Corporation, Tokyo, Japan). Recordings were performed from the sagittal plane for five trials where participants walked along a 5-m walkway at a comfortable speed. The camera, positioned 5 m from the participants with a lens height of 1 m above the floor, recorded video at 120 fps with a shutter speed of 1/500 s. Based on the video data, the pose estimation library [OpenPose version 1.7.0 (GPU release)] was used to estimate joint feature points (hip, knee, and ankle) during gait. OpenPose, a library that uses deep-learning to estimate human joint positions from image and video data,¹⁰ and its accuracy in motion analysis has been validated against three-dimensional motion analysis systems.¹¹

Knee joint angles during gait were calculated as the angles formed by lines connecting the feature points of hip and knee joints and the feature points of knee and ankle joints. Joint angle data were preprocessed using a 6-Hz low-pass filter.¹¹ A foot switch (DC-F01, Delsys Inc., Natick, MA, USA) was used to determine the gait cycle. It was fixed with tape to the center of the participant's heel, with one gait cycle defined as the period from heel contact to next heel contact on the same side. Knee flexion excursion was calculated by subtracting the maximum knee extension angle from the maximum knee flexion angle within a gait cycle. For each trial, we analyzed two gait cycles recorded at the center of the walkway, averaging data from 5 trials (10 gait cycles). Referring to previous studies, participants with knee flexion excursions of <50° and ≥50° were classified into the SKG+ and SKG− groups, respectively.¹²

2.3 Evaluation items

2.4 Statistical analysis

An unpaired t-test was used to compare evaluation item results between the SKG+ and SKG− groups for normally distributed data, whereas the Mann–Whitney U test was used for non-normally distributed data. Normality was assessed using the Shapiro–Wilk test. Fisher's exact probability test was used to compare KL grades, and the chi-square test was employed to compare gender ratios, the duration and treatment methods of outpatient rehabilitation, and the presence or absence of posterior osteophytes in the knee joint. For evaluation items showing significant differences between groups, correlation analysis was performed to examine their relationships with knee flexion excursion. Pearson's product-moment correlation coefficients were used for normally distributed data, whereas Spearman's rank correlation coefficients were employed for non-normally distributed data. R 4.3.1 (The R Foundation for Statistical Computing, Vienna, Austria) was used to conduct statistical analysis, with significance set at 5 %.

3.1 Comparison of participants’ basic information, KOA severity, and treatment status

Among the 21 participants, 9 and 12 were in the SKG+ and SKG− groups, respectively. Table 1 displays the participants’ basic information and KOA severity. No significant differences were found between the groups in age, gender ratio, height, weight, and BMI. KL grade, indicating KOA severity, and FTA angle, representing lower limb alignment, showed no significant differences between groups. Posterior osteophytes were present in some participants (1 in the SKG + group and 3 in the SKG− group), with no significant difference observed between groups (p = 0.81). Outpatient rehabilitation duration did not differ significantly between the groups [SKG + group: 4.0 months (2.0, 16.0); SKG− group: 7.0 months (3.75, 9.75); p = 0.89]. Analgesic use (SKG + group: 22.2 %; SKG− group: 16.7 %; p = 1.00), anti-inflammatory analgesic use (SKG + group: 44.4 %; SKG− group: 41.7 %; p = 1.00), and hyaluronic acid injections (SKG + group: 22.2 %; SKG− group: 25.0 %; p = 1.00) did not differ between the groups.

3.2 Comparisons of evaluation items

Table 2 presents the results of each evaluation item. Pain intensity was significantly higher in the SKG + group compared with the SKG− group (p = 0.042), whereas quadriceps muscle strength and passive knee flexion ROM were significantly lower in the SKG + group (p = 0.007 and p = 0.022, respectively). No significant differences were found in passive knee extension ROM, pain duration, gait speed, and FES-I scores. WOMAC scores indicated significantly lower function scores and total scores in the SKG + group compared with the SKG− group (p = 0.017 and p = 0.036, respectively), with no significant difference in pain scores.

3.3 Correlation between knee flexion excursion and evaluation items

Results of correlation analysis for evaluation items with significant differences between the SKG+ and SKG− groups are shown in Fig. 1. Pain intensity showed a moderate negative correlation with knee flexion excursion (r = −0.59, p = 0.01; Fig. 1a). Conversely, quadriceps muscle strength displayed a moderate positive correlation with knee flexion excursion (r = 0.57, p = 0.01; Fig. 1b), whereas knee flexion ROM exhibited a strong positive correlation with knee flexion excursion (r = 0.74, p = 0.01; Fig. 1c).

Close

Fig. 1 Correlations between knee flexion excursion and (a) pain intensity, (b) quadriceps muscle strength, and (c) knee flexion ROM. (a–c) Data points are differentiated as follows: open circles (○) and filled circles (●) represent SKG+ and SKG−, respectively. Trend lines indicate correlations between knee flexion excursion and (a) pain intensity, (b) quadriceps muscle strength, and (c) knee flexion ROM.

Export to PPT

4.1 Pain and SKG

This study revealed that pain intensity was significantly higher in the SKG + group. Pain, a common symptom in KOA, results from osteophyte formation, subchondral bone loading, and inflammation or damage to surrounding tissues, including the periosteum, ligaments, muscles, synovium, and joint capsule.¹⁶ Some patients with KOA alter their gait to avoid pain.⁸ Those experiencing severe pain may unconsciously adopt a gait pattern with reduced knee flexion to minimize discomfort. Conversely, SKG may increase mechanical stress within the joint due to reduced movement and a smaller load-bearing area, potentially exacerbating KOA symptoms and joint deformity progression.² This study was cross-sectional; thus, it could not determine whether SKG causes or results from pain. However, pain and SKG may interact bidirectionally, creating a “vicious cycle,” highlighting the need for interventions to break this cycle.

Pain was evaluated using both WOMAC and VAS pain scores. A significant difference was found in the VAS pain scores between groups but not in the WOMAC pain scores. This discrepancy may be attributed to the VAS focusing solely on pain during gait, whereas WOMAC pain scores incorporate pain under static conditions, such as lying down or standing. Capturing gait-related pain is most crucial for assessing SKG risk.

4.2 Quadriceps muscle strength and SKG

Quadriceps muscle strength was significantly weaker in the SKG + group, consistent with a previous study linking reduced knee flexion angle during gait with muscle weakness.⁷ In patients with KOA, quadriceps muscle weakness may be attributed to degenerative changes in joint proprioceptors and pain-induced neural inhibition.¹⁷ This weakness can exacerbate KOA-induced joint instability,¹⁸ prompting patients to cocontract their quadriceps and hamstrings to compensate.¹⁹ Although this cocontraction enhances dynamic knee joint stability, it may contribute to SKG.²⁰ The present study does not establish a causal relationship between muscle weakness and SKG; nevertheless, increased cocontraction accompanying quadriceps weakness may affect SKG. Additionally, some patients with KOA may reduce quadriceps activity to avoid pain during gait,⁸ suggesting that SKG adopted to avoid pain could lead to secondary quadriceps muscle weakness. Further research is needed to explore the causal relationship between quadriceps muscle weakness and SKG.

4.3 ROM and SKG

The significantly reduced knee flexion ROM observed in the SKG + group indicates a relationship between limited ROM and SKG. In KOA, knee flexion ROM is often affected by structural changes, such as posterior osteophyte formation²¹ and joint space narrowing.²² Severe limitations in knee flexion ROM may cause SKG. However, no significant differences in knee joint deformity severity were found between the groups. Although a small number of participants had posterior osteophytes, all participants exhibited ROM sufficient for normal gait (<67°).⁹ Therefore, the reduced knee flexion excursion in the SKG + group may not primarily arise from structural changes but could be related to soft tissue characteristics, including quadriceps tendon stiffness, which affects knee flexion ROM and excursion during gait.²³ Additionally, SKG could cause secondary ROM limitations owing to repeated use of a limited knee flexion gait pattern, leading to soft tissue shortening around the knee joint. Thus, further research is required to clarify the causal relationship between ROM limitations and SKG.

4.4 Health-related QOL and SKG

This study used the Japanese version of WOMAC to assess health-related QOL, revealing significant differences in both function scores and total scores between the SKG+ and SKG− groups. This suggests that SKG affects daily living activities and overall health-related QOL. Although this study focused on level walking, daily activities involve more complex movements potentially impacted by SKG. Future research should investigate how SKG influences these complex movements beyond level walking.

4.5 KOA severity and SKG

We assessed KOA severity using the KL grade, based on radiographic findings, and observed no significant differences between the SKG+ and SKG− groups. Previous studies assessing KOA severity and knee flexion angle during gait^5,6 used different approaches, basing assessment on surgical indications beyond structural changes visible in imaging, i.e., not only joint deformity severity but also pain intensity and daily activity limitations. No previous research has specifically explored the link between SKG and KOA severity using only radiographic findings, such as severity quantified via KL grade. Our results provide preliminary evidence suggesting no direct correlation between deformity severity and SKG, although further validation is necessary given the limited sample size.

4.6 Limitations

This study has some limitations. As a pilot study, the relatively small sample size may limit the generalizability of our findings. Although our results offer valuable preliminary evidence regarding SKG-associated factors in patients with KOA, larger-scale studies with more participants are needed to confirm and extend these findings. Additionally, although questionnaire-based subjective assessments of pain duration are commonly used, variability in individual perceptions should be considered when interpreting these findings. Although we thoroughly investigated factors related to SKG using clinical evaluation methods, such as pain, ROM, and muscle strength, our assessment of muscle strength focused solely on the quadriceps muscle, which is considered crucial for SKG. Nevertheless, other muscles, such as the hamstrings (antagonists of the quadriceps), may also influence SKG. Future studies should assess muscle strength more comprehensively, including other lower limb muscles that affect gait patterns and knee kinematics. Additionally, previous studies have shown that insufficient push-off due to ankle plantarflexion during terminal stance is associated with SKG.²⁴ Knee flexion during the swing phase is also related to hip joint flexion.²⁵ This suggests that adjacent joints, such as the hip and ankle joints, may influence SKG. Therefore, future research should evaluate the angles of adjacent joints during gait for a more comprehensive analysis.