By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
To assess immediate effects of laterally wedged insoles on walking pain, external knee adduction moment, and static alignment, and whether these immediate effects together with age, body mass index, and disease severity predict clinical outcome after 3 months of wearing insoles in medial knee osteoarthritis.
Forty volunteers (mean age 64.7 years, 16 men) were tested in random order with and without a pair of 5° full-length lateral wedges. Immediate changes in static alignment were measured via radiographic mechanical axis and changes in adduction moment via 3-dimensional gait analysis. After 3 months of treatment with insoles, changes in pain and physical functioning were assessed via the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and patient-perceived global change scores.
Reductions in the adduction moment occurred with insoles (first peak mean [95% confidence intervals (95% CI)] −0.22 [−0.28, −0.15] Nm/body weight × height %), accompanied by a reduction in walking pain of ∼24% (mean [95% CI] −1.0 [−4.0, 2.0]). Insoles had no mean effect on static alignment. Mean improvement in WOMAC pain (P = 0.004) and physical functioning (mean [95% CI] −6 [−11, −1]) was observed at 3 months, with 25 (69%) and 26 (72%) of 36 individuals reporting global improvement in pain and functioning, respectively. Regression analyses demonstrated that disease severity, baseline functioning, and magnitude of immediate change in walking pain and the first peak adduction moment with insoles were predictive of clinical outcome at 3 months.
Lateral wedges immediately reduced knee adduction moment and walking pain but had no effect on static alignment. Although some parameters predicted clinical outcome, these explained only one-third of the variance, suggesting that other unknown factors are also important.
Knee osteoarthritis (OA) is a common disease, affecting ∼30% of individuals over age 65 years (1). It is associated with pain and physical disability (2) and imposes a significant personal, societal, and economic burden (3). Contemporary management aims to reduce pain and optimize physical function while minimizing adverse side effects of therapy (4, 5). Because expensive surgical interventions are reserved for end-stage disease, conservative treatments are desirable. Although efficacious, drug therapies are associated with adverse effects (6–8). Accordingly, nonpharmacologic measures are considered the cornerstone of OA management. Laterally wedged shoe insoles are one such strategy recommended by rheumatology health professionals for managing knee OA (4, 5).
Osteoarthritic changes are common in the medial tibiofemoral compartment of the knee (9–11). Progressive loss of cartilage and joint space in this compartment typically results in varus malalignment, which causes the ground reaction force vector to pass more medially to the knee joint center during gait. This results in increased loads across the medial compartment, as indicated by the external knee adduction moment measured during gait analysis (12). Cross-sectional studies have demonstrated that patients with knee OA have a higher knee adduction moment when compared with healthy age-matched controls (13, 14). One study has shown that a higher adduction moment is associated with more severe knee pain (15). Furthermore, a relationship between malalignment and symptom severity (knee pain and physical function) has also been demonstrated (16). Sasaki and Yasuda (17, 18) were the first to propose laterally wedged insoles for medial knee OA. They demonstrated that the insole shifted the calcaneus into a valgus position relative to the tibia, thereby leading to a more vertically aligned lower limb. They concluded that this helped reduce excessive loading of the medial knee, leading to mitigation of knee pain.
Biomechanical studies have since evaluated the effects of laterally wedged insoles on static alignment and medial compartment loading. Most agree that insoles immediately reduce the peak adduction moment by 5–10% during walking (19–21); however, the effect of insoles on static alignment is less clear (18, 22, 23). Interpretation of findings from studies to date is limited by small sample sizes, methodologic issues including use of imprecise measurement techniques, and failure to evaluate clinical outcome concurrently. Furthermore, most studies report the effect of insoles on the peak knee adduction moment only. Typically, 2 peaks occur throughout the stance phase of gait, yet few studies report the effect of insoles on the second peak in knee OA. Furthermore, as one study has demonstrated that the adduction moment at midstance has a greater correlation with static alignment than does the peak moment (24), wedged insoles may have their greatest biomechanical effect on this parameter. This has not been evaluated.
Although case series report decreased pain in 53–82% of patients with knee OA with laterally wedged insoles (25–27), a review (28) concluded that there is only limited evidence from 3 randomized controlled trials attesting to the efficacy of wedged insoles. Given disease heterogeneity, it is possible that insoles may have differential therapeutic effects in patient subgroups. For example, it may be that a good clinical outcome over time is only possible in patients who immediately demonstrate improved knee alignment with insole use. Little is known about which factors may predict clinical outcome with laterally wedged insoles. Although some authors have demonstrated a relationship between clinical outcome with insoles and patient age (29), anthropometry (29, 30), and disease severity (17, 25, 26, 30), findings are inconsistent across studies. Furthermore, no study has investigated whether immediate clinical and biomechanical effects of the insoles predict longer-term clinical outcome.
The primary aim of the present study was to evaluate immediate effects of laterally wedged insoles on walking pain, knee adduction moment, and static alignment in knee OA. Exploratory analyses were also performed to evaluate whether age, baseline pain and function, disease severity, and immediate changes in pain and biomechanical parameters with insoles could predict clinical outcome following 3 months of treatment.
PATIENTS AND METHODS
Forty community volunteers age >50 years with knee OA were recruited into the study by advertisements. Diagnosis was based on clinical and radiographic criteria (31). Participants reported knee pain on most days and demonstrated medial tibiofemoral joint osteophytes on radiograph. Other inclusion criteria were average pain score >3 on an 11-point Likert scale and knee pain when walking 2 blocks and/or climbing stairs. Exclusion criteria included a walking aid, poor English, body mass index ≥36 kg/m2, hip or lumbar spine arthritis, hip or knee replacement, knee surgery or injection (past 6 months), use of insoles or orthotics (past 6 months), foot or ankle problem precluding insole use, and usual footwear incompatible with insoles.
The University of Melbourne Human Research Ethics Committee approved the study and all participants provided written informed consent.
Participants underwent a baseline assessment of pain and function. Following this assessment, a biomechanical analysis was conducted both with and without insoles applied in random order. Block randomization (randomly alternating blocks of 4 and 6) was employed using a computer-generated table of random numbers. All participants were then provided with insoles and instructed to wear them inside their footwear for 3 months. Following the intervention, participants underwent reassessment of pain and function.
Bilateral, standardized, laterally wedged (∼5°) insoles made of high-density ethyl-vinyl acetate were evaluated. Insoles were wedged along the lateral edge of the full length of the foot, and were trimmed to fit the shoes. A 5° wedge was selected because greater wedging is associated with discomfort (19). Participants started wearing the insoles for 1 hour, thereafter increasing use by 1 hour per day until wearing them full time. Compliance was assessed by means of a logbook in which participants recorded the number of hours per day that the insoles were worn.
Knee adduction moment.
A Vicon motion analysis system with six M2 CMOS cameras (1,280 × 1,024) operating at 120 Hz (Vicon, Oxford, UK) was utilized. The standard Plug-In-Gait marker set (Vicon) was used (anterior superior iliac spine, posterior superior iliac spine, midlateral thigh, lateral knee joint, lateral shank, lateral malleolus, on the shoe over the second metatarsal head, and over the posterior calcaneus). Additional medial knee and ankle markers were used during the static standing trial to determine tibial torsion. Individual markers remained in situ throughout all test conditions and were not moved. Ground reaction forces were measured by two 0R6-6-2000 force plates at 1,080 Hz (Advanced Mechanical Technology, Watertown, MA), in synchrony with the cameras.
Participants wore their usual footwear with and without insoles inserted in random order. Average pain during walking for each condition was measured by an 11-point Likert scale numbered from 0 to 10 (higher scores indicating worse pain). Participants walked at a comfortable pace and data were collected from 5 trials for each condition. Walking speed was monitored by 2 photoelectric beams and verbal feedback was provided to ensure that speed during the second condition varied not more than 10% from the average speed of the first.
Joint moments were calculated via inverse dynamics (Vicon Plug-In-Gait version 1.9). The knee adduction moment was normalized for body weight and height (Nm/body weight × height %). Variables of interest were peak adduction moment in the first half of the stance, peak adduction moment in the second half of the stance, and the moment at 50% of stance phase (midstance). Data were averaged over 5 trials for each condition. Test–retest reliability in our laboratory was excellent (intraclass correlation coefficients [ICC(3,5)] 0.92–0.97 in 11 elderly patients with knee pain tested 1 week apart).
Full-leg, anteroposterior, weight-bearing radiographs were obtained as previously reported (32). Radiographs were taken with participants barefoot and standing on a pair of wedged insoles (in the same random order used for gait analysis), with foot maps to control foot position across conditions. Participants stood with the knee in full extension, and were positioned with the tibial tuberosity facing the X-ray beam (16). The X-ray tube was positioned at a distance of 2.44 meters from the cassette, and depending on individual limb characteristics, settings of approximately 25 mA/second and 100 kV were applied.
Alignment was defined as the angle of intersection of the femoral and tibial mechanical axes according to the method by Sharma et al (16). To determine the mechanical axis of the femur, a line was drawn from the center of the femoral head to the center of the femoral intercondylar notch. The femoral head center was located using a series of concentric circular hip templates. A second line from the tibial spine center to the ankle talus center established the tibial mechanical axis. Varus malalignment was indicated by values <180° and valgus malalignment by values >180°. The mechanical axis was determined by 1 investigator (RSH; ICC 0.98) (32) blinded to the test condition.
Pain and physical function.
The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) was used to assess pain (score range 0–20, higher scores indicated worse pain) and physical function (score range 0–68, higher scores indicated worse function) (33). Patient-perceived global change in 1) pain (since commencing the intervention) and 2) physical function was recorded on 5-point Likert scales with possible scores of 1 (much worse), 2 (slightly worse), 3 (no change), 4 (slightly better), and 5 (much better).
Radiographic severity of tibiofemoral OA was assessed with the Kellgren/Lawrence system (where 0 = normal; 1 = possible osteophytes; 2 = minimal osteophytes and possible joint space narrowing; 3 = moderate osteophytes, some narrowing, and possible sclerosis; and 4 = large osteophytes, definite narrowing, and severe sclerosis) (34).
Statistical analyses were performed using SPSS (Norusis/SPSS, Chicago, IL). For the primary analyses, immediate effects of insoles on pain and biomechanical parameters were determined by analysis of mean change scores and their associated 95% confidence intervals (95% CIs). WOMAC scores at baseline and 3 months were evaluated similarly. For the exploratory analyses evaluating predictors of outcome at 3 months, univariate correlations and multiple regression analyses were performed. For regression, WOMAC pain and function scores at 3 months were dependent variables. Initially, participant age, disease severity, and baseline WOMAC score were forced into the model. Immediate changes in 1) the first peak adduction moment, 2) static alignment, and 3) walking pain with insoles were then forced into the model to determine if additional variance in outcome could be explained by these factors.
Participant characteristics are presented in Table 1. The cohort comprised more women than men. Participants were generally overweight (35). A range of radiographic severity was evident, with the majority of participants (68%) demonstrating moderate-severe changes on radiograph (grade 3 or 4). Participants demonstrated a mechanical axis of 174.5° on average (e.g., 5.5° of varus) (Table 2); however, alignment ranged from 186.0° to 164.5°. Participants reported moderate levels of pain, stiffness, and physical limitation.
Table 1. Presenting characteristics of the cohort*
Values are the mean ± SD unless otherwise indicated. WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.
Higher scores indicate worse symptoms.
Radiographic severity according to the Kellgren/Lawrence grading system, where higher grades indicate more severe disease.
Table 2. Immediate effect of laterally wedged insoles*
Mean ± SD
Values are the mean ± SD unless otherwise indicated. 95% CI = 95% confidence interval; BW = body weight; HT = height.
External knee adduction moment (Nm/BW × HT %)
4.04 ± 1.05
3.82 ± 1.00
−0.22 ± 0.21
2.89 ± 0.83
2.63 ± 0.84
−0.25 ± 0.22
2.56 ± 0.79
2.43 ± 0.75
−0.15 ± 0.17
Static alignment, degrees
174.5 ± 4.7
174.5 ± 4.7
0.1 ± 0.7
Pain during walking
4.2 ± 2.3
3.2 ± 2.1
−1.0 ± 1.4
Immediate effects of insoles.
Insoles resulted in an immediate reduction in all knee adduction moment parameters (Table 2). Mean reductions were 5–9% of the moment measured without insoles. Walking speed was similar across conditions (mean ± SD 1.13 ± 0.14 meters/second without insoles versus 1.13 ± 0.15 meters/second with insoles; mean difference −0.001 [95% CI −0.016, 0.012]). Figure 1 depicts representative data from a single participant demonstrating changes in the adduction moment with the insoles. However, the effect of the insoles on the knee adduction moment was not consistent across the cohort (Figure 2). While most participants demonstrated a reduction in the first peak moment, the magnitude of the reduction ranged from 0.1% to 18.2%, and 5 individuals actually demonstrated an increase. The range of change scores in Table 2 indicates similar results for other parameters of the adduction moment. Changes in the adduction moment were accompanied by a reduction in walking pain (Table 2); however, the insoles had no effect on static alignment (Table 2).
Outcome at 3 months.
Data from 34 (85%) of 40 participants revealed that insoles were worn for a mean ± SD of 38 ± 25 hours per week. WOMAC pain scores decreased from a mean ± SD of 9 ± 3 at baseline to 7 ± 5 at 3 months (mean change −2 [95% CI −4, −1]), representing a 22% reduction. Similarly, WOMAC function scores declined from a mean ± SD of 30 ± 11 at baseline to 24 ± 17 at 3 months (mean change −6 [95% CI −11, −1]), representing a reduction of 20% from baseline. Regarding global change in pain, 13 (36%) of 36 participants were much better, 13 (36%) were slightly better, 7 (20%) were unchanged, 1 (3%) was slightly worse, and 2 (5%) were much worse. Regarding global change in function, 11 (31%) of 36 participants were much better, 14 (39%) were slightly better, 7 (20%) were unchanged, 2 (5%) were slightly worse, and 2 (5%) were much worse.
Predictors of outcome at 3 months.
Correlations between the variables are reported in Table 3. Few relationships were observed. Age correlated with change in WOMAC pain score, such that older participants reported less improvement. Immediate change in the first peak adduction moment with insoles was associated with change in WOMAC function score, whereby participants who demonstrated a greater reduction in the adduction moment with wedges reported less physical impairment.
Table 3. Correlation coefficients describing the univariate association between predictor variables and outcome variables*
Inclusion of age, disease severity, and baseline WOMAC score accounted for 24% (adjusted r2 = 0.238) of the variance in WOMAC pain score at 3 months. Forced entry of change in 1) the first peak adduction moment, 2) static alignment, and 3) walking pain with insoles explained only a further 5% of the variance (adjusted r2 = 0.292). Only disease severity and change in walking pain were significant predictors (Table 4).
Table 4. Regression coefficients obtained from exploratory regression analyses, with Western Ontario and McMaster Universities Osteoarthritis Index pain and function scores at 3 months as dependent variables*
Pain at 3 months
Function at 3 months
95% CI = 95% confidence interval; KAM = knee adduction moment; BW = body weight; HT = height.
Radiographic severity according to Kellgren/Lawrence grading system, where higher grades indicate more severe disease.
Inclusion of age, disease severity, and baseline WOMAC score accounted for 21% (adjusted r2 = 0.211) of the variance in WOMAC function at 3 months. Forced entry of change in 1) the first peak adduction moment, 2) static alignment, and 3) walking pain with insoles explained a further 13% of the variance (adjusted r2 = 0.347) (Table 4). Baseline WOMAC score, change in walking pain, and change in first peak adduction moment were significant predictors (Table 4).
Our study demonstrated that laterally wedged insoles resulted in an immediate reduction in walking pain and knee adduction moment, but did not change static alignment. Improvements in pain and physical function were reported by the cohort after 3 months of treatment with insoles. Factors predictive of clinical outcome included disease severity, baseline function score, and magnitude of immediate change in walking pain and the first peak adduction moment with insoles, although the majority of variance in outcome remained unexplained.
Our findings concur with those of most other biomechanical studies evaluating immediate effects of insoles on the knee adduction moment in knee OA. In 15 patients with knee OA, Kerrigan et al (19) demonstrated that lateral wedges reduced the first peak by 5.3% and the second peak by 6.5%. Our similar insoles had comparable effects, with mean reductions of 5.4% and 9.0%, respectively. In another small group (n = 13), Kakihana et al (21) demonstrated that lateral wedges attached to the bare feet reduced the mean ± SD stance-phase adduction moment from 0.36 ± 0.02 Nm/kg to 0.34 ± 0.02 Nm/kg, again a reduction of ∼5.6%. Similar to our findings, Kakihana et al reported inconsistent insole effects, with 2 participants demonstrating an increased adduction moment with the insoles. In contrast, Maly et al (23) failed to show any significant effect of a 5° lateral heel wedge on the adduction moment in 9 persons with medial OA. It may be that wedging the entire length of the foot, and not just the heel, is a key insole feature necessary for the knee adduction moment to be reduced.
A longitudinal study has demonstrated that a 1-unit increase in the knee adduction moment results in a 6.46-fold increase in the risk of medial disease progression (15). Given their one-quarter–unit reduction of the adduction moment, laterally wedged insoles have the potential to slow disease progression. The only long-term study in OA evaluating the effect of laterally wedged insoles on disease progression failed to demonstrate any structural effect (36). However, that study tested an insole that wedged the rearfoot only, and did not include any biomechanical gait analysis. Thus it is possible that the insoles utilized did not significantly reduce the knee adduction moment, which may explain their lack of structural effect. It is unknown what effect laterally wedged insoles have on other knee joint compartments, in particular the lateral tibiofemoral and patellofemoral compartments. Although efficacious in reducing the load medially, further research is warranted to evaluate whether they have adverse effects on other compartments that may be concomitantly affected by OA.
We were unable to demonstrate any significant effect of insoles on static alignment. Similar results have been reported by other researchers (18, 23). This result may be because we constrained foot position across conditions, thus preventing participants from altering their base of support with wedges (as a likely adaptation to the imposed foot valgus) and thereby limiting their ability to alter knee position. Interestingly, in a previous study, a lateral heel wedge attached to the foot with subtalar strapping resulted in less varus knee alignment (mean change of 3.4°) compared with a lateral heel wedge inserted into the shoe, which had no effect on alignment (37). The authors contend that restricted talar motion by strapping forces the calcaneum into a lateral position when the wedge is used, thus correcting a varus knee deformity (38). It is possible that variations in subtalar joint stiffness may mediate the effect of wedged insoles. It is also possible that in some individuals malalignment is rigid in nature and the fixed bony position is not amenable to change with lateral wedges.
We demonstrated a clinically relevant level of improved pain and function (39) following treatment with insoles. Not all individuals improved, however. We conducted a preliminary analysis of predictors of clinical outcome. This analysis was only exploratory and hypothesis-generating in nature, rather than conclusive, due to our small sample size. Participants reporting a greater immediate pain-relieving effect of insoles tended to report better outcome at 3 months. This finding has important implications. For instance, if a patient does not report immediate benefit of insoles, then it is less likely that such a patient will receive any benefit from wearing the insoles over the longer term. Although it seems unlikely that a delayed symptomatic benefit will eventuate over time, a long-term benefit of reduced risk of disease progression cannot be excluded, given that insoles reduced the adduction moment in most individuals.
It is not clear how laterally wedged insoles reduce the knee adduction moment. It may be minimized by reducing the magnitude of the ground reaction force and/or its varus moment arm relative to the knee center. There does not appear to be a cushioning effect of the insole on the ground reaction force, as lateral wedges significantly reduce the adduction moment even when compared with unwedged control insoles (19). It seems that wedges increase the valgus moment arm at the subtalar joint, causing a lateral shift in the center of pressure location (21). This lateral shift likely decreases the length of the knee joint moment arm.
A simplified muscle model has been used to estimate the effect of laterally wedged insoles on the medial knee compartment load in healthy persons (20). A significant reduction in medial compartment load occurred with laterally wedged insoles, and most of this change resulted from a reduction in the knee adduction moment. This suggests that pain relief and improved function reported by patients with knee OA following treatment with insoles are probably due to unloading of the medial compartment. This hypothesis is supported by our finding that immediate change in the first peak adduction moment with wedges was predictive of change in WOMAC function score at 3 months.
Disease severity was also predictive of change in pain at 3 months; participants with worse disease tended to report more pain at 3 months. Numerous studies support our finding that persons with milder disease report a better clinical outcome (17, 25, 26, 30). It is likely that patients with more advanced disease demonstrate greater impairment in parameters such as muscle strength. Our findings may be explained by the fact that quadriceps weakness influences pain severity in knee OA (40), which is unlikely to improve with insoles. Toda et al (29) support this hypothesis by demonstrating that patients with decreased lower limb lean mass report poorer outcome with insoles and subtalar strapping. Thus, it may be appropriate to combine the use of wedges with muscle strengthening for older patients and those with severe OA.
Because it was not the aim of this study to conduct a clinical trial in order to establish efficacy, it is unclear how much benefit observed at 3 months is attributable to the insoles themselves, or to other factors such as a placebo effect. The fact that the evaluated predictors only explained one-third of the variance in outcome demonstrates that other unknown factors are important. We used radiographic mechanical axis to measure static alignment. Although this is accepted as the gold standard, its static nature is a drawback given the dynamic nature of gait. Further research evaluating the effect of lateral wedges on dynamic knee alignment is needed. Although we demonstrated that changes in the first peak adduction moment and static alignment are predictive of clinical outcome, the complicated nature of their measurement renders them impractical for routine use by clinicians to predict prognosis with wedged insoles. While surrogate clinical measures are available to assess static alignment (32), it is unknown whether changes in these surrogates are sensitive enough to predict outcome with wedged insoles. Currently, there is no clinical surrogate for 3-dimensional gait analysis to measure the adduction moment. Further research in this area is required.
In conclusion, we demonstrated that laterally wedged insoles immediately reduced the knee adduction moment and walking pain in a large group of patients with medial knee OA, but had no effect on static alignment. Several factors predicted clinical outcome at 3 months including disease severity, baseline function score, and magnitude of immediate change in walking pain and the first peak adduction moment with insoles, although the majority of variance in outcome remained unexplained.
Dr. Hinman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Hinman, Bennell.
Acquisition of data. Metcalf, Wrigley.
Analysis and interpretation of data. Hinman, Payne, Metcalf, Wrigley, Bennell.