Effect of Biomechanical Footwear on Knee Pain in People With Knee Osteoarthritis: The BIOTOK Randomized Clinical Trial | Geriatrics | JAMA | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Participant Recruitment, Randomization, and Follow-up
Participant Recruitment, Randomization, and Follow-up

aThe Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) is a self-administered questionnaire with a score range from 0 to 10 (0, no symptoms; 10, extreme symptoms).

bThe Kellgren-Lawrence grade ranges from 0 to 4; a grade of 2 or greater indicates definite osteoarthritis on anteroposterior weight-bearing radiograph.

cSTEADI indicates Stopping Elderly Accidents, Deaths, and Injuries; a score greater than 4 at the screening visit was considered to indicate a high risk of falls. The score range is 0 to 14.

dOf the 7 participants in the biomedical footwear group and 13 participants in the control footwear group who discontinued treatment, 2 and 5, respectively, did not have data for the primary outcome and were counted as part of those who discontinued treatment. Multiple imputation was used for missing outcome data.

Figure 2.  Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Scores During 24 Weeks of Follow-up
Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Scores During 24 Weeks of Follow-up

The box represents the median and interquartile range, the whiskers represent the most extreme values within 1.5 times of the interquartile range beyond the 25th and 75th percentile, and the circles represent the more extreme values. The WOMAC is a self-administered questionnaire including 5 questions on pain, 17 questions on physical function, and 2 questions on stiffness. All 4 composite scores were standardized to range from 0 to 10 (0, no symptoms; 10, extreme symptoms).

Table 1.  Participant Characteristics at Baseline
Participant Characteristics at Baseline
Table 2.  Primary and Secondary Outcomesa
Primary and Secondary Outcomesa
Table 3.  Adverse Eventsa
Adverse Eventsa
Supplement 3.

Investigators and committees

eMethods. Detailed eligibility criteria, selection of index knee, and biomechanical footwear (experimental), control footwear, trial information provided to participants (translated from German), detailed biomechanical footwear therapy methodology, definitions of study outcomes, additional statistical methods, and discussion of CIOMS criteria for trials that withhold information or use deception

eFigure 1. The biomechanical footwear device

eFigure 2. Elements of the biomechanical footwear device

eFigure 3. Screenshot of webpage presenting the control footwear

eResults. Additional baseline characteristics and missing data, additional prespecified secondary outcomes, other prespecified outcomes, subgroup analyses, sensitivity analyses, per-protocol analysis, complete case analysis, adjustment for potential procedural confounders, analyses of all knees with baseline WOMAC pain ≥3, additional adverse events, and post-hoc analyses

eTable 1. Additional baseline characteristics

eTable 2. Number and percentage of missing data, baseline characteristics

eTable 3. Number and percentage of missing data, outcomes

eTable 4. Additional prespecified secondary outcomes

eTable 5. Other prespecified outcomes

eTable 6. Subgroup analyses of the primary outcome

eTable 7. Reasons for exclusion from per-protocol set

eTable 8. Per-protocol analysis

eTable 9. Complete case analysis

eTable 10. WOMAC pain after adjustment for potential procedural confounders

eTable 11. Analysis of all knees with WOMAC pain ≥3

eTable 12. Serious adverse events and relationship to intervention

eTable 13. Effects from linear mixed-effects regression including all time-points

eFigure 4. WOMAC pain subscores vs time spent wearing the footwear 38

eReferences

Original Investigation
May 12, 2020

Effect of Biomechanical Footwear on Knee Pain in People With Knee Osteoarthritis: The BIOTOK Randomized Clinical Trial

Author Affiliations
  • 1Institute for Social and Preventive Medicine, University of Bern, Bern, Switzerland
  • 2Department of Rheumatology, Immunology, and Allergology, Bern University Hospital, Bern, Switzerland
  • 3Arthritis Research UK Centre for Epidemiology, Centre for Musculoskeletal Research, Faculty of Biology, Medicine, and Health, University of Manchester, Manchester, England
  • 4NIHR Manchester Musculoskeletal Biomedical Research Centre, Manchester University Hospitals NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, England
  • 5Clinical Epidemiology Unit, Boston University, Boston, Massachusetts
  • 6Applied Health Research Centre, Li Ka Shing Knowledge Institute, St Michael’s Hospital, Toronto, Ontario, Canada
  • 7Department of Chiropractic Medicine, Faculty of Medicine, University of Zurich and Balgrist University Hospital, Zurich, Switzerland
  • 8Clinical Trials Unit, University of Bern, Bern, Switzerland
  • 9Department of Medicine and Institute of Health Policy, Management, and Evaluation, University of Toronto, Toronto, Ontario, Canada
  • 10Department for Diagnostic, Interventional, and Pediatric Radiology, Bern University Hospital, Bern, Switzerland
  • 11Centre for Health Sciences Research, School of Health Sciences, University of Salford Manchester, Manchester, England
JAMA. 2020;323(18):1802-1812. doi:10.1001/jama.2020.3565
Key Points

Question  Is an individualized biomechanical footwear therapy effective for reducing knee pain in people with knee osteoarthritis?

Findings  In this randomized clinical trial that included 220 participants with knee pain due to knee osteoarthritis, treatment with an individualized biomechanical footwear therapy compared with control footwear resulted in a lower Western Ontario and McMaster Universities Osteoarthritis Index pain subscore (range, 0-10) after 24 weeks of follow-up (1.3 vs 2.6, respectively), a difference that was statistically significant.

Meaning  Although use of biomechanical footwear compared with control footwear resulted in an improvement in knee pain at 24 weeks of follow-up that was statistically significant, the difference was of uncertain clinical importance, and further research is needed to assess long-term efficacy and safety.

Abstract

Importance  Individually calibrated biomechanical footwear therapy may improve pain and physical function in people with symptomatic knee osteoarthritis, but the benefits of this therapy are unclear.

Objective  To assess the effect of a biomechanical footwear therapy vs control footwear over 24 weeks of follow-up.

Design, Setting, and Participants  Randomized clinical trial conducted at a Swiss university hospital. Participants (N = 220) with symptomatic, radiologically confirmed knee osteoarthritis were recruited between April 20, 2015, and January 10, 2017. The last participant visit occurred on August 15, 2017.

Interventions  Participants were randomized to biomechanical footwear involving shoes with individually adjustable external convex pods attached to the outsole (n = 111) or to control footwear (n = 109) that had visible outsole pods that were not adjustable and did not create a convex walking surface.

Main Outcomes and Measures  The primary outcome was knee pain at 24 weeks of follow-up assessed with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscore standardized to range from 0 (no symptoms) to 10 (extreme symptoms). The secondary outcomes included WOMAC physical function and stiffness subscores and the WOMAC global score, all ranging from 0 (no symptoms) to 10 (extreme symptoms) at 24 weeks of follow-up, and serious adverse events.

Results  Among the 220 randomized participants (mean age, 65.2 years [SD, 9.3 years]; 104 women [47.3%]), 219 received the allocated treatment and 213 (96.8%) completed follow-up. At 24 weeks of follow-up, the mean standardized WOMAC pain subscore improved from 4.3 to 1.3 in the biomechanical footwear group and from 4.0 to 2.6 in the control footwear group (between-group difference in scores at 24 weeks of follow-up, −1.3 [95% CI, −1.8 to −0.9]; P < .001). The results were consistent for WOMAC physical function subscore (between-group difference, −1.1 [95% CI, −1.5 to −0.7]), WOMAC stiffness subscore (between-group difference, −1.4 [95% CI, −1.9 to −0.9]), and WOMAC global score (between-group difference, −1.2 [95% CI, −1.6 to −0.8]) at 24 weeks of follow-up. Three serious adverse events occurred in the biomechanical footwear group compared with 9 in the control footwear group (2.7% vs 8.3%, respectively); none were related to treatment.

Conclusions and Relevance  Among participants with knee pain from osteoarthritis, use of biomechanical footwear compared with control footwear resulted in an improvement in pain at 24 weeks of follow-up that was statistically significant but of uncertain clinical importance. Further research would be needed to assess long-term efficacy and safety, as well as replication, before reaching conclusions about the clinical value of this device.

Trial Registration  ClinicalTrials.gov Identifier: NCT02363712

Introduction

Knee osteoarthritis affects approximately 265 million people worldwide and was estimated in 2017 to account for 8.3 million years lived with disability.1 The prevalence of knee osteoarthritis is rising due to population aging and the increasing prevalence of obesity. Acetaminophen, nonsteroidal anti-inflammatory drugs, and opioids are most commonly used to treat the pain associated with osteoarthritis,2 but have limited effectiveness3,4 and are associated with adverse effects.3,5,6 In the US, the rates of knee replacement surgery, almost all related to osteoarthritis, have been increasing, in part because of ineffective nonsurgical treatments.

Quiz Ref IDBiomechanical treatments for knee osteoarthritis have been developed to reduce pain, improve physical function, and potentially slow disease progression,7 but evidence of their effectiveness has been inconclusive.8,9 Two small prospective, nonrandomized clinical studies suggested that an individualized biomechanical footwear system may improve pain and physical function in people with symptomatic knee osteoarthritis.10,11 In those studies, the footwear system consisted of shoes with 2 convex pods on the outsoles, individually calibrated based on findings from detailed gait studies. Adjustment of the location of the pods may alter limb biomechanics and reduce stress on osteoarthritic knee compartments.12-14 Walking on the convex pods results in gait alterations, and these alterations are hypothesized to induce reconditioning of the neuromuscular system and improve pathological gait patterns.15

The objective of this study, the Biomechanical Therapy for Osteoarthritis of the Knee (BIOTOK) randomized clinical trial, was to determine whether the use of biomechanical footwear was more effective than the use of control footwear for improving knee pain in participants with knee osteoarthritis.

Methods
Study Design and Participants

The trial protocol and statistical analysis plan appear in Supplement 1 and Supplement 2, respectively. This was an investigator-initiated single-center randomized clinical trial conducted among participants with symptomatic knee osteoarthritis, which compared a biomechanical footwear therapy using shoes with 2 individually calibrated convex pods on the outsoles (AposTherapy, Apos Medical Assets; eFigures 1 and 2 in Supplement 3) with a similarly appearing control footwear.

We enrolled men and nonpregnant women aged 40 years or older who had symptomatic, radiologically confirmed knee osteoarthritis according to criteria from the American College of Rheumatology.16 At the screening visit, participants had knee pain lasting 6 months or longer and a score of 3 or greater on the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscale17 (standardized to range from 0-10). The full eligibility criteria and description for the selection of the index knee appear in Supplement 3.

Individuals were excluded if they had a history of inflammatory rheumatic disease, had undergone knee surgery during the prior 6 months, had a planned hip or knee surgery within 24 weeks of the baseline assessment, had received glucocorticoid knee injections within the prior 3 months, or had a high risk for falls.

The trial was approved by the independent research ethics committee of Canton Bern (KEK-BE 041/215). All participants provided written informed consent.

Randomization and Blinding

Participants were randomized 1:1 to the biomechanical footwear group or to the control footwear group using a concealed and secure web-based system (Figure 1). Randomization was computer-generated using varied block sizes of 2 and 4 and stratified by unilateral vs bilateral knee disease and predominantly affected compartment (medial vs lateral) in the index knee.

The biomechanical footwear device consisted of 2 shoes with 2 convex adjustable rubber pods screwed to the outsole at the heel and forefoot (eFigures 1 and 2 in Supplement 3). Quiz Ref IDThe control footwear was specifically designed by the manufacturer for this trial to have a similar appearance to the biomechanical footwear, but with pods embedded in the transparent outsole so that they were visible yet did not create a convex walking surface (eFigure 3 in Supplement 3).

To try to maintain blinding of the participants, they were kept unaware of the study design and the use of the control footwear. Participants were informed in a neutral manner that 2 different types of footwear were being compared (Supplement 3). Both the biomechanical and control footwear were presented on the manufacturer’s website, and the control footwear was described as a device with a novel design of the sole (eFigure 3 in Supplement 3).

The technicians and study nurses who coordinated the clinical visits could not be blinded to treatment group but were asked not to disclose the treatment or the nature of the control footwear study component to participants. The technicians were from Israel and did not speak German, therefore, direct interaction between the technicians and participants was limited. Verbal communication was carried out through the translating study nurses, who were independent of the manufacturer and were encouraged to facilitate unbiased participant interaction. The remaining study personnel (performing data entry, management, and data cleaning) and the statistician were blinded to the intervention until all primary and secondary analyses were completed.

The consent form did not state that the control footwear was intended to be ineffective; rather, it implied that both types of shoes may be effective. Furthermore, the manufacturer’s website was altered to imply potentially therapeutic benefits of both the intervention and the control footwear. Therefore, the trial could be considered potentially deceptive according to international guidelines.18

Both the biomechanical and control footwear included some therapeutic elements. Both were high-top shoes, which provided more stability and proprioceptive input than loose shoes or sandals. Furthermore, proposed mechanisms of the biomechanical footwear were hypothetical at the initiation of this trial, and the trial was considered to entail no more than minimal risk and burden to participants according to Article 2 of the Swiss clinical trials ordinance.19

Therefore, the responsible research ethics committee did not classify the trial as involving incomplete participant information (ie, did not consider the study procedures to be deceptive) according to Article 18 of the Swiss human research act.20 Nonetheless, because the trial may have been considered deceptive by some individuals, participants were debriefed after the trial was completed.

Participants were advised of the rationale of the placebo-controlled design, informed of differences between the biomechanical footwear and the control footwear, and informed about their group allocation, and were given the opportunity to withdraw consent to participate. The criteria specified in the international guidelines18 for trials that withhold information or use deception with respect to this trial appear in Supplement 3.

Procedures

Participants in both groups underwent an initial fitting of their assigned footwear by technicians at baseline and during recalibration at 4, 8, 12, and 16 weeks of follow-up. The positioning of the external pods was individually adjusted on the biomechanical footwear, in accordance with gait patterns and reported pain intensity during walking, with the aim of decreasing clinically observed malalignment and reported pain intensity, and increasing gait symmetry13,14,21,22 as determined by 2-dimensional computerized spatiotemporal gait analysis software (Zeno walkway version Z-216T and PKMAS version 5.07C3, both manufactured by ProtoKinetics; Supplement 3). Participants in the control footwear group received a simulated calibration that mimicked calibration of the biomechanical footwear. The technicians provided by the manufacturer performed all device-related procedures (gait analyses, fitting, and calibrations of the biomechanical and control footwear).

Participants were instructed to use the footwear during indoor activities for a half hour each day during the first week of the intervention, with subsequent increases of 10 minutes per week on average but were not given explicit instructions to perform specific home-based exercises. After 6 weeks of follow-up, the participants were advised to use the footwear to walk outdoors. The participants were asked to discontinue their regular pain medication and advised that other interventions, such as physical therapy, should be avoided during the trial. They were permitted daily therapy as needed with acetaminophen at a maximum dose of 2 g and the amounts taken were recorded at each visit.

Outcomes

The prespecified primary outcome was pain at the end of treatment (24-week follow-up) in the index knee assessed with the WOMAC pain subscore (visual analog version) (standardized range, 0-10; 0 = no symptoms and 10 = extreme symptoms).17

The secondary outcomes prespecified in the protocol were WOMAC global score and WOMAC physical function and stiffness subscores (all 3 with standardized range, 0-10; 0 = no symptoms) at 12 and 24 weeks of follow-up; WOMAC pain subscore at 4, 8, 12, and 16 weeks of follow-up; the physical and mental component summary scores of the Medical Outcomes Study 36-item Short Form Health Survey (SF-36) (standardized to have a mean of 50 [SD, 10] for the general population; theoretical range, 0-100; 100 = best)23 at 12 and 24 weeks of follow-up; gait velocity, step length, and single limb support measured by 2-dimensional computerized gait analysis when walking barefoot at 4, 8, 12, 16, and 24 weeks of follow-up; self-reported time spent wearing the footwear per day; self-reported health care use; and self-reported analgesic use (because between-group differences in analgesic use could result in performance bias).24 The minimal clinically important differences were not considered when planning the trial.

Other prespecified outcomes were treatment response defined as decreases of 30% and 50% in WOMAC pain subscore from baseline.25 A treatment response defined as a 50% decrease in WOMAC pain subscore from baseline was not prespecified in the protocol (Supplement 1), but was included in the statistical analysis plan (Supplement 2).

The adverse events prespecified in the protocol were falls, any adverse events, serious adverse events, dropped out for any reason, and dropped out due to adverse events (Supplement 1). The WOMAC scores,17 analgesic intake, and gait analysis parameters were recorded at baseline and at 4, 8, 12, 16, and 24 weeks of follow-up; and the SF-36 scores and health care use were recorded at baseline and at 12 and 24 weeks of follow-up. The adverse events and the time spent wearing the footwear were recorded at each follow-up visit. Two investigators (S.R. and P.J.) blinded to the assigned treatment adjudicated all potential adverse events based on notes by participants and nurses, and, in cases of potential serious adverse events, based on relevant medical records.

Statistical Analysis

A sample size of 100 participants per group yielded 80% power to detect a between-group difference of 1.05 on a standardized WOMAC pain scale ranging from 0 to 10 at a 2-sided α level of .05. This between-group difference corresponds to a moderate effect size of 0.4-SD units, assuming a typical SD of 2.65.4 The protocol prespecified the use of analyses of covariance for all continuous outcomes, adjusted for the outcome’s baseline values. For this approach, a sample size of 100 participants per group would yield approximately 90% power, assuming a correlation of 0.5 between baseline and 24 weeks of follow-up. Anticipating an attrition rate of 10%, the target sample size was 220 participants.

Continuous outcomes were analyzed using analysis of covariance adjusted for the outcome’s baseline values and variables used for stratified randomization, considering only the assessments of the index knee for each participant. Binary outcomes were analyzed using Cochran-Mantel-Haenszel tests stratified by stratification variables.26 Participants were included in the analyses according to their randomized allocation,27 using multiple imputation for missing outcome data, using all baseline characteristics (age, sex, body mass index, blood pressure, medical history, WOMAC scores, SF-36 scores, and parameters of gait analysis), outcomes at all time points, the treatment indicator, and stratification variables to generate 20 imputed data sets (Supplement 2 and Supplement 3).

Prespecified subgroup analyses of the primary outcome were performed according to the predominantly affected compartment and the presence or absence of symptomatic contralateral knee osteoarthritis and were accompanied by tests for interaction. A post hoc subgroup analysis was performed according to WOMAC pain intensity at baseline.28 Prespecified sensitivity analyses of the primary outcome included a per-protocol analysis, a complete case analysis, adjustments for potential procedural confounders, and a linear mixed-effects model to analyze all knees (ie, index or both index and contralateral knee) with a baseline WOMAC pain subscale score of 3 or greater. Post hoc sensitivity analyses of WOMAC scores, SF-36 scores, and parameters of gait analyses were performed using all time points in a linear mixed-effects regression model (Supplement 2 and Supplement 3).

The P values and 95% CIs were 2-sided and P values ≤.05 were considered statistically significant. Because of the potential for type I error due to multiple comparisons, the findings for the analyses of the secondary outcomes should be interpreted as exploratory. All analyses were performed using R version 3.3.2 (R Foundation for Statistical Computing)29 by an independent statistician located within an academic clinical trials unit (in Bern, Switzerland) who was unaware of group assignment. The statistical analysis plan was finalized after completion of follow-up, but before examination of the data. Data were interpreted and conclusions formulated prior to the unblinding of the investigators.

Results

Between April 20, 2015, and January 10, 2017, 220 participants were randomized. There were 111 participants randomized to the biomechanical footwear group and 109 participants randomized to the control footwear group (Figure 1). One participant in the biomechanical footwear group refused treatment and did not receive the intervention. Seven participants in the biomechanical footwear group and 13 participants in the control footwear group discontinued treatment during follow-up. The last participant visit occurred on August 15, 2017. There were complete data for the primary outcome at 24 weeks of follow-up for 109 participants (98.2%) in the biomechanical footwear group and 104 participants (95.4%) in the control footwear group.

After trial completion, 217 of the 220 randomized participants were reached and advised of the potential for deception in the study design. Of the 3 participants who were not reached (all in the biomechanical footwear group), 1 had died and 2 were lost to follow-up. None of the remaining 217 participants withdrew consent after learning that the trial involved randomization to either the biomechanical footwear group or the control footwear group that was expected to be ineffective.

Baseline characteristics were similar between the participants randomized to each group (Table 1 and eTable 1 in Supplement 3). The study population had a mean age of 65.2 years (SD, 9.3 years), 104 were women (47.3%), and the population had a mean body mass index (calculated as weight in kilograms divided by height in meters squared) of 28.0 (SD, 4.6). Medial knee osteoarthritis was present in 90.9% of participants and unilateral disease in 67.7%. The number of participants with missing data was between 0 and 3 (1.4%) for baseline characteristics (eTable 2 in Supplement 3) and between 2 (0.9%) and 29 (13.2%) for outcomes (eTable 3 in Supplement 3).

Primary Outcome

The biomechanical footwear group had a larger decrease in standardized WOMAC pain subscore at 24 weeks of follow-up than the control footwear group (mean score, 1.3 vs 2.6, respectively; between-group difference, −1.3 [95% CI, −1.8 to −0.9]; P < .001) (Figure 2 and Table 2).

Secondary Outcomes

The biomechanical footwear group had larger declines in the secondary outcomes of WOMAC physical function and stiffness subscores and WOMAC global score at 24 weeks of follow-up (Figure 2 and Table 2). Between-group differences in velocity, step length, and single limb support emerged in favor of the biomechanical footwear group between 12 and 24 weeks of follow-up (Table 2). The mean self-reported time spent wearing the footwear at 24 weeks of follow-up was 209 minutes/day in the biomechanical footwear group vs 174 minutes/day in the control footwear group (between-group difference, 35 minutes/day [95% CI, 4-67 minutes/day]).

For the SF-36 physical component summary score, there was no statistically significant between-group difference for the biomechanical footwear group vs the control footwear group (mean score, 45.9 vs 44.5, respectively; between-group difference 1.4 [95% CI, −0.5 to 3.2]). There were no significant between-group differences in the SF-36 mental component summary score, for analgesic use, or for health care use. Additional prespecified secondary outcomes, including types of analgesics, types of clinicians visited, use of corticosteroid injections, and performed or planned knee replacement surgeries appear in eTable 4 in Supplement 3.

The other prespecified outcomes appear in eTable 5 in Supplement 3, including participants with a treatment response achieving a 30% or 50% reduction for the WOMAC pain subscore from baseline to 24 weeks of follow-up. Among participants in the biomechanical footwear group, 92% achieved a 30% reduction for the WOMAC pain subscore vs 58% in the control footwear group (risk difference, 34% [95% CI, 23%-45%]), and 83% vs 42%, respectively, achieved a 50% reduction for the WOMAC pain subscore (risk difference, 41% [95% CI, 28%-52%]); the corresponding numbers needed to treat were 3 (95% CI, 2-5) and 3 (95% CI, 1-4).

The prespecified subgroup analyses of the primary outcome according to the predominantly affected compartment and symptomatic contralateral disease did not show significant treatment × subgroup interactions (eTable 6 in Supplement 3). The sensitivity analyses of the primary outcome, including a per-protocol analysis, a complete case analysis, adjustments for potential procedural confounders, and a linear mixed-effects model to analyze all knees with a baseline WOMAC pain subscore of 3 or greater were consistent with the primary outcome analysis (eTables 7-11 in Supplement 3).

Adverse Events

Twenty-six participants (23.4%) in the biomechanical footwear group and 38 participants (34.9%) in the control footwear group experienced an adverse event and 3 (2.7%) and 9 (8.3%), respectively, experienced serious adverse events (Table 3). None were considered to be related to treatment. Of the serious adverse events, there were 0 in the biomechanical footwear group vs 4 in the control footwear group that were musculoskeletal, 1 vs 3, respectively, that were circulatory, and 2 vs 2 that were in other categories (eTable 12 in Supplement 3). One or more falls occurred in 2 participants (1.8%) in the biomechanical footwear group and in 4 participants (3.7%) in the control footwear group. One participant in the control group fell while wearing the control footwear.

Post Hoc Analyses

A post hoc subgroup analysis of the primary outcome by WOMAC pain intensity at baseline did not show significant treatment × subgroup interactions (eTable 6 in Supplement 3). The post hoc use of a mixed-effects model simultaneously including all time points showed results similar to those of the main analyses. In the mixed-effects models, there were significant between-group differences in the WOMAC pain and physical function subscores and in the WOMAC global score at 12, 16, and 24 weeks of follow-up, and in the WOMAC stiffness subscore at 16 and 24 weeks of follow-up. Significant between-group differences in parameters of gait analysis were observed for velocity and step length at 12, 16, and 24 weeks of follow-up and for single limb support at 24 weeks of follow-up (eTable 13 in Supplement 3).

eFigure 4 in Supplement 3 contrasts WOMAC pain subscore with the time spent wearing the footwear over the duration of the trial. The maximal between-group difference in time spent wearing the footwear occurred at 16 weeks of follow-up, whereas the maximum between-group difference in the WOMAC pain subscore was observed 8 weeks later at 24 weeks of follow-up.

Discussion

Quiz Ref IDIn this randomized clinical trial, the biomechanical footwear system with individually calibrated outsole convex pods was significantly more effective than the control footwear at reducing pain at 24 weeks of follow-up in participants with knee pain from symptomatic knee osteoarthritis. The results were consistent for the secondary outcomes of WOMAC physical function and stiffness subscores and the WOMAC global score at 24 weeks of follow-up. There were no significant between-group differences in the physical and mental components of the SF-36.

Quiz Ref IDThere are 2 differences between the biomechanical footwear system tested in this trial and other biomechanical devices such as shoes9 or wedges.8 First, in this trial, the individualized calibration of proximal and distal pods of the biomechanical device in the coronal and sagittal planes shifts the trajectory of the foot’s center of pressure, thereby specifically changing the direction of the ground reaction force vector as appropriate for each individual.12,13,30 Second, the convexity of the pods in the biomechanical footwear results in repetitive gait perturbation, with mild destabilization of the knee during walking, which in turn may elicit neuromuscular responses.

To our knowledge, no other published randomized clinical trials have investigated the effectiveness of this biomechanical footwear system in people with symptomatic knee osteoarthritis. Of 6 published clinical studies,10,11,15,31-33 4 were uncontrolled studies conducted by the manufacturer,15,31-33 the remaining 2 were prospective and controlled, but not randomized.10,11 The most rigorous investigation was a prospective nonrandomized controlled study among 57 participants with symptomatic knee osteoarthritis,10 which found improved pain and physical function with the biomechanical footwear system compared with the control footwear. However, the between-group difference in the WOMAC pain subscore at 8 weeks of follow-up was not consistent with the negative 8-week result in the current study. The reason for this between-group difference is unclear but may be due to lack of randomization in the prior trial.10

Limitations

This study has several limitations. First, there were differences in the appearance of the biomechanical footwear and the control footwear. To overcome this limitation and minimize the likelihood that participants would correctly guess that they were not receiving the active intervention, participants were kept unaware that the control footwear was not expected to have therapeutic benefits. Participants were informed in a neutral fashion that 2 different types of footwear were being compared. The manufacturer’s website described the control footwear as a device with a novel design of the sole, and the participants randomized to the control group received a simulated calibration that mimicked the actual calibration.

Second, the use of a blinding index34 to determine the success of blinding was not performed because such an index assumes indistinguishable interventions. Third, the self-reported time per day wearing the footwear was longer in the biomechanical footwear group than in the control footwear group. It is possible that the greater benefit in the biomechanical footwear group was due to longer wear time.

Fourth, analgesic treatment for pain was allowed during the trial; however, the rates of analgesic use did not differ between groups. Fifth, it was not possible to explore changes in knee adduction moments using 3-dimensional gait analyses.

Quiz Ref IDSixth, the trial was conducted at a single center, potentially limiting generalizability. Seventh, the between-group differences occurred only late during follow-up and were smaller than the observed within-group change from baseline in the control group. Therefore, the clinical importance of these findings remains uncertain. Eighth, the findings from this trial are not generalizable to people at high risk for falls because these individuals were not eligible to participate. Ninth, the findings are not generalizable to people with severe knee pain because these individuals were underrepresented in the trial.

Conclusions

Among participants with knee pain from osteoarthritis, use of biomechanical footwear compared with control footwear resulted in an improvement in pain at 24 weeks of follow-up that was statistically significant but of uncertain clinical importance. Further research would be needed to assess long-term efficacy and safety, as well as replication, before reaching conclusions about the clinical value of this device.

Back to top
Article Information

Corresponding Author: Peter Jüni MD, Applied Health Research Centre, Li Ka Shing Knowledge Institute, St Michael’s Hospital, 250 Yonge St, Toronto, ON M5B 2L7, Canada (peter.juni@utoronto.ca).

Accepted for Publication: March 3, 2020.

Author Contributions: Drs Reichenbach and Jüni had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Reichenbach, Felson, da Costa, Jones, Jüni.

Acquisition, analysis, or interpretation of data: Reichenbach, Felson, Hincapié, Heldner, Bütikofer, Lenz, da Costa, Bonel, Hawker, Jüni.

Drafting of the manuscript: Reichenbach, Felson, Hincapié, Jüni.

Critical revision of the manuscript for important intellectual content: Reichenbach, Hincapié, Heldner, Bütikofer, Lenz, da Costa, Bonel, Jones, Hawker, Jüni.

Statistical analysis: Bütikofer, Lenz, da Costa, Jüni.

Obtained funding: Jüni.

Administrative, technical, or material support: Felson, Heldner, da Costa, Jones, Jüni.

Supervision: Reichenbach, Felson, Bonel, Jüni.

Conflict of Interest Disclosures: Dr Jüni reported receiving grants from AstraZeneca, Biotronik, Biosensors, St Jude Medical, the Medicines Company, and the Canadian Institutes of Health Research; receiving honoraria from Amgen and Fresenius paid directly to his institution for participation on advisory boards; and serving as an unpaid member on steering committees for cardiovascular trials funded by AstraZeneca, Biotronik, Biosensors, St Jude Medical, and the Medicines Company. No other disclosures were reported.

Funding/Support: The trial was sponsored by Bern University Hospital and coordinated by CTU Bern, the University of Bern’s clinical trials unit. The trial was funded by the Mäxi Foundation. Dr Jüni is a tier 1 Canadian research chair in clinical epidemiology of chronic diseases; this research was completed, in part, with funding from the Canada Research Chairs Programme. Apos Medical Assets provided the biomechanical footwear system and the control footwear, and provided the technicians trained to install and calibrate the external pods on the biomechanical footwear without charge.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 4.

Additional Contributions: We thank the participants who participated in this trial. We acknowledge Avi Elbaz, MD, and Amit Mor, MD, for their contributions in the development of the control footwear; both were paid members of the board of AposTherapy Israel. We thank Madeleine Dähler and Zina Heg, MSc, for their contributions to data collection. At the time of data collection, both were paid staff members at CTU Bern and their salaries were funded by the Mäxi Foundation.

References
1.
James  SL, Abate  D, Abate  KH,  et al.  Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017.   Lancet. 2018;392(10159):1789-1858.PubMedGoogle ScholarCrossref
2.
Gore  M, Tai  K-S, Sadosky  A,  et al.  Use and costs of prescription medications and alternative treatments in patients with osteoarthritis and chronic low back pain in community-based settings.   Pain Pract. 2012;12(7):550-560. PubMedGoogle ScholarCrossref
3.
da Costa  BR, Nüesch  E, Kasteler  R,  et al.  Oral or transdermal opioids for osteoarthritis of the knee or hip.   Cochrane Database Syst Rev. 2014;(9):CD003115. PubMedGoogle Scholar
4.
da Costa  BR, Reichenbach  S, Keller  N,  et al.  Effectiveness of non-steroidal anti-inflammatory drugs for the treatment of pain in knee and hip osteoarthritis.   Lancet. 2017;390(10090):e21-e33. PubMedGoogle ScholarCrossref
5.
Trelle  S, Reichenbach  S, Wandel  S,  et al.  Cardiovascular safety of non-steroidal anti-inflammatory drugs.   BMJ. 2011;342:c7086. PubMedGoogle ScholarCrossref
6.
Bhala  N, Emberson  J, Merhi  A,  et al.  Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs.   Lancet. 2013;382(9894):769-779. PubMedGoogle ScholarCrossref
7.
McAlindon  TE, Bannuru  RR, Sullivan  MC,  et al.  OARSI guidelines for the non-surgical management of knee osteoarthritis.   Osteoarthritis Cartilage. 2014;22(3):363-388. PubMedGoogle ScholarCrossref
8.
Parkes  MJ, Maricar  N, Lunt  M,  et al.  Lateral wedge insoles as a conservative treatment for pain in patients with medial knee osteoarthritis.   JAMA. 2013;310(7):722-730. PubMedGoogle ScholarCrossref
9.
Hinman  RS, Wrigley  TV, Metcalf  BR,  et al.  Unloading shoes for self-management of knee osteoarthritis.   Ann Intern Med. 2016;165(6):381-389. PubMedGoogle ScholarCrossref
10.
Bar-Ziv  Y, Beer  Y, Ran  Y,  et al.  A treatment applying a biomechanical device to the feet of patients with knee osteoarthritis results in reduced pain and improved function.   BMC Musculoskelet Disord. 2010;11:179. PubMedGoogle ScholarCrossref
11.
Bar-Ziv  Y, Debbi  EM, Ran  Y,  et al.  Long-term effects of AposTherapy in patients with osteoarthritis of the knee.   Arthritis. 2013;2013:689236. PubMedGoogle Scholar
12.
Haim  A, Rozen  N, Dekel  S,  et al.  Control of knee coronal plane moment via modulation of center of pressure.   J Biomech. 2008;41(14):3010-3016. PubMedGoogle ScholarCrossref
13.
Haim  A, Wolf  A, Rubin  G,  et al.  Effect of center of pressure modulation on knee adduction moment in medial compartment knee osteoarthritis.   J Orthop Res. 2011;29(11):1668-1674. PubMedGoogle ScholarCrossref
14.
Haim  A, Rubin  G, Rozen  N,  et al.  Reduction in knee adduction moment via non-invasive biomechanical training.   J Biomech. 2012;45(1):41-45. PubMedGoogle ScholarCrossref
15.
Elbaz  A, Mor  A, Segal  G,  et al.  APOS therapy improves clinical measurements and gait in patients with knee osteoarthritis.   Clin Biomech (Bristol, Avon). 2010;25(9):920-925. PubMedGoogle ScholarCrossref
16.
Altman  R, Asch  E, Bloch  D,  et al.  Development of criteria for the classification and reporting of osteoarthritis.   Arthritis Rheum. 1986;29(8):1039-1049. PubMedGoogle ScholarCrossref
17.
Bellamy  N, Buchanan  WW, Goldsmith  CH,  et al.  Validation study of WOMAC.   J Rheumatol. 1988;15(12):1833-1840.PubMedGoogle Scholar
18.
Council for International Organizations of Medical Sciences; World Health Organization.  International Ethical Guidelines for Health-Related Research Involving Humans. Council for International Organizations of Medical Sciences; 2016.
19.
Swiss Federal Council. Ordinance on clinical trials in human research. Accessed December 24, 2019. https://www.admin.ch/opc/en/classified-compilation/20121176/index.html
20.
Swiss Federal Council. Federal act on research involving human beings (human research act). Accessed December 24, 2019. https://www.admin.ch/opc/en/classified-compilation/20061313/index.html
21.
Khoury  M, Wolf  A, Debbi  EM,  et al.  Foot center of pressure trajectory alteration by biomechanical manipulation of shoe design.   Foot Ankle Int. 2013;34(4):593-598. PubMedGoogle ScholarCrossref
22.
Khoury  M, Haim  A, Herman  A,  et al.  Alteration of the foot center of pressure trajectory by an unstable shoe design.   J Foot Ankle Res. 2015;8:67. PubMedGoogle ScholarCrossref
23.
Laucis  NC, Hays  RD, Bhattacharyya  T.  Scoring the SF-36 in orthopaedics.   J Bone Joint Surg Am. 2015;97(19):1628-1634. PubMedGoogle ScholarCrossref
24.
Jüni  P, Altman  DG, Egger  M.  Systematic reviews in health care.   BMJ. 2001;323(7303):42-46. PubMedGoogle ScholarCrossref
25.
Dworkin  RH, Turk  DC, Wyrwich  KW,  et al.  Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials.   J Pain. 2008;9(2):105-121. PubMedGoogle ScholarCrossref
26.
Klingenberg  B.  A new and improved confidence interval for the Mantel-Haenszel risk difference.   Stat Med. 2014;33(17):2968-2983. PubMedGoogle ScholarCrossref
27.
Nüesch  E, Trelle  S, Reichenbach  S,  et al.  The effects of excluding patients from the analysis in randomised controlled trials.   BMJ. 2009;339:b3244. PubMedGoogle ScholarCrossref
28.
Kapstad  H, Hanestad  BR, Langeland  N,  et al.  Cutpoints for mild, moderate and severe pain in patients with osteoarthritis of the hip or knee ready for joint replacement surgery.   BMC Musculoskelet Disord. 2008;9:55. PubMedGoogle ScholarCrossref
29.
R Core Team. R: a language and environment for statistical computing. Accessed April 7, 2020. https://www.R-project.org/
30.
Haim  A, Rozen  N, Wolf  A.  The influence of sagittal center of pressure offset on gait kinematics and kinetics.   J Biomech. 2010;43(5):969-977. PubMedGoogle ScholarCrossref
31.
Drexler  M, Elbaz  A, Mor  A,  et al.  Effects of a customized biomechanical therapy on patients with medial compartment knee osteoarthritis.   Ann Phys Rehabil Med. 2012;55(4):213-228. PubMedGoogle ScholarCrossref
32.
Elbaz  A, Mor  A, Segal  G,  et al.  Patients with knee osteoarthritis demonstrate improved gait pattern and reduced pain following a non-invasive biomechanical therapy.   J Orthop Surg Res. 2014;9:1. PubMedGoogle ScholarCrossref
33.
Lubovsky  O, Mor  A, Segal  G,  et al.  A novel self-care biomechanical treatment for obese patients with knee osteoarthritis.   Int J Rheum Dis. 2017;20(7):818-824. PubMedGoogle ScholarCrossref
34.
James  KE, Bloch  DA, Lee  KK,  et al.  An index for assessing blindness in a multi-centre clinical trial.   Stat Med. 1996;15(13):1421-1434. PubMedGoogle ScholarCrossref
×