Reduced knee joint loading with lateral and medial wedge insoles for management of knee osteoarthritis: a protocol for a randomized controlled trial

Ethics approval for this study has been obtained from the University of Calgary Conjoint Health Research and Ethics Board (CHREB). In accordance with CHREB guidelines, written consent will be obtained from each individual prior to being enrolled in the study formally as a participant, and prior to any testing.

The trial will have two parallel groups, where participants are randomly assigned to a waitlist control or experimental arm for a study duration of 3 months (Figure 1). Patients with painful radiographic medial knee osteoarthritis will be recruited from either the University of Calgary Sport Medicine Centre patient population (a Knee Osteoarthritis clinic is part of the services provided by this centre) or the Alberta Hip and Knee Arthroplasty clinics. Identified patients will be approached for inclusion into the study based on the inclusion/exclusion criteria described below. If recruitment is too slow to meet study timelines, advertisements of the project will be put into local newspapers and to other media. In these cases, a phone assessment of the volunteer will be undertaken for preliminary screening, and the individual will be invited for a detailed assessment by the study physician to determine their eligibility.

All patients considered eligible for the study must have physician-diagnosed knee OA primarily localized to the medial tibiofemoral compartment, based on the American College of Rheumatology clinical and radiographic diagnostic criteria [17]. Patients showing lateral compartment or patellofemoral OA in addition to medial knee OA will still be considered eligible as long as the medial compartment is the primary disease location and symptoms are attributed to the medial compartment. In addition, patients must be between the ages of 40–85 years of age, have no history of viscosupplement injections within the past 6 months, cortisone injection within the past 3 months, narcotic pain medication within the past 3 months, or use of other footwear insoles or unloader knee braces within the past 2 months, and must have a Kellgren-Lawrence grade (a measure of OA severity) of 1 to 3, as determined by a clinician, to avoid inclusion of end-stage OA patients that may be more appropriate for surgery [18]. Patients may continue to use anti-inflammatory medication or acetaminophen for mild pain management if needed. Those who meet the intrarticular injection criteria must also be willing to not undergo further treatment during the 3-month study. Patients must also present with a KOOS pain subscale score of 75 points or lower. This will help ensure that all participants have at least mildly painful knee OA. Any participant with recent (within 2 months) musculoskeletal or neuromuscular injury to the lower extremity that could bias pain assessments or gait analysis will not be considered eligible. For this reason, patients who typically use a cane will also be excluded.

Two footwear insole interventions will be utilized within the experimental arm of this study: (1) the first is a medial wedge insole placed within the participants own shoe. The insole will be developed using a 3D printer, where the wedge thickness is 0.5 mm on the lateral edge, and 6.5 mm on the medial edge. The insole will taper at the metatarsal heads to a thickness of 4.5 mm to ensure a comfortable fit within the participant’s shoe. A wedge thickness of about 6 mm has been used previously and has been shown to be large enough to influence knee mechanics, yet small enough to maintain comfort [14, 19‐21]. During testing and long term use, this insole will be placed under the existing sock liner of the participant’s shoe. (2) The second condition is a lateral wedge insole placed within the participants own shoe. The insole will be developed using a 3D printer, where the thickness of the medial edge is 0.5 mm and the thickness of the lateral edge is 6.5 mm, tapering to 4.5 mm beyond the metatarsal heads. During testing and long term use, this insole will be placed under the existing sock liner of the participant’s shoe. For construction of the medial and lateral wedge, the wedge will extend along the full length of the foot, creating approximately 5–6 degree medial and lateral wedges [22]. All insoles will be produced by New Balance Athletic Shoe Inc. (Boston, MA). The material utilized for the insoles is stiff in compression, such that the wedge shape will be maintained throughout the 3 month study; however, the material is also flexible to ensure comfort within the shoe during flexion of the metatarsal joints.

The authors must emphasize the relationship between the intervention we intend to utilize (wedged insoles), and the mechanical effect we aim to achieve (reduced KAMs). It is this mechanical result (reduced KAMs) that is suggested in the literature to affect knee OA – not wedged insoles on their own. Thus it is critical to bear in mind that the wedged insoles are simply a means to achieve the desired mechanical outcome.

Recently, it has been reported by a meta-analysis that future knee OA trials should utilize a flat insole as a control condition [16]. The motivation for this claim was based on the notion that this would control for placebo effects. However, recent evidence has suggested that even flat insoles can alter pressure distributions under the foot during gait [23], and thus it remains a possibility that KAMs could be influenced as well. Therefore, while flat insoles may control for placebo effects, they may not control for gait biomechanics, which is the primary concern of our trial. Consequently, the control condition for our study will be the participant’s own footwear, as this will control for gait biomechanics changes. Therefore, for biomechanical reasons, the trial will not be placebo controlled. In effort to ensure recruitment, compliance and participant retention in our control condition, we will utilize a waitlist control strategy, where we will indicate that a 3 month wait is required prior to providing an insole. After 3 months, we will provide control participants with access to a wedged insole, and continue to monitor them for safety.

All eligible participants will be invited to the laboratory for baseline testing. At baseline, body metrics such as height (m), mass (kg) and age (years) will be recorded, and also basic clinical diagnostic data (eg. Kellgren-Lawrence grade).

Following the baseline laboratory and survey testing, participants will be randomized into one of two study arms: (1) a KAM-reduction experimental group (KAM-R), or (2) a waitlist control group. In the KAM-R group, participants will be given either a medial or lateral wedge insole (whichever insole reduces KAMs more for each individual participant) to put inside their own shoe and use for 3 months. In the waitlist control group, participants will be asked to continue using their own shoes for a period of 3 months, and following this will be provided with access to a wedged insole (whichever reduces KAMs more based on baseline data collection).

Participants will be randomized using a multi-block-randomization procedure into either the KAM-R or waitlist control group. Specifically, a separate block sequence will be used for males and females to ensure an even sex distribution across groups. A block size of 6 will be used for each, and the sequence will be determined from a random number generator.

The one exception to this block randomization procedure is that if neither wedged insole produces a KAM reduction, the participant will be excluded from the main study altogether. This scenario will not be common, as most people do experience a large change in KAMs with one of the two wedged insole types, but this strategy will ensure that all participants in the KAM-R and waitlist control groups are all biomechanically suitable for a wedged insole intervention. Thus, neither group will contain biomechanical non-responders to wedged insoles.

Knowledge of the participant’s KAM change is required to ensure that the study sample consists entirely of those who experience a reduction in KAMs with at least one insole type. We will randomize patients to either the KAM-R or control group prior to knowing the KAM results and prior to any data collection. After collecting biomechanical data, KAM results will be assessed to verify that the participant experienced a KAM reduction in at least one wedged insole type. As described previously, in cases where the participant does not experience a KAM reduction with either insole, they will be excluded from the main study. Thus our knowledge of KAM changes will not affect our initial randomization, and only be used as a verification tool. Although researchers will be aware of each participant’s study group following randomization (in order to assign the participants in the KAM-R group an insole to take home), the researchers will be blinded during future survey data collection and data analysis as identifying information will be stripped by an independent researcher and replaced with unique patient ID numbers and numerical values for group allocation. Those involved in statistical analysis will be blinded to the group allocation and outcomes during data analysis. All KAM-R participants will be aware as to which condition they have been assigned, based on knowing that they have received an insole. All control condition participants will be aware that they have not received an insole at baseline, but will receive an insole at 3 months. Thus, the study will not be blinded at baseline data collection; however, follow-up data collection and data analysis will be blinded to the researchers.

The primary outcome measure in this study will be the KOOS pain score (0–100 points). This measure will allow us to identify the overall treatment effect of the wedged insole intervention over the 3-month follow-up period.

Secondary outcome measures include the remaining four KOOS subsection scores (symptoms, function (daily living), function (sport & recreation activities), quality of life; 0–100 points each), PASE (0-400+ points) and UCLA (0–10 points) physical activity scores, peak KAM (Nm), KAM impulse (Nms), varus thrust (mm), adiposity (%), footwear comfort (0–100 mm) and the general intervention adherence survey. In addition, we will use Kellgren-Lawrence grade (grade 1–3) in secondary analyses.

Sample size calculations were performed based on our primary outcome variable, the KOOS pain score, with a desired power of 80% and a significance level of 0.05 using the statistical software Stata version 11.2. Given the early stage of research concerning the effects and interactions of load reduction, and different OA subtypes on treatment outcome, we chose to only conduct sample size calculations on a repeated measures basis for within group comparisons. Based on the results of this trial, sample sizes may be expanded to increase statistical power in order to make comparisons between the control and KAM-R group.

The minimum perceptible clinical improvement and minimum detectible change in KOOS score has been reported to be about 13.4 points [33, 41]. Using a baseline KOOS pain score of 57.3 points [42] and a standard deviation of 15 points [41], to detect a minimal clinically important improvement of 13.4 points within each group, 20 participants are required in each group, or 40 total participants. To account for a potential dropout rate of 15%, we will recruit a minimum of 46 participants.

Descriptive analyses will be performed on all variables. Outcome measures data will be plotted to assess trends of change over the 3 months.

This randomized controlled trial will help provide insight to optimize clinical management strategies for patients with mild to moderate medial knee OA by evaluating the effects of a knee joint load-reducing intervention. These results could have immediate implications on clinical practice, specifically the efficacy of knee joint load-reducing insoles, and also help identify new areas of study for basic scientists.