Association between changes in the knee adduction moment and changes in knee pain and function in response to biomechanical interventions for medial knee osteoarthritis; a systematic review

Abstract

Purpose: Knee osteoarthritis (KOA) has been closely associated with knee biomechanics, mainly during gait. The knee adduction moment (KAM) is widely accepted as a surrogate measure of load in knee OA and therefore a target for therapeutic interventions. While reducing the KAM is the target of several biomechanical interventions for medial knee OA patients, there is a lack of understanding of the benefits of KAM reductions in terms of improving pain and function. The objective of this systematic review is to investigate the relationship between KAM changes induced by non-surgical biomechanical interventions and changes in pain/function, in KOA patients. Methods: We considered eligible all prospective randomised controlled trials using any type of non-surgical biomechanical intervention aimed to reduce KAM in KOA patients. We excluded trials focusing on surgical, pharmacological, physical exercise (i.e. muscle strengthening, proprioception, physical activity (i.e. yoga, Nordic-walking, tai-chi)) physical agents (i.e. laser-therapy, ultra-sound), manual therapy and/or alternative medicines (i.e. acupuncture, moxibustion) interventions. We searched 9 different databases using a formula prepared by the librarians of the University of Lausanne- Faculty of Medicine and Biology. We included all the literature published in the last 30-years (1.1.1988-14.6.2018). Given the possibilities of modern technologies for translation, we did not consider language restrictions in our search. Two authors firstly scrutinised the blinded titles and abstracts of all identified publications and subsequently evaluated all selected articles in full-text using specific software. Disagreements were resolved by discussion and mediated by a third reviewer. Data extraction was performed by the main author. The methodological quality of RCTs for between-group comparisons was assessed through the Cochrane tool for risk of bias tool. We used Cohen’s d effect size to standardise the differences between the means and graphically represent the association between changes in the independent and dependent variables (positive values indicating an improving in KAM, pain and function). The association between the variables was evaluated through logistic regression statistics (R2). Results: A total 2452 references were screened and 119 full-text studies were assessed for eligibility. Of these, 14 papers reporting outcomes of 11 studies were retained for analysis. After further scrutiny 8 trials qualified to the present systematic review, comprising a total of 17 arms and 9 interventions. The final selection involved a total of 470 patients, mean age 51 to 66 years old, diagnosed with medial KOA Kellgren & Lawrence I to IV. Arm size (n) ranged from 6 to 80. The interventions used consisted on different modalities of insoles, knee braces, shoes and gait retraining. For studies using an orthosis as an intervention, the range of reported usage ranged from 7,33 (SD 0,88) to 8,26 (SD not reported) hours daily. Two studies had a crossover design while 6 remaining had a parallel design. One paper focused on a single gender (female) and the remaining were having mixed genders. The quality assessment showed overall low risk of bias, with exception to items “performance bias” and “detection bias” as most of the studies were not double blinded. The WOMAC sub-scales for pain and function were the most commonly reported outcome measurements for pain and function, respectively. VAS-pain was the second most commonly used measurement of knee pain. Pain was measured in all arms of selected trials, function in 12 of the 17. The effect sizes for changes in KAM ranged from -1,36 to 3,81, for pain from 0,00 to 6,80 and for function from -0,35 to 7,95. We could not observe an association between the effect size of the changes in KAM and the effect size of changes in WOMAC pain or function (Figures 1 and 2). There is an association between the effect size of changes in KAM and the effect size of changes in VAS-pain with a logistic regression indicating a R2 of 0,67 (Figure 3). Conclusions: There is insufficient number of high-quality prospective data to establish a formal association between changes in KAM induced by specific non-surgical, non-pharmacological biomechanical interventions and changes in knee pain or knee function. The available evidence suggests that there is an association between changes in KAM and changes in pain rated on the VAS. We did not find any association between changes in KAM and changes in function measured by the WOMAC. The small number of trials matching our criteria and the amount of reported data makes it very difficult to perform sensitivity analysis. Further primary research on this subject is deemed necessary.View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)