Exercise for lower limb osteoarthritis: systematic review incorporating trial sequential analysis and network meta-analysis

We would like to thank Ida Svege and colleagues for their appreciation of our systematic review, and their important comments regarding the imbalance of evidence for the effectiveness of exercise for lower limb osteoarthritis, which - as we emphasised in our review - is dominated by trials in knee osteoarthritis. This imbalance was indeed compounded in our meta-analysis by the fact that we had to exclude two of the small number of available trials in patients with hip OA (Fernandes et al. (1) and Juhakoski et al. (2)). Exclusion of these two trials was not due to diagnostic criteria used for selecting study participants, as all trials using accepted, standardised clinical or radiological criteria for osteoarthritis were eligible for inclusion. We understand the confusion as we indeed referred to the ACR guidelines in our paper, but this was meant to be an example which we should have clarified.

The reason for exclusion of these two trials was related to the fact that both these trials investigated exercise in addition to another intervention (GP care or patient education) versus the GP care or patient education alone. These designs are highly useful in direct comparisons, as they provide an estimate of the effectiveness of exercise as an add-on to another commonly used intervention for osteoarthritis. In a network meta-analysis, however, where direct as well as indirect comparisons are included in the analysis, it is difficult to disentangle the effect of exercise from such combined treatments. Therefore, we decided to exclude trials which explicitly stated that exercise was used in combination with another intervention package, and also excluded such arms from the analysis for trials with more than two treatment arms. We realise this distinction can be difficult to make in circumstances where investigators have combined exercise with other co-interventions, but designed and presented this as a single package of care (e.g. Hurley et al. (3)). However, the decision to exclude trials/arms that explicitly offered combinations of treatments was made a priori and applied to all potentially eligible trials, regardless of the location of osteoarthritis.

We fully agree with the authors of this letter that there is a strong need for high quality trials in patients with hip osteoarthritis. A network meta-analysis focusing on exercise interventions was not yet feasible for hip osteoarthritis, but the two highlighted trials do provide an important contribution to the growing body of evidence. Future systematic reviews and meta-analysis should provide more conclusive evidence on the effectiveness of exercise interventions for hip osteoarthritis. With the impending publication of several new trials there will also be an opportunity for earlier application of sequential analysis which may help investigators and funders recognise when sufficient evidence on the basic question of effectiveness vs no exercise control has been accrued for hip osteoarthritis.

1. Fernandes L, Storheim K, Sandvik L, Nordsletten L, Risberg MA. Efficacy of patient education and supervised exercise vs patient education alone in patients with hip osteoarthritis: a single blind randomized clinical trial. Osteoarthritis Cartilage 2010;18(10):1237-1243.
2. Juhakoski R, Tenhonen S, Malmivaara A, Kiviniemi V, Anttonen T, Arokoski JP. A pragmatic randomized controlled study of the effectiveness and cost consequences of exercise therapy in hip osteoarthritis. Clin Rehabil 2011;25(4):370-83.
3. Hurley MV, Walsh NE, Mitchell HL, Pimm TJ, Patel A, Williamson E, Jones RH, Dieppe PA, Reeves BC. Clinical effectiveness of a rehabilitation program integrating exercise, self-management, and active coping strategies for chronic knee pain: a cluster randomized trial. Arthritis Rheum 2007;57(7):1211-9.