Background Several systematic reviews have studied the effects of exercise therapy in people with rheumatoid arthritis (RA). Although overlapping reviews on the same topic are often a waste of money and effort, we used a new approach to pool only the best available evidence. It is common practice for systematic reviews to focus just on methodological validity as a marker for quality. However, perhaps we should also address the therapeutic validity (i.e., potential for effectiveness) of the studied interventions as a marker for quality. After all, interventions with low potential for effectiveness are more likely to result in negative study findings and therefore should not be pooled in a best-evidence synthesis.
Objectives The purpose of this study was to assess the effectiveness of therapeutically valid exercise therapy interventions in individuals with RA on physical functioning.
Methods RCTs that evaluate the effectiveness of structured, supervised exercise therapy on self-reported and observed physical functioning in adults with RA were identified through four electronic databases (through Oct 2013). Hand programs were excluded. Two reviewers extracted data and assessed methodological and therapeutic validity (using the PEDro and CONTENT scale  respectively; scores of ≥6 reflect high validity). Studies with both methodological and therapeutic validity were included in the best evidence synthesis. A sensitivity analysis was performed, in which we included all papers, regardless of their quality. Effects were pooled using random-effect models.
Results Of the 1181 articles retrieved, 10 studies describing 11 interventions (612 patients) were included. One study demonstrated therapeutic validity (median score (IQR) 3 (2-5)), six demonstrated methodological validity (6 (5-6)), and none demonstrated both. The latter made it impossible to perform the planned best-evidence synthesis. The sensitivity analysis on the effectiveness of low quality studies showed that exercise therapy was associated with self-reported (SMD (95%>CI), 0.19 (0.03; 0.36)) and observed (0.59 (0.01; 1.18)) functional improvements on directly after the intervention. The largest association was found for weight exercise and improvements in short-term observed functioning (-1.00 (-1.46;-0.54)). There were insufficient studies to pool long-term outcomes; individual studies showed no significant benefits.
Conclusions Low quality evidence suggests that therapeutic exercise has beneficial effects on short-term physical functioning for individuals with RA. However, the magnitude of the effect of exercise therapy is limited. Perhaps the latter is explained by the poor therapeutic validity of the exercise programs. One of the major concerns was the lack of a rationale for the exercise interventions, resulting in unclear patient selection (i.e. most studies included patients without functional limitations) and unclear exercise dosing (i.e. mostly low intensity without a clear explanation). The latter should at the very least be addressed in future studies on the (long-term) effectiveness of therapeutic exercise in people living with RA.
Siontis, et al. BMJ. 2013:19;347.
Hoogeboom, et al. PLoS One. 2012;7(5):e38031.
Acknowledgements Thank you to the Dutch Arthritis Foundation for financially supporting this study.
Disclosure of Interest : None declared