OARSI recommendations for the management of hip and knee osteoarthritis: part III: Changes in evidence following systematic cumulative update of research published through January 2009.

OBJECTIVE To update evidence for available therapies in the treatment of hip and knee osteoarthritis (OA) and to examine whether research evidence has changed from 31 January 2006 to 31 January 2009. METHODS A systematic literature search was undertaken using MEDLINE, EMBASE, CINAHL, AMED, Science Citation Index and the Cochrane Library. The quality of studies was assessed. Effect sizes (ESs) and numbers needed to treat were calculated for efficacy. Relative risks, hazard ratios (HRs) or odds ratios were estimated for side effects. Publication bias and heterogeneity were examined. Sensitivity analysis was undertaken to compare the evidence pooled in different years and different qualities. Cumulative meta-analysis was used to examine the stability of evidence. RESULTS Sixty-four systematic reviews, 266 randomised controlled trials (RCTs) and 21 new economic evaluations (EEs) were published between 2006 and 2009. Of 51 treatment modalities, new data on efficacy have been published for more than half (26/39, 67%) of those for which research evidence was available in 2006. Among non-pharmacological therapies, ES for pain relief was unchanged for self-management, education, exercise and acupuncture. However, with new evidence the ES for pain relief for weight reduction reached statistical significance, increasing from 0.13 [95% confidence interval (CI) -0.12, 0.36] in 2006 to 0.20 (95% CI 0.00, 0.39) in 2009. By contrast, the ES for electromagnetic therapy which was large in 2006 (ES=0.77, 95% CI 0.36, 1.17) was no longer significant (ES=0.16, 95% CI -0.08, 0.39). Among pharmacological therapies, the cumulative evidence for the benefits and harms of oral and topical non-steroidal anti-inflammatory drugs, diacerhein and intra-articular (IA) corticosteroid was not greatly changed. The ES for pain relief with acetaminophen diminished numerically, but not significantly, from 0.21 (0.02, 0.41) to 0.14 (0.05, 0.22) and was no longer significant when analysis was restricted to high quality trials (ES=0.10, 95% CI -0.0, 0.23). New evidence for increased risks of hospitalisation due to perforation, peptic ulceration and bleeding with acetaminophen >3g/day have been published (HR=1.20, 95% CI 1.03, 1.40). ES for pain relief from IA hyaluronic acid, glucosamine sulphate, chondroitin sulphate and avocado soybean unsponifiables also diminished and there was greater heterogeneity of outcomes and more evidence of publication bias. Among surgical treatments further negative RCTs of lavage/debridement were published and the pooled results demonstrated that benefits from this modality of therapy were no greater than those obtained from placebo. CONCLUSION Publication of a large amount of new research evidence has resulted in changes in the calculated risk-benefit ratio for some treatments for OA. Regular updating of research evidence can help to guide best clinical practice.