SAT0565 Adjusting for the intra-articular placebo effect in knee osteoarthritis therapies

Background Currently, there is a large debate regarding the appropriateness of intra-articular (IA)-saline injection as a “placebo” comparator in knee osteoarthritis (OA) trials and meta-analyses. There is substantial evidence to suggest that the injection of saline into the joint is not without treatment effect. Objectives This study aimed to assess the current literature’s estimates of the IA-saline treatment effect against a range of appropriate minimal clinically important difference (MCID) values to identify if IA-saline provides a therapeutic effect that is not indicative of a null-effect. Methods The treatment effect estimates of IA-saline and topical placebo for knee OA pain, relative to oral placebo, were derived from a published network meta-analysis Bannuru et al, 2015 and compared across a range of plausible MCID values. Effect estimates of pharmacologic knee OA treatments were also extracted from recent high-quality meta analyses, and the effect of IA-saline or topical placebo was used as an adjustment for meta-analyses that used IA-saline or topical placebo as a comparator, respectively. This was done to estimate the therapeutic effect of these treatment options when compared to a truly null treatment option, as IA-saline has been shown to not be a null-effect intervention. The unadjusted and adjusted treatment effect values were compared across an MCID value range of 0.2 to 0.5 standard deviation units representing an effect size range. This range was used to determine if the adjustment for the effect of IA-saline would affect the potential clinical interpretation of the previously published meta-analysis results. Results IA-saline provides a therapeutic benefit that is potentially clinically meaningful to patients based on a range of MCID values (figure 1). Across the same range of MCID values, the effects of high molecular weight IA-hyaluronic acid products and IA-corticosteroid treatments were not conclusively clinically significant. However, when reassessing these treatments while accounting for the treatment effect of IA saline, they were found to have a clinically significant effect at the strictest MCID value (figure 2). Conclusions The use of IA-saline as a placebo treatment within RCTs of IA injectable therapies is inappropriately underestimating the true effect of these treatments. When the potential therapeutic effect that IA-saline demonstrates is accounted for, these IA therapies show a considerably larger therapeutic effect. Reference [1] Bannuru RR, et al. Ann Intern Med2015;163:365–372. Disclosure of Interest I. Simsek Shareholder of: Samumed, LLC, Employee of: Samumed, LLC, T. Phalen Shareholder of: Samumed, LLC, Employee of: Samumed, LLC, A. Bedenbaugh Shareholder of: Samumed, LLC, Employee of: Samumed, LLC, M. Phillips Grant/research support from: Samumed, LLC, J. Tambiah Shareholder of: Samumed, LLC, Employee of: Samumed, LLC