Exploring reasons for the observed inconsistent trial reports on intra-articular injections with hyaluronic acid in the treatment of osteoarthritis: Meta-regression analyses of randomized trials

Objective: The aim was to identify factors explaining inconsistent observations concerning the efficacy of intra-articular hyaluronic acid compared to intra-articular sham/control, or non-intervention control, in patients with symptomatic osteoarthritis, based on randomized clinical trials (RCTs). Methods: A systematic review and meta-regression analyses of available randomized trials were conducted. The outcome, pain, was assessed according to a pre-specified hierarchy of potentially available outcomes. Hedges's standardized mean difference [SMD (95% CI)] served as effect size. REstricted Maximum Likelihood (REML) mixed-effects models were used to combine study results, and heterogeneity was calculated and interpreted as Tau-squared and I-squared, respectively. Results: Overall, 99 studies (14,804 patients) met the inclusion criteria: Of these, only 71 studies (72%), including 85 comparisons (11,216 patients), had adequate data available for inclusion in the primary meta-analysis. Overall, compared with placebo, intra-articular hyaluronic acid reduced pain with an effect size of -0.39 [-0.47 to -0.31; P < 0.001], combining very heterogeneous trial findings (I-2 = 73%). The three most important covariates in reducing heterogeneity were overall risk of bias, blinding of personnel and trial size, reducing heterogeneity with 26%, 26%, and 25%, respectively (Interaction: P <= 0.001). Adjusting for publication/selective outcome reporting bias (by imputing "null effects") in 24 of the comparisons with no data available reduced the combined estimate to -0.30 [-0.36 to -0.23; P < 0.001] still in favor of hyaluronic acid. Conclusion: Based on available trial data, intra-articular hyaluronic acid showed a better effect than intra-articular saline on pain reduction in osteoarthritis. Publication bias and the risk of selective outcome reporting suggest only small clinical effect compared to saline. (C) 2016 Elsevier Inc. All rights reserved.