Intra-Articular Hyaluronic Acid in the Symptomatic Treatment of Knee Osteoarthritis: A Meta-Analysis of Single-Injection Products

BACKGROUND: Viscosupplementation of the synovial fluid with intra-articular hyaluronic acid (IA HA) is a well-known symptomatic treatment of knee osteoarthritis. The question arises whether a monoinjection (ie, single injection) could be as efficient as multi-injection (ie, 3–5 injections) regimens. METHODS: A meta-analysis of published studies relating to IA HA monoinjection trials was performed. The efficacy criterion was the Western Ontario and MacMaster Universities pain subscore. Any study design was accepted, from randomized control trials to single-arm observational open-label studies. An extensive search was performed using PubMed, Google, Google Scholar, and references found in recent meta-analyses, for all articles published before end of April 2018. Population profiles were analyzed in terms of age, sex, body mass index (BMI), and Kellgren-Lawrence (KL) radiology grades. Results of intra-articular single injection of placebo were collected to create a database allowing post hoc comparisons. Each IA HA study arm was compared to an IA placebo arm (either pooled or not), to present a similar KL profile controlled with the χ(2) test. The effect size (ES) (95% CI) and P values were calculated and synthesized for each follow-up visit at 1, 2, 3, and 6 months. In parallel, a global approach was used to represent the variations from baseline for each group or subgroup studied. RESULTS: From 1547 citations, 28 studies were included in the meta-analysis, representing 4129 patients treated with monoinjection: 3360 received IA HA and 769 patients received IA placebo. The mean (SD) IA HA patient was 61.2 (9.6) years, 63% women, BMI 28.0 (4.1), 47% KL III, and 3% KL IV. A good placebo KL profile matching was obtained for 26 of the 31 IA HA arms. For the whole IA HA population, ES = 0.30 (95% CI, 0.25–0.35) at 3 months and ES = 0.39 (95% CI, 0.33–0.44) at 6 months. In a restricted analysis, after removal of outliers, poorly KL matched and active arms <30 patients, results remained unchanged, ES = 0.29 (95% CI, 0.23–0.34) and ES = 0.40 (95% CI, 0.34–0.45) at 3 and 6 months respectively, whilst heterogeneity was improved. CONCLUSIONS: There are certainly limits to the post hoc placebo comparison method, for individual studies. But for each synthesis per subgroup or group, the results were properly confirmed using multiple statistical approaches and weighing methods. This meta-analysis suggests that monoinjections produce results similar to multi-injections of IA HA in terms of pain relief in the treatment of knee osteoarthritis. (Curr Ther Res Clin Exp. 2019; 80:XXX–XXX)