Nine double-blind RCTs (at least 1,000 patients) were included. The 9 studies considered 11 separate trials.
Five studies were classified as good quality. Only 2 studies reported an intention-to-treat analysis.
At 1 week, hyaluronic acid was associated with a significant improvement in pain compared with placebo; the mean difference was 4.4 (95% confidence interval, CI: 1.1, 7.2) and significant heterogeneity was found (p<0.001). After restricting the analysis to good-quality RCTs, there was no significant difference between treatments but the significant heterogeneity persisted (p<0.001).
At 5 to 7 weeks, hyaluronic acid was associated with a significant improvement in pain compared with placebo for all studies (mean difference 17.6, 95% CI: 7.5, 28.0); significant heterogeneity was found (p<0.001). It was also associated with a significant improvement in pain for good-quality studies (mean difference 7.2, 95% CI: 2.4, 12.0; based on 2 studies); no significant heterogeneity was found (p=0.688).
At 8 to 12 weeks, hyaluronic acid was associated with a significant improvement in pain compared with placebo for all studies (mean difference 18.1, 95% CI: 6.3, 29.9); significant heterogeneity was found (p<0.001). It was also associated with a significant improvement in pain for good-quality studies (mean difference 7.1, 95% CI: 3.0, 11.3; based on 3 studies); no significant heterogeneity was found (p=0.332).
At 15 to 22 weeks, there was no significant difference in pain between hyaluronic acid and placebo for all studies (3 good-quality studies); significant heterogeneity was found (p<0.001).
Egger's test suggested the absence of publication bias (p=0.096).
The meta-regression showed no significant association between outcome and type of pain. Studies using hylan GF-20 showed significantly better outcomes than studies using hyaluronon at 5 to 7 and 8 to 12 weeks. Poorer quality studies showed larger treatment effects than good-quality studies, but the difference was only significant at 1 week.