Twenty-two RCTs (n=2,902) were included in the meta-analyses.
The overall quality of the primary studies was considered unsatisfactory.
Pain at rest (8 RCTs).
A beneficial effect of hyaluronic acid compared with the control group was shown at 2 to 6 weeks, as assessed by a visual analogue scale (WMD -8.7 mm, 95% CI: -17.2, -0.2); due to a large degree of unexplained heterogeneity (94%), the summary estimate was not used. Individual study WMD estimates ranged from -29.00 to 8.30 mm. No statistically significant effect was found between hyaluronic acid and controls at 10 to 14 weeks, 22 to 30 weeks, or at 44 to 60 weeks in 2 high-quality trials. Poorer quality trials were shown to favour the intervention.
Pain during or immediately after exercise (9 RCTs).
No statistically significant difference between the treatment groups was shown at 2 to 6 weeks, as assessed by a visual analogue scale (WMD -3.8 mm, 95% CI: -9.1, 1.4); unexplained heterogeneity was high (81%). A statistically significant effect in favour of hyaluronic acid (WMD -4.2 mm, 95% CI: -7.5, -0.8) was shown with the removal of one trial with a significant interaction between treatment effect and severity of osteoarthritis. A statistically significant effect in favour of hyaluronic acid was also found at 10 to 14 weeks (WMD -4.3 mm, 95% CI: -7.6, -0.9) and at 22 to 30 weeks (WMD -7.1 mm, 95% CI: -11.8, -2.4); no unexplained heterogeneity was shown. Only one trial at 44 to 60 weeks was found; this showed no treatment effect. One trial of high quality was found; this was not found to influence the effect size at any point.
Joint function (9 RCTs).
No statistically significant differences between the groups were found at 2 to 6 weeks' follow-up (SMD 0.00, 95% CI: -0.23, 0.23), 10 to 14 weeks' follow-up (SMD -0.11, 95% CI: -0.31, 0.09), or at 22 to 30 weeks' follow-up (SMD -0.16, 95% CI: -0.45, 0.13); unexplained heterogeneity was found (66%, 59% and 62%, respectively). Two trials followed patients until 44 to 60 weeks, but no significant difference between the treatment groups was shown. No statistically significant between-group differences were found at 2 to 6 weeks' follow-up and 10 to 14 weeks' follow-up when only high-quality trials were considered. Poorer quality trials were shown to favour the intervention.
Adverse events (15 RCTs).
Adverse events occurred more frequently among individuals receiving hyaluronic acid, although this difference was not statistically significant (RR 1.08, 95% CI: 1.01, 1.15).
The funnel plot analysis indicated an over-selection or over-reporting of trials reporting adverse events in people treated with hyaluronic acid; this could be explained by lack of blinding.
The effect size did not appear to be associated with molecular mass.