Thirteen RCTs (n=1,983) were included.
None of the RCTs reported the method of randomisation. The withdrawal rates ranged from 1 to 23%. The funnel plot was asymmetrical, suggesting the possibility of publication bias.
Topical NSAIDs versus placebo.
Compared with placebo, topical NSAIDs significantly improved pain relief at 1 week (ES 0.41, 95% CI: 0.16, 0.66; 7 RCTs, n=1,000) and 2 weeks (ES 0.40, 95% CI: 0.15, 0.65; 6 RCTs, n=893). Statistically significant heterogeneity was found for both meta-analyses. There was no statistically significant difference at 3 or 4 weeks.
Topical NSAIDs significantly improved function compared with placebo at 1 week (ES 0.37, 95% CI: 0.20, 0.53; 4 RCTs, n=566) and 2 weeks (ES 0.35, 95% CI: 0.19, 0.53; 4 RCTs, n=540). Statistically significant heterogeneity was found for both meta-analyses. There was no statistically significant difference at 3 or 4 weeks.
Topical NSAIDs significantly increased the clinical response rate at 1 week but not at 4 weeks. Adverse event rates were similar for topical NSAIDs and placebo (see URL).
Topical NSAID versus oral NSAIDs.
Topical NSAIDs were less effective than oral NSAIDs for pain, stiffness and function at 1, 2, 3 and 4 weeks. The difference was statistically significant only at week 1 for pain (ES -0.38, 95% CI: -0.66, -0.10; 1 RCT, n=208) and function (ES -0.32, 95% CI: -0.60, -0.04; 1 RCT, n=208). There was no difference in clinical response rate between topical and oral NSAIDs (see URL).
Compared with oral NSAIDs, topical NSAIDs reduced the proportion of patients with any adverse event, withdrawal due to adverse events and gastrointestinal events, but statistically significantly increased the proportion of patients with local adverse events (rate ratio 5.29, 95% CI: 1.14, 24.51).
In the sensitivity analyses, only the type of topical NSAID used significantly influenced the ES.