Seven RCTs (n=606 patients, 610 knees) were included in the meta-analysis. Overall, only two trials were considered to be higher quality. One trial was double-blinded, one reported allocation concealment and five used intention-to-treat analysis.
Statistically significant pooled effects were found in favour of corticosteroids at week two (-0.39, 95% CI -0.65 to -0.12). There was no statistically significant difference between the treatments at week four. At week eight, the trend was in favour of hyaluronic acid (0.22, 95% CI -0.05 to 0.49); this became statistically significant at week 12 (0.35, 95% CI 0.03 to 0.66) and week 26 (0.39, 95% CI 0.18 to 0.59). The results were consistent in sensitivity analysis and meta-regression showed no significant interactions.
There was moderate statistical heterogeneity in all analyses (I2 range: 37% to 49%).