Knee osteoarthritis (OA) is a frequent degenerative disease representing an important health and economic burden. Symptomatic medical treatments available include intra-articular (IA) injections of corticosteroids (GC) but their efficacy and safety profile are debated.

We performed a systematic literature review (SLR) and a meta-analysis (MA) of randomized controlled trials (RCTs) assessing the effect of IA GC injections for knee OA. The effect of the interventions on pain and function was extracted from the single studies and pooled. Standardized mean differences (SMD) are reported.

Of 520 studies screened, 23 were included in the SLR and 15 subsequently included in the MA. IA GC showed a trend towards a superior effect compared to control on both pain (SMD −0.61 (95% CI: −1.25, 0.03)) and function (SMD −1.02 (95% CI: −2.14, 0.10)) in short term follow-up (≤6 weeks), while long term follow-up (≥24 weeks) analysis showed a trend towards superiority of controls (IA HA, IA NSAID, physiotherapy) for pain (SMD 0.68 (95% CI: −0.11, 1.47)) and function (SMD 0.88 (95% CI: −0.36, 2.12). There were no differences between interventions in medium term (>6 weeks &<24 weeks).

In this work, IA GC injections reduced pain and improved function early after administration (≤6 weeks) compared to placebo; while this result was no longer statistically significant with other comparators (IA hyaluronic acid or physiotherapy). Other interventions seem to be more efficient in the long term (≥24 weeks) but this effect was largely driven by single studies with large effect sizes.