Background: Osteoarthritis of the knee affects up to 10% of the elderly population. The condition is frequently treated by intra-articular injection of hyaluronic acid. We performed a systematic review and meta-analysis of randomized controlled trials to assess the effectiveness of this treatment.

Methods: We searched MEDLINE, EMBASE, CINAHL, BIOSIS and the Cochrane Controlled Trial Register from inception until April 2004 using a combination of search terms for knee osteoarthritis and hyaluronic acid and a filter for randomized controlled trials. We extracted data on pain at rest, pain during or immediately after movement, joint function and adverse events.

Results: Twenty-two trials that reported usable quantitative information on any of the predefined end points were identified and included in the systematic review. Even though pain at rest may be improved by hyaluronic acid, the data available from these studies did not allow an appropriate assessment of this end point. Patients who received the intervention experienced a reduction in pain during movement: the mean difference on a 100-mm visual analogue scale was -3.8 mm (95% confidence interval [CI] -9.1 to 1.4 mm) after 2-6 weeks, -4.3 mm (95% CI -7.6 to -0.9 mm) after 10-14 weeks and -7.1 mm (95% CI -11.8 to -2.4 mm) after 22-30 weeks. However, this effect was not compatible with a clinically meaningful difference (expected to be about 15 mm on the visual analogue scale). Furthermore, the effect was exaggerated by trials not reporting an intention-to-treat analysis. No improvement in knee function was observed at any time point. Even so, the effect of hyaluronic acid on knee function was more favourable when allocation was not concealed. Adverse events occurred slightly more often among patients who received the intervention (relative risk 1.08, 95% CI 1.01 to 1.15). Only 4 trials explicitly reported allocation concealment, had blinded outcome assessment and presented intention-to-treat data.

Interpretation: According to the currently available evidence, intra-articular hyaluronic acid has not been proven clinically effective and may be associated with a greater risk of adverse events. Large trials with clinically relevant and uniform end points are necessary to clarify the benefit-risk ratio.