Evidence coverage: A BMJ network meta-analysis restricted to large randomized trials (>200 patients), the highest-quality pooled evidence the glucosamine question has attracted.
Statistically almost something; clinically nothing. certainty: Moderate-high (network meta-analysis of large randomized trials)
Across 10 trials in 3803 patients, glucosamine’s pain effect versus placebo (−0.4 cm VAS) never reaches the −0.9 cm clinically-important threshold, joint-space changes are minute with intervals overlapping zero, and industry-independent trials show smaller effects than funded ones.
- Pain versus placebo: −0.4 cm (95% credible interval −0.7 to −0.1) on a 10 cm scale (source-reported)
- No estimate crosses the prespecified −0.9 cm clinical-importance boundary (source-reported)
- Joint-space narrowing: minute changes, credible intervals overlapping zero (source-reported)
- Independent trials show smaller effects than commercially funded ones (P = 0.02 interaction) (source-reported)
Product-form chaos (sulfate vs hydrochloride, salts vs hydrolysates) muddies cross-trial comparison — but no formulation escapes the category-level network result.