Glucosamine and chondroitin failed their largest trial, and the funder mattered

In the largest randomised trial of glucosamine and chondroitin ever run, neither supplement, alone or together, reduced knee pain significantly more than placebo [s1]. A later network meta-analysis of the biggest trials found effects that never reached the threshold set in advance for clinical importance — and found that industry-independent trials produced smaller effects than commercially funded ones [s2].

That second finding is the one worth carrying away. It is not a claim that the industry-funded results were dishonest. It is a measured, statistically significant interaction, and it is the kind of check that almost never appears in the marketing of a supplement category.

GAIT

The Glucosamine/chondroitin Arthritis Intervention Trial randomly assigned 1,583 patients with symptomatic knee osteoarthritis to 1,500 mg of glucosamine daily, 1,200 mg of chondroitin sulfate daily, both, 200 mg of celecoxib daily, or placebo, for 24 weeks [s1]. It was multicentre, double-blind, and controlled against both placebo and an active drug — a design that lets a reader check whether the trial was capable of detecting an effect at all. Patients were stratified by baseline pain severity, with 1,229 in the mild stratum and 354 in the moderate-to-severe stratum; mean age was 59 and 64% were women [s1]. Up to 4,000 mg of acetaminophen daily was permitted as rescue analgesia [s1].

The primary outcome was a 20% decrease in knee pain from baseline to week 24. The placebo response rate was 60.1% [s1]. Against that, the response rate was 3.9 percentage points higher with glucosamine (P = 0.30), 5.3 points higher with chondroitin (P = 0.17), and 6.5 points higher with the combination (P = 0.09) [s1]. The celecoxib control arm was 10.0 points higher than placebo (P = 0.008) [s1] — which is how we know the trial could detect a real analgesic when one was present.

One exploratory finding is the one supplement marketing has leaned on ever since. In the moderate-to-severe pain stratum, the response rate was 79.2% with combined therapy against 54.3% with placebo (P = 0.002) [s1]. The trial's authors label this an exploratory analysis in a subgroup of 354 patients that was not the primary comparison [s1]. It is a hypothesis, not a result.

Note the placebo response rate: 60.1% of patients on nothing met the trial's improvement threshold [s1]. In a condition with that much movement in the control arm, an active treatment has to be substantial to separate from it.

The pooled analysis, and the funding interaction

A network meta-analysis published in the BMJ restricted itself to large trials — more than 200 patients each — in knee or hip osteoarthritis, and combined direct within-trial comparisons with indirect evidence in a Bayesian model. It included 10 trials covering 3,803 patients [s2].

The minimal clinically important difference was prespecified at 0.9 cm on a 10 cm visual analogue scale [s2]. Against placebo, the overall difference in pain intensity was −0.4 cm (95% credible interval −0.7 to −0.1) for glucosamine, −0.3 cm (−0.7 to 0.0) for chondroitin and −0.5 cm (−0.9 to 0.0) for the combination [s2]. For none of the three did the credible interval cross the boundary of clinical importance [s2]. Differences in minimal joint space width — the structural outcome, the one that would justify the phrase "protects your joints" — were described as minute, with credible intervals overlapping zero [s2].

And industry-independent trials showed smaller effects than commercially funded trials, with P = 0.02 for the interaction [s2]. The authors' conclusion is blunt: health authorities and health insurers should not cover the costs of these preparations, and new prescriptions to untreated patients should be discouraged [s2].

Added to something that does work

A different question is whether the supplements add anything on top of exercise, which is the first-line treatment for knee osteoarthritis. A meta-analysis of six randomised trials covering 297 participants — rated good methodological quality, with a mean PEDro score of 8.2 — tested exactly that [s3].

Adding glucosamine, or glucosamine with chondroitin, to an exercise programme produced no significant benefit over exercise alone on WOMAC pain (SMD −0.18, 95% CI −0.47 to 0.11, P = 0.23), VAS pain (SMD −0.34, −0.85 to 0.17, P = 0.20) or physical function (SMD −0.13, −0.95 to 0.69, P = 0.76) [s3]. The sample is small and the confidence intervals are wide enough to accommodate a modest effect, but the point estimates are unimpressive and none reaches significance.

Where the current literature sits

The supplements have not disappeared from comparative analyses, and a reader searching will find recent papers that rank them respectably. A 2026 network meta-analysis of 23 randomised trials comparing metformin, probiotics, krill oil, celecoxib and glucosamine with chondroitin in knee osteoarthritis put celecoxib first on VAS pain improvement (mean difference −0.48, 95% CI −0.78 to −0.17, SUCRA 0.93) and on WOMAC total score (SUCRA 0.912), with the glucosamine–chondroitin combination second at 0.756 [s4].

A second-place ranking sounds better than it is. The authors state that network certainty for most of these auxiliary comparisons was low or very low, driven by severe imprecision and the absence of direct head-to-head trials, and warn explicitly that the SUCRA rankings should be interpreted cautiously in clinical settings [s4]. A ranking statistic will always produce an order, including when the underlying differences are noise.

What this leaves

The safety record is the one consistent positive: adverse events in GAIT were mild, infrequent and evenly distributed across groups [s1], and later reviews describe both compounds as well tolerated. What the evidence does not support is the efficacy claim — not for pain at a clinically meaningful size [s1] [s2], not for joint structure [s2], and not as an addition to exercise [s3].

What to watch

The funding interaction in the BMJ analysis is the finding most worth revisiting as trials accumulate [s2]. If independent trials continue to produce smaller effects than commercially sponsored ones in this category, that is a fact about the evidence base rather than about the molecules, and it applies well beyond these two supplements.

This article is informational and is not medical advice. Doses stated here are trial parameters, not recommendations.

Sources

• Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis — New England Journal of Medicine, 2006-02-22
• Effects of glucosamine, chondroitin, or placebo in patients with osteoarthritis of hip or knee: network meta-analysis — BMJ, 2010-09-16
• Effects of adding glucosamine or glucosamine combined with chondroitin to exercise on pain and physical function in adults with knee osteoarthritis — European Journal of Translational Myology, 2023-11-23
• Network meta-analysis of metformin, probiotics, krill oil, celecoxib, and glucosamine/chondroitin in the treatment of knee osteoarthritis — Frontiers in Medicine, 2026-07-22