Jan 28

Glucosamine and/or chondroitin for osteoarthritis

Journal reference: Wandel S, Jüni P, Tendal B. Effects of glucosamine, chondroitin, or placebo in patients with osteoarthritis of hip or knee: network meta-analysis. BMJ 2010;341:c4675 [1]

Link: http://dx.doi.org/10.1136/bmj.c4675

Evidence cookie says...

Glucosamine, chondroitin and their combination do not improve pain or reduce joint space loss in osteoarthritis of the hip or knee.

The evidence is convincing. It is very probable that these agents are therapeutically inert for osteoarthritis.  They should not be recommended.

Nevertheless, patients who choose to take glucosamine and/or chondroitin can be reassured that these therapies are safe.

More details:

Article details

Study design:

systematic review, network meta-analysis of randomised controlled trials

Study aim:

determine the effect of glucosamine, chondroitin, or the two in combination on joint pain and on radiological progression of disease in osteoarthritis of the hip and knee

Methods summary:

Literature search:

  • Cochrane Controlled Trials Register
  • Medline
  • Embase
  • and Cumulative Index to Nursing and Allied Health Literature (CINAHL)

Study selection:

  • randomised controlled trials
  • average of at least 100 participants with knee or hip osteoarthritis per arm
  • comparison of chondroitin sulphate, glucosamine sulphate, glucosamine hydrochloride, or the combination of any two with placebo, or head to head
  • two reviewers evaluated trials independent for eligibility
    • disagreements were resolved by consensus
  • trial arms with subtherapeutic doses were excluded, i.e.,:
    • chondroitin < 800 mg/day
    • glucosamine < 1500 mg/day

Outcome measure:

  • Primary: absolute pain intensity reported in any of nine time windows in increments of three months
  • Secondary:
    • changes in the minimum radiographic joint space between baseline and the end of treatment
    • number of individuals withdrawn because of adverse event
    • number of patients experiencing adverse events
  • Trial quality was formally assessed by two reviewers independently
  • Data was extracted by two reviewers independently and cross checked by a third
  • an “extension of multivariable Bayesian hierarchical random effects models for mixed multiple treatment comparisons with minimally informative prior distributions” was used for statistical analysis.
  • pooled results were translated to differences on a 10 cm visual analogue scale (VAS)
    • this was based on a median pooled standard deviation (SD) of 2.5 cm found in large scale osteoarthritis trials
  • prespecified minimal clinically important difference of 0.37 SD, corresponding to 0.9 cm on a 10 cm VAS
    • appropriate referencing for this decision is given

Results summary:

  • 10 trials met inclusion criteria and were included in the meta-analysis
  • total of 3803 patients

Overall difference in pain intensity vs. placebo:

  • Glucosamine:
  • Chondroitin:
    • -0.3 cm (95% credible interval -0.7 to 0.0 cm) on 10 cm VAS
  • Combination:
    • -0.5 cm (95% credible interval -0.9 to 0.0 cm) on 10 cm VAS
  • low heterogeneity between trials: τ2=0.04
  • no pain-time correlation detected

Effect of industry funding vs independent funding had a significant effect on trial results (P = 0.02):

  • Glucosamine:
    • Independent: -0.1 cm (-0.5 to 0.2)
    • Industry: -0.6 cm (-0.9 to -0.3)
  • Chondroitin:
    • Independent:  0.1 cm (-0.3 to 0.6)
    • Industry: -0.4 cm (-0.7 to -0.1)
  • Combination:
    • Independent: -0.3 cm (-0.7 to 0.2)

Effect on joint space narrowing:

  • 6 trials reported data used in the meta-analysis
  • low heterogeneity in the results: τ2=0.04
  • no significant effect detected (clinical or statistical)
  • Glucosamine vs placebo:
    • -0.2 mm (-0.3 to 0.0 mm)
  • Chondroitin vs placebo:
    • -0.1 mm (-0.3 to 0.1 mm)
  • Combination vs placebo:
    • 0.0 mm (-0.2 to 0.2 mm)

Study conclusion:

Compared with placebo, glucosamine, chondroitin, and their combination do not reduce joint pain or have an impact on narrowing of joint space. Health authorities and health insurers should not cover the costs of these preparations, and new prescriptions to patients who have not received treatment should be discouraged.


  • 3803 patients
  • mix of participants examined for chondroitin, glucosamine or both
  • probable excess of female participants but grouped demographic data not given
  • mean age of participants in most included trials in early 60s (median 62 years)
  • most trials for knee osteoarthritis only (8 of 10)

Methodological weaknesses

  • Most included studies only directed at knee osteoarthritis
  • Substantial variation in the length of follow up in the included trials and the average duration if symptoms in the participants (though admittedly there is little heterogeneity in results)

Methodological strengths

  • Inclusion criteria limited to large randomised control trials
  • Extensive search for studies; unlikely any will be missed
  • Trials arms (data) with subtherapeutic doses of chondroitin and/or glucosamine were excluded.
    • it is unlikely that the meta-analysis would under-estimate the effect of these agents
  • Predefined levels of minimal clinical significance given and adequately justified.
  • Majority of trials had adequate concealment of allocation (6 of 10) and blinding of patients (9 of 10)
    • again, the meta-analysis is more likely to over-estimate rather than under-estimate the effect of these agents

Biases and conflicts of interests

Nil significant declared and none seem obvious.

Clinical relevance to primary health care

The use of glucosamine, with or without chondroitin, for treatment of osteoarthritis is controversial. It has been promoted as a safe over-the-counter therapy with some patients claiming substantial improvements in their symptoms. It is recommended by some clinicians (primary care physicians, rheumatologists and orthopaedic surgeons) but denounced as ineffective (albeit harmless) therapy by others.  Older and smaller studies have demonstrated beneficial effects but newer and better designed studies have generally been disappointing.  This study limited the analysis to the best quality large randomised controlled trials.

It demonstrates that glucosamine, chondroitin and their combination do not have any effect on joint space in hip and knee osteoarthritis compared to placebo.

It is interesting to see a statistically significant effect of glucosamine on pain, with a 0.4 cm improvement on the 10 cm visual analogue scale. There were similar effect sizes with chondroitin and their combination though statistical significance was not reached (wide 95% credible intervals were present).  However, even if the true effect was near the upper end of the credible intervals, these are not clinically significant effects.  Moreover, there was a significant difference in effect from industry sponsored studies compared to independently funded studies.  When the meta-analysis is restricted to independently funded studies, statistical significance for glucosamine, chondroitin and their combination is completely lost.  Moreover, the effect sizes and their credible intervals of these interventions compared to placebo centres around zero.

It is very likely that glucosamine, chondroitin and their combination is therapeutically inert for treatment of osteoarthritis. They should not be recommended. Being therapeutically inert, it is unlikely that these agents will cause any harm though their cost is not insubstantial.


  1. Wandel S, Jüni P, Tendal B. Effects of glucosamine, chondroitin, or placebo in patients with osteoarthritis of hip or knee: network meta-analysis. BMJ 2010;341:c4675
Print Friendly, PDF & Email


    • Joel Rhee on 8 February 2011 at 9:47 PM
    • Reply

    Interesting. I haven’t read the article but I wonder whether they did a separate analysis of Glucosamine Sulphate. My understanding was that there was some weak evidence that suggests benefit for this, whereas there is no evidence for Glucosamine Hydrochloride?

  1. Stratified analysis compared between glucosamine sulphate and glucosamine hydrochloride and no significant difference was found between the two.

    Specifically: difference in means on 10 cm visual analogue scale (95% confidence interval):
    – glucosamine sulphate: -0.5 cm (-0.7 to -0.2)
    – glucosamine hydrochloride: -0.1 cm (-0.8 to 0.7)

    That is, the results for glucosamine sulphate only were similar to the combined result for glucosamine. Even if the actual effect was at the upper limit of difference in the confidence interval (-0.7 cm), it still remains below the accepted level of clinical significance.

Leave a Reply

%d bloggers like this: