Saturday, October 2, 2010

Glucosamine, Chondroitin Useless for OA, Data Say

EBPM Logo Understanding Evidence-Based Pain Management (EBPM)

The latest and most extensive investigation to date found no clinically relevant benefits of glucosamine, chondroitin, or their combination for osteoarthritis (OA) pain or disease progression. Put simply, they do not work any better than placebo and claims of their efficacy may reflect deficiencies and biases of prior research evidence.

Writing in the British Medical Journal, researchers at the University of Bern, Switzerland conducted an exhaustive review and sophisticated analysis of trials examining glucosamine and/or chondroitin for improvements of joint pain and radiological progression in osteoarthritis of the hip or knee [Wandel et al. 2010]. Relevant databases through June 2010 were searched for larger-scale randomized controlled trials enrolling patients with osteoarthritis of the knee or hip and comparing glucosamine, chondroitin, or their combination with placebo or head to head.

Wandel et al. discovered 10 eligible trials, including a total 3,803 patients. On a 10 cm visual analog scale (VAS) the overall improvement in pain intensity compared with placebo for glucosamine was only −0.4 cm (95% CI: −0.7 to −0.1 cm), for chondroitin −0.3 cm (−0.7 to 0.0 cm), and for the combination −0.5 cm (−0.9 to 0.0 cm). Therefore, none of the estimated 95% confidence intervals exceeded a minimal clinically worthwhile difference of –0.9 on VAS measures. Radiological differences in minimal joint space width compared with placebo were all nonsignificant, with 95% confidence intervals overlapping zero.

In rather strong but direct language, the investigators conclude: “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.” At the same time, however, Wandel and colleagues concede that some patients are convinced that these preparations are helpful, and neither of the agents is dangerous; so, there would be no harm in having patients continue to take the agents as long as they are willing to spend their own money.

COMMENTARY: Research on glucosamine and chondroitin is controversial, since many studies have been equivocal or contradictory, and certain guidelines still favorably mention use of the supplements. However, we have previously noted failures of these agents presented in research reports [here], [here], [here], and [here] (all open in new browser windows). The Wandel team acknowledges that some critics would argue that patients in the trials they examined were either too ill with arthritis to benefit from the supplements or not ill enough to show clinical improvements in pain or other outcomes. However, this ambivalence itself casts doubt over just who, if anyone, might benefit from these agents.

Wandel and colleagues were careful to select longer-term trials that included therapeutic doses of glucosamine (>1500 mg/day, almost all glucosamine sulphate) and chondroitin sulphate (>800 mg/day). Furthermore, the investigators set the bar very low in terms of clinical improvement (ie, −0.9 cm on a 10 cm VAS). As we have previously discussed [here], a 10% to 20% reduction in pain attributed to a therapy is the minimally acceptable threshold of clinical significance to patients; whereas, a 50% or more reduction in pain is generally considered by patients as satisfactory improvement, which some may describe as “treatment success.” The failure of glucosamine and/or chondroitin to achieve even the minimal threshold of pain improvement is remarkable.

The authors also selected only randomized trials having on average at least 100 subjects in each treatment arm. This would provide more than 80% power to detect a small, –0.40, but significant (p=0.05) effect size comparable to a difference of 1 cm on a 10 cm visual analog pain scale. We have previously noted [here] that trial size can make a great deal of difference in accuracy of results, and a great many trials in the pain management field do not enroll sufficient numbers of patients to produce valid outcomes.

Essentially this same team of Swiss researchers had previously observed that small studies, with fewer than 100 patients per treatment arm, are prone to numerous methodological deficiencies and reporting biases [Nuesch et al. 2010]. This is especially evident in systematic reviews or meta-analyses in which estimated treatment effects become more inaccurate as the sizes of included trials become smaller. For example, they found that meta-analyses of prior research on glucosamine and chondroitin that primarily included small, underpowered trials did show statistically significant therapeutic benefits of the agents.

Of further interest, Nuesch and colleagues found that reports of small-scale trials in osteoarthritis research did not even discuss how they calculated the study sample sizes a majority of the time (61%). And, not unsurprisingly, Wandel et al. noted that trials supported by public funding showed significantly smaller effects than manufacturer-funded trials (p=0.02). Indeed, this is big business; the supplements are not inexpensive and in 2008 the worldwide market for glucosamine alone was $2 billion, up 60% from 2003, according to Wandel and colleagues.

There are several messages of importance for readers from these studies…
  • Based on the preponderance of best evidence it appears inappropriate to recommend glucosamine and/or chondroitin to patients for osteoarthritis, even though the agents appear to have favorable safety profiles and might not do any harm.

  • Some patients may insist that glucosamine/chondroitin supplements are helpful, which may only reflect placebo effects; however, this might be acceptable as long as patients can readily afford to pay for the supplements.

  • The studies on these agents emphasize that practitioners and patients need to be more cautious and critical consumers of medical research. For example…

    1. Individual trials of a therapy, no matter how compelling the outcomes, cannot be accepted as definitive evidence of therapeutic benefit until the results are reliably confirmed by additional and valid studies.

    2. Smaller-sized trials may demonstrate statistically significant results but they are subject to numerous methodological flaws and biases, such as in subject selection, randomization, blinding, and imprecisions of estimated effects.

    3. Pain trials with fewer than 100 subjects in each arm of a trial (eg, active treatment vs placebo) may have inherent biases and not have sufficient power to detect meaningful treatment effects.

    4. In pain-therapy studies it has been suggested that minimally important improvement would be a 1 point reduction on a 0 to 10 numerical rating scale (NRS), or a 10% to 20% reduction on a visual analog scale (VAS). A more substantially meaningful improvement favored by patients would be a 4 point NRS reduction or 50% or more on VAS.

    5. Prior research and commentary have suggested that placebo effects may be present in more than 30% of patients treated for pain conditions, providing clinically significant reductions in pain of 50% or greater [full discussion here]. While such placebo effects may be worthwhile in themselves, the true benefit of any therapy requires subtracting outcomes in a placebo-treated group from those in the active therapy group.

    6. Throughout medical research in general it has been demonstrated that studies supported by industry (eg, manufacturers) tend to produce more favorable outcomes than investigations of the same therapies supported by public funding from impartial organizations or agencies.
Using the above considerations as a guide, we have observed that a great deal of published research in the pain management field may be suspect of having questionable validity for everyday clinical practice — to put it mildly. As we stated earlier in this series on EBPM (Evidence-Based Pain Medicine), “the most reasonable response seems to be for consumers of medical research literature to develop a healthy skepticism about everything they read and very cautiously accept research conclusions as guides for patient care.” Caveat Lector; reader beware.

> Nuesch E, Trelle S, Reichenbach S, et al. Small study effects in meta-analyses of osteoarthritis trials: meta-epidemiological study. BMJ. 2010(Jul 16);341:c3515 [
full article here].
> Wandel S, Juni P, Tendal B, et al. Effects of glucosamine, chondroitin, or placebo in patients with osteoarthritis of hip or knee: network meta-analysis. BMJ. 2010(Sep 16);341:c4675 [
full article here].