PRACTICE CHANGER
Tell patients with large joint arthritis that glucosamine and chondroitin have been found to be little better than placebo.1
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;314:c4675.
STRENGTH OF RECOMMENDATION
A: Based on a good-quality meta-analysis.
ILLUSTRATIVE CASE
A 64-year-old woman with osteoarthritis (OA) of both knees reports that acetaminophen does not relieve the pain, and both ibuprofen and naproxen give her an upset stomach. She wonders if glucosamine and chondroitin would help relieve the pain. How should you respond?
Degenerative joint disease is a common and frustrating problem for patients and clinicians. Symptomatic knee OA has a prevalence of 16% among adults older than 45 years, and is one of the top 5 reasons for disability in noninstitutionalized adults.2 With no highly effective treatment for OA of the hip or knee other than joint replacement surgery, patients often turn to unproven over-the-counter remedies. Individuals with OA spend about $2600 per year out-of-pocket on disease-related expenses.2
Trials of these supplements have had mixed results
Glucosamine and chondroitin have been touted as beneficial, and sales have grown rapidly over the last decade, reaching nearly $900 million in the United States in 2008 alone.3 There have been many randomized trials of these supplements, with inconsistent results.
Larger and higher quality studies have found little or no effect, while smaller studies reported that glucosamine and chondroitin helped to relieve joint pain. A meta-analysis published in 2000 found 15 studies and reported moderate to large effect sizes, but the authors noted that quality issues and publication bias probably exaggerated the benefit.4 An updated Cochrane meta-analysis of 25 randomized controlled trials (RCTs), published in 2009, found little benefit from glucosamine. A subgroup analysis found that one company’s preparation appeared to be beneficial, but all 14 studies of that particular formulation had some connection with the manufacturer.5
STUDY SUMMARY: Effects of glucosamine and chondroitin, alone or together, were small
The meta-analysis we review in this PURL only included RCTs with an average of ≥100 patients with hip or knee OA in each group.1 This was based on the minimum sample size needed to detect a small or moderate difference between the 2 groups (roughly 1 cm on a 10-cm visual analogue scale [VAS]). The authors found 10 eligible RCTs with a total of 3803 patients; the average age of participants ranged from 58 to 66 years. Most of the trials studied knee arthritis, and most were sponsored by pharmaceutical firms.
Included studies had to compare glucosamine sulphate, glucosamine hydrochloride, chondroitin sulphate, or a combination, either with a placebo or head-to-head. Minimum daily doses were 800 mg chondroitin and 1500 mg glucosamine. The primary outcome was absolute pain intensity over the duration of the study. The authors summarized pain scores every 3 months for up to 2 years; they also analyzed changes in joint space narrowing in the studies reporting that measure.
The authors used a sophisticated framework that adjusted for comparisons over time and between studies, allowing them to increase the power, and likely the accuracy, of their comparisons. They reported outcomes as effect sizes, then translated the findings to a real-world outcome by converting results to a 10-cm VAS. Typically, an effect size of 0.2 standard deviation (SD) units is considered small, 0.5 SD units is a moderate difference, and 0.8 SD units is large. The authors set their threshold for a clinically important difference at 0.37 SD units, which translated to a 0.9 cm change on a 10-cm VAS—a generally accepted minimal clinically significant difference in pain.
They found that all 3 interventions (glucosamine alone, chondroitin alone, and a combination) were statistically better than placebo, with very little difference in outcomes over time. Compared with placebo, VAS improvements were 0.4 cm for glucosamine (95% confidence interval [CI], 0.1-0.7), 0.3 cm for chondroitin (95% CI, 0-0.7) and 0.5 cm for the combination (95% CI, 0-0.9). All of these improvements in pain were less than the authors’ defined minimum clinically significant improvement of 0.9 cm on a 10-cm scale.
Among the 6 trials that reported on joint space narrowing, the changes were minute and not statistically significant. There was a net difference between treatment and placebo groups of less than 0.2 mm (an effect size ≤0.16 SD units). There was no evidence of increased risk of adverse effects or increased dropout rates with any of the substances.