Interview with Medical Observer

I was interviewed today by Emily Dunn, a health journalist from Medical Observer regarding my thoughts on the new interesting study by Kwoh et al. (2014), on the effect of glucosamine on knee osteoarthritis progression, as assessed using MRI, which was just published in the journal, Arthritis and Rheumatology.

My perspective

The following is an except of an e-mail that I sent to Emily:

Basically, this is an interesting study insofar that it is examining the effect of glucosamine on knee osteoarthritis, but using a number of biological markers, specifically, a scoring system on MRI.

As far as I can tell, the study methods seemed okay. The major limitation to this study is that it is very underpowered. The authors in the planning stages used an assumption that the incidence of worsening in subregions on MRI would be 5% when in actuality, it was only 1.4% in the placebo group.

As such, any conclusions made on the results need to be done with caution – the confidence interval of the estimate of effect is wide (i.e., looking at this study alone, we can’t really make any firm conclusions on whether glucosamine was beneficial or harmful).

However, in the planning (statistical analysis, page 7), it looked like that the investigators wanted to power the study to detect an odds-ratio of 2.0 for decreased worsening of cartilage damage. Assuming that this was what the authors considered a clinically significant result (I don’t know if this is a reasonable assumption as I’m not a radiologist, but the authors seem to indicate this), then this study does suggest that it is very unlikely that the true effect will be at this magnitude. The researchers discuss this on page 16.

Summary:

This study doesn’t give us much precision on what the effect of glucosamine is on progression of cartilage damage in knee osteoarthritis (as imaged using MRI), as compared to placebo. However, the estimate of the effect was near that of the placebo, and far away from a clinically significant effect. Thus, this study suggests that glucosamine is not associated with benefit for the knee, insofar as less deterioration on MRI.

In the broader context, glucosamine +/- chondroitin is associated with some improvement in pain scores for knee and hip osteoarthritis, but this does not appear to be at a clinically significant magnitude. When analyses are limited to the best studies (in terms of methodological quality), even this improvement vanishes.

IMHO, the best interpretation from the evidence from clinical studies is that glucosamine is probably therapeutically inert for knee/hip osteoarthritis (i.e., it doesn’t do anything).