After my last posts, a few people have sent me links to reports that the NIH study did not look very positive for G & C. Basically, they’re saying that while the percentage of subjects reducing pain by 20% or more after six months was higher, it wasn’t statistically different for each of the supplements compared to the placebo, even though the Glucosamine & Chondroitin Sulfate group showed a statistical trend to be better than placebo.

It’s a bit confusing.

So I talked to my doctor about this—he’s a fan of glucosamine & chondroitin—and this is how he explained it.


  • This comparison does not mean much, since the 20% reduction in pain is a very low hurdle to jump over. It is no surprise that the placebo group, which received as much acetaminophen (Tylenol®) as they wanted, was able to cross the 20% threshold in 60% of the subjects.

  • However, when the subjects were divided into those that hurt more and those that hurt less, the results were very encouraging, and the NIH investigators concluded: “Combination Glucosamine & Chondroitin Sulfate is effective in treating moderate to severe knee pain due to OA.” (emphasis mine)

  • One thing the media did not pick up on is the real results of this study using updated responder rates of 50% reduction in pain after six months, the same criteria used for pharmaceutical agents (and thus, much more meaningful than the 20% response threshold). Apparently, these results were presented by the GAIT investigators, but were not put into print in the abstract.

  • When the hurdle for reducing pain was raised to a 50% reduction or better, the Glucosamine & Chondroitin group was significantly better than the placebo group for all subjects, not just a subset that hurt more. This means that Glucosamine & Chondroitin worked very well across the spectrum of knee osteoarthritis pain.



In other words, if you just looked at small decreases in pain, it was hard to tell what was working. But when the bar was raised, the NIH study showed that G & C was clearly effective. Mystery solved.