As coauthors of the review of therapy in osteoarthritis who are not funded by companies producing COX-2 inhibitors or any brands of glucosamine or chondroitin, we can comment impartially on Dr. Lindow's concerns about any bias in advocating COX-2 inhibitors compared with glucosamine or chondroitin. We consider glucosamine and chondroitin promising agents for osteoarthritis. Our meta-analysis of trials of these agents suggested several biases, including probable publication bias, that should inject caution into interpretation of the positive trial results. We should note that since the publication of our meta-analysis, several new randomized, placebo-controlled trials have been published that have not shown efficacy of either glucosamine or chondroitin (1-3). An additional null trial was presented at the 2000 American College of Rheumatology meeting. We know of no published or unpublished data suggesting that COX-2 inhibitors have no significant effect relative to placebo in patients with osteoarthritis. Furthermore, COX-2 inhibitors underwent critical scrutiny by the U.S. Food and Drug Administration before approval; glucosamine and chondroitin have not been so evaluated. Thus, we believe that there is stronger evidence for relief of symptoms in osteoarthritis with COX-2 inhibitors than with glucosamine and chondroitin, for which the data are genuinely conflicting. It is to be hoped that the large National Institutes of Health–funded trial of glucosamine and chondroitin will help definitively resolve the issue of their place in osteoarthritis management. We concur with Dr. Lindow that these agents are safe and did not mean to imply that COX-2, glucosamine, or chondroitin clearly affects structural disease progression. However, such an effect for glucosamine has recently been suggested by a large-scale, long-term randomized trial.

Dr. Waddell advocates viscosupplementation in the form of hyaluronic injections for the treatment of osteoarthritis. This treatment is controversial, and the largest trial's intention-to-treat analysis has shown no evidence of efficacy compared with placebo saline injections in the knee (4) (Table). In this large multicenter trial, no difference was seen between placebo and hyaluronic acid. Other multicenter trials of hyaluronic acid have been similarly null (5).

Dr. Rosch advocates pulsed electromagnetic fields for osteoarthritis, a treatment that he contends has been shown to be efficacious. Two reasonably well-done although small-scale clinical trials suggest that pulse electromagnetic fields may relieve symptoms in patients with osteoarthritis. A small trial (6) of 27 patients reported improvement in those treated with pulsed-signal therapy, although the improvement in the active versus placebo groups was not compared statistically and the placebo group also experienced modest improvement. In a larger trial reported by some of the same investigators (7), the improvement experienced by treated patients versus placebo recipients reached significance for some measures of outcome at some time points but not others. Multiple outcomes were evaluated at several time points. Results of both trials suggested that pulsed-signal therapy had a modestly greater effect than placebo. Thus, we would characterize pulsed-signal therapy as an experimental therapy whose efficacy must be demonstrated by additional data.