To the Editor:

We thank Drs. Gregory and Toda for their comments about the American College of Rheumatology recommendations for the use of glucosamine and acetaminophen, respectively, in the management of OA. The Technical Expert Panel, comprised of a diverse group of physicians and other health care providers, was provided with summary of findings tables for glucosamine and acetaminophen and then voted on their recommendations after discussing this body of evidence. The recommendations for glucosamine were based on the results of the meta-analyses cited by Dr. Gregory in his letter as well as an evidence report from the Agency for Healthcare Research and Quality on the treatment of OA of the knee (Samson DJ, Grant MD, Ratko TA, Bonnell CJ, Ziegler KM, Aronson N. Treatment of primary and secondary osteoarthritis of the knee: evidence report/technology assessment no. 157. AHRQ publication no. 07-E012. Rockville (MD): Agency for Healthcare Research and Quality; 2007); the summary of evidence tables are available as supplementary files (see the online version of the article at The conditional recommendation that patients with hip and knee OA should not take glucosamine was based largely on the heterogeneity of the results of randomized placebo-controlled clinical trials (the evidence of efficacy was limited to the preparation of crystalline glucosamine sulfate produced and marketed by Rottapharm/Madaus) and the lack of availability of prescription-quality preparations evaluated and approved for the indication of OA by the Food and Drug Administration.

Dr. Toda cites a number of epidemiologic observational studies suggesting that patients receiving high-dose acetaminophen have increased odds of serious upper gastrointestinal events compared to those not receiving acetaminophen and questions whether patients with OA who have been recommended to take acetaminophen at a dosage of ≥2,000 mg/day should also be given a PPI to reduce this perceived risk. We believe that the most likely explanation for the findings in these observational studies is a channeling bias that persists in some of the studies despite the best efforts of investigators to adjust for patient-level confounders with propensity scores (Schneeweiss S, Gagne JJ, Glynn RJ, Ruhl M, Rassen JA. Assessing the comparative effectiveness of newly marketed medications: methodological challenges and implications for drug development. Clin Pharmacol Ther 2011;90:777–90). Furthermore, there are no prospective randomized trials to address Dr. Toda's question of whether PPIs would reduce the alleged risk of serious upper gastrointestinal events in those taking high-dose acetaminophen with acceptable safety.

Marc C. Hochberg MD, MPH*, * University of Maryland School of Medicine, Baltimore, MD.