We appreciate Dr. Boers' interest in our report. Table 1 provides the dosage information and results he requests. Calculating dosage from an observational cohort study is not straightforward. We calculated dosage by treating each prescription as a separate exposure period. Thus, if a patient's dosage escalated over time, successive prescriptions at higher dosages were considered separate exposure periods. This method accounts for the varying dosages during different periods of exposure and avoids mixing low- and high-dosage exposure periods, a methodologic flaw inherent in calculating an average daily dose.
|Drug, daily dose||n||Myocardial infarction or stroke|
|Events||Person-years||Incidence rate*||Adjusted rate ratio (95% CI)†|
|≤200 mg||21,956||952||8,710||10.9||0.89 (0.82–0.97)|
|201–400 mg||3,997||368||2,946||12.5||0.89 (0.77–1.03)|
|>400 mg||413||22||112||19.6||0.91 (0.50–1.65)|
|≤25 mg||16,435||765||5,981||12.8||1.15 (1.06–1.29)|
|>25 mg||1,532||147||765||19.2||1.05 (0.81–1.37)|
|≤10 mg||2,105||55||569||9.7||0.96 (0.78–1.17)|
|>10 mg||955||57||416||13.7||0.74 (0.18–2.95)|
The lack of a dosage gradient in this study is somewhat surprising. We observed such a dosage gradient in a prior observational study of rofecoxib (Solomon DH, Schneeweiss S, Glynn RJ, Kiyota Y, Levin R, Mogun H, et al. Relationship between selective COX-2 inhibitors and acute myocardial infarction. Circulation 2004;109:2068–73). In addition, the Adenoma Prevention with Celecoxib study showed a clear dosage gradient between placebo and high-dose celecoxib (either 200 mg twice daily or 400 mg twice daily) (Solomon SD, McMurray JJ, Pfeffer MA, Wittes J, Fowler R, Finn P, et al. Cardiovascular risk associated with celecoxib in a clinical trial for colorectal adenoma prevention. N Engl J Med 2005;352:1071–80). While we do not have a clear explanation for this apparent discrepancy, it is possible that there may have been channeling of high-risk patients away from higher-dose treatment in our study cohort. As well, with small sample sizes and relatively wide confidence intervals in the higher dose groups, we cannot exclude the possibility of a dosage gradient.