Letter by Cen Regarding Article, “Effect of Celecoxib on Cardiovascular Events and Blood Pressure in Two Trials for the Prevention of Colorectal Adenomas”
To the Editor:
I read with interest the paper by Solomon et al1 in which the authors have combined the analysis results of 2 independent trials, Adenoma Prevention With Celecoxib (APC)2 and Prevention of Colorectal Sporadic Adenomatous Polyps (PreSAP),3 in an attempt to estimate cardiovascular events (CVs) associated with celecoxib across 3 different dose regimens. I am concerned about the validity of their combined analysis and their conclusion that celecoxib at 200 or 400 mg twice daily (APC) or at 400 mg once daily (PreSAP) showed a nearly 2-fold increase in CV risk.
First, patients in APC differed significantly from patients in PreSAP regarding some very important baseline clinical characteristics—including prevalence of hypertension, smoking, low-dose aspirin use, and lipid-lowering drug use—that are known to be possible predictors of CV events. For example, 414 patients (31%) in APC were on low-dose aspirin compared with 159 patients (17%) in PreSAP. Five hundred sixteen patients (38%) in APC were on lipid-lowering drugs compared with 159 patients (17%) in PreSAP. We do not know whether increased CV events in APC were caused by drug interactions with celecoxib. We do know that CV risk was the greatest for patients with a history of CVs at baseline.
To justify a combined hazard ratio, the authors ran 2 Cox models to compare the 2 dose arms in APC and to compare APC with PreSAP. The large differences among the hazard ratios (3.4, 2.6, and 1.3) made it unconvincing that one could combine those results, even though the heterogeneity tests were insignificant. The relatively small probability value (0.13) suggests some difference. It may not be appropriate to apply the conclusion drawn from the combined estimate to both patient populations.
My other concerns involve modeling issues. The authors do not report the interactions between low-dose aspirin use with celecoxib or between lipid-lowering drugs with celecoxib. Furthermore, we do not know how their stratified model was derived and which covariates were included. We also do not know which covariates were used for the 2 Cox models to generate the 2 hazard ratios that they combined.
With access to complete data from both trials, the authors could have performed a much more rigorous and comprehensive statistical analysis by fitting a Cox regression model to the entire data from both trials. In this model, with complete data, the authors could have formally tested the validity of combining the 3 dose regimens and derived a more statistically sound combined hazard ratio for the entire dataset. Furthermore, the authors could have potentially investigated the dose effect, and this might have provided a better summary of the drug safety.
The authors have made an important contribution to our understanding of the adverse effects of celecoxib at 400- and 800-mg doses. I look forward to additional results from their collaboration.
Solomon SD, Pfeffer MA, McMurray JJV, Fowler R, Finn P, Levin B, Eagle C, Hawk E, Lechuga M, Zauber AG, Bertagnolli MM, Arber N, Wittes J; APC and PreSAP Trial Investigators. Effect of celecoxib on cardiovascular events and blood pressure in two trials for the prevention of colorectal adenomas. Circulation. 2006; 114: 1028–1035.
Bertagnolli MM, Eagle CJ, Zauber AG, Redston M, Solomon SD, Kim K, Tang J, Rosenstein RB, Wittes J, Corle D, Hess TM, Woloj M, Boisserie F, Anderson WF, Viner JF, Bagheri D, Gordon GB, Burn J, Chung DC, Dewar T, Folwy TR, Hoffman N, Macrae, F, Pruitt RE, Salzman B, Tylwestrowicz T, Hawk ET; for the APC study Investigators. Celecoxib for the prevention of sporadic colorectal adenomas. N Engl J Med. 2006; 355 (9): 873–884.
Arber N, Eagle CJ, Spicak J, Racz I, Dite P, Hajer J, Zavoral M, Lechuga MJ, Gerletti P, Tang J, Rosenstein RB, Macdonald K, Bhadra P, Fowler R, Wittes J, Zauber AG, Solomon SD, Levin B; for PreSAP Trial Investigators. Celecoxib for the prevention of colorectal adenomatous polyps. N Engl J Med. 2006; 355 (9): 885–895.