A Comparison of Reported Gastrointestinal and Thromboembolic Events Between Rofecoxib and Celecoxib Using Observational Data

Abstract

Rofecoxib, a selective cyclo-oxygenase (COX)-2 inhibitor, was a widely marketed drug that was used for relief of pain and inflammation in arthritic conditions. It was withdrawn from the market worldwide in September 2004 because of an increased risk of cardiovascular events. Celecoxib, which belongs to the same class of drugs, is now under scrutiny for the risk of similar events. The objective of our study was to compare the incidence of gastrointestinal (GI) [symptomatic and complicated upper GI] events and thromboembolic (cardiovascular, cerebrovascular and peripheral venous) events reported for patients prescribed rofecoxib or celecoxib in primary care. A retrospective analysis of selected events was conducted using data collected from previously conducted prescription-event monitoring (PEM) studies for rofecoxib and celecoxib. PEM is an observational cohort technique. Exposure data were derived from dispensed prescriptions for rofecoxib (July-November 1999) and celecoxib (May-December 2000) that were written by primary care general practitioners in England. Outcome data were clinical events and information on potential risk factors reported on simple questionnaires (posted to prescribers approximately 9 months after the date of the first prescription). Incidence rates of the first event during treatment within each thromboembolic and GI group were calculated during the 270 days after the patient started receiving either of the drugs; crude and adjusted rate ratios (RR) were calculated for rofecoxib compared with celecoxib using Poisson regression modelling. The rofecoxib and celecoxib PEM cohorts contained 15 268 and 17 458 patients, respectively. For the GI event groups, the adjusted RRs for rofecoxib compared with celecoxib were: 1.21 (95% CI 1.09, 1.36) for symptomatic upper GI events and 1.60 (95% CI 0.95, 2.70) for complicated upper GI conditions. For the thromboembolic event groups, the adjusted RRs were: 1.04 (95% CI 0.50, 2.17) for cardiovascular thromboembolic events; 1.43 (95% CI 0.86, 2.38) for cerebrovascular thromboembolic events; and 0.36 (95% CI 0.01, 1.34) for peripheral venous thromboembolic events. This study was a retrospective comparison of PEM studies conducted for rofecoxib and celecoxib. For symptomatic upper GI events, a 21% increase in the relative rate was found for rofecoxib users compared with celecoxib users after adjusting for identified risk factors. For complicated upper GI events, no statistically significant difference in the incidence was observed between rofecoxib and celecoxib users after adjusting for identified risk factors. For the three thromboembolic event groups, no evidence of a statistically significant difference between rofecoxib and celecoxib users was found after adjusting for identified risk factors. This study contributes to the understanding of the association between COX-2 inhibitors and thromboembolic events. However, it should be borne in mind that we had information on only a limited number of confounding variables for thromboembolic events. Further research is required to fully understand the risks and benefits of using celecoxib and other COX-2 inhibitors. Meanwhile, doctors should be cautious when prescribing these products, particularly to patients with risk factors for developing thromboembolic events.