Rofecoxib Does Not Appear to Increase the Risk of Venous Thromboembolism: A Systematic Review of the Literature,

Abstract 3334

Rofecoxib, a selective cyclo-oxygenase-2 (COX-2) inhibitor, has been associated with increased arterial thrombosis. It is unknown whether it is associated with venous thromboembolism (VTE). We investigated, using a systematic review of the literature, the association between rofecoxib and venous thrombosis. A search strategy was developed with the assistance of a research librarian and was implemented in 2 electronic databases. We identified all English language studies in which rofecoxib was compared with placebo, irrespective of the primary outcome of the study. Only placebo controlled trials were included in the review as NSAIDs (the comparator in many rofecoxib studies) may have anti-thrombotic effects. We also limited studies to those with 3 or months of follow-up since we presume that the effects of rofecoxib on the coagulation system would take time to develop. Two reviewers assessed study eligibility, with disagreements resolved by consensus. Data was extracted from each of study in a standardized manner using RefMan software. Quality of the studies was assessed according to pre-established criteria: (1) randomization process (2) blinding, and (3) documentation of losses to follow up. Pooled incidence rates were calculated from the reported number of VTEs and total person years at risk. Confidence intervals and risk difference were calculated using a Poisson distribution. The search strategy identified 1339 papers. After review, a total of 15 studies in 14 populations met inclusion criteria. The majority of trials were short in duration (∼12 weeks). All studies met at least two of the three quality criteria. In 15160 (9217 person years follow up) patients allocated to rofecoxib there were 8 VTEs reported, compared with 9 VTEs in 13147 (9092 person years) patients allocated to placebo (relative risk 0.87, 95% CI 0.29–2.56, p = NS). The estimated incidence of VTE was 86.8 per 100,000 (95% CI 37.5 –171.2) person years with rofecoxib, and 99.1 per 100,000 person years with placebo (95%CI 45.3 – 188). This difference is statistically insignificant (p=0.78). Our findings are limited by the relatively small number of events, although the contributing sample size 28307 subjects (18309 person years) is reasonable. We are also unsure that all venous events were captured, although the fact that many of the contributing studies were performed in support of regulatory filings of rofecoxib suggests that events were unlikely to be missed. All included trials had processes for adjudication of VTE. Keeping these limitations in mind, our findings do suggest that there is no increase in the risk of VTE with rofecoxib use.