Incremental cost-effectiveness of cyclooxygenase 2-selective versus nonselective nonsteroidal anti-inflammatory drugs in a cohort of coumarin users: a pharmacoeconomic analysis linked to a case-control study.

BACKGROUND

A previous case-control study involving concomitant users of coumarin and nonsteroidal anti-inflammatory drugs (NSAIDs) found that cyclooxygenase 2 (COX-2)-selective NSAIDs were associated with fewer bleeding complications than nonselective NSAIDs.

OBJECTIVE

The goal of this study was to determine the incremental cost-effectiveness of COX-2-selective versus nonselective NSAIDs in relation to the occurrence of bleeding complications in a cohort of concomitant coumarin users.

METHODS

The pharmacoeconomic evaluation was linked to a case-control analysis (patients with and without bleeding complications) based on data from the earlier study in users of concomitant coumarin and NSAIDs. Medical costs associated with NSAID use and bleeding complications were estimated according to Dutch guidelines for pharmacoeconomic analyses, based on Dutch drug prices and national averages for health care costs. Rofecoxib, meloxicam, and nabumetone were considered COX-2 selective. Total costs were calculated and compared for 2 hypothetical scenarios in which patients used either COX-2-selective or nonselective NSAIDs. Sensitivity analyses were performed in which both the odds ratios (ORs) and the costs of NSAIDs and bleeding episodes were varied.

RESULTS

A total of 1,491 bleeding complications occurred in 4400 coumarin users: among the 221 (15%) NSAID users with a bleeding episode, 96% used a nonselective NSAID and 4% used a COX-2-selective NSAID. The adjusted OR of a bleeding episode for nonselective compared with COX-2-selective NSAIDs was 3.07 (95% CI, 1.18-8.03). The estimated mean cost of a bleeding episode was 478 per patient. Factoring in the excess cost of a COX-2-selective NSAID compared with a nonselective NSAID, as well as the cost savings in averted bleeding episodes, it was determined that there would be net medical cost savings of 53,800 and 162 averted bleeding episodes if the entire patient group received COX-2-selective NSAIDs rather than nonselective NSAIDs. The sensitivity analysis showed these results to be robust.

CONCLUSION

In this study population of concomitant coumarin and NSAID users, the reduction in bleeding complications with the use of more expensive COX-2-selective inhibitors was associated with net medical cost savings compared with nonselective NSAIDs.