Rofecoxib, but not celecoxib, increases the risk of thromboembolic cardiovascular events in young adults-a nationwide...

To examine the risk of thromboembolic cardiovascular events in users of coxibs and NSAIDs in a nationwide cohort.

Data were synchronised from three nationwide databases, the Icelandic Medicines Registry (IMR), The Icelandic National Patient Registry (INPR) and the Registry for Causes of Death at Statistics Iceland (RCD), for prescriptions for NSAIDs or coxibs with respect to hospitalisation for unstable angina pectoris, myocardial infarction and cerebral infarction over a 3-year period. The Cox proportional hazards model and Poisson regression were used to analyse the data.

A total of 108,700 individuals received prescriptions for NSAIDs or coxibs (ATC code M01A), of whom 78,539 received one drug only (163,406 person-years). Among those receiving only one drug 426 individuals were discharged from hospital with endpoint diagnoses. In comparison to diclofenac, the incidence ratios, adjusted for age and gender, were significantly higher for cerebral infarction (2.13; 95% CI 1.54-2.97; P < 0.001), for myocardial infarction (1.77; 95% CI 1.34-2.32; P < 0.001) and for unstable angina pectoris (1.52; 95% CI 1.01-2.30; P = 0.047) for patients who used rofecoxib. For naproxen users, the incidence ratio was 1.46 for myocardial infarction (95% CI 1.03-2.07; P = 0.03), but was reduced in ibuprofen users (0.63; 95% CI 0.40-1.00; P = 0.05). The youngest users of rofecoxib (< or =39 years) had the highest hazard ratio (HR) for cardiovascular events (8.34; P < 0.001), while those > or =60 years had a lower but still significantly elevated HR (1.35; P = 0.001).

This Icelandic nationwide registry-based study amounting to 163,406 patient-years showed increased risk of cardiovascular events, i.e. cerebral infarction, myocardial infarction and unstable angina pectoris, among rofecoxib and naproxen users in comparison to diclofenac users. The added risk was most pronounced in young adults using rofecoxib.