OBJECTIVE: To assess the cardiovascular risk factors at baseline in Dutch patients prescribed cyclooxygenase-2 inhibitors (coxibs) compared to nonselective nonsteroidal anti-inflammatory drugs (NSAIDs). METHODS: Data were obtained from the PHARMO database, which includes linked drug-dispensing records and hospital records of more than one million subjects in defined areas in the Netherlands. All users of celecoxib, rofecoxib or NSAIDs in the period January, 2001 - December, 2003 were included in this retrospective population based cohort study. The presence of cardiovascular risk factors based on hospital admissions and drug use was compared in the year before the first prescription between users of coxibs or NSAIDs using logistic regression analyses. RESULTS: The cohort consisted of approximately 4,000 users of celecoxib, 34,000 users of rofecoxib and 326,000 users of NSAIDs. Users of coxibs more often were female (68% vs. 56%) and of higher age (average of 56 years vs. 45 years) compared to users of NSAIDs. Multivariate analyses including gender, age and prescriber showed that patients using celecoxib 1.3 times (95%CI: 1.05-1.59) more often had a history of hospital admissions for cardiovascular diseases compared to patients using NSAIDs. This association was not significant for users of rofecoxib. Other factors positively associated with the use of coxibs were previous hospitalizations for RA, OA and other diagnoses and previous use of gastroprotective agents. CONCLUSIONS: In general, users of coxibs seem to be less healthy then users of NSAIDs with higher cardiovascular risk factors at baseline. This difference should be taken into account when investigating the possible association between cardiovascular adverse events and use of coxibs.