Comparison of the baseline cardiovascular risk profile among hypertensive patients prescribed COX-2-specific inhibitors or nonspecific NSAIDs: data from real-life practice.

OBJECTIVE To evaluate the baseline cardiovascular (CV) risk of hypertensive patients newly starting cyclooxygenase (COX)-2-specific inhibitors (celecoxib or rofecoxib) or nonspecific nonsteroidal anti-inflammatory drugs (NSAIDs). METHODS Cross-sectional analysis was performed based on real-life practice data contained in the LifeLink Integrated Claims Solutions employer claims database. Patients who newly received treatment of celecoxib, rofecoxib, ibuprofen, naproxen, or diclofenac between January 1, 1999, and September 30, 2000, were identified from the database. Among them, only those who had a stable hypertensive condition for at least 3 consecutive months before the index prescription were included. Baseline characteristics were determined from claims data at the time of the index prescription. RESULTS A total of 55 396 index prescriptions were identified, which consisted of 20,915 (37.8%) prescriptions for celecoxib, 12,952 (23.4%) for rofecoxib, 10 789 (19.5%) for ibuprofen, 8,840 (16.0%) for naproxen, and 1,900 (3.4%) for diclofenac. Both univariate and multivariate analyses showed that the patients prescribed COX-2-specific inhibitors were older and more likely to be female than those given nonspecific NSAIDs. Patients prescribed COX-2-specific inhibitors had a significantly higher baseline history of and/or current CV conditions, including ischemic heart disease, heart failure, other forms of heart disease, and cerebrovascular diseases or disorders, than patients prescribed nonspecific NSAIDs. The baseline proportion of patients with rheumatoid arthritis was also higher among COX-2-specific inhibitor users. In addition, COX-2-specific inhibitor users at baseline had higher prescription rates for medications that influence blood pressure, including estrogens, certain types of antidepressants, and corticosteroids. CONCLUSION COX-2-specific inhibitors were prescribed preferentially to patients who, at the time of their index COX-2-specific inhibitor prescription, were at an increased baseline risk of CV events compared with patients prescribed nonspecific NSAIDs. Researchers aiming to compare the incidence of CV events between COX-2-specific inhibitors and nonspecific NSAIDs using observational study designs should take into account the underlying baseline CV risk of the populations being compared.