The search for less gastrointestinal toxic nonsteroidal anti-inflammatory drugs (NSAIDs) led to the introduction of the selective cyclooxygenase-2 (COX-2) inhibitors. However, with this introduction into the market, there have been concerns regarding their safety, particularly cardiovascular safety. OBJECTIVE: The purpose of this study was to assess the cardiovascular risk (events included: myocardial infarction, stroke, and myocardial infarction-related deaths) associated with long-term use (after 180 days of exposure) of nonselective NSAIDs and COX-2 inhibitors. METHODS: A retrospective analysis of the Veterans Integrated Service Network 17 Veteran's Affairs (VA) database was conducted. Medicare data and Department of Health mortality data were incorporated to capture events occurring outside the VA health care network. Patients 35 years of age and older receiving celecoxib, rofecoxib, ibuprofen, etodolac, and naproxen from January 1, 1999 through December 31, 2001, were included. Multivariate Cox proportional hazard models were used to analyze the relationship between cardiovascular risk and NSAID/COX-2 inhibitor use while adjusting for various demographic and co-morbidity factors. RESULTS: We identified 6814 long-term exposure periods and 42 cardiovascular events over the study period. Compared to long-term ibuprofen use, long-term use of celecoxib was associated with a 3.64 fold (95% CI 1.36–9.70; p=0.01) increase in cardiovascular risk. Long-term use of rofecoxib was associated with a 6.64 fold (95% CI 2.17–20.28; p=<0.01) increased risk when compared to long-term users of ibuprofen. Long-term exposure to naproxen and etodolac was not associated with a cardionegative or cardioprotective effect when compared to long-term ibuprofen users. CONCLUSIONS: The findings of this observational study along with recent clinical trial results suggest that long-term exposure to COX-2 inhibitors is associated with an increased cardiovascular risk. In addition, the study results do not show that naproxen or etodolac provide a cardioprotective or cardionegative effect.