OBJECTIVES: With withdrawal of arthritics painkiller Vioxx from the market due to reported events of heart attack and stroke, the risk of Cox2 drugs has been investigated and challenged. The study is designed to retrospectively evaluate the heart attack/stroke occurrence of arthritic patients treated with COX2 and other painkillers in the commercially insured outpatient setting. METHODS: Arthritic patients(n=7035) who were continuously enrolled from January 2000 to December 2004 and had been treated with arthritic painkillers were extracted from the CDW outpatient data warehouse and grouped into 3 groups: Vioxx group (n=2636); Cox2-Vioxx: Cox2 drugs excluding Vioxx (n=2520) and Others group: others arthritic drugs (OTC, Preferential cox2, Salicylate Salts, and NSAIDs, n=1,879). The adverse events were identified as patient diagnosis of either heart attack or stroke in patient visits. Six months were defined as clearance time period to ensure there was no such adverse event prior to receiving the drugs. Chi square test and Logistic regression were used to compare differences among incident rates of the three groups. RESULTS: Patients with rheumatoid and osteoarthritis showed statistically significant differences (Chi=7.4, p=0.0065) among groups Vioxx, Cox2-Vioxx and Others in the adverse events rates of 5.84%, 5.00% and 3.83% respectively. Further Logistic regression anlaysis revealed that Vioxx and Cox2-Vioxx patients experienced higher events rates (group Vioxx vs. Others: OR=1.218 , 95%CI= 0.905-1.640; group Cox2-Vioxx vs. Others OR=1.13 , 95%CI= 0.834-1.532) with influencing factors adjusted, gender ( male OR=1.555, 95% CI= 1.190-2.031) , age (OR=1.080, 95%CI= 1.063-1.097) and comorbidities. CONCLUSION: The study demonstrated that Vioxx before withdrawn from market had highest heart attack/stroke events as compared to other Cox2 drugs or drug class. The other Cox2 drugs, Celebrex and Bextra, had higher adverse events than other painkiller classes such as NSAIDs. The follow up study with extended data is expected to confirm the finding.