Background: Lumiracoxib is a selective cyclooxygenase-2 inhibitor effective in the treatment of osteoarthritis (OA) with a superior gastrointestinal (GI) safety profile as compared to traditional nonsteroidal anti-inflammatory drugs (NSAIDs, ibuprofen and naproxen). This safety study compared the GI tolerability, the blood pressure (BP) profile and the incidence of oedema with lumiracoxib and rofecoxib in the treatment of OA. Rofecoxib was withdrawn worldwide due to an associated increased risk of CV events and lumiracoxib has been withdrawn from Australia, Canada, Europe and a few other countries following reports of suspected adverse liver reactions. Methods: This randomised, double-blind study enrolled 309 patients (aged greater than or equal to 50 years) with primary OA across 51 centres in Europe. Patients were randomly allocated to receive either lumiracoxib 400 mg od (four times the recommended dose in OA) (n = 154) or rofecoxib 25 mg od (n = 155). The study was conducted for 6 weeks and assessments were performed at Weeks 3 and 6. The primary safety measures were the incidence of predefined GI adverse events (AEs) and peripheral oedema. The secondary safety measures included effect of treatment on the mean sitting systolic and diastolic blood pressure (msSBP and msDBP). Tolerability of lumiracoxib 400 mg was assessed by the incidence of AEs. Results: Lumiracoxib and rofecoxib displayed similar GI safety profiles with no statistically significant difference in predefined GI AEs between the two groups (43.5% vs. 37.4%, respectively). The incidence and severity of individual predefined GI AEs was comparable between the two groups. The incidence of peripheral oedema was low and identical in both the groups (n = 9, 5.8%). Only one patient in the lumiracoxib group and three patients in the rofecoxib group had a moderate or severe event. At Week 6 there was a significantly lower msSBP and msDBP in the lumiracoxib group compared to the rofecoxib group (p < 0.05). A similar percentage of patients in both groups showed an improvement in target joint pain and disease activity. The tolerability profile was similar in both the treatment groups. Conclusion: Lumiracoxib 400 mg od (four times the recommended dose in OA) provided a comparable GI safety profile to rofecoxib 25 mg od (therapeutic dose). However, lumiracoxib was associated with a significantly better BP profile as compared to rofecoxib. Trial registration number -: NCT00637949 [ABSTRACT FROM AUTHOR]