The influence of cyclo-oxygenase specificity of non-steroidal anti-inflammatory drugs on bleeding complications in concomitant coumarine users

Background: Concomitant use of coumarines and non-steroidal anti-inflammatory drugs (NSAIDs) may induce bleeding complications, due to the inhibition of both coagulant factors and platelet function. Unlike non-selective NSAIDs, cyclooxygenase-2 (COX-2)-selective NSAIDs interfere very little with platelet aggregation. Aim: To determine whether COX-2-selective NSAIDs are associated with less bleeding complications in coumarine users, compared with non-selective NSAIDs. Design: Prospective, nested case-control study. Methods: We studied concomitant coumarine and NSAID users over two years. Patients with bleeding (cases), and frequency-matched patients without bleeding (controls), were sent questionnaires regarding possible risk factors for bleeding. International normalized ratio (INR) values were recorded. Univariate and multivariate analyses were used to detect factors contributing to bleeding. Results: There were 1491 reported bleeds. NSAIDs were involved in 14.8%; 3.9% involving COX-2-selective NSAIDs. In non-bleeders, 2601 prescriptions with a coumarine/NSAID combination were detected; 9.7% were COX-2-selective. Adjusted ORs (95%Cl) for a bleeding complication were 3.07 (1.18-8.03) for non-selective NSAID use, 3.01 (1.42-6.37) for NSAID use > 1 month, and 1.89 (1.03-3.49) for INR 4.0. Discussion: In coumarine users, COX-2-selective NSAIDs are associated with less bleeding complications than non-selective NSAIDs are. Duration of NSAID use, as well as intensity of coumarine treatment, plays an important additional role. When the coumarine-NSAID combination is inevitable in an individual patient, a COX-2-selective NSAID may be preferred, with careful monitoring of the INR.