Risk of clinically relevant bleeding in warfarin-treated patients--influence of SSRI treatment

Purpose To investigate the risk of clinically relevant bleeding in warfarin-treated patients with or without concomitant treatment with selective serotonin reuptake inhibitors (SSRIs). Methods A cohort study was performed in patients treated with warfarin due to atrial fibrillation. Exposed patients were defined as patients treated with SSRI at any time between January 1999 and September 2005 (n = 117). Unexposed patients without SSRI-treatment were randomly selected and matched for age and sex (1:1). The primary endpoint was hospital admission due to bleeding during the same time period. Results Bleeding occurred in 17 exposed patients (totally 23 bleedings) and in two unexposed patients (totally two bleedings) (p = 0.0003). A total of 11 bleedings occurred during treatment with the combination of warfarin and SSRI, and 14 during treatment with warfarin only. The total incidences of bleedings per 1000 treatment years were 51.4 (25.7–92.0) and 23.9 (13.1–40.1), respectively, and the unadjusted incidence rate ratio (IRR) 2.15 (0.88–5.11). Cox regression analysis including first bleedings revealed an adjusted hazard ratio of 3.49 (1.37–8.91) for bleeding during treatment with a combination of SSRI and warfarin compared with treatment with warfarin only. Initiation of SSRI therapy was not associated with a change in dose of warfarin or with a change in international normalized ratio (INR) (p = 0.48 and p = 0.31, respectively). Conclusion Addition of SSRI to warfarin-treated patients may be associated with an increased risk of clinically relevant bleeding. The effect seems not to be associated with a direct influence of SSRI on the anti-coagulant activity of warfarin. Copyright © 2009 John Wiley & Sons, Ltd.