The Effect of Flucloxacillin on Warfarin Anticoagulation: A Swedish Register-Based Nationwide Cohort Study.

Abstract

Data indicate that codispensing flucloxacillin to patients already on warfarin may result in decreased warfarin efficacy. This article investigates the effect of flucloxacillin on warfarin anticoagulation. In a retrospective cohort study of warfarin users, using three nationwide registers we included 5,848 patients receiving 10 days flucloxacillin treatment and 201 with ≥30 days treatment. To assess the potential for confounding by indication, we also identified 21,430 individuals initiating phenoxymethylpenicillin. International normalized ratio (INR) values and warfarin doses were calculated day-by-day and proportion of patients with a subtherapeutic INR week-by-week during cotreatment. Following initiation of flucloxacillin with a planned treatment duration of 10 days and ≥30 days, the mean INR decreased from 2.36 (95% confidence interval [CI] 2.34; 2.37) to 2.20 (95% CI 2.19; 2.21) and from 2.24 (95% CI 2.16; 2.32) to 1.96 (95% CI 1.89; 2.02), respectively. Consequently, for individuals with 10 days treatment the proportion of patients with a subtherapeutic INR of < 2 increased from 22% in the week preceding flucloxacillin initiation to 35% in the third week after initiation of flucloxacillin. In patients with 30 days treatment, the proportion increased from 34 to 63% by week 6. In individuals initiating phenoxymethylpenicillin, INR levels did not decrease. One in three patients with 10 days flucloxacillin and almost two in three patients initiating long-term treatment, was exposed to a subsequent subtherapeutic anticoagulant effect. To avoid unnecessary thromboembolic complications, the initiation of flucloxacillin should be accompanied by closer INR monitoring which may be especially important among individuals with lengthy treatments.