Outcomes of Dabigatran and Warfarin for Atrial Fibrillation in Contemporary Practice: A Retrospective Cohort Study

Abstract

Background: Dabigatran (150 mg twice daily) has been associated with lower rates of stroke than warfarin in trials of atrial fibrillation, but large-scale evaluations in clinical practice are limited.

Objective: To compare incidence of stroke, bleeding, and myocardial infarction in patients receiving dabigatran versus warfarin in practice.

Design: Retrospective cohort.

Setting: National U.S. Food and Drug Administration Sentinel network.

Patients: Adults with atrial fibrillation initiating dabigatran or warfarin therapy between November 2010 and May 2014.

Measurements: Ischemic stroke, intracranial hemorrhage, extracranial bleeding, and myocardial infarction identified from hospital claims among propensity score-matched patients starting treatment with dabigatran or warfarin.

Results: Among 25 289 patients starting dabigatran therapy and 25 289 propensity score-matched patients starting warfarin therapy, those receiving dabigatran did not have significantly different rates of ischemic stroke (0.80 vs. 0.94 events per 100 person-years; hazard ratio [HR], 0.92 [95% CI, 0.65 to 1.28]) or extracranial hemorrhage (2.12 vs. 2.63 events per 100 person-years; HR, 0.89 [CI, 0.72 to 1.09]) but were less likely to have intracranial bleeding (0.39 vs. 0.77 events per 100 person-years; HR, 0.51 [CI, 0.33 to 0.79]) and more likely to have myocardial infarction (0.77 vs. 0.43 events per 100 person-years; HR, 1.88 [CI, 1.22 to 2.90]). However, the strength and significance of the association between dabigatran use and myocardial infarction varied in sensitivity analyses and by exposure definition (HR range, 1.13 [CI, 0.78 to 1.64] to 1.43 [CI, 0.99 to 2.08]). Older patients and those with kidney disease had higher gastrointestinal bleeding rates with dabigatran.

Limitation: Inability to examine outcomes by dabigatran dose (unacceptable covariate balance between matched patients) or quality of warfarin anticoagulation (few patients receiving warfarin had available international normalized ratio values).

Conclusion: In matched adults with atrial fibrillation treated in practice, the incidences of stroke and bleeding with dabigatran versus warfarin were consistent with those seen in trials. The possible relationship between dabigatran and myocardial infarction warrants further investigation.

Primary funding source: U.S. Food and Drug Administration.