Incident thrombocytopenia and bleeding risk in elderly patients with atrial fibrillation on direct oral anticoagulants: insights from the ATHEROsclerosis in Atrial Fibrillation study

Background: The bleeding risk of patients with atrial fibrillation (AF) changes over time. Most studies thus far evaluated only the baseline bleeding risk with discordant results. The impact of incident thrombocytopenia during direct oral anticoagulant (DOAC) therapy and its relation to bleeding has not been previously investigated. Objectives: To investigate the incidence rate of thrombocytopenia and major bleeding (MB) risk in AF patients on DOACs. Methods: Prospective ongoing ATHEROsclerosis in Atrial Fibrillation study including patients with nonvalvular AF on DOACs. Incident thrombocytopenia was defined as a platelet count <150 × 109/L. MB events were recorded at each follow-up visit. Gray estimator for competing risk data was used. Estimates are expressed in terms of subdistributional hazard ratios (sHR) and relative 95% CI for MB. Results: We enrolled 957 AF patients treated with DOACs (mean age, 77.3 ± 9.0 years; 49.1% women). During a follow-up (median time to censoring 1330 days; 95% CI, 1246-1443), 139 patients developed thrombocytopenia (3.08 per 100 person-years; 95% CI, 2.27-3.89) with no difference between direct thrombin and factor Xa inhibitors. Overall, 179 bleedings occurred, of which 80 were major (3.17 per 100 person-years; 95% CI, 2.34-3.99). Patients sustaining bleedings were more frequently affected by arterial hypertension, heart failure, anemia and had higher CHA2DS2-VASc and HAS-BLED scores. On multivariable Cox analysis, independent risk factors for MB were incident thrombocytopenia (sHR, 12.77; 95% CI, 8.880-18.360; P < .001), and age (sHR, 1.030 per year; 95% CI, 1.010-1.040; P = .002). Conclusion: Patients developing thrombocytopenia have an increased risk of MB. Dynamic evaluation of platelet count during follow-up may provide better prognostic value than baseline assessment only.