AIMS: Pharmacological effect of direct oral anticoagulants (DOACs) is influenced by binding status with albumin. This study aimed to assess the association between albumin levels and bleeding risk in atrial fibrillation (AF) patients treated with DOACs.

METHODS: We conducted DIRECT-Extend registry (N=7,512), a pooled database combining three large-scale observational study of AF patients treated with DOAC. The primary endpoint was major bleeding as defined by International Society on Thrombosis and Haemostasis criteria. Multivariable Cox hazard model was used to assess the impact of albumin level on major bleeding.

RESULTS: Out of the overall cohort, 2,523 patients [73 (IQR 66-80) years, 1,620 (64.2%) males] with albumin data available at enrollment were analyzed in this study. Median follow-up duration was 532 days (IQR 94-1405 days). The entire cohort was divided into tertiles based on albumin levels (Lower tertile: <3.7 g/dL, Middle tertile: 3.7-4.1 g/dL, and Higher tertile: ≥4.1 g/dL). The incidences of major bleeding increased as albumin levels decreased; 56 patients (6.8%), 81 patients (9.7%), and 113 patients (13.1%) in the higher, middle, and lower tertiles, respectively. (Log-rank test p value < 0.0001). A lower albumin level was independently associated with a higher incidence of major bleeding (adjusted hazard ratio 0.61, 95% confidence interval 0.47-0.80, P value < 0.01), which was consistently observed in all DOACs (P value for interaction > 0.05).

CONCLUSIONS: A lower albumin level was independently associated with a higher bleeding risk in AF patients using DOACs. Careful attention should be paid to hypoalbuminemia when prescribing DOACs.