Rates of Bleeding and Discontinuation of Direct Oral Anticoagulants in Patients With Decompensated Cirrhosis

Studies of the effects of direct oral anticoagulants (DOACs) in patients with cirrhosis have been limited by their small sample size, inclusion of patients with well-compensated cirrhosis, short follow-up times, inadequate validation of cirrhosis diagnoses, and non-standard definitions of bleeding. We aimed to systematically determine the characteristics, indications, and outcomes of patients with cirrhosis of all severity classes who received DOACs. We performed a retrospective study of 138 patients with confirmed cirrhosis (93 with Child-Turcotte-Pugh scores of B or C) at a single center who started DOAC therapy (58,984 person-days; median, 181 days per patient) from September 2011 through April 2019. We collected data on clinical characteristics, indications for DOAC use, and outcomes. Standardized and validated definitions for bleeding complications were used. Twenty-nine patients (21%) stopped therapy due to a diagnosis of or perceived bleeding. The most common bleeding events were non-variceal upper and lower intestinal bleeding. No pretreatment laboratory parameters were associated with bleeding while patients received treatment, including platelet count (P = .50), international normalized ratio (P = .34), creatinine (P = .27), and model for end-stage liver disease score (P = .22). Frequency of bleeding events related to DOAC did not differ significantly among patients of different Child-Turcotte-Pugh classes (P = .81), DOAC indications (P = .60), or DOAC dosages (P = .10). Higher proportions of patients with hepatocellular carcinoma (P = .01) had major bleeding while receiving. Patients with decompensated cirrhosis have significant bleeding and rates of discontinuation of DOACs when they take them long term. Pretreatment laboratory parameters, DOAC dose, and Child-Turcotte-Pugh class were not associated with bleeding; hepatocellular carcinoma was associated with major bleeding.