Common P-glycoprotein (ABCB1) polymorphisms do not seem to be associated with the risk of rivaroxaban-related bleeding events: Preliminary data.

ABSTRACT

Introduction: Considering conflicting previous reports, we aimed to evaluate whether the common ABCB1 polymorphisms (rs1128503, rs2032582, rs1045642, rs4148738) affected the risk of bleeding in rivaroxaban-treated patients. Materials and methods: We report preliminary data from a larger nested case-control study. Consecutive adults started on rivaroxaban for any indication requiring > 6 months of treatment were followed-up to one year. Patients who experienced major or non-major clinically relevant bleeding during the initial 6 months were considered cases, whereas subjects free of bleeding over > 6 months were controls. The polymorphisms of interest (rs1128503, rs2032582, rs1045642, rs4148738) were in a strong linkage disequilibrium, hence patients were classified regarding the "load" of variant alleles 0-2, 3-5 or 6-8. The three subsets were balanced regarding a range of demographic, comorbidity, comedication and genetic characteristics. A logistic model was fitted to probability of bleeding. Results: There were 60 cases and 220 controls. Raw proportions of cases were similar across the subsets with increasing number of ABCB1 variant alleles (0-2, N = 85; 3-6, N = 133; 6-8, N = 62) 22.4%, 21.8%, and 19.4%, respectively. Fully adjusted probabilities of bleeding were also similar across the subsets 22.9%, 27.5% and 17.7%, respectively. No trend was observed (linear, t = -0.63, df = 273, P = 0.529; quadratic, t = -1.10, df = 273, P = 0.272). Of the 15 identified haplotypes, the completely variant (c.1236T_c.2677T(A)_c.3435T_c.2482-2236A) (40.7%) and completely wild-type (C_G_C_G) (39.5%) haplotypes prevailed, and had a closely similar prevalence of cases 21.1% vs. 23.1%, respectively. Conclusions: The evaluated common ABCB1 polymorphisms do not seem to affect the risk of early bleeding in patients started on rivaroxaban.