Withdrawal of aspirin in patients with left ventricular assist device treated with vitamin K antagonists: impact of anticoagulation quality in the randomized ARIES-HM3 trial

BACKGROUND AND AIMS: Left ventricular assist devices (LVADs), including the HeartMate 3 (HM3), have improved outcomes in patients with advanced heart failure. Use of vitamin K antagonists (VKA) is mandated to reduce the risk of thrombotic events, but there is heterogeneity in management. Time in therapeutic range (TTR) is a crucial metric for assessing the quality of VKA management. The ARIES-HM3 trial demonstrated that aspirin can be safely omitted from the antithrombotic regimen, resulting in reduced bleeding without increased thrombosis. This pre-specified trial analysis explores the relationship of quality of VKA management assessed by TTR with haemocompatibility-related outcomes. METHODS: ARIES-HM3 was an international, randomized, double-blind, placebo-controlled study of aspirin (100 mg/day) or placebo (1:1) with VKA therapy in patients with de-novo HM3 placement. Participants were stratified into low TTR or high TTR groups based on median levels (n = 554). Primary endpoint success and secondary endpoint rates were stratified based on TTR groups. Bleeding rates at 12 months were estimated using an Andersen-Gill model with TTR as a single continuous variable, and multivariable regression analysis was performed. RESULTS: The percentage of patients with a TTR above or below the median of 56 was similar between the aspirin and placebo groups. More participants achieved primary endpoint success with TTR ≥56% (77% vs 66.9%, P = .01). Higher TTR was associated with lower bleeding rates at 12 months (26.4 vs 49.2 events per 100 participant-years; rate ratio 1.84, 95% confidence interval CI 1.37-2.53) without stroke increase (3.2 vs 2.8 events per 100 participant-years; rate ratio 0.88 95% CI: 0.30-2.53). No interaction was observed between the assigned treatment group and TTR. Modelling demonstrated a constant decrease in bleeding as a function of increasing TTR. Female sex and Black race were independent predictors of low TTR (odds ratio: 1.70 95% CI: 1.12-2.57; 1.62 95% CI: 1.11-2.35, respectively), with more frequent INRs below the therapeutic range. Multivariable modelling identified age ≥65 years, aspirin use, TTR <56%, and blood urea nitrogen ≥30 mg/dL as predictors of non-surgical bleeding. CONCLUSIONS: The quality of VKA management as measured by TTR correlates with the occurrence of non-surgical bleeding in patients with the HM3 LVAD, with a lower TTR associated with an increased bleeding risk. These data provide new clinical direction to define a benchmark TTR to achieve further mitigation of residual risk of bleeding and enhance haemocompatibility with the HM3 LVAD.