To describe adoption of aspirin avoidance after ARIES-HM3, assess whether aspirin use is concentrated among patients with vascular disease, and compare routine-care outcomes with randomized-trial findings.
Approach:
Data source: Retrospective analysis of deidentified electronic health records from 30 US health systems in the Truveta database.
Cohorts: The overall implementation cohort includes 3,365 adults undergoing de novo durable left ventricular assist device (LVAD) implantation from January 2016 through March 5, 2026. The contemporary HeartMate 3 (HM3) outcomes cohort includes 1,577 eligible patients implanted from June 3, 2021, through December 31, 2025, with documented warfarin therapy within 14 days. Primary outcome comparisons focus on 739 post-ARIES-HM3 recipients.
Exposure and follow-up: Early aspirin use requires a recorded administration, dispense, or active order on postoperative days 2–7. Follow-up begins at day 14 among patients alive, supported with the index device, under observation, and free of major nonsurgical hemocompatibility-related adverse events.
Key Findings:
Recorded aspirin use in the overall LVAD cohort declines from 90.2% in 2023 to 55.7% in 2024 and 52.7% in 2025.
Among post-ARIES-HM3 recipients, 54.7% have recorded early aspirin use: 67.1% of patients with prior coronary intervention, coronary bypass surgery, stroke, or peripheral vascular disease versus 43.4% without these conditions.
The 12-month composite of death or first major nonsurgical hemocompatibility-related adverse event occurs in 29.3% with aspirin and 21.7% without recorded aspirin. The association is not significant after adjustment (HR for no recorded aspirin versus aspirin, 0.86; 95% CI, 0.61–1.22).
Major nonsurgical bleeding occurs in 22.0% and 15.0%, respectively; the adjusted association is not significant (HR, 0.82; 95% CI, 0.54–1.26). No excess stroke or peripheral arterial thromboembolism is observed without recorded aspirin; pump thrombosis is rare.
Interpretation:
Aspirin avoidance increases after ARIES-HM3, but aspirin persists particularly among patients with vascular disease. Outcomes reveal no safety signal discordant with the randomized evidence; they do not establish comparative effectiveness.
Limitations:
Nonrandomized treatment selection and residual confounding preclude causal conclusions.
Early medication records do not establish long-term adherence, discontinuation, crossover, or over-the-counter use.
Outcomes rely on electronic-record algorithms without patient-level adjudication. Rare thrombotic events limit precision.
Unequal follow-up, exclusion of events before day 14, unavailable timing of prior vascular events, and incomplete event capture limit interpretation. Findings are not nationally representative.
Conclusion:
Aspirin use declines substantially but remains common after HM3 implantation. Early aspirin avoidance is not associated with increased adjusted composite risk. Findings support continued implementation with individualized management for recent vascular events and surveillance for rare thrombotic complications.
