Comparing Anticoagulation Strategies in Left Ventricular Assist Devices - Direct Oral Anticoagulants, No Anticoagulation, and Warfarin: A Case Series and Systematic Review of Literature

Introduction

Left ventricular assist devices (LVADs) are being used increasingly for medical therapy-resistant heart failure, reducing mortality but associated with bleeding complications due to anticoagulation (AC) use. Vitamin K Antagonists (VKAs) are the standard of care in patients with LVADs, while the safety and efficacy of direct oral anticoagulants (DOACs) remains uncertain. We report our experience of using DOACs or discontinuing AC in patients with LVADs, followed by a systematic review of the literature.

Methods

We conducted a retrospective, single-center review, identifying patients with HeartMate III (HM3) LVAD, prescribed a DOAC or being monitored off AC after experiencing warfarin failure between 2018 and 2024. Demographics, therapy duration, reasons for change in AC, adverse events such as stroke, thromboembolism, pump thrombosis, major bleeding (MB), and life-threatening bleeding were recorded. A systematic search of PubMed, CINAHL, and OVID databases using the PRISMA guidelines identified studies evaluating the use of DOACs in patients with LVADs, excluding animal or ex-vivo studies.

Results

Six patients (4 males, average age of 71 ± 6 years, and BMI of 29.7 kg/m²) are included from our institution. Following warfarin failure, 3 patients were switched to apixaban 2.5 mg twice daily, 1 patient to rivaroxaban 20 mg once daily, and AC was discontinued for 2 patients. They remained on warfarin for 850 ± 558 days, DOACs for 345 ± 302 days (all 4 continue on DOAC to present), and were off AC for 646 ± 365 days. Previously, time within therapeutic range on warfarin averaged 37%. Reasons for change in AC included recurrent gastrointestinal bleeding (GIB), intracranial hemorrhage (ICH), and supra-therapeutic INR (more than 10 with no overt bleeding). Nine bleeding events occurred in patients while on warfarin (7 were life-threatening including ICHs and GIBs) whereas only 1 patient had a major GIB and ICH while on a DOAC. AC was discontinued for 2 patients after life-threatening GIBs on warfarin, with only 1 recurrence. Only 1 transient ischemic attack occurred on warfarin, with no other thrombotic events.

49 studies were screened, and 15 were included in the review (8 case reports with 1-7 patients, 4 retrospective single-center analyses (RSCA) with 7-63 patients, and 3 randomized controlled trials (RCTs) with 16-45 patients, totaling 222 patients). Three studies involved HeartWare LVAD patients, of which an RCT was halted early due to 50% (4/8) pump thrombosis on dabigatran vs. 12.5% (1/8) on VKA. Five studies involved the HeartMate II LVAD; 2 RSCAs with 7 patients each assessed dabigatran vs. VKA and apixaban/rivaroxaban vs VKA with similar incidence of bleeding and thrombotic complications in each group. Seven studies involved the HM3 LVAD. 2 RSCAs with 35 and 63 patients reported similar incidence of thrombotic events, 6% vs 10%, p = 1.0 and 24% vs. 15%, p=0.75, and clinically lower bleeding events or related admission, 5% vs. 30%, p=0.1, 2.2 vs. 9.7 admission per 1000 days on therapy, in patients on apixaban vs. VKA, respectively. 2 RCTs in HM3 LVAD patients reported no thrombosis on a VKA or a DOAC but demonstrated a clinical reduction in rates and severity of bleeding events (1/47 vs. 6/28) in patients on apixaban vs. VKA.

Conclusion

In our experience, albeit a very small sample size, the use of apixaban or rivaroxaban in LVAD patients following VKA failure was not associated with increase in thrombosis but with a clinical reduction in number and severity of bleeding events. In 2 LVAD patients for whom AC was discontinued due to life-threatening bleeding on a VKA, no thrombotic complications occurred for 2.5 and 0.75 years. Furthermore, review of literature supports our findings though total number of patients is relatively small. Nonetheless, there is growing evidence of the safety and efficacy of administering DOACs in patients with LVADs, although more extensive trials and registry studies are required to confirm this and further reassure clinicians given the prevailing standard of care of “never” administering a DOAC in a LVAD patient.

No relevant conflicts of interest to declare.