Apixaban Anticoagulation in HeartMate 3 Left Ventricular Assist Device: A Meta-Analysis of Randomized Controlled Trials

As the burden of advanced heart failure continues to escalate, the role of left ventricular assist devices (LVADs) became increasingly prominent. Consequently, the latest guidelines stated a grade IA recommendation for patients classified as INTERMACS 1–3 and a IIb–B recommendation for those classified as INTERMACS 4.1 Given the excellent LVAD performance and organ donation shortages, LVADs are now implanted as both bridge-to-transplant (BTT) and destination therapy (DT), with similar evidence. Nevertheless, one major drawback negatively affecting patients’ quality of life is the occurrence of hemocompatibility-related adverse events (HRAEs), involving thrombotic and hemorrhagic complications that are secondary to and/or associated with LVAD therapy. Indeed, while antithrombotic therapy is needed to prevent thromboembolic complications (stroke, pump thrombosis), it also raises the bleeding risk. Since the ARIES-HM3 trial,2 anticoagulant therapy (vitamin-K antagonist [VKA]) without antiplatelet therapy (low-dose aspirin) became the gold standard in LVAD anticoagulation management. Nevertheless, VKAs pose important challenges such as noncompliance, bleeding, drug/drug–food interactions, and the need for regular monitoring to stay within the therapeutic range.1,3 Direct oral anticoagulants (DOACs) have proven to address these issues, primarily by reducing intracranial hemorrhages. Nevertheless, caution is warranted about switching every indication from VKAs to DOACs, as the sole trial on DOACs in mechanical heart valves, showed no benefit and increased bleeding risk.4 The direct thrombin inhibitor dabigatran was the first DOAC tested in a randomized controlled trial (RCT) in LVAD patients, resulting in 50% thromboembolic events in the dabigatran arm leading to a contraindication of dabigatran, and by extension every DOAC in LVAD patients.5 However, more recently, some ex vivo and observational studies suggested that the use of apixaban (factor-Xa inhibitor) could be safe and noninferior to VKAs.6 Moreover, two small-scale RCTs, DOAC-LVAD7 and DOT-HM3,8 evaluated apixaban versus warfarin with the primary endpoint being the occurrence of HRAEs, a composite criterion encompassing thromboembolic (stroke, pump thrombosis) and hemorrhagic events (any location) (see Supplementary Table 1, Supplemental Digital Content, https://links.lww.com/ASAIO/B412), concluding that anticoagulation with apixaban appears feasible in LVAD patients without increasing HRAEs, although larger RCTs are warranted before adoption in clinical practice. However, both studies may be subject to biases due to the small scale of both patient and center numbers. Therefore, we performed a pooled data meta-analysis to reassess safety and efficacy of apixaban in LVAD patients compared with VKAs. Methods In our meta-analysis, heterogeneity, assessed by Higging’s I²-index, deemed acceptable at I² < 50%, led to a fixed-effects model application. Analyses were two-tailed, with statistical significance at p < 0.05. Statistical analysis used R software with the “meta” (v. 7.0.0) package. Subgroup analyses were conducted to compare patients who received apixaban or warfarin with add-on therapy of aspirin, and to compare apixaban alone versus warfarin and aspirin. Results Within two trials, 75 LVAD patients were randomized to either apixaban (n = 47) (31 receiving concomitant aspirin) or warfarin (all of them in combination with aspirin) (n = 28), followed for a weighted mean of 175.2 days amounting to 36.0 patient-years of follow-up (see Supplementary Table 2, Supplemental Digital Content, https://links.lww.com/ASAIO/B412). All patients were supported with a fully magnetically levitated LVAD (HeartMate 3; Abbott [North Chicago, IL]). Survival-free from HRAEs was comparable between apixaban and warfarin groups in the overall analysis (relative risk [RR]: 1.11 [0.95–1.30], p = 0.55), as well as in the subgroup analysis including patients receiving apixaban or warfarin with add-on therapy of aspirin (RR: 1.11 [0.94–1.31], p = 0.45), and in the subgroup analysis comparing apixaban alone versus warfarin and aspirin (RR: 1.13 [0.97–1.33], p = 0.72) (Figure 1). In both studies, no thromboembolic event was reported. Major gastrointestinal bleeding occurred in one patient in the apixaban group and in two patients in the warfarin group. In addition, in the DOT-HM3 study, six heart transplants were performed in apixaban-treated patients and one in the warfarin group, without excessive postoperative bleeding. No deaths occurred in either group, however, one patient was withdrawn from DOT-HM3 after being transferred to palliative care.Figure 1.: Meta-analysis of survival free from HRAEs for apixaban vs. warfarin in HeartMate 3 patients. CI, confidence interval; DOAC, direct oral anticoagulant; HRAE, hemocompatibility-related adverse event; LVAD, left ventricular assist device; RR, relative risk.Discussion This meta-analysis supports evidence that treatment with apixaban appears feasible for HeartMate 3 patients, without increased thromboembolic events. Notably, the only other existing RCT to date assessing dabigatran included patients with HeartWare, an older-generation LVAD associated with a higher risk of HRAEs compared with HeartMate3 in large-scale registries.5 Conversely, apixaban works as a Factor-Xa inhibitor, more proximally in the coagulation cascade, making it more efficient in preventing thrombin generation.6 Although no excessive bleeding was reported for the six transplanted patients in the apixaban group in DOT-HM3, larger studies focusing on this point are necessary to draw firm conclusions.8 Importantly, both trials included were conducted before publication of the ARIES-HM3 study results,2 showing a reduced bleeding rate after elimination of aspirin and leading to anticoagulant alone as the standard treatment for LVAD patients. In our meta-analysis, almost all patients included were treated with aspirin in addition to the anticoagulant (apixaban or warfarin), which constitutes a bias in the collection of hemorrhagic events. This meta-analysis is limited by the single-center nature of the RCTs and the small number of LVAD patients included. The ongoing ApixiVAD study,6 a multicenter, international, randomized, controlled trial, plans to enroll 50 LVAD patients to assess efficacy and safety outcomes for apixaban, even though it remains underpowered to demonstrate superiority/noninferiority of DOACs versus VKA. Our meta-analysis further supports conducting large, international multicenter RCTs to demonstrate the safety of apixaban as compared with VKAs. Finally, although no thromboembolic events were reported with apixaban in trials, case reports highlight this risk,3 underscoring the necessity to further clarify the selection criteria for LVAD candidates for DOAC treatment. Conclusion In conclusion, an apixaban-guided anticoagulation strategy appears feasible and safe in selected BTT or DT patients supported by a HeartMate 3 LVAD. Further large-scale RCTs are needed to assess whether DOACs can decrease the incidence of hemorrhagic events (safety) while maintaining equivalent antithrombotic efficacy as VKAs.