New Antithrombotic Strategies to Improve Outcomes With the HeartMate 3

Left ventricular assist devices (LVADs) have progressively broadened the expectations of survival for patients with heart failure, being de facto permanent therapy with scarce adherence to the previous labeling with strategies of implant. To date, the 2 year survival with an LVAD is competitive with heart transplant1 and the number of patients who remain on support for 5 years and more is continuously increasing. Technological innovations of modern devices (miniaturized, intrapericardial pumps) have largely contributed to this clinical success. The most recent advances have been achieved in hemocompatibility. In particular, low risk of pump thrombosis has been reported in patients implanted with the HeartMate 3 (HM3, Abbott, Chicago, IL)2; moreover, retrospective analysis of the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) database revealed that the HM3 is associated with a significant lower hazard of major neurologic adverse events as compared with other continuous-flow pumps.3 However, nonsurgical bleeding remains a major concern also with this pump. Importantly, the rate of bleeding complications is consistent across all series of HM3 support. The Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy With HeartMate 3 (MOMENTUM 3) trial reported that that >40% of HM3 patients suffered from bleeding events over a 2 year follow-up, 10% of them requiring surgery.2 In the Evaluating the HeartMate 3 with Full MagLev Technology in a Post-Market Approval Setting (ELEVATE) registry, major bleeding was observed in >30% of patients.4 Accordingly, the identification of reliable strategies to prevent and safely treat hemorrhagic events with the HM3 is urgently needed. Pathophysiology of bleeding in HM3 patients is still not fully understood. Netuka et al. reported attenuated hemostatic defects driven by the pump, namely von Willebrand factor (vWf) degradation.5 Nevertheless, although reduced with respect to previous generation devices, vWf degradation in HM3 patients was associated with significant vWf functional deficiency and impaired hemostatic activity.6 On the other hand, recent data highlight potential inferences of the current recommended antithrombotic (AT) strategy (warfarin titrated to a target international normalized ratio [INR] of 2.0–3.0 and aspirin 81–100 mg/die) to magnify the bleeding risk in HM3 patients. In particular, data questioning the need for aspirin in HM3 patients are accumulating, and the observation of a contribution of aspirin to bleeding in the setting of HM3 support is being increasingly reported.6–10 The safety of aspirin-free AT regimen and the poor evidence for aspirin requirement in HM3 patients is consistent with the low thrombogenicity of the pump: previous studies have indeed shown no differences in the rate of ischemic stroke in HM3 patients managed with or without antiplatelet drugs.2,11 Supporting clinical evidences, we have proposed a mechanistic rationale for aspirin-free AT regimen in LVAD patients, showing that platelet activation is not/minimally influenced by aspirin in this peculiar clinical setting.9 Reduced AT regimen with the HM3 has also been approached with a look at INR: Netuka et al. have suggested to titrate warfarin therapy to a lower INR target (1.5–1.9).12 In the Minimal Anticoagulation Evaluation To Augment Hemocompatibility study, the authors reported that >90% of the enrolled patients met the primary end-point of survival free of pump thrombosis, disabling stroke, or major bleeding at 6 months of follow-up.12 To date, available data are fragmented and protocols for reduced AT regimen with the HM3—aspirin discontinuation, lower INR target, or the combination of both—are extremely variable among institutions. In particular, while some studies portend toward primary warfarin monotherapy (no aspirin at discharge) in selected patients,7 others suggest to use aspirin in the early postimplant period because of a hypothesized higher thrombotic risk, and then to switch to aspirin-free regimen8 or low-intensity anticoagulation12; however, no reliable criteria have been proposed on when to start reduced AT strategy. Moreover, despite overall consensus on the need for a rationale rethinking of current AT strategy with the HM3, whether to adopt short-term or definitive cessation of aspirin after a bleeding event has been poorly analyzed. In addition, although recurrence of bleeding has been observed in some patients who discontinued aspirin after a primary bleeding event,9,10 few data are available on how to treat patients who suffer from recurrent bleeding. Here, we synthesize contemporary clinical data on long-term outcomes with reduced anticoagulation with the HM3 and derive a new practical algorithm for rationale management of anticoagulation in HM3 patients, which applies for the prevention of primary bleeding events as well as postevent treatment strategy to avoid their recurrence. The first clinical concern is how to manage anticoagulation to safely treat a bleeding event and avoid recurrence without ending up to thrombotic events. We have suggested the safety and efficacy of definitive aspirin withdrawal in HM3 patients after a bleeding event, showing low rate of recurrence and no differences in the rate of thromboembolic complications in the long term as compared with patients managed with aspirin [7,9]. Safety and efficacy of aspirin discontinuation was also reported by Lim et al., who showed that warfarin monotherapy was not associated with long-term increased risk of thrombosis in patients with the HM38. The article by Marshall et al. further adds for a revision of AT therapy after a bleeding event, as the authors confirm low rate of bleeding recurrence and thromboembolic events in patients switched to a reduced AT regimen.10 However, Marshall et al.10 also enrolled patients who had definitive discontinuation of warfarin or cessation of any anticoagulation after a bleeding event; while we agree that full withhold of anticoagulation may be of help during the acute phase of a bleeding event, to the best of our knowledge, no data exist supporting its safety in the long term. On the basis of these considerations, we introduce a new algorithm for reduced anticoagulation management with the HM3 after a bleeding event (Figure 1), which challenges the paradigm of aspirin use according to six main evidences:Figure 1.: Algorithm for rationale management of reduced anticoagulation in patients with the HeartMate 3 (HM3) after a bleeding event. GI, gastrointestinal; INR, international normalized ratio. We must recognize that different bleeding events might be treated with a different approach depending on INR value at the time of the event: in the case of supratherapeutic INR (e.g., INR > 4), its stabilization must be pursued before any change in AT regimen. Conversely, bleeding events not related to overanticoagulation (INR < 4) should trigger aspirin discontinuation. We must acknowledge that aspirin discontinuation might be adopted irrespective of the bleeding severity, as we know that minor events (such as epistaxis) are an important signal. However, this strategy should be pursued with different temporal scales—that is, short-term cessation vs. definitive—as a function of the “type” of bleeding. In the case of retroperitoneal bleeding, we suggest to stop aspirin and adopt warfarin monotherapy titrated to a lower INR target (INR 1.5–2). Lower anticoagulation levels (INR 1.5-2) should be also considered after recurrence of any type of bleeding in patients who previously discontinued aspirin. Stopping aspirin is suggested after intracranial hemorrhage; on the other hand, the risk of hemorrhagic conversion of an embolic stroke is very high in LVAD patients and accordingly aspirin should be withheld in the acute phase similarly. Although recent data showed lower incidence of bleeding events in a small cohort of HM3 patients treated with apixaban with respect to patients treated with warfarin,13 we highlight high risk of thromboembolic complications associated with novel oral anticoagulants in the setting of LVAD support.14 The prevention of first bleeding events is a further compelling clinical need, which requires thorough evaluation on whether/how to expand previous concepts to primary AT strategies. Preliminary data suggest that warfarin monotherapy (i.e., no aspirin) at discharge in patients stratified according to preimplant higher risk of bleeding is reasonable.7 Conversely, the use of aspirin in the early postimplant period and delayed switch to an aspirin-free regimen might have lower efficacy to prevent bleeding episodes.8 Efficacy of lower INR target (1.5–1.9) was also reported to prevent bleeding in HM3 patients.12 However, possible detrimental effects of lower INR target in patients at high risk of thromboembolic complications restrict the clinical applicability of this protocol to only a few patients.12 As a result, we remark that preimplant stratification of patients at high bleeding or thrombotic risk is essential to properly tailor the AT regimen and should become routine in clinical practice; yet, specific validated stratification tools are urgently required for the LVAD population, aspirin dose (if any) should be 81–100 mg/die, and there is no rationale for dual antiplatelet therapy,15 patients with a significant prohemorrhagic profile might be safely managed without aspirin at discharge (warfarin monotherapy), tailoring the INR target to 2–2.5,7 or 2.5–316 in case of concomitant high risk of stroke (e.g., A-fib, LV thrombus). In case of a bleeding event with INR in the therapeutic target, the intensity of anticoagulation should be further decreased to 1.8–2 and thereafter 1.5–1.8 if bleeding recurs, omega-3 fatty acids, digoxin, or angiotensin-converting enzyme inhibitors/angiotensin receptor blockers have been shown to contribute to reduce the risk of gastrointestinal bleeding in routine clinical care and stabilize INR and deserve further evaluation,17–19 In case of subtherapeutic INR (<1.5), bridging with low molecular weight heparin should be only pursued for patients at high thrombotic risk.20,21 In conclusion, according to contemporary real-world experiences, the clinical need to prevent bleeding events with the HM3 might be addressed with new paradigms of anticoagulation strategy. In particular, while waiting for randomized trials to definitely determine the clinical need for aspirin after HM3 implantation,22 preliminary but consistent and reproducible data suggest that aspirin-free AT regimen might be a reliable strategy both in primary and rescue settings. In this background, we defined a new algorithm for rationale management of reduced anticoagulation with the HM3, which challenges the paradigm of aspirin use in this setting, and calls for new tools to stratify patients according to bleeding and thrombotic risk profile and eventually tailor the anticoagulation regimen according to patient-specific clinical characteristics. Indeed, we highlight a gap of knowledge in terms of reliable criteria to identify patients at high risk of bleeding in the setting of LVAD support who might tolerate and benefit from aspirin-free AT regimen. Accordingly, while we call for new markers of bleeding predisposition relevant to LVAD-hemocompatibility, as we know that the device per se triggers peculiar pathologic mechanistic pathways that ultimately lead to adverse events, we must acknowledge that the issue of anticoagulation in LVAD patients goes beyond the “simple” need to counteract complications driven by the pump. In other words, the choice of AT regimen should be contextualized according to the wide spectrum of collateral—and often concomitant—prothrombotic and hemorrhagic factors typical of LVAD patients. These features might be present at baseline or emerge later (still unpredictable) after LVAD implantation (e.g., infection). To date, standardized strategies to properly balance bleeding risk with that of thrombosis on a patient-based level are elusive. Despite previous studies identified demographic patients’ characteristics (age at implant),6 baseline right ventricular dysfunction,23 or specific postimplant therapeutic targets24 as potential markers of high bleeding risk, we warn that none of these markers might alone comprehensively address the complex scenario depicted previously. Moreover, previous experiences with multivariable scores of preoperative bleeding risk are limited to the translation of risk scores from different clinical scenarios and require validation in the setting of LVAD implantation.7,25,26 According to those considerations, the proposed protocol should not be interpreted as a recommendation and should not drive changes in standard AT management until validated. Prospectively, the effectiveness of strategies and procedures that reduce the risk of thromboembolic events (e.g., left atrial appendage occlusion27) in terms of higher tolerance (safety) of a reduced AT regimen should be also acknowledged. Furthermore, the role of systematic evaluation of prothrombotic platelet function as a marker of higher thrombotic risk9,28,29 as well as of novel models recently introduced to predict gastrointestinal bleeding30 warrants further investigation. Of note, quantitative estimates of the efficacy of new protocols or therapeutic approaches, that is, objective assessment of freedom from hemocompatibility-related adverse events, are currently available.31