Long-term outcomes of apixaban as main anticoagulant in patients with HeartMate 3 left ventricular assist devices
Abstract
Aims
HeartMate 3 (HM3) left ventricular assist devices (LVADs) offer improved haemocompatibility-related outcomes for end-stage heart failure patients, facilitating the exploration of alternative anticoagulation therapies beyond warfarin. This study presents a long-term evaluation of thrombotic and bleeding outcomes in HM3 LVAD patients transitioned from warfarin to apixaban.
Methods
We retrospectively identified HM3 LVAD patients at our single centre who transitioned from warfarin to apixaban. Baseline characteristics were described at discharge from implant hospitalization and at the last follow-up on each anticoagulation regimen. We reported survival, thrombo-embolic events (including LVAD pump thrombosis, stroke, arterial thrombo-embolic events and pump exchange) and bleeding events on both warfarin and apixaban.
Results
Eight patients were identified between May 2018 and June 2022 who transitioned from warfarin to apixaban 5 mg twice daily. Patients were followed for a mean of 1233 days after LVAD implantation and 789 days after transition to apixaban. All patients were transitioned due to difficulty maintaining a therapeutic international normalized ratio (INR), including five patients who experienced bleeding complications on warfarin. No patients encountered LVAD pump thrombosis, stroke events, arterial thrombo-embolic event, pump exchange or death. While on warfarin, five patients had eight bleeding events: one major [requiring 2 units of packed red blood cells (pRBCs)] and seven minors (five gastrointestinal bleeds, one episode of haematuria and one episode of haemoptysis). After switching to apixaban, one patient with angioectasia had a major gastrointestinal bleed requiring two pRBCs and endoscopic clipping.
Conclusions
Apixaban demonstrated safe and favourable long-term outcomes in a cohort of HM3 LVAD patients over a mean follow-up of more than 2 years. To our knowledge, our report provides the longest follow-up duration for this patient population to date. Larger prospective studies are needed before this can be adopted as the standard of care.
Introduction
Left ventricular assist devices (LVADs) have proved to be a lifesaving platform for patients with end-stage heart failure who are not candidates for heart transplantation, providing recipients with an average 5 year survival rate of nearly 60%.1 The HeartMate 3 (HM3) LVAD represents the pinnacle of this evolution by offering improved haemocompatibility, attributed to its frictionless fully magnetically levitated rotor system and enhanced blood flow pathways.1 This has led to a reduced risk of thrombotic events and merits exploration of less aggressive anticoagulation regimens than the current mainstay of therapy, warfarin. Apixaban is one of the direct oral anticoagulant (DOAC) therapies that can be considered, though research in this domain remains limited.
Aims
We present a small-scale, long-term retrospective evaluation of thrombotic and bleeding outcomes in HM3 LVAD patients who transitioned from warfarin to apixaban therapy.
Methods
We report our experience with eight patients supported by the HM3 LVAD system who were transitioned to apixaban 5 mg twice daily due to warfarin resistance or intolerance. Our study was conducted with appropriate institutional review board approval. We reviewed patients' characteristics at LVAD implantation, during warfarin therapy and after transitioning to apixaban. The reasons for transitioning to apixaban and clinical outcomes on both therapies were documented. The following outcomes were assessed: LVAD pump thrombosis, stroke events, vascular arterial thrombo-embolic events, pump exchange requirements, major and minor bleeding events, and death. Major bleeding events were defined as any fatal bleeding, symptomatic bleeding in a critical area or organ (e.g., intracranial, retroperitoneal or pericardial) or any drop in haemoglobin level requiring a transfusion of at least 2 units of packed red blood cells (pRBCs). Minor bleeding events were defined as non-major.
Results
We identified eight patients between May 2018 and June 2022 who transitioned from warfarin to apixaban for post-LVAD anticoagulation therapy. Apixaban was administered at a dose of 5 mg twice daily with no dose adjustments. Patients were followed for a mean of 1233 days (range: 586–2092) since LVAD implantation and a mean of 789 days (range: 525–1475) since transitioning to apixaban.
Baseline characteristics and transitioning to apixaban
The average age at the time of LVAD implantation was 58 years (range: 24–71), and the average body mass index (BMI) was 29.48 kg/m2 (range: 21.13–43). Four patients were on anticoagulation prior to the procedure (warfarin n = 2, apixaban n = 1 and rivaroxaban n = 1) for other indications. None were on digoxin, octreotide or thalidomide at any time. Five patients were Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) Profile 3 (62.5%). The procedure was performed via sternotomy (n = 6, 75%) or thoracotomy (n = 2, 25%). Three patients underwent simultaneous left atrial appendage (LAA) ligation; one had patent foramen ovale (PFO) closure, and another had cryoablation for ventricular tachycardia. One patient's post-operative course was complicated by right ventricular failure requiring temporary right ventricular assist device (RVAD) support. Intraoperatively, four patients required an average of 6.25 units of cryoprecipitate (range: 0–10), 1 unit of platelets (range: 0–6), 1.25 unit of fresh frozen plasma (range: 2–4) and 1 unit of pRBCs (range: 0–6). Ventricular assist device (VAD) speed ranged between 5000 and 6000 rpm, with a mode of 5000 rpm, and lactate dehydrogenase levels averaged 294 U/L (range: 138–657). No change in pump power was noted at any given point. Table 1 describes patients' characteristics at discharge, on warfarin before transitioning to apixaban and on apixaban at the last follow-up.
Patients' characteristics (n = 8) | |
---|---|
Baseline demographics and comorbidities (n) | |
Age, mean (range) | 58 (24–71) |
Female sex | 2 (25%) |
BMI (kg/m2), mean (range) | 29.48 (21.13–43) |
Hypertension | 6 (75%) |
Diabetes mellitus | 4 (50%) |
Chronic kidney disease | 2 (25%) |
Type of cardiomyopathy (n) | |
Ischaemic | 1 (12.5%) |
Non-ischaemic | 7 (87.5%) |
INTERMACS profile (n) | |
INTERMACS 1 | 1 (12.5%) |
INTERMACS 2 | 2 (25%) |
INTERMACS 3 | 5 (62.5%) |
At discharge from implantation | On warfarin | On apixaban | |
---|---|---|---|
VAD speed (rpm), mode (range) | 5000 (5000–6000) | 5000 (5000–6000) | 5000 (5000–6000) |
Medications (n) | |||
Warfarin | 7 (87.5%) | 7 (87.5%) | 0 |
Apixaban | 1 (12.5%) | 1 (12.5%) | 8 (100%) |
Aspirin | 5 (62.5%) | 5 (62.5%) | 4 (50%) |
ACEi or ARB | 3 (37.5%) | 2 (25%) | 1 (12.5%) |
ARNi | 1 (12.5%) | 3 (37.5%) | 5 (62.5%) |
Laboratory, mean (range) | |||
Haemoglobin (g/dL) | 9.65 (7.8–12.1) | 11.89 (8.8–12.6) | 12.5 (9.3–16.7) |
Platelets (K/μL) | 284.13 (175–396) | 207.25 (139–252) | 205.25 (154–248) |
eGFR (mL/min/1.73 m2) |
90.75 (39–150) CKD (n = 1) |
69.13 (34–100) CKD (n = 2) |
62.13 (35–99) CKD (n = 3) |
Lactate dehydrogenase (U/L)a | 294 (138–657) | 245.75 (192–304) | 213.75 (158–286) |
Echocardiogram (n) | |||
Right ventricular dysfunction | 5 (62.5%) | 8 (100%) | 8 (100%) |
Moderate tricuspid regurgitation | 1 (12.5%) | 2 (25%) | 2 (25%) |
Mild aortic regurgitation | 1 (12.5%) | 3 (37.5%) | 4 (50%) |
- Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ARNi, angiotensin receptor/neprilysin inhibitor; BMI, body mass index; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; LDH, lactate dehydrogenase; VAD, ventricular assist device.
- a The reported LDH values represent the LDH range during the time the patient was on the relative strategy.
All patients were started on warfarin therapy after LVAD implantation and then transitioned to apixaban due to difficulty maintaining a therapeutic international normalized ratio (INR), including five who experienced bleeding complications on warfarin. INR was reported as labile in four (50%) patients, sub-therapeutic in three (37.5%) and supra-therapeutic in one (12.5%). Additionally, aspirin 81 mg daily was prescribed to all but three patients due to various reasons: one had von Willebrand disease and a large pericardial haematoma; another had gastrointestinal bleeding due to angioectasia; and the third had a pericardial haematoma causing tamponade, requiring evacuation and multiple blood transfusions. These events took place during implant hospitalization. The average duration of support from implantation to apixaban transition was 444 days (range: 61–1477).
Clinical outcomes
Clinical outcomes on both anticoagulation therapies are shown in Table 2 and detailed individually in Table 3.
Clinical outcomes (n) | On warfarin | On apixaban |
---|---|---|
Major bleeding | 1 | 1 |
Gastrointestinal bleed | 0 | 1 |
Haemoglobin drop requiring at least 2 pRBCs | 1 | 0 |
Minor bleeding | 7 | 0 |
Gastrointestinal bleed | 4 | 0 |
Haematuria | 1 | 0 |
Haemoptysis | 1 | 0 |
Haemoglobin drop requiring 1 pRBC | 1 | 0 |
- Abbreviations: n, number of events; pRBCs, packed red blood cells.
Patient | AC on discharge | AP on discharge | Warfarin complications | Medication adherence | INR on warfarin | Days from implant until transition | Apixaban complications | Days on apixaban | Last LDH on warfarin | Last LDH on apixaban |
---|---|---|---|---|---|---|---|---|---|---|
1 | Warfarin | Aspirin | - Recurrent LGIBs from haemorrhoids | Adherent | Supra-therapeutic | 90 | - UGIB from an oozing gastric ulcer requiring 2 pRBCs | 644 | 304 | 286 |
2 | Warfarin | No aspirin due to VWD | None | Adherent | Sub-therapeutic | 637 | None | 1063 | 192 | 203 |
3 | Warfarin | No aspirin due to bleeding |
- Haemoglobin drop requiring 2 pRBCs - Two LGIBs requiring 1 pRBC each time |
Adherent | Labile | 258 | None | 640 | 233 | 158 |
4 | Apixabana | No aspirin due to bleeding | - Haematuria | Adherent | Labile | 61 | None | 525 | 201 | 182 |
5 | Warfarin | Aspirin |
- Multiple LGIBs - UGIB |
Non-adherent | Labile | 1477 | None | 615 | 282 | 230 |
6 | Warfarin | Aspirin | - Haemoptysis | Adherent | Sub-therapeutic | 170 | None | 1475 | 282 | 200 |
7 | Warfarin | Aspirin | None | Adherent | Labile | 615 | None | 533 | 220 | 211 |
8 | Warfarin | Aspirin | None | Adherent | Sub-therapeutic | 245 | None | 817 | 252 | 240 |
- Abbreviations: AC, anticoagulant; AP, antiplatelet; INR, international normalized ratio; LDH, lactate dehydrogenase; LGIBs, lower gastrointestinal bleeds; pRBCs, packed red blood cells; UGIB, upper gastrointestinal bleed; VWD, von Willebrand disease.
- a Patient transitioned to apixaban during index hospitalization.
No patients encountered LVAD pump thrombosis, stroke events, arterial thrombo-embolic events, pump exchange or death. All eight patients remained alive on LVAD support at the time of writing. While on warfarin therapy, one patient experienced a major bleeding event that required 2 U of pRBCs. There were seven minor bleeding events, including five gastrointestinal bleeds, one instance of haematuria and one episode of haemoptysis. No thrombo-embolic events or deaths were reported. After transitioning to apixaban, one patient had a major gastrointestinal bleed due to angioectasia causing an oozing gastric ulcer, requiring 2 U of pRBCs and endoscopic clipping.
Discussion
We report our experience using apixaban as the main anticoagulant in a small cohort of patients supported with HM3 LVAD who were intolerant or resistant to warfarin. The enhanced haemocompatibility and significantly reduced risk of thrombotic events associated with the HM3 LVAD system made it feasible to consider an alternative anticoagulation regimen. The decision to use apixaban over other DOACs was based on the agent's lower gastrointestinal and intracranial bleeding risk.2 This strategy has proved beneficial for our patients, as no thrombotic events have been encountered to date while on apixaban. Several studies have reported the potential benefits of using apixaban in patients supported with LVAD systems.
Whitehouse et al. compared 20 patients on HM3 LVAD support on warfarin to 15 patients on apixaban as initial or transitional therapy. Patients on apixaban experienced one minor bleed, one deep vein thrombosis (DVT) and no stroke events. In contrast, patients on warfarin had three minor bleeding events, two DVTs, one inferior vena cava thrombosis, three haemorrhagic strokes and one ischaemic stroke. No significant differences were found between the two therapies regarding thrombotic and bleeding complications, stroke or death at 6 months.3 In our cohort, patients were exposed to both anticoagulation strategies, serving as their own controls and experiencing fewer bleeding events on apixaban.
Kos et al. compared LVAD patients who were switched to a DOAC (n = 8) after a trial of warfarin to those who remained on warfarin (n = 260). They found no significant difference in mortality, gastrointestinal or intracranial bleeding, ischaemic stroke or pump thrombosis between the two groups.4 They did not specify the type of LVAD, DOAC used or reasons for transitioning to a DOAC. Similarly, Soares et al. reported their experience with 12 patients on LVAD support treated with apixaban compared with 60 patients who remained on warfarin. In the apixaban group, 10 out of 12 patients were previously on warfarin. Among these 10 patients, the incidence rates for bleeding-related hospitalizations were 4.2 per 1000 days while on warfarin, compared with 0.76 per 1000 days after transitioning to apixaban [relative risk (RR) 81.9%]. Additionally, there were no statistically significant differences in bleeding or thrombotic complications between the two groups nor in mortality rates.5 They did not report the type of LVAD system used either.
In the recent DOAC LVAD single-centre trial, patients were randomized to apixaban 5 mg twice daily (n = 16) or warfarin (n = 14), both with aspirin 81 mg daily. At 24 weeks of enrolment, there were no deaths or thrombo-embolic complications in either group. However, while no apixaban patients experienced major gastrointestinal bleeding, two patients in the warfarin group did, accounting for three bleeding events (one recurrent) (P = 0.12). These events occurred in one patient enrolled within 3 months of LVAD implantation and one patient after 3 months.6 Comparable results were reported by the DOT-HM3 trial, where no thrombo-embolic events occurred in either therapy group over 6 months. Warfarin patients experienced two uterine bleeds and one episode of epistaxis, whereas one patient on apixaban and aspirin 100 mg experienced a gastrointestinal bleed.7
Similar to recent reports, we found a lower incidence of bleeding on apixaban compared with warfarin. One patient with angioectasia, who had recurrent episodes of haematochezia on warfarin and a labile INR, experienced gastric bleeding on apixaban therapy. He required 2 units of pRBCs and esophagogastroduodenoscopy with clipping. Since then, he has remained event-free on apixaban therapy for over 9 months.
It is important to note that a few reports have linked apixaban therapy to LVAD pump thrombosis.8, 9 However, these cases were mostly associated with the HeartWare LVAD system, which was removed from the market due to increased risk of pump thrombosis and neurological adverse events. A case series by Kobayashi et al. described one thrombotic event in five paediatric and young adult patients on HM3 LVAD support treated with apixaban and aspirin. This non-occlusive outflow graft thrombus was found 2 years after apixaban initiation in a non-compliant patient with chronic VAD infection. Otherwise, no major bleeding, death or stroke events were encountered over a median of 170 days on apixaban therapy (range: 90–1120).10 This highlights the challenges of managing non-adherent LVAD patients, whether on warfarin or apixaban.
Our experience, although small, contributes to the literature by providing the longest follow-up outcomes of apixaban therapy in LVAD patients. Our patients have been followed for an average of 789 days after transition to apixaban, serving as their own controls on both anticoagulation therapies.
Larger scale long-term studies are needed to evaluate apixaban as the anticoagulation of choice in LVAD patients. An important related question is whether aspirin is necessary in patients treated with apixaban. The ARIES-HM3 trial demonstrated that an aspirin-free regimen is non-inferior to the combination of warfarin and aspirin. This regimen was associated with a lower incidence of bleeding events and did not result in a significant increase in thrombo-embolic risk.11
Conclusions
This brief single-centre experience supports the safety and efficacy of using apixaban as an alternative to warfarin in patients supported by HM3 LVAD systems. This approach is particularly beneficial and convenient for patients with difficulty maintaining a therapeutic INR and/or recurrent episodes of bleeding. Larger prospective studies are needed before this can be adopted as the standard of care.
Conflict of interest statement
None declared.