Keywords
left ventricular assist device - hemocompatibility-related adverse event - antithrombotic
treatment - antiplatelet therapy
Introduction
Heart failure (HF) has an estimated worldwide prevalence of 56.2 million individuals,[1] and it is still increasing. In 2019, The Heart Failure Association ATLAS reported
data for 13 European countries, with a median prevalence of HF of 17.20 (interquartile
range [IQR]: 14.30–21) per 1,000 people and a median annual incidence of 3.20 (IQR:
2.66–21) per 1,000 person-years, ranging from < 2.0 in Italy to ≧6.0 in Germany.[2] The 5-year survival rate is less than 50% after diagnosis.[3]
[4] Heart transplantation (HT) is considered the optimal therapy for advanced HF, but
donor shortage and strict selection criteria limit its use.[1]
[5]
In this context, left ventricular assist devices (LVADs) have emerged as a therapeutic
alternative. The intended goals of therapy include short-term assistance, either “bridge
to transplant” (BTT), the most frequent indication in Europe (EU) (66%),[6] or “bridge to candidacy,” and long-term assistance, also known as “destination therapy”
(DT) for patients who are not eligible for HT, the latter accounting for 73% of the
indications in the United States from 2018 to 2022.[5]
The use of continuous-flow LVAD (CF LVAD) in appropriate candidates has improved survival,
functional capacity, and quality of life for patients with end-stage HF.[6] HeartMate 3 (HM 3; Abbott, Chicago, Illinois, United States) is nowadays the only
available device for clinical use approved by the regulatory agencies in the European
Union (EU) and the United States.[7] According to the observational extended follow-up study of the MOMENTUM 3 trial,[8] it has a 5-year survival of 58.4% in the United States and 54 to 63.3% in European-based
cohorts.[6]
[9] Nevertheless, hemocompatibility-related adverse events (HRAEs; pump thrombosis,
hemorrhagic or ischemic stroke, and gastrointestinal [GI] bleeding)[10]
[11] continue to contribute to mortality and morbidity in assisted patients.[12] Therefore, optimal antithrombotic treatment is one of the important goals to be
reached in the future.
LVAD Generations
Mechanical circulatory support (MCS)-implanted devices include LVADs, VADs used as
right ventricular support (RVAD), biventricular support (BiVAD), and total artificial
hearts (TAHs).[13] This review will focus on LVADs, as they are the most commonly implanted devices.
The goal of MCS research, starting in 1964 with the formation of the Artificial Heart
Program,[14] was initially to develop a mechanical pump able to assist the left ventricle (LV)
by pumping blood from the LV to the aorta, without major thrombosis, and enable patients
to survive until a donor heart became available.[15]
The first LVADs were large pulsatile, pneumatically driven devices designed to mimic
rhythmic cardiac activity and had limited durability.[15] Due to the limitations in miniaturization, the focus of research shifted toward
developing electrically powered devices.[16] In 1999, the REMATCH trial[17] compared optimal medical therapy with a vented electric LVAD in advanced HF patients
ineligible for HT. The improvement of survival at 1 year in the LVAD group (52 vs.
25% in the medical therapy group, p = 0.002) led to the approval of the HeartMate vented electric LVAD (HeartMate VE)
for DT[15] after its former approval as BTT. The HeartMate XVE was then successfully designed
to improve pump reliability and durability.[18] Despite improvements in survival with the HeartMate XVE (61% at 1 year),[19] adverse events (infection, neurological dysfunction, or pump failure) still affected
about half of the patients.[15]
The second generation of LVADs was designed as continuous-flow pumps, categorized
into axial-flow devices (HeartMate II, Thoratec Corporation, San Diego, California,
and Jarvik 2000, Jarvik Heart, Inc., New York, New York, United States) and centrifugal-flow
devices (HeartWare, HeartWare International, Framingham, Massachusetts, United States).
Continuous-flow pumps were much smaller (suitable for smaller patients including women),
and they had a single internal rotor less prone to dysfunction.[15] They showed superior durability and better neurological outcomes compared with pulsatile
pumps.[20] This generation contributed to improved survival of LVAD-implanted patients, with
1- and 2-year survival of 73.1 and 69.1%, respectively, and a median patient survival
of 46.5 months (95% confidence interval [CI]: 44.7–48.2 months).[21] These devices were the most frequently implanted between 2010 and 2014 in the United
States[21] before their use declined (1.8% of implanted devices in the United States in 2019).
Axial-flow pumps improved patient comfort with less pump failure and improved clinical
outcomes, including lower risk of thromboembolic events.[22] Initially, centrifugal-flow devices were found to be noninferior to axial-flow pumps
with respect to the incidence of disabling stroke or the need for device replacement,
but they were associated with a higher risk of ischemic and hemorrhagic stroke (29.7
vs. 12.1%, p < 0.001).[23] Due to safety issues (pump malfunction), increased risk of mortality, and neurological
adverse events, Medtronic stopped the sale of the HeartWare in 2021 after the implantation
of several thousand patients worldwide.[24]
The HM 3 represents the third generation of LVADs. It received the CE mark in Europe
for short- and long-term support in 2015, followed by FDA approval in 2017. The HM
3 utilizes centrifugal flow technology but with a fully magnetically levitated rotor
(MagLev LVAD), eliminating mechanical contact between internal components and blood,
thereby reducing the risk of mechanical failure and shear stress.[25] The HM 3 incorporates intrinsic pulsatility into the continuous flow, simulating
periodic pulses every two seconds by briefly slowing down and speeding up the rotor
speed. A shift to nearly exclusive implantation of MagLev LVAD occurred, representing
77.7% of LVAD implants in 2019[21] and 99.8% in 2022.[5]
In the MOMENTUM 3 randomized controlled trial,[26] Mehra et al compared the mechanical-bearing axial continuous-flow pump, Heartmate
II (HM II) (n = 512), with the MagLev centrifugal continuous-flow pump, HM 3 (n = 516). Results showed that the HM 3 was superior with regard to 2-year survival
free of disabling stroke or need for reoperation to replace or remove a malfunctioning
device (76.9 vs. 68.8%, relative risk, 0.84; 95% CI: 0.78–0.91; p < 0.001).[26] Also, the number of events per patient-year for stroke, major bleeding and GI bleeding
were lower in the MagLev pump group.[26]
Survival, again, improved, as showed in the 2023 annual report from the Society of
Thoracic Surgeons (STS) Interagency Registry for Mechanically Assisted Circulatory
Support (Intermacs). During the period 2018 to 2022, MagLev LVAD recipients exhibited
significantly higher 1- and 5-year survival of 86 and 64%, respectively, than those
receiving non-MagLev devices.[5]
Despite these encouraging results, the patient's quality of life remains affected
by adverse events and mortality.[8] Rehospitalization occurs in more than half of the LVAD patients.[5]
The next part of this review will focus only on MagLev LVAD, namely, HM 3. It will
cover advancements in hemocompatibility, HRAEs, and current strategies for antithrombotic
and antiplatelet treatments.
Hemocompatibility
Hemocompatibility is a key factor in the interaction between a foreign material or
device and the patient's blood. To achieve optimal hemocompatibility, the design goals
of the HM 3 were to minimize the degree of shear force acting on blood components,
to reduce the biomaterial–blood interface area, and to enhance continuous blood flow
with an artificial pulse.[25] Despite these advances, the nonphysiological blood flow dynamic (continuous laminar
flow with minimal pulsatility) contributes to HRAEs. The reduced pulsatility is associated
with numerous alterations such as endothelial dysfunction, the release of proinflammatory
factors, the rise of reactive oxygen species, a decreased nitric oxide bioavailability,
platelet activation, and changes in the vascular bed (increased permeability, vascular
smooth muscle proliferation, and dysregulated tone).[27]
[28] The high shear stress associated with continuous flow can also lead to von Willebrand
factor (vWF) abnormalities. These abnormalities have been described in some studies
as a decrease of large vWF multimers (acquired von Willebrand disease [vWD] type 2A)
due to mechanical destruction,[29] cleavage of large vWF multimers by ADAMTS-13 (as indicated by a drop of ADAMTS-13
level), and platelet-dependent mechanisms.[30] These factors contribute to the increased hemorrhagic risk associated with the use
of LVAD.
HRAEs during LVAD Assistance
HRAEs during LVAD Assistance
Thrombosis and bleeding are among the most frequent deleterious complications occurring
in patients with LVAD. Major bleeding is the second most common adverse event (after
infection), affecting 17% of patients with MagLev devices in the first 90 days postimplantation
and another 17% in the period beyond 90 days.[5] The most frequent bleeding complications include GI bleeding and hemorrhagic stroke.
Thrombotic complications predominantly include ischemic stroke and pump thrombosis.
HRAEs were reported as the third cause of death in the 5-year follow-up of the MOMENTUM
3 trial patients, with an attributed mortality rate of 3.9%,[8] and are also responsible for rehospitalization and morbidity.[5]
Hemorrhagic and Ischemic Stroke
Hemorrhagic and Ischemic Stroke
Ischemic and hemorrhagic strokes remain significant causes of death in LVAD patients.[27] Even though continuous-flow devices have drastically improved patients' outcomes
over the last two decades, 9.6% of MagLev patient's deaths are still attributed to
neurological dysfunction.[5] The MOMENTUM 3 trial showed a significantly improved stroke-free survival at 2 years
for HM 3 compared with HM II, with stroke rates of 9.9 versus 19.4%, respectively
(hazard ratio, 0.42; 95% CI, 0.30–0.57, p < 0.001).[26] A secondary analysis of the MOMENTUM 3 trial confirmed the better neurologic outcome
at 5 years with an occurrence of any stroke of 0.050 events/patient-year in the HM
3 group compared with 0.136 events/patient-year in the HM II group (rate ratio: 0.37,
95% CI: 0.27–0.50, p < 0.01).[8] In the 2023 Intermacs report, stroke occurred in 5% in the early period (≦90 days
following implantation) and in another 4% in the late period.[5]
Under continuous flow, the arterial baroreceptors are unloaded, leading to an increase
in neurohumoral and sympathetic activation, with consequently a chronic elevation
of mean arterial pressure, a rightward shift of the cerebral autoregulation curve,
and a reduction of the dilator capacity of the cerebral vascular bed.[27] However, the altered cerebral autoregulation in the context of continuous-flow devices
is not well established, and there is conflicting evidence showing a preserved autoregulation
during implantation and in the early postoperative period[31] as well as later.[32]
The multiple detrimental effects of reduced pulsatility and vWF degradation are predisposing
factors for stroke.[27] Intravascular hemolysis can also lead to the activation of platelets (ADP release
by damaged red blood cells) and the hemostatic system and may enhance clot stability.[33] Finally, infectious complications increase the risk of stroke through inflammation
and septic emboli.[27]
Strategies to minimize both hemorrhagic and ischemic stroke risks are adequate antiplatelet
therapy, close monitoring of anticoagulation, and strict blood pressure control (mean
arterial pressure: 75–90 mm Hg).[34]
Pump Thrombosis
Pump thrombosis is a severe HRAE, and it is associated with significant morbidity
and mortality.[35] The thrombotic risk is partly linked to the type of LVAD pump and its design. The
rate of pump thrombosis with HM II began to increase in 2011. A pooled analysis from
three experienced LVAD centers confirmed this trend, reporting an 8.4% (95% CI, 5.0–13.9)
incidence of confirmed pump thrombosis within 3 months.[35] Contributing factors to this rise were thought to be the adoption of lower anticoagulation
strategies (to prevent GI bleeding which was also a concern), particular pump-related
settings (lower speed), and lack of adherence to recommended implantation techniques
(i.e., pump below the diaphragm, inflow cannula parallel to the septum, outflow graft
position right of the sternal midline, and pump fixation).[36] In the PREVENT trial (prospective, multicenter, single-arm), strict adherence to
antiplatelet therapy combined with anticoagulation, and to other components of a structured
surgical implant technique and postoperative hemodynamic management, was associated
with a significant reduction in pump thrombosis risk at 6 months (1.9 vs. 8.9%; p < 0.01).[36]
Innovations in pump design allowed for a decrease in the rate of pump thrombosis.[37] As described in a computational fluid dynamics model, HM 3 compared with HM II minimizes
the shear force on blood components due to the wide blood-flow gaps, has a lower hemolysis
index, and alternates the speed every 2 seconds, changing blood flow to eliminate
the presence of recirculation and stasis zones.[25] The design of HM 3 is also believed to allow the preservation of vWF by reducing
shear stress and platelet activation.[30] In the MOMENTUM 3 trial, suspected or confirmed pump thrombosis at 2 years occurred
in 1.4% of patients (n = 7), significantly lower than for axial pump (13.9%, n = 70).[26] Similarly, in the 2023 Intermacs report, 5-year freedom from device malfunction/pump
thrombosis was significantly higher for HM 3 compared with non-MagLev devices (83
vs. 54%, p < 0.0001).[5]
There are several other thrombotic risk factors, non-modifiable and modifiable. Non-modifiable
factors include age at implant, female gender, higher body mass index, non-O blood
type, some psychosocial issues (e.g., limited support, repeated non-compliance), the
presence of a prothrombotic state, atrial fibrillation, dysfunction of the right ventricle,
pulmonary disease, and finally history of GI bleeding.[38] Modifiable factors include tobacco use, bacteremia, pump infection, and hypertension.[38] Shear-mediated platelet activation and cytokine-mediated endothelial cell inflammatory
activation may contribute to thrombosis by enhancing the adhesion of platelets to
the inflamed endothelium and platelet prothrombotic function.[39]
Pump thrombosis can be highlighted by biological (elevated lactate dehydrogenase,
plasma-free hemoglobin rise, and hemoglobinuria) or technical anomalies (increased
pump power), or by clinical symptoms suggestive of new ongoing HF.[36] Interestingly, the detection of loss of circadian fluctuations of the pump power
using time-frequency analysis of the LVAD logfiles[40] or accelerometer signal changes in the third-harmonic and non-harmonic amplitude[41] have been identified as new tools allowing early detection of pump thrombosis, prior
to clinical manifestation or symptoms. Initial treatment, in case of hemodynamic stability,
should consist of systemic intravenous anticoagulation with unfractionated heparin
(Class I, level of evidence C).[34]
[36] For patients who are candidates for surgery, pump replacement is the definitive
therapy (Class I, level of evidence C), with urgent transplantation as an alternative
if the expected wait time for HT is short (Class IIa, level of evidence C).[34] In carefully selected patients, systemic or intraventricular thrombolytic therapy
can be considered as an initial strategy over pump replacement (Class IIa, level of
evidence C).[34] The safety and efficacy of glycoprotein IIb/IIIa inhibitors have not been established
in this setting.[34]
Gastrointestinal Bleeding
Gastrointestinal Bleeding
Approximately 50% of cases of bleeding have a GI etiology.[21] In the MOMENTUM 3 trial, 24.5% of patients suffered from GI bleeding, a significantly
lower rate than for HM II patients (30.9%).[26]
The etiology of GI bleeding is mostly attributed to the presence of vascular architecture
abnormalities (angiodysplasias) in the small bowel and vWF degradation.[42] The pathophysiology related to GI bleeding in the context of CF LVAD consists of
extrinsic (pharmacological) and intrinsic factors (related to LVAD).[43] Anticoagulation and antiplatelet therapy (extrinsic factors) increase the bleeding
risk, but they are not the underlying cause of GI bleeding.[43] Intrinsic factors are secondary to the non-physiological, continuous blood flow
and the foreign interface of the LVAD.[43] The presence of angiodysplasias is the result of a constant hypoperfusion, due to
lower intraluminal pressure combined with an increased sympathetic tone, and hypoxia
of the GI tract mucosa, stimulating the release of angiogenic factors.[43]
[44]
Acquired vWD has been attributed to the degradation of the large vWF multimers by
ADAMTS-13[30]
[43]
[45]; however, current evidence does not completely support this mechanism. High shear
stress seems the primary mechanism for inducing a conformational change in vWF that
becomes hyperadhesive and activated. This results in the binding and activation of
platelets, and also the promotion of angiogenesis.[46] Acquired vWD contributes to the high prevalence of bleeding during long-term support
and at the time of transplantation.[47] Finally, acquired platelet dysfunction (antiplatelet agents and high shear stress
exposure) and modulation of platelet microRNAs induced by LVAD are other contributing
factors.[47]
[48]
In the 2023 Intermacs report, GI bleeding occurred in 8% of patients in the early
period following implantation (≦90 days), and in 11% in the late period.[5] Over 5 years, patients implanted with a MagLev device showed a higher freedom from
GI bleeding compared with the non–MagLev-implanted patients (72 vs. 60%, p < 0.0001).[5]
Future strategies to reduce the rate of GI bleeding will rely not only on technological
advancements in LVADs but also on the pharmacological properties of specific drugs
and innovative therapeutic approaches.
The EVAHEART (EVAHEART Inc, Houston, Texas, United States) is an LVAD, approved in
2010 by the Japanese Pharmaceuticals and Medical Devices Agency, designed to reduce
vWF degradation by increasing pulsatility and reducing shear stress.[27] This device caused significantly less vWF degradation than the HM II in a mock circulatory
loop with whole human blood.[49] A prospective observational study evaluated its effectiveness between 2011 and 2013.[50] Interestingly, over the 96 patients implanted with EVAHEART, no GI bleeding occurred.
This new evolution in pump design seems promising to reduce GI bleeding, and the ongoing
COMPETENCE trial will provide further insights by comparing EVAHEART 2 with HM 3.[51]
Pharmacologically available interesting agents include digoxin, octreotide, thalidomide,
angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and hormone
therapy. A detailed description of their potential to reduce GI bleeding is beyond
the scope of this review, but interestingly, the use of the readily available and
cheap drug, digoxin, was associated with a reduction of GI bleeding in retrospective
trials testing its impact on clinical outcomes in LVAD patients.[52]
[53]
[54] One potential explanation is the suppression of neo-angiogenesis through inhibition
of the hypoxia-inducible factor-1α, which is a mediator of angiopoietin-2-induced
angiodysplasias.[53]
Finally, a pilot study assessed the safety of infusing umbilical cord lining stem
cells to improve vascular stability by addressing angiogenesis dysregulation, which
is believed to contribute to bleeding in LVAD patients.[55] Despite including only nine patients, the study indicated a trend toward reduced
GI bleeding.[55]
[56] This study highlights cell therapy as a promising research direction to improve
hemocompatibility.
Antithrombotic Treatment Following Implantation
Antithrombotic Treatment Following Implantation
As stated, bleeding and thrombosis are still of the two most common complications
in patients assisted with LVAD. The primary concern in the early postoperative period
is bleeding; hence, “hemostasis management is a key priority.”[34] Coagulopathy should be corrected early in the postoperative phase (Class IIa, level
of evidence C), and targeted hemostatic intervention algorithms should be applied
to achieve appropriate haemostasis.[34] Careful introduction of anticoagulant and antiplatelet therapy is necessary to achieve
the desired therapeutic effect without increasing the risk of bleeding. Evidence regarding
the introduction of anticoagulants is low and recommendations are based mostly on
expert opinion. According to the 2023 International Society of Heart and Lung Transplantation
guidelines, a heparin bridge (with unfractionated heparin) should be started within
the first 24 hours after surgery if hemostasis is adequate and chest tube output is
less than 50 mL/hour for several hours.[34] A targeted activated partial thromboplastin time (aPTT) between 40 and 60 seconds
is recommended; however, as the normal range of aPTT is dependent on the reactive
used by the laboratory, this range could vary depending on institutional and local
protocols. Anti-Xa activity monitoring may be useful in specific situations where
a prolonged aPTT is not associated with increased bleeding risk (preanalytical issue,
factor XII deficiency, pre-kallikrein deficiency, interaction with C-reactive protein).[57] Vitamin K antagonist (VKA) therapy should be initiated when chest tubes are removed
or by postoperative day 2, according to local hospital protocols.[34]
Long-Term Antithrombotic Treatment
Long-Term Antithrombotic Treatment
As for the immediate postoperative period, evidence concerning long-term antithrombotic
treatment following LVAD implantation is lacking, and most recommendations are based
on expert consensus.
Long-term anticoagulation is recommended with VKA to maintain the international normalized
ratio (INR) within a specific range specified by the manufacturer for each device
(Class I, level of evidence B).[34]
[38] For HM 3 devices, the recommended target of INR is 2.0 to 3.0.[26] As for other indications for anticoagulation, maximizing the time interval during
which the patient is within the INR target is challenging. It has been demonstrated
in patients with atrial fibrillation that the mean time in therapeutic range (TTR)
is ∼65% ± 20%.[58] In a 2017 meta-analysis on LVAD patients receiving warfarin, Martinez et al analyzed
a total of 270 patients with follow-up ranging from 9 to 76 person-years. The weighted
mean TTR was only 46.6% (95% CI: 36–57.3%, I
2 = 94%), illustrating the difficulty of managing VKA anticoagulation in LVAD patients.[59] A retrospective study on HeartWare devices showed that a low TTR was associated
with significantly lower 2-year survival (61.7%) compared with moderate and high TTR
(72.4 and 75.1%, respectively; p = 0.001), and with higher rates of HRAEs.[60] Still, VKA therapy remains the recommended long-term treatment following LVAD implantation.[34]
[38]
As mentioned in the consensus document developed in accordance with the International
Society of Heart and Lung Transplantation, the INR target may need to be adjusted
in response to bleeding or thrombotic events occurring during LVAD support.[38]
In the pilot study MAGENTUM I, Netuka et al tested low-intensity warfarin anticoagulation
(INR: 1.5–1.9) combined with aspirin in patients with HM 3 device, following an initial
6-week post-implantation period of standard anticoagulation (INR: 2.0–3.0) combined
with aspirin.[61] No thrombotic events occurred during the 6-month follow-up, suggesting that low-intensity
anticoagulation is achievable and may be safe in the context of HM 3 during the first
6 months of postimplantation. Further large-scale trials are necessary to confirm
this finding. Interestingly, the TTR was 75.3 ± 8.6%, during the low-intensity phase,
which was much higher than the TTR reported by Martinez et al.[59]
[61] The strict follow-up performed in clinical trials may have contributed to the improved
TTR and outcome in this pilot study.
Direct oral anticoagulants (DOACs) have several advantages over VKA, including rapid
onset and offset of action, fixed dosing, less drug or dietary interactions, and no
need for repetitive coagulation monitoring.[62] DOACs have a direct and targeted anticoagulant effect by inhibiting free and complex-bound
clotting factor X (-xabans) or II (dabigatran). One of the first DOACs to be tested
in a prospective randomized, open-label, single-center study in the context of LVADs
was dabigatran.[63] Patients assisted with HeartWare devices were randomized to receive either VKA or
dabigatran in addition to aspirin. The study had to be interrupted for safety reasons
because of a higher rate of thromboembolic events in the dabigatran arm. DOACs became
contraindicated for LVADs and for mechanical heart valves following another phase
II study that was prematurely interrupted for the same issue.[64]
Even so, researchers have not given up on the potential of DOACs in this field, and
have concentrated their efforts on another, possibly more promising molecule, the
anti-Xa agent apixaban. In the ARISTOTLE trial, apixaban was compared with warfarin
in patients with atrial fibrillation. Apixaban was superior in preventing stroke and
systemic embolism and caused less bleeding (massive bleeding, hemorrhagic stroke).[65] In a retrospective study, apixaban was also associated with lower rates of GI bleeding
compared with warfarin,[66] which may be particularly relevant to LVAD-associated bleeding.
Apixaban's anticoagulant effect was assessed in an ex-vivo mock loop model with a
HeartWare device. It performed similarly to warfarin and better than dabigatran.[67] VKA and apixaban target coagulation before the initiation of the common pathway,
as opposed to dabigatran which acts more distally in the pathway of the clotting cascade.
Remembering that each molecule of factor Xa generates ∼1,000 molecules of thrombin,
a more proximal inhibition on the pathway might be more effective.[68] One interesting proposition from Aimo and colleagues, to improve -xaban anticoagulation
in the field of LVAD, would be to measure factor Xa activity in animals receiving
warfarin (with an INR target of 2–3), and to search for the anti-factor Xa dose that
achieves similar trough-and-peak factor Xa levels.[68]
In a phase 2, open-label trial, Shah et al compared apixaban to warfarin in patients
assisted with a MagLev LVAD, and assessed the safety and freedom from a composite
primary outcome of death or major HRAEs (stroke, device thrombosis, major bleeding,
aortic root thrombus, and arterial non-central nervous system thromboembolism).[12] LVAD recipients were randomized in a 1:1 ratio to receive low-dose aspirin with
either apixaban (5 mg, twice daily) or warfarin. A total of 30 patients were included
(14 warfarin and 16 apixaban). The primary outcome did not occur in any patient in
the apixaban group at 24 weeks and did occur in two patients (14%) in the warfarin
group. This study showed that alternative anticoagulation with apixaban is feasible
without an excess of HRAEs or mortality and supports the development of an appropriately
powered clinical trial to assess the efficacity and safety of apixaban for patients
with LVADs.
The DOT-3 randomized trial also demonstrated the safety of using apixaban (with or
without aspirin) in a small number of HM 3 patients for up to 6 months, with no thromboembolic
events reported in the DOAC groups. Moreover, patients on apixaban had successful
HT.[69] However, the associated hemorrhagic risk on DOAC compared with VKA in the case of
transplantation should be assessed on a prospective and broad basis before considering
its inclusion in the management of BTT patients.
The ongoing ApixiVAD study (registered in the Australian New Zealand Clinical Trials
Registry ACTRN12621000956808) is a multicenter, international, open-label, randomized,
controlled, non-inferiority pilot trial. It aims to include 50 BTT or DT HM 3 patients.
This trial will assess the safety of apixaban 2.5 mg twice daily compared with the
standard of care with VKA.[70] Mortality, thromboembolic events, major bleeding (including operative bleeding),
immediate transplant outcomes, and patients' quality of life related to anticoagulation
will be assessed and the results should provide further information on the safety
and feasibility of apixaban anticoagulation.
Antiplatelet Therapy
Antiplatelet therapy is recommended with VKA following LVAD implantation, even if
solid evidence of efficacy and safety is lacking.[71] The 2023 International Society of Heart and Lung Transplantation guidelines state
that antiplatelet therapy should be initiated in the postoperative period in the intensive
care unit (Class I, level of evidence C) and that chronic antiplatelet therapy with
aspirin may be used in addition to VKA (Class I, level of evidence C).[34] Aspirin is the most common antiplatelet agent used in this context.[34] Therapy is started between postoperative days 1 and 3 with the dosage ranging from
81 to 325 mg, depending on the local hospital practice.
Platelet function assays may be used to direct the dosing and number of antiplatelet
drugs (Class IIb, level of evidence C).[34] A retrospective study in the context of HM 3 investigated the usefulness of aspirin
titration based on the antiplatelet effect monitored by the Verify Now Aspirin test
(Accumetrics Inc., San Diego, California, United States). The study found that increasing
aspirin doses in non-responders to achieve responsiveness demonstrated a similar rate
of pump thrombosis and freedom from thrombotic complications compared with the patients
who were initially responsive.[72] It may be reasonable to consider escalation of antiplatelet therapy in patients
who have thrombotic events with documented compliance to VKA and aspirin (Class IIb,
level of evidence C).[34] Dual-antiplatelet therapy with aspirin and a P2Y12 inhibitor (clopidogrel, prasugrel,
or ticagrelor) is not routinely indicated for LVAD patients unless there is a markedly
increased thrombotic risk, a prior history of pump thrombosis or very recent coronary
revascularization.[38]
Regarding the dosage of aspirin, there is no evidence of benefit on the rate of HRAEs
in HM 3 patients receiving a high dose of aspirin compared with a low dose of aspirin;
thus, it is recommended to use the lower doses.[38] In an exploratory analysis of the patients implanted with HM 3 from the MOMENTUM
3 trial, two groups were compared based on the aspirin dose: a low-dose group (81 mg,
n = 180) and a usual-dose group (325 mg, n = 141).[73] The primary endpoint was survival free from HRAEs, including non-surgical bleeding,
pump thrombosis, stroke, and peripheral arterial thromboembolic events. At 2 years,
the proportion of patients who survived without HRAEs was similar between the low-dose
and usual-dose groups (45.3 vs. 43.4%, p = 0.94) and it was also the case for freedom from hemorrhagic events (51.7 vs. 54.4%,
p = 0.42).[73] This suggests either that the full aspirin effect is already achieved with an 81-mg
dose or that aspirin is not required to prevent thrombosis in patients implanted with
HM 3.[74]
The usefulness of aspirin has been questioned. First, the impact of aspirin with VKA
compared with VKA alone on thrombin generation seems of low intensity in the context
of LVAD.[75] Then, there is growing evidence of the increased bleeding risk associated with aspirin
treatment in LVAD patients, without a positive effect on thromboembolic events. In
a prospective study, including 53 LVAD patients with a median duration of support
of 324 days (IQR: 226–468), 25 bleeding events were recorded (47% of the patients).[76] Coagulation tests showed that the INR was in the targeted interval, with a median
of 2.51 (IQR: 1.98–2.97), and that there was a significant decrease of vWF:Ag and
vWF:CB after the implant. Aspirin contributed significantly to bleeding events in
the background of acquired vWD and its withdrawal significantly reduced bleeding recurrence.[76] Other studies supported the contribution of aspirin to bleeding in HM 3 patients
and the safety and efficacy of an aspirin-free antiplatelet regimen in reducing HRAEs.[77]
[78]
Recently, an international, double-blind, randomized controlled trial, ARIES-HM3,
investigated this issue.[71] HM 3-implanted patients were randomized in a 1:1 ratio to receive aspirin 100 mg
combined with VKA or placebo combined with VKA. The primary composite outcome to assess
the non-inferiority of placebo was survival free of major nonsurgical HRAEs (including
stroke, pump thrombosis, major bleeding, or arterial peripheral thromboembolism) at
12 months. The principal secondary endpoint was the incidence of nonsurgical bleeding
events. A total of 589 patients were analyzed, 271 in the placebo group and 273 in
the aspirin group. At 12 months, 74% of the patients in the placebo group reached
the primary outcome compared with 68% in the aspirin group. These findings demonstrate
the non-inferiority of the placebo with an absolute between-group difference of 6.0%
improvement in event-free survival with placebo ([lower 1-sided 97.5% CI, −1.6%];
p < 0.001). Very interestingly, Mehra et al also showed that the avoidance of aspirin
was associated with reduced nonsurgical bleeding events (relative risk, 0.66 [95%
CI, 0.51–0.85]; p = 0.002) with no increase in stroke or other thromboembolic events. These results
challenge the actual guideline of antiplatelet treatment in the context of LVAD, and
the use of aspirin in this setting will likely decrease in the future.[38]
Finally, in the case of thrombotic or hemorrhagic HRAEs with or without recurrence,
treatment may be tailored based on the patient's bleeding or thrombotic risk, even
though no validated algorithm currently exists to guide clinicians.[38] Shah et al developed a multistate model to help clinicians estimate the dynamic
risk of an adverse event (GI bleeding, stroke, and death) occurring in HM 3 patients
in the next 30 days. The model includes 39 variables, to predict risk of GI bleeding
(16 variables), stroke (10 variables), and death (19 variables). It has been developed
and validated using the population of HM 3 patients included in the MOMENTUM 3 trial.
During the validation process, the model was able to predict GI bleeding and death
with moderate accuracy (AUC: 0.73 and 0.76, respectively), but accuracy was low for
stroke prediction (AUC: 0.6).[79] This model might help the clinician decide on the modification of ongoing treatment,
but it has yet to be validated in real-world HM 3 patients and its impact on the incidence
of adverse outcomes will need to be studied.
An interesting treatment algorithm has been proposed by Consolo et al to reduce anticoagulation
or antiplatelet therapy to avoid bleeding recurrence in HM 3 patients.[80] Depending on the anticoagulation status (INR <4 vs. INR >4), and the type of bleeding
(mucosal/GI/occult anemia vs. retroperitoneal/muscle hematoma vs. intracranial), different
adjustments of the antithrombotic regimen are suggested with the interruption of aspirin
combined or not with a stepwise reduction of the INR target.[80]
Conclusions
The evolution in LVAD design has led to a decrease in HRAEs for the latest generation,
represented by the HM 3. Bleeding adverse events, mostly GI bleeding, remain a major
concern. In HM 3 patients, the use of aspirin had not been proven to be beneficial
and there is growing evidence supporting its removal. VKA anticoagulation with strict
and frequent INR monitoring is recommended within a specific INR range specified by
the manufacturer. Clinical trials on apixaban like the ongoing ApixiVAD study will
provide further information regarding the safety and feasibility of treating HM 3
implanted patients with this DOAC.
Finally, clinicians should remember that one treatment does not fit all LVAD recipients
and that anticoagulation and antiplatelet therapy should be adapted based on the individual
patient's bleeding and thrombotic risk.
