Fully magnetically centrifugal left ventricular assist device and long-term outcomes: the ELEVATE registry

Abstract

Background and Aims

HeartMate 3 (HM3) is a fully magnetically levitated continuous flow left ventricular assist device, which received CE marking in 2015. The ELEVATE Registry was initiated to collect real-world outcomes in patients treated with HM3 post-CE Mark approval.

Methods

A total of 540 subjects implanted at 26 centres between March 2015 and February 2017 were included in this registry. Of these, 463 received the device as a primary implant (primary implant cohort, PIC), 19 as a pump exchange (pump exchange cohort), and in 58 patients, only anonymized survival data were collected (anonymized cohort, AC). Patients in the PIC contributed to the baseline demographics, survival, adverse events, quality of life (QoL) (EuroQoL-5 Dimensions-5 Levels visual analogue scale), and functional capacity (6 min walk distance) assessments, while patients in the AC contributed only to survival.

Results

Primary implant cohort patients had a mean age of 56 years and were predominantly male (89%) with 48% ischaemic aetiology. The majority of subjects was designated bridge to transplant (66%) and had INTERMACS Profiles 1–3 (70%). At baseline, the subjects had poor functional capacity (104 ± 140 m) and impaired QoL (35 ± 19 points). The overall survival rate of the PIC was 63.3% and survival free of stroke was 58.1% at 5 years. Significant improvements in functional capacity and QoL were observed and maintained for 5 years (301 ± 131 m and 64 ± 20 points, respectively).

Conclusions

Real-world data from the ELEVATE registry demonstrate an overall survival rate for primary implants of 63.3%. In the PIC, reductions in adverse events for patients in the extended follow-up and improved QoL and functional capacity were observed at 5 years in this patient population with advanced heart failure.

Structured Graphical Abstract

Structured Graphical Abstract.

Long-term survival and outcomes after Heartmate 3 implantation

See the editorial comment for this article ‘The transformative potential of left ventricular assist devices in advanced heart failure: no more a therapeutic orphan', by M.R. Mehra  et al., https://doi.org/10.1093/eurheartj/ehad555.

Introduction

Chronic heart failure (HF) remains a major clinical, socio-economic, and health problem. The incidence and prevalence of HF are steadily increasing worldwide, and the increase with age is often integrated into the complex context of multi-morbidity.^1,2 Multiple different treatment options exist for patients with advanced stages of the disease, but their success rate often remains limited. The treatment of advanced HF with left ventricular assist devices (LVADs), especially the newer generation of continuous flow (CF) LVADs, has demonstrated improved survival over optimal medical management for advanced HF over the last decade^3,4 The HeartMate 3™ LVAD (HM3) is a new-generation, magnetically levitated centrifugal CF-LVAD, which was first introduced clinically during the European CE Mark trial in June 2014.^5,6 The HM3 LVAD includes technical innovations that are advantageous over prior CF-LVADs such as full magnetic levitation of the rotor, large blood-flow gaps ensuring less trauma to blood components, modular driveline cable, and speed modulation for the generation of an artificial pulse. Multiple mid-term 1- and 2-year results in previously conducted studies with HM3 have been published and demonstrated survival rates higher than 80%, as seen, e.g. in the CLEAR-LVAD study, which reported a better survival rate (85% at 12 months post implant), fewer hospitalizations, and less days spent in the hospital when compared with other commercially available ventricular assist devices (VADs).^7–13 Based on the short- and mid-term results of the multi-centre, prospective, single-arm HM3 CE Mark trial, the CE Mark approval was given in October 2015.⁶ Since this period, the use of HM3 has expanded considerably throughout Europe, the Middle East, and Asia. In 2018, the Food and Drug Administration (FDA) granted approval for long-term use of HM3 based on outcomes from the MOMENTUM 3 trial conducted in North America.^13–16

More recently, 5-year survival results from the European CE Mark trial presented an overall survival rate of 61% [95% confidence interval (CI): 44%–74%] and the MOMENTUM 3 randomized trial reported an overall survival rate of 58.4% (95% CI: 52.8%–63.6%) at 5 years.^17,18 Next to these encouraging survival rates and even more importantly, is a reduction in the haemocompatibility-related adverse events in long-term follow-up, especially the near elimination of reoperations for confirmed pump thrombosis in HM3.^{6,11–13,15–18} However, real-world, long-term data on survival and incidence of adverse event rates associated with HM3 are still rare, especially outside the USA,^19,20 with the INTERMACS (US focused) and EuroMACS (EU focused, but open to all devices) registries being the main source of information. The ELEVATE registry, a prospective, predominantly European registry (with additional contributing centres in the Middle East and Asia) of consecutive, commercial HM3 implants, was implemented upon CE Mark acquisition to collect clinical data and assess the post-market experience with this device by evaluating clinically relevant endpoints in a real-world scenario.^21–24 The follow-up duration of the ELEVATE registry was initially 2 years and was expanded to 5 years after the 5-year expanded follow-up was requested by the FDA in the MOMENTUM 3 trial. The final results of the post-approval experience with HM3 in a real-world clinical setting are reported in this analysis.

Methods

The ELEVATE registry (Evaluating HeartMate 3 with Fully Magnetically Levitated Technology in a Post-Market Approval Setting; ClinicalTrials.gov Identifier: NCT02497950) is a prospective, observational, multi-national registry including patients implanted with the HM3 LVAD after commercial approval in Europe (CE Mark, October 2015) and Kazakhstan (January 2015). The registry was conducted in accordance with the Declaration of Helsinki, and its protocol and informed consent form were approved by the ethics committee of each institution. Informed consent was obtained for all the consecutively enrolled patients included between March 2015 and February 2017. The study was initiated upon the commercial approval in each region and included patients with a primary LVAD implantation and also patients in whom an HM3 replaced another LVAD; only patients who declined participation were excluded from the registry. While the initial sample size of the ELEVATE registry was set at 500 patients in up to 50 centres, successive verifications determined that a proportion of patients were not included in the registry because of a lack of consent prior to implant (for instance, emergency or comatose patients). For these patients, consent could not be obtained. However, permission from ethics committees was obtained in order to include anonymized outcomes from patients from whom consent could not be obtained, with the aim of achieving best survival and event rates in real-world conditions, thus establishing a final study size of 540 subjects.

Patients were initially followed up for 24 months post implant or until one of the prespecified outcomes was defined as transplant, explant, or death, or until withdrawal occurred. Upon successful completion of the first 2 years of follow-up in February 2019, the ELEVATE registry was extended to 5 years and 247 subjects consented to participate in the long-term follow-up. The ELEVATE registry used INTERMACS definitions for outcomes and adverse events, as previously done in both the CE Mark trial and the MOMENTUM 3 trial, although in the ELEVATE registry, these events were not adjudicated by an independent Clinical Events Committee. The post-operative management did not follow a specified protocol but was performed at the discretion of the investigators, depending on the Standard of Care practice at the enrolling centre. Data from patients with primary implants and pump upgrades included baseline data such as demographics, comorbidities, previous cardiovascular history, as well as assessments such as echocardiographic and haemodynamic parameters profile, New York Heart Association (NYHA) classification, and INTERMACS profile. Functional capacity and quality of life (QoL) were measured via 6 min walk distance (6MWD) and the EuroQoL-5 Dimensions-5 Level (EQ-5D-5L) questionnaire, respectively. At implant, information on surgical approach, complete procedure, and bypass durations, use of extracorporeal life support, outflow graft, and pump placement was collected. Patients were evaluated at 6 months and then yearly after 1, 2, 3, 4, and 5 years post implant for their clinical and functional status. Adverse events according to the protocol-specified INTERMACS definitions were reported as they occurred as well as outcomes such as transplant, explant for recovery, death, and withdrawal. Data were monitored and queried for inconsistencies and missing entries.

Statistical analysis

Continuous data are presented as the mean with a standard deviation unless otherwise specified. Categorical data are reported as frequencies and percentages. Survival data are presented using the Kaplan–Meier product method. Patients were censored at the time of transplant or last-known follow-up without a reported event (death or adverse event). Competing risks were analysed using the method of Fine and Grey. Adverse events were presented as per cent of patients and events per patient-year (EPPY). There was no imputation of missing data, except for the 6MWD. For patients not performing the 6MWD because of HF, the distance was imputed as 0 m. Six-minute walk distance and QoL visual analogue scale (VAS) at each visit were compared with baseline using the Wilcoxon signed-rank test. Improvement in NYHA class was analysed with McNemar’s test. Statistical analysis was performed using SAS version 9.4.

Role of the funding source

The ELEVATE registry was funded by Abbott and its data were collected in an electronic database developed by the sponsor. The sponsor also performed the analysis for this manuscript following the author’s directives and decisions. The ELEVATE Steering Committee and the listed authors of this manuscript had access to registry data and contributed to its preparation by determining the modality of analysis and presentation of the results.

Results

Overall, 540 patients were implanted at 26 centres [median implants per centre = 15.5 patients (interquartile range, IQR 6–30)] between March 2015 and February 2017. Out of the 540 patients (full cohort), complete data sets were obtained for 482 patients, of whom 463 received HM3 as their primary implant [primary implant cohort (PIC), median per centre = 13 patients (IQR 6–27)], 19 patients received HM3 as a replacement or upgrade from another durable device [pump exchange cohort (PEC)], and in 58 patients, only anonymized survival data were collected (anonymized cohort). Upon completion of the first 2 years of follow-up, the ELEVATE registry was extended to 5 years and 254 subjects (247 in the PIC and 7 in the PEC) consented to participate in the long-term follow-up and were available for this analysis. Patient and cohort dispositions are illustrated in the consort diagram reported in Figure 1. Additional baseline data are provided for completeness of information in the Supplementary data online, Table S1.

Figure 1.

ELEVATE consort diagram. The consort diagram reports the disposition of patients at each follow-up interval. The ELEVATE registry included three cohorts of patients: 463 patients who received HeartMate 3 as a de novo implant (primary implant cohort), 19 patients who received HeartMate 3 as a replacement of another durable device such as an HeartMate II (HMII) or an HeartWare HVAD (pump exchange cohort), and an anonymized cohort composed of 58 patients who had an outcome prior to having the possibility of signing the informed consent form, for which data on only type of outcome and duration of support were collected. Upon completion of the first 2 years of follow-up, 247 patients in the primary implant cohort and 7 in the pump exchange cohort consented to participate in the study’s extended follow-up. Patients in the primary implant cohort contributed to the baseline demographics, survival, adverse events, quality of life (EuroQoL-5 Dimensions-5 Levels visual analogue scale), and functional capacity (6 min walk distance) assessments, while patients in the anonymized cohort contributed only to survival analysis

Primary implant cohort

Baseline characteristics of the 463 patients in the PIC have been previously published^21–24 and are reported in Table 1. Ninety-two per cent (426 out of 463) of the implanted patients were successfully discharged with an index hospitalization median length of stay of 29 days, while 7.1% (33 out of 463) expired in hospital, and 0.8% were withdrawn either due to transplant (2 out of 463 patients, 0.4%) or due to consent withdrawal (2 out of 463, 0.4%, Supplementary data online, Table S2). The 24-month survival rate in the PIC was 83.4% (95% CI: 79.9%–86.8%) and the median duration spent out of hospital was 671.5 days (range 592–698). The majority of patients (73.2%) was alive on their initial device at 2 years.²⁴

Table 1.

Baseline data: primary implant cohort

	ELEVATE primary implant cohort N = 463
Age mean (years) median (IQR)	55.6 ± 11.7 58 (50–64)
BSA mean (m2) median (IQR)	2.02 ± 0.23 2.01 (1.87–2.16)
BMI (kg/m2) mean median (IQR)	27.3 ± 5 26.7 (23.9–30.2)
Sex, n (%)
Male	412	89%
Female	51	11%
Indication, n (%)
Bridge to transplant	305	66%
Destination therapy	122	26%
Bridge to candidacy or recovery	36	8%
INTERMACS profile, n (%)
Profile 1	43	9%
Profile 2	102	22%
Profile 3	176	39%
Profile 4	125	27%
Profile 5	8	2%
Profile 6	2	<1%
Not provided	7
NYHA class, n (%)
IIIA	42	10%
IIIB	165	38%
IV	231	53%
Not provided	25
Respiratory disorder, n (%)	159	34.3%
COPD	49	10.6%
Diabetes Type 1, n (%)	18	3.9%
Diabetes Type 2, n (%)	101	21.8%
Bleeding disorder, n (%)	28	6.0%
Neuro history, n (%)	76	16.4%
TIA	16	3.5%
Ischaemic stroke	32	6.9%
Haemorrhagic stroke	6	1.3%
Seizure	6	1.3%
Other	23	5.0%
Renal insufficiency, n (%)	141	30.5%
Hypertension, n (%)	220	47.5%
Coronary artery disease, n (%)	235	50.8%
PCI	179	38.7%
CABG	48	10.4%
Myocardial infarction, n (%)	194	41.9%
Carotid artery disease, n (%)	11	2.4%
Arrhythmia, n (%)	280	60.5%
AF	179	38.7%
AFL	16	3.5%
VF	43	9.3%
VT	96	20.7%
Valve disease, n (%)	331	71.5%
Valve repair/replace, n (%)	44	9.5%
Pacemaker/defibrillator, n (%)	310	67.0%
Pacemaker	10	2.2%
CRT	9	1.9%
CRT-D	117	25.3%
ICD	212	45.8%
Transplant, n (%)	0	0.0%
MCS, n (%)	59	12.7%
ECLS	44	9.5%
RVAD	1	0.2%
LVAD	8	1.7%
Other MCS	16	3.5%
IABP, n (%)	32	6.9%
HF duration, n (%)
<1 year	119	25.7%
≥1 year	344	74.3%
BP systolic (mmHg), n (mean ± SD) median (IQR)	451	102.3 ± 14.9 100 (92–110)
BP diastolic (mmHg), n (mean ± SD) median (IQR)	451	65.4 ± 11 65 (59–70)
Mean BP (mmHg), n (mean ± SD) median (IQR)	451	77.7 ± 10.8 77 (70–83)
PCWP (mmHg), n (mean ± SD) median (IQR)	314	24.8 ± 9.4 26 (18–31)
PAS (mmHg), n (mean ± SD) median (IQR)	354	50.8 ± 15.9 51 (40–63)
PAD (mmHg), n (mean ± SD) median (IQR)	341	24.9 ± 9.2 25 (19–31)
mPAP (mmHg), n (mean ± SD) median (IQR)	352	34.6 ± 11.2 36 (27–42)
CVP (mmHg), n (mean ± SD) median (IQR)	247	10.8 ± 6.7 10 (5–15)
Cardiac index (L/min/m2), n (mean ± SD) median (IQR)	339	1.92 ± 0.56 1.87 (1.52–2.20)
Cardiac output (L/min), n (mean ± SD) median (IQR)	330	3.81 ± 1.15 3.61 (3.00–4.34)
PVR (dynes cm−5), n (mean ± SD) median (IQR)	309	393.6 ± 1297.8 227.0 (143.0–325.6)
LVEF (%), n (mean ± SD) median (IQR)	428	18.3 ± 6.3 19.5 (15.0–23.0)

AF, atrial fibrillation; AFL, atrial flutter; BMI, body mass index; BP, blood pressure; BSA, body surface area; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; CRT, cardiac resynchronization therapy; CRT-D, cardiac resynchronization therapy with defibrillation; CVP, central venous pressure; ECLS, extracorporeal life support; HF, heart failure; IABP, intra-aortic balloon pump; ICD, implantable cardiac defibrillator; IQR, interquartile range; LVAD, left ventricular assist device; MCS, mechanical circulatory support; LVEF, left ventricular ejection fraction; mPAP, mean pulmonary artery pressure; NYHA, New York Heart Association; PAD, pulmonary artery diastolic pressure; PAS, pulmonary artery systolic pressure; PCI, percutaneous coronary intervention; PCWP, pulmonary capillary wedge pressure; PVR, pulmonary vascular resistance; RVAD, right ventricular assist device; SD, standard deviation; TIA, transient ischaemic attack; VF, ventricular fibrillation; VT, ventricular tachycardia.

INTERMACS profiles: These profiles describe the health status of the HF patient in seven categories in decreasing order of severity. Profile 1 is used for patients in a life-threatening situation, such as cardiogenic shock, while Profile 7 describes clinically stable patients with a reasonable level of comfortable activity. For a complete description of the profiles, please see Supplementary data online, Table S8—INTERMACS profiles description.

At 5 years post implant, 204 of the 463 primary implant subjects (44.1%) were still ongoing. Seventy-eight patients (16.8%) were transplanted within the first 5 years after HM3 implantation, 146 expired (31.5%), and 35 (7.6%) were withdrawn for different reasons, including refusal to participate in the extended study follow-up (Figure 2).

Figure 2.

Competing outcomes primary implant cohort. The competing outcomes graph represents the percentages of patients who were ongoing supported with a HeartMate 3 or had an outcome during the study. At each point, the sum of all percentages plotted equals to 100% of the primary implant cohort population. Forty-four (44.1%) per cent of the primary implant cohort patients were on support at 5 years, while 16.8% were transplanted and 31.5% expired. Almost 7.6% of the patients discontinued their participation for other reasons, mainly due to the refusal to participate in the longer follow-up

At the conclusion of the study, the Kaplan–Meier survival rate in the PIC was 63.3% (95% CI: 58.2%–67.9%; Figure 3A) and 56.3% (95% CI: 51.5%–60.8%) for patients from the combined primary implant and anonymized cohorts (Figure 3B), while the overall survival for the pump exchange cohort is reported in Supplementary data online, Figure S1. The rate of composite survival free of non-surgical bleeding, stroke, or pump thrombosis was 43.8% (95% CI: 38.8%–48.8%), while the rates of survival free of stroke or bleeding were 58.1% (95% CI: 52.9%–62.8%) and 40.6% (95% CI: 35.7%–45.4%), respectively (Figure 4).

Figure 3.

Overall survival at 5 years—primary implant (A); primary implant and anonymized cohorts (B). Overall survival for the primary implant cohort (n = 463): the survival rate at 5 years was 63.3%. (B) Overall survival of the primary implant (n = 463) and anonymized cohort (n = 58) combined: the overall survival rate was 56.3% at 5 years

Figure 4.

Survival free of non-surgical bleeding, stroke, or pump thrombosis (A), survival free of stroke (B), survival free of major bleeding (C)—primary implant cohort. (A) A combined endpoint of non-surgical bleeding, stroke, or pump thrombosis: at 5 years, 43.8% of patients did not experience any of these events. (B) Survival free of stroke and (C) survival free of major bleeding: 58.1% and 40.6% of subjects were event-free at 5 years, respectively. All graphs are for the primary implant cohort population (n = 463)

The majority of patients showed sustained improvements in NYHA functional class, 6MWD, and EQ-5D-5L VAS during the first 2 years of follow-up. Similarly, the majority of patients remained in NYHA Classes I and II (18.2% and 46.1%, respectively; P < .0001 vs. baseline, Supplementary data online, Table S3) and maintained significant improvements in both 6MWD, 301 ± 131 m (median 308.5, IQR 230–398.5, Supplementary data online, Table S4), and QoL with the EQ-5D-5L VAS 64 ± 20 points (median 70, IQR 50–80, Supplementary data online, Table S5) at 5 years (P < .001) against baseline for both assessments (Figure 5). INR changes over time are reported in Supplementary data online, Table S6

Figure 5.

New York Heart Association class (A), 6 min walk distance (B) and quality of life (C) over time—primary implant cohort. Functional and quality-of-life assessments. (A) New York Heart Association class across the 5-year study duration: at baseline, the majority of patients was New York Heart Association Class III or IV, while at 5 years, the majority of patients upgraded to New York Heart Association Class I or II. (B) Dramatic improvement in the 6 min walk distance in the first year post implant, which was then sustained throughout the long-term follow-up to 5 years. (C) Visual analogue scale, a part of the EuroQoL-5 Dimensions-5 Levels questionnaire, in which the patients indicate current health status on a scale from 0 (extremely bad) to 100 (extremely good): during the first year post implant, the perceived health status improved considerably in the first year and this maintained throughout the long-term follow-up to 5 years. The lines in B and C depict median. All graphs are for the primary implant cohort population (n = 463)

The rates of adverse events reduced as patients were maintained on support for a longer duration. In the primary implant group, 32.8% of subjects experienced a major bleeding event in the first 2 years of follow-up (EPPY = 0.36), which reduced to 18.2% in the successive 2- to 5-year follow-up period (EPPY = 0.07), with a total of 0.18 EPPY at 5 years. More specifically, gastrointestinal bleeding was reported in 10.2% of patients in the first 2 years (EPPY = 0.08) and in 8.1%, between 2 and 5 years of follow-up (EPPY = 0.03) with an EPPY = 0.05 between 0 and 5 years. Event rates between early and later observation periods also reduced for major infections (56.4% vs. 48.6%, EPPY 0.67 vs. 0.23) with an EPPY of 0.40 between 0 and 5 years, sepsis (15.1% vs. 12.1%, EPPY 0.11 vs. 0.03; EPPY 0–5 years = 0.06), ischaemic stroke (5.38% vs. 0.8%, EPPY 0.04 vs. 0.00; EPPY 0–5 years = 0.02), and haemorrhagic stroke (5.0% vs. 2.8%, EPPY 0.03 vs. 0.01; EPPY 0–5 years = 0.02). At 5 years, 0.21 EPPY were reported for haemocompatibility-related adverse events such as major bleeding and stroke, while infection was reported with EPPY = 0.40 and right HF EPPY = 0.04 (Table 2). The full adverse event data set and event definitions are shown in Supplementary data online, Tables S7 and S8. Patients in the PIC, who expired or had an outcome and as such could not contribute to the freedom from major bleeding analysis, have been accounted for in a competing risk analysis for bleeding using the method of Fine and Grey. The cumulative incidence rate of bleeding within 5 years is slightly <40% (see Supplementary data online, Figure S2).

Table 2.

Haemocompatibility-related adverse events at 5 years

	0–2 years (0–730 days)				2–5 years (731–1825 days)				0–5 years
	(N = 463)				(N = 247)
	n	%	# Events	EPPY	n	%	# Events	EPPY	EPPY
HRAE	174	37.6	326	0.42	52	21.1	87	0.08	0.21
Major infection	261	56.4	515	0.67	120	48.6	261	0.23	0.40
Right heart failure	71	15.3	73	0.09	9	3.6	13	0.01	0.04
Other events	321	69.3	950	1.23	124	50.2	350	0.30	0.68

HRAE, haemocompatibility-related adverse events (major bleeding and stroke); EPPY, events per patient-year.

Echocardiography examinations collected at 6 or 12 months (n = 406), 24 months (n = 420), and 60 months (n = 164) post implant were analysed to identify aortic regurgitation. The study did not include a specific echo protocol. However, a new onset of moderate-to-severe aortic regurgitation was observed in 19 subjects (4 severe and 15 moderate) at 1 year, while an additional 17 moderate cases were reported for a total of 36 moderate-to-severe cases by 60 months.

While the rate for freedom from re-hospitalization at 2 years was 30.9% (95% CI: 26.4%–35.4%), the re-hospitalization rate between 2 and 5 years was 37.9%, with the majority of re-hospitalizations caused by adverse events (see Supplementary data online, Table S7). The freedom from re-hospitalization curve between 2 and 5 years is reported in Supplementary data online, Figure S3.

Over the observation time of 5 years, suspected device malfunctions were reported for 32 patients (6.9%) regarding internal components (outflow graft and pump) and 44 (9.5%) regarding external components. Within reported device malfunctions, suspected pump thrombosis (1.1%), intrinsic outflow graft complications, such as outflow graft bend relief disconnections (0.5%), outflow graft twists (4.3%), and kink (0.5%) or the newly described extrinsic compression of the outflow graft phenomenon (2.8%) were included. After 5 years of VAD support, the rate of freedom from intrinsic outflow graft complications was 93.8% (95% CI: 90.6%–95.9%).

As haemocompatibility-related adverse events can significantly influence the outcome of VAD patients, a specific analysis based on survival free of stroke, pump thrombosis, or non-surgical bleeding was performed. At 5 years, the rate of overall survival free from stroke, pump thrombosis, and non-surgical bleeding was 43.8% (Figure 4A).

Primary implant and anonymized cohorts combined

The rate of 5-year survival of the combined primary implant and anonymized cohorts was 56.3% (95% CI: 51.5%–60.6%). The majority of patients was alive on their initial device or transplanted (39.2% and 15.0%, respectively) at 5 years after HM3 implantation.

Thirty-nine per cent (38.96%) of patients expired on their device, the device was explanted in 1.5% of patients, and 1% of patients were lost to follow-up. In addition, 3.3% of patients did not consent to participate in the study’s extended follow-up.

In this cohort, infection/sepsis (23.9%) was the most frequent cause of death, followed by multi-organ failure (20.6%) and stroke (11.5%). The causes of death for this cohort are also reported in Table 3.

Table 3.

Causes of death, primary implant cohort, and primary implant + anonymized cohorts

	Primary implant cohort deaths (n = 146)		Primary implant + anonymized cohorts deaths (n = 203)
	n	%	n	%
Bleeding	7	4.8	8	3.9
COVID-19	3	2.1	3	1.5
Cardiac arrest	3	2.1	4	2.0
Cardiac decompensation	8	5.5	12	5.9
Right heart failure	9	6.2	16	7.9
Cardiac infarction	0	0.0	1	0.5
Circulatory failure	3	2.1	3	1.5
Infection/sepsis	36	24.7	47	23.2
Multi-organ failure	18	12.3	43	21.2
Renal failure	2	1.4	2	1.0
Respiratory failure	3	2.1	4	2.0
Stroke	20	13.7	24	11.8
Suicide	5	3.4	5	2.5
Other	13	8.9	15	7.4
Unknown	16	11.0	16	7.9

Discussion

The ELEVATE registry provides a thorough evaluation of a ‘real-world population’ of HM3-supported patients followed for up to 5 years. As 70% of the study population was categorized as INTERMACS Profiles 1–3, and 12% of patients were on pre-LVAD temporary mechanical circulatory support, the ELEVATE population is comparable with other registries and represents a valuable illustration of a ‘real-world’ LVAD cohort.^25–27 Data from this registry demonstrated a 5-year survival rate of 63.3% in advanced HF patients with advanced congestive HF who received HM3 as their primary LVAD at implantation (Structured Graphical Abstract). These results compare favourably with the 5-year survival rate reported in the HM3 CE Mark study (61%),¹⁷ with the rate provided in the INTERMACS 12th report (44.2%), and with the rate reported in the recently published long-term results of the MOMENTUM 3 trial (58.4%).¹⁸ This important result was achieved despite a low transplant rate below 15% and the liberal inclusion of patients in INTERMACS Profile 1 or on pre-LVAD temporary mechanical support.

Additionally, a sustained improvement in NYHA functional class and QoL was observed throughout the 5-year follow-up duration. Stroke, bleeding, and pump thrombosis rates in ELEVATE remained low over the study period. The low transplant rate observed in ELEVATE reflects the severe shortage of donor organs in the participating countries and underlines the opportunities provided by a stable LVAD platform suitable for long-term support.

In the ELEVATE registry, HM3 implantation was associated with a sustained improvement in the 6MWD, and 64.3% of patients were in NYHA Classes I and II after 5 years. At baseline, the majority of patients was unable to walk because of HF. By 6 months post-HM3 implant, the absolute mean distance walked had improved by 203 m, plateauing thereafter. The plateau after 6 months, however, might reflect a limitation of the current assist devices and fixed LVAD speeds with limited response to patient activities, warranting further investigation. Quality of life also improved significantly after HM3 implantation. The improvement in QoL sustained up to 5 years and did not decrease with the accumulation of adverse events. This is in line with the 5-year results of the CE Mark study.^12,17

Historical reports of VAD implantation suggested that post-thromboembolic and bleeding complications were frequent, which could limit the merits of therapy.²⁸ Stroke has been one of the most devastating complications of these types of events and is a major driver of mortality and morbidity²⁸: in the ELEVATE PIC, it was observed in 15% of patients within the first 2 years and in 3.6% of patients between 2 and 5 years, accounting for 13.7% of deaths at 5 years. This is comparable with the stroke rate in the MOMENTUM 3 trial as well as IMACS and EUROMACS.^15,18,25,26 The reduction in stroke events over time reflects the potential haemocompatibility superiority of the HM3 system over previous CF devices.^16,29,30 but could also reflect improved medical management. Given that almost half of stroke events in ELEVATE were haemorrhagic, adjusted protocols for reduced anticoagulation might further improve stroke rates with the HM3 system.³¹ The HM3 design features, such as speed modulation and wide blood-flow paths, contribute to improved pump haemocompatibility⁶ and reduced gastrointestinal bleeding events, which have been linked to a degradation of high-molecular-weight multimers of the von Willebrand factor and continuous blood flow.³¹ As a result, the amount of high-molecular-weight multimers of von Willebrand factor degradation with HM3 is lower than that of other devices³² and is associated with an overall decrease in the gastrointestinal bleeding rate.^{10,13,15,16,19,20}

Outflow graft twist is an HM3-specific complication³³ and affected 4.3% of patients in ELEVATE. This issue, however, has since been resolved and a solution has been found for new HM3 implants. A recently published report of 2108 HM3 patients showed that while improved outflow graft attachments have solved outflow graft twist, there remains a small percentage of HM3 patients (2.9%) who may still experience outflow graft obstruction due to external compression resulting from the accumulation of a gelatinous substance between the outflow graft and the bend relief.³⁴

Suspected pump thrombosis was reported only in 1.1% of patients (n = 5), which echoes the extremely low pump thrombosis rates observed in the CE Mark study (0%)¹⁷ and MOMENTUM 3 trial (2.1%),¹⁸ demonstrating superiority over previous axial pumps.^13–16,18

Limitations

A limitation of the ELEVATE registry is the presence of the AC of 58 patients, for whom only survival data could be collected, as well as the group of patients who did not consent to the extended follow-up (6.9%). The large size of the registry counteracts but does not eliminate the importance of this source of bias as it introduces a lack of comprehensive data for these patients, which could have impacted the rates presented. As the data collection of a registry is less strict than that of traditional clinical studies, it yields more missing data in some of the assessments. However, it was possible to present accurate survival rates with the introduction of the AC, as approved by the participating centre’s ethics committees. For presenting the rates of overall survival and composite survival free of non-surgical bleeding, stroke, or pump thrombosis analysis, subjects who withdrew from the study were censored at the time each subject was last known to be alive. This implies that individuals with unobserved events are assumed to have the same hazard of failure as those patients still in the study after the censor time, which can be difficult to verify. Another limitation is the absence of an independent adjudication of clinical events, which are collected as reported by the investigators, and this may have introduced mis-categorizations of events. However, independent adjudications of events are not a standard practice in a real-world scenario, which is the main focus of the ELEVATE study. Post-operative management was performed at the discretion of the investigators, depending on the Standard of Care practice at the enrolling centre; this inherently caused differences in data collection and availability from centre to centre. The protocol did not include a specific echo protocol, and assessments such as QoL and NYHA class could be potentially influenced by a personal interpretation of either the patients completing the QoL or the physician assessing NYHA functional class. Although patient and physician interpretation cannot be extricated from these measurements, they are widely accepted as useful tools to describe HF populations.

Conclusions

The real-world ELEVATE registry long-term 5-year data demonstrate a survival rate exceeding 60% in the PIC and significant improvements in both QoL and functional capacity. There was a reduction in complications in Years 3–5 compared with the first 2 years for patients in the extended follow-up, with the majority of events related to bleeding and infection. These results underline the need for defining and adopting best management practices and striving for optimal biocompatibility in next-generation pumps.

Supplementary data

Supplementary data are available at European Heart Journal online.

Declarations

Disclosure of Interest

The submitting author declares grants and personal fees from Abbott during the conduct of study; Steven Shaw has nothing to disclose; Jen Garbade has nothing to disclose, Finn Gustafsson declares consulting fees from Abbott, Corwave, FineHeart; Michiel Morshuis has nothing to declare; Daniel Zimpfer declares grants and speaker's fees from Medtronic, Abbott, Edwards, Berlin Heart and support for attending meeting and/or travel from Medtronic, Edwards, Abbott, Berlin Heart, Abiomed; Jacob Lavee has nothing to disclose; Yuriy Pya has nothing to disclose; Michael Bertchtold-Herz has nothing to disclose; AiJia Wang is an Abbott employee; Carlo Gazzola is an Abbott employee; Evgenij Potapov declares grants, consulting fees, expert testimony, speaker's fees, advisory board/data safety monitoting board fees, support for attending meeting and/or travel from Abbott, Medtronic, Abiomed; Diyar Saeed has nothing to disclose.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

The ELEVATE registry was funded by Abbott.

Ethical Approval

The registry was conducted in accordance with the Declaration of Helsinki and its protocol and the informed consent form was approved by the ethics committee of each institution.

Pre-registered Clinical Trial Number

The pre-registered clinical trial number is NCT02497950.