Adverse outcomes in nontrial like patients on commercial HeartMate 3 LVAD support

Abstract

Background

Patient selection for real-world use of HeartMate 3 left ventricular assist device (HM3 LVAD) differs from that of clinical trials. We assessed adverse-event rates between “nontrial like” (NTL) and “trial like” (TL) patients who underwent commercial HM3 implantation.

Methods

We conducted a multicenter, retrospective cohort study of 231 consecutive patients who underwent primary HM3 implant from October 2017 to December 2021. Patients were identified as NTL if they met ≥1 exclusion criteria from the Mulitcenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial and TL if they had no exclusion criteria. The primary outcome was freedom from death, heart failure hospitalization, LVAD-related infection, and hemocompatibility-related adverse events including major gastrointestinal bleed, all-cause stroke, and pump thrombosis.

Results

Of the 231 patients (median age 59 years, 24% female), 105 (45%) were categorized as NTL, most commonly due to low albumin and/or prealbumin levels alone (n = 48) followed by the use of temporary mechanical circulatory support alone (n = 10) and having ≥2 exclusion criteria (n = 26). Compared to the TL group, the NTL group had a lower 1-year event-free survival (34% vs 48%, adjusted hazard ratio (HR) [95% confidence interval (CI)]: 1.5[1.01-2.0], p = 0.043). One-year overall survival was 82% in NTL group vs 95% in TL group (adjusted HR [95%CI]: 3.3[1.1-8.7], p = 0.013). Post-HM3 adverse events were associated with the presence of ≥2 exclusion criteria (adjusted HR [95%CI]: 1.8[1.06-3.1], p = 0.028); and veno-arterial extracorporeal membrane oxygenation support prior to HM3 (adjusted HR [95%CI]: 1.9[1.08-3.4], p = 0.026).

Conclusions

In this multicenter cohort of HM3 patients, those outside of traditional clinical trial eligibility had a significantly lower 1-year event-free survival.

Graphical abstract

Left ventricular assist device (LVAD) therapy helps patients with advanced heart failure beyond what could be achieved by optimal medical therapy alone.1, 2 The HeartMate 3 (HM3; Abbott Labs, Chicago, IL) is currently the only clinically available LVAD.³ In the Mulitcenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial and its prespecified longitudinal analyses, patients on HM3 support had a significantly greater freedom from death, disabling stroke, and pump removal/replacement when compared to those on HeartMate II support.4, 5

However, in real-world practice, LVAD therapy is utilized in a broader patient population outside of clinical trial eligibility.6, 7, 8 Exclusion criteria often overlaps with clinical risk factors, which would preclude patients from trial enrollment. For example, patients with cardiogenic shock on temporary mechanical support (tMCS), other than intra-aortic balloon pumps (IABPs), would have been excluded from MOMEMTUM 3. As a result, the frequency of post-HM3 adverse events may be under-recognized in real-world populations.

A recent analysis of the International Registry for Mechanically Assisted Circulatory Support (INTERMACS) registry demonstrated that 42% of patients were “trial-ineligible” at the time of HM3 implantation.⁸ This cohort had a lower survival when compared to trial-eligible patients.⁸ Nevertheless, there is limited data on the concurrent risk of nonfatal complications such as hemocompatibility-related adverse events (HRAE), heart failure hospitalization (HFH), and LVAD-related infection (LVADI), all of which have differential effects on survival and quality of life.9, 10, 11, 12 For patients with a “nontrial like” profile, these facets may affect shared decision-making on whether to proceed with LVAD surgery.

In this study, we evaluated a composite endpoint of mortality and nonfatal adverse events between “nontrial like” (NTL) and “trial like” (TL) patients who underwent commercial HM3 implantation.

Methods

Study population and design

This multicenter, retrospective cohort study was approved by the institutional review board and was conducted at Advocate Christ Medical Center (Oak Lawn, IL) and Aurora St. Luke’s Medical Center (Milwaukee, WI). All patients who underwent primary, commercial HM3 implantation from October 2017 to December 2021 were included. Patients in MOMENTUM 3 were excluded from this study. Follow-up data were through December 2022.

During this period, 231 patients met these criteria. We divided our cohort into 2 groups: (1) TL patients (N = 126): HM3 patients who met all inclusion criteria and had no exclusion criteria according to the MOMENTUM 3 trial; and (2) NTL patients (N = 105): HM3 patients who met all inclusion criteria and had at least one exclusion criteria (Supplemental Protocol and Supplemental Figure 1). The exclusion criterion of “use of tMCS” applied to all devices except IABPs.

The primary composite endpoint was a composite of all-cause mortality, LVADI, HFH, and HRAEs including major gastrointestinal bleed, all-cause stroke, and pump thrombosis at 1 year following HM3 implantation. Patients who underwent heart transplantation in the first year of HM3 support were censored at the time of transplantation. In patients who had multiple events, we evaluated the time to first adverse event. Outlined in the Supplemental Protocol, adverse event definitions were in accordance with INTERMACS version 5.0.¹³ Secondary endpoints were the individual adverse events that made up the primary endpoint and post-HM3 outcomes listed in Supplemental Protocol. The association between the individual exclusion criteria and primary endpoint was also examined. Patients with 1 and ≥2 criteria were separately compared to the TL group. Secondary analyses of event-free survival were performed in patients without veno-arterial extracorporeal membrane oxygenation (VA-ECMO) support and in those with an uncomplicated index hospitalization, defined as having no post-HM3 adverse events during implant hospitalization.

Trial like vs nontrial like designation

The research coordinators from each institution (C.G. and K.S.) who previously enrolled patients in MOMENTUM 3 identified NTL patients by using the same published exclusion criteria.⁴ To reduce potential bias, the research coordinators only evaluated clinical data prior to HM3 implant and were blinded to the post-HM3 outcomes in both groups.

Statistical analysis

Continuous variables (median and interquartile range or mean and standard deviation) were compared using the Mann-Whitney U-test because the data were not normally distributed. Discrete variables (absolute and percent values) were compared using Fisher’s exact test. Survival was estimated using Kaplan-Meier log-rank test. Competing risks cumulative incidence was done to model the primary composite endpoint and all-cause mortality with heart transplantation as a competing event. Cox regression analysis was used to determine the hazard ratio (HR) and risk factors for the primary and secondary endpoints (time interval was time to first adverse event). Collinearity diagnostics for linear relationship were performed on risk factors significant on univariable analysis. INTERMACS profiles 1 and 2 were compared individually to the profile 3 to 4 group. After univariable analysis, variables with p < 0.05 were placed in a multivariable model to evaluate their association with the primary and secondary endpoints. Cases with the missing variable were omitted, and the remaining cases were analyzed. Center effect was examined by stratifying by center in the Cox analyses. Statistical significance was set at p < 0.05. GraphPad Prism version 9.2 (San Diego, CA), SPSS Statistics version 25.0 (Armonk, NY), and SAS version 9.4 (Cary, NC) were used for statistical analyses.

Results

Patient characteristics and trial eligibility

Clinical characteristics of the studied 231 patients (24% female, 44% White) are detailed in Table 1. The median age was 59 (50−66) years. There were 105 (45%) NTL patients, in which 79 patients met 1 exclusion criterion and 26 patients met 2 or more criteria (Figure 1). Of the 19 patients needing tMCS, 17 (89%) patients were on VA-ECMO support while 2 patients were on percutaneous LVAD support.

Table 1.

Clinical Characteristics of Trial and Nontrial Like HM3 Patients

	Overall cohort (N = 231)	Trial like (N = 126)	Nontrial like (N = 105)	p-value
Age, years	59 (50-66)	60 (49-66)	58 (50-66)	0.72
Female sex	56 (24%)	25 (20%)	31 (30%)	0.09
Race
White	103 (44%)	56 (44%)	47 (45%)	>0.99
Black	104 (45%)	57 (45%)	47 (45%)	>0.99
Hispanic (non-Caucasian)	18 (8%)	11 (9%)	7 (7%)	0.63
Other	6 (3%)	2 (2%)	4 (4%)	0.41
Ischemic cardiomyopathy	99 (43%)	46 (37%)	53 (50%)	0.045
History of stroke	37 (16%)	19 (15%)	18 (17%)	072
History of atrial fibrillation	98 (41%)	45 (36%)	53 (50%)	0.032
Body-surface area, m2	2.0 (1.9-2.2)	2.0 (1.9-2.3)	2.0 (1.8-2.2)	0.45
Left ventricular ejection fraction, %	16 (14-20)	17 (14-20)	15 (14-20)	0.77
CVP, mm Hg	11 (7-16)	11 (7-16)	12 (7-16)	0.80
PCWP, mm Hg	24 (18-29)	24 (17-29)	25 (18-30)	0.21
Cardiac index, liter/min/m2	1.8 (1.5–2.2)	1.8 (1.5-2.0)	1.8 (1.5-2.3)	0.42
PVR, Wood units	2.9 (1.9-4.4)	2.8 (1.9-3.9)	3.1 (1.8-4.9)	0.32
Serum creatinine, mg/dl	1.2 (1.0-1.5)	1.2 (1.0-1.5)	1.1 (0.94-1.4)	0.21
Estimated GFR, ml/min/1.73 m2	62 (47-78)	62 (48-78)	66 (44-80)	0.65
Total bilirubin, mg/dl	0.8 (0.5-1.2)	0.7 (0.5-1.1)	0.8 (0.5-1.4)	0.023
INR	1.1 (1.1-1.2)	1.1 (1.1-1.2)	1.1 (1.1-1.3)	0.021
Albumin, g/dl	3.1 (2.7-3.4)	3.3 (3.0-3.5)	2.8 (2.5-3.1)	<0.0001
Prealbumin, mg/dl	18 (13–23)	20 (17-25)	14 (11-19)	<0.0001
Pre-HM3 LOS, days	13 (8–20)	13 (7-19)	13 (8-20)	0.55
Intended goal of HM3 support
Destination therapy	171 (74%)	82 (65%)	89 (85%)	0.0008
Bridge to transplant	52 (26%)	44 (35%)	16 (15%)
HM3 risk score	2.56 (2.17-3.06)	2.52 (2.05-2.94)	2.63 (2.21-3.22)	0.13
HM3 risk score ≥2.97	70 (30%)	30 (24%)	40 (38%)	0.021
HM3 risk score componentsa
CABG or prior valvular surgery	55 (24%)	29 (23%)	26 (25%)	0.76
Serum sodium, mmol/liter	135 (132-138)	136 (134-138)	134 (131-136)	0.0008
BUN, mg/dl	24 (18-32)	24 (18-31)	25 (17-34)	0.47
LV EDD ≤5.5 cm	19 (8%)	4 (3%)	15 (14%)	0.0031
CVP to PCWP ratio ≥0.6	60 (26%)	36 (28%)	24 (23%)	0.36
Pre-HM3 hemodynamic support
Inotrope(s) alone	70 (30%)	47 (37%)	23 (22%)	0.014
IABP	142 (62%)	79 (63%)	63 (60%)	0.68
tMCS	19 (8%)	0 (0%)	19 (18%)	<0.0001
VA-ECMO	17 (7%)	0 (0%)	17 (16%)	<0.0001
INTERMACS profiles
1-2	82 (35%)	34 (27%)	48 (45%)	0.0037
3-4	149 (65%)	82 (73%)	46 (55%)

Abbreviations: BUN, blood urea nitrogen; CABG, coronary artery bypass graft; CVP, central venous pressure; GFR, glomerular filtration rate; IABP, intra-aortic balloon pump; INR, international normalized ratio; LOS, length of hospital stay; LV EDD, left ventricular end-diastolic diameter; PCWP, pulmonary capillary wedge pressure; PVR, pulmonary vascular resistance; tMCS, temporary mechanical circulatory support.

^aAge is not in this section because it is detailed elsewhere in the Table.

Figure 1.

Exclusion criteria that were present in the nontrial like (NTL) cohort. PAD, peripheral arterial disease; tMCS, temporary mechanical circulatory support. Hematology = thrombocytopenia and/or hypo- or hypercoagulable state. Pulmonology = Severe chronic obstructive pulmonary disease. Exclusion criteria are detailed in Supplemental Protocol.

When compared to the TL group, the NTL group had a greater proportion of patients with a designation of “destination therapy” (85% vs 65%, p = 0.0008). The NTL group had lower serum sodium levels and more INTERMACS profile 1 or 2 (45% vs 27%, p = 0.0037). Compared to TL group, the NTL group had a greater proportion of HM3 risk score ≥2.97 (38% vs 24%, p = 0.021), the cutoff value designating high-risk status.¹⁴ Both groups were similar in age, gender, race, and body mass index. There were no differences in presurgery hospital length of stay and history of prior coronary artery bypass graft or valvular surgery.

NTL patients and adverse outcomes

Compared to TL patients, the NTL group had a lower event-free survival at 1 year after HM3 implantation on Kaplan-Meier estimates (34% vs 48%, p = 0.0077, Figure 2A-C). The cumulative incidence of the primary endpoint with heart transplant as a competing risk was higher in the NTL vs TL group (63% vs 50%, p = 0.01, Supplemental Figure 2A). The NTL group had 3.5 events per patient-year while the TL group had 2.0 events per patient-year (p = 0.0056, Table 2). The NTL group had a lower 1-year overall survival (82% vs 95%, p = 0.0025, Supplemental Figure 3A), a trend toward higher rates of HRAEs and LVADI events, and similar HFH rates (Supplemental Figure 3B-E). No pump thrombosis was observed. The trend was similar when using competing risks to model 1-year mortality in the NTL vs TL cohorts (17% vs 4%, p < 0.01, Supplemental Figure 2B). Eighteen of the 24 (75%) patients who died experienced at least one incident of HFH, HRAEs, or LVADI prior to death.

Figure 2.

Primary composite endpoint of freedom from death, heart failure hospitalization (HFH), hemocompatibility-related adverse events (HRAE), and left ventricular assist device-related infection (LVADI) at 1 year on commercial HM3 support among nontrial and trial like patients (A). Competing incidences of overall survival vs HFH, HRAE, and LVADI in the trial (B) and nontrial like (C) groups.

Table 2.

Clinical Outcomes Following HM3 Implantation

	Overall cohort (N = 231)	Trial like (N = 126)	Nontrial like (N = 105)	p-value
Bypass time, minutes	58 (44-79)	57 (44-73)	62 (45-94)	0.046
OR time, minutes	224 (165-310)	206 (156-268)	248 (175-375)	0.0019
Total index LOS, days	42 (28-57)	39 (26-55)	44 (31-62)	0.052
Post-HM3 LOS, days	24 (17-38)	22 (17-35)	29 (17-41)	0.09
Post-HM3 duration of inotropes, days	11 (6-22)	11 (6-20)	13 (6-25)	0.29
Post-HM3 early acute RHF needing temporary RVAD, no.	42 (18%)	12 (10%)	29 (28%)	0.005
Duration of RVAD support, days	30 (20-46)	31 (18-35)	29 (20-58)	0.75
Adverse events per patient-year	2.6 ± 7.3	2.0 ± 6.4	3.5 ± 8.1	0.0056
LVADI and HRAE per patient-year	1.7 ± 5.0	1.3 ± 4.4	2.1 ± 5.7	0.016
HRAE per patient-year	1.1 ± 3.5	0.92 ± 3.0	1.5 ± 4.1	0.069
HFH events per patient-year	0.24 ± 0.63	0.27 ± 0.64	0.21 ± 0.62	0.37
Timing of initial adverse event
During index hospitalization, no.	53 (23%)	19 (15%)	34 (32%)	0.0026
After index hospitalization, no.	79 (34%)	45 (35%)	34 (32%)	0.67
30-day Readmission after index hospitalization, no.	61 (28%)	35 (28%)	26 (27%)	0.88
Readmissions per patient-year	1.9 ± 2.0	1.9 ± 2.0	1.8 ± 2.0	0.58
Prolonged readmissions per patient-yeara	0.50 ± 0.80	0.47 ± 0.63	0.54 ± 0.96	0.95
Proportion of Post-HM3 Days in hospital, %	11 (7–22)	10 (7–17)	13 (7–26)	0.069

Abbreviations: HRAE, hemocompatibility-related adverse event; LOS, length of hospital stay; no., number of; OR, operating room; RHF, right heart failure; RVAD, right ventricular assist device.

^aProlonged readmission is defined as being hospitalized for ≥7 days.

Additional clinical outcomes following HM3 implant are detailed in Table 2. When compared to the TL patients, the NTL group had a greater proportion of patients who required temporary right ventricular assist device (RVAD) support for early acute right heart failure (28% vs 10%, p = 0.005); suffered an initial adverse event during their implant hospitalization (32% vs 15%, p = 0.0026); and a trend to a longer index hospital stay (44 vs 39 days, p = 0.052). The NTL group had a similar rate of readmissions per patient-year to the TL group (1.8 vs 1.9, p = 0.58).

NTL status, exclusion criteria, and adverse events

We further analyzed the association between NTL and the primary endpoint. On univariable analysis, age (every 1-year increase), “destination-therapy” status, HM3 risk score ≥2.97, pre-HM3 length of stay (every 1-day increase), and serum sodium (every 1 mmol/liter increase) were associated with the primary endpoint (Table 3). No collinerarity between the variables was observed.

Table 3.

Cox Regression Analysis of Clinical Risk Factors for Primary Outcome of Adverse Events

	Univariable analysis		Multivariable analysis
	HR 95% CI	p-value	HR 95% CI	p-value
Nontrial like	1.56 (1.1-2.2)	0.01	1.4 (1.01-2.0)	0.043
Age, per year	1.01 (1.0-1.02)	0.036	1.0 (0.99-1.02)	0.20
Female sex	0.99 (0.66-1.4)	0.96
Black or Hispanic race	1.07 (0.75-1.5)	0.70
Ischemic cardiomyopathy	1.1 (0.77-1.5)	0.57
History of atrial fibrillation	1.1 (0.81-1.6)	0.45
History of stroke	1.4 (0.72-1.8)	0.56
Prior sternotomy	1.1 (0.76-1.6)	0.53
Left ventricular ejection fraction, per %	1.02 (0.99-1.05)	0.11
Destination therapy	1.7 (1.1-2.7)	0.01	1.4 (0.92-2.2)	0.10
HM3 risk score ≥2.97	1.7 (2.2-2.4)	0.003	1.4 (0.87-2.1)	0.17
LV EDD <5.5 cm	1.6 (0.94-2.8)	0.081
INTERMACS profile	0.89 (0.72-1.09)	0.28
INTERMACS profile 1	1.4 (0.72-2.8)	0.29
INTERMACS profile 2	1.1 (0.75-1.6)	0.62
INTERMACS profile 3-4	ref	n/a
Pre-HM3 hemodynamic support
Inotrope(s)	1.2 (0.76-1.9)	0.39
IABP	1.1 (0.80-1.6)	0.44
Pre-HM3 LOS, per day	1.01 (1.0-1.03)	0.037	1.02 (1.0-1.03)	0.037
Body mass index, per kg/m2	1.0 (0.97-1.02)	0.88
Body-surface area, per m2	0.77 (0.45-1.3)	0.36
RAP, per mm Hg	1.0 (0.97-1.03)	0.93
PCWP, per mm Hg	0.98 (0.96-1.0)	0.19
Cardiac index, per liter/min/m2	1.1 (0.80-1.6)	0.45
PVR, per Wood units	1.0 (0.92-1.09)	0.88
Serum sodium, per mmol/liter	0.95 (0.91-0.99)	0.022	0.97 (0.93-1.02)	0.35
EGFR, per ml/min/1.73 m2	0.99 (0.99-1.00)	0.59
Alanine aminotransferase, per U/liter	1.0 (0.99-1.0)	0.97
Aspartate aminotransferase, per U/liter	1.0 (0.99-1.0)	0.38

Abbreviations: CI, confidence interval; EGFR, estimated glomerular filtration rate; IABP, intra-aortic balloon pump; LOS, length of hospital stay; LV EDD, left ventricular end-diastolic diameter; PCWP, pulmonary capillary wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure.

Covariates with p < 0.05 in univariate analysis were included in the multivariate analysis model to ascertain significance.

After adjusting for these covariates, the NTL group was associated with the primary endpoint (adjusted HR [95%CI]: 1.5[1.01-2.0], p = 0.043, Figure 3 and Table 3). We did not observe a center effect on this association. Furthermore, NTL status was associated with an increased risk of all-cause mortality at 1 year after HM3 implantation (adjusted HR [95%CI]: 3.3[1.1-8.7], p = 0.013), but the associations between NTL and the individual nonfatal outcomes were statistically insignificant (Figure 3, Supplemental Table 1). There was an association between being NTL and the secondary outcome of all-cause mortality or stroke (adjusted HR [95%CI]: 2.2[1.07-4.5], p = 0.030) and the combined LVAD-related outcome LVADI or HRAE (adjusted HR [95%CI]: 1.5[1.04-2.3], p = 0.029, Figure 3).

Figure 3.

Forest plot showing the adjusted regression analysis of the association between nontrial like group and the primary and secondary outcomes. The models were adjusted for age (every 1-year increase), “destination-therapy” status, HM3 risk score ≥2.97, pre-HM3 length of stay (every 1-day increase), and serum sodium (every 1 mmol/liter increase). GIB, gastrointestinal bleeding; HRAE, hemocompatibility-related adverse event; LVADI, left ventricular assist device related infection.

We observed that meeting 2 or more exclusion criteria was associated with meeting the primary endpoint (adjusted HR [95%CI]: 1.8[1.06-3.1], p = 0.028, Supplemental Figure 4). Patients who met only a single exclusion criterion had a statistically insignificant association with the primary endpoint (adjusted HR [95%CI]: 1.3[0.90-1.9], p = 0.15, Supplemental Figure 4). The influence of each individual exclusion criteria on the primary endpoint was heterogenous. The exclusion criterion of low prealbumin/albumin was not significantly associated with the primary endpoint. However, the primary endpoint was associated with the use of pre-HM3 tMCS and ≥1 criteria (adjusted HR [95%CI]: 2.1[1.03-4.6], p = 0.041, Supplemental Figure 4), especially with the use of pre-HM3 VA-ECMO (adjusted HR [95%CI]: 1.9[1.08-3.4], p = 0.026).

A secondary analysis in patients without VA-ECMO support was done to evaluate further the relationship between NTL and the primary outcome (Supplemental Figure 5A-E). On Kaplan-Meier analysis, the NTL group had a lower event-free survival (38% vs 48%, p = 0.042, Supplemental Figure 5A) and lower overall survival (86% vs 95%, p = 0.023, Supplemental Figure 5B) when compared to TL patients. There were 2.6 ± 4.6 events per patient-year in the NTL group vs 2.0 ± 6.4 events per patient-year in the TL group (p = 0.034). When adjusted for baseline covariates, there was a statistically insignificant association between NTL patients without VA-ECMO support and the primary outcome (adjusted HR [95%CI]: 1.3[0.90-1.9], p = 0.15) and with all-cause mortality (adjusted HR [95%CI]: 2.6[0.95-7.1], p = 0.064).

Clinical course after uncomplicated index hospitalization

We examined patients with an uncomplicated index hospitalization, defined as having no post-HM3 adverse events during the implant hospitalization. Of the 231 patients, 178 (77%) patients had an uncomplicated index hospitalization. The TL group had a greater proportion of such patients than the NTL group (85% vs 67%, p = 0.0026). The 1-year event-free survival was similar among the 2 groups (56% vs 52%, respectively, p = 0.39; Figure 4A-E) as well as a similar overall survival (98% vs 94%, respectively, p = 0.17; Figure 4B).

Figure 4.

Outcomes of nontrial and trial-like patients with an uncomplicated index hospitalization, defined as no post-HM3 adverse events during the implant hospital stay. (A) Freedom from composite endpoint of death, heart failure hospitalization (HFH), hemocompatibility-related adverse events (HRAE), and left ventricular assist device-related infection (LVADI), (B) overall survival, (C) freedom from HRAE, (D) freedom from LVADI, and (E) freedom from HFH at 1 year on commercial HM3 support.

Discussion

In this multicenter cohort of patients on commercial HM3 support, 45% of patients were classified as NTL. The most common exclusion criterion was low prealbumin/albumin levels, followed by pre-HM3 tMCS. Compared to the TL group, the NTL cohort had a lower 1-year survival free of HRAE, HFH, and LVADI. Having 2 or more exclusion criteria as well as the need for VA-ECMO support was significantly associated with the primary endpoint. NTL patients who survived their index hospitalization without post-HM3 complications had a similar event-free survival as TL patients.

This real-world representation of commercial HM3 implants demonstrated a higher risk of adverse events in those who did not meet traditional clinical trial eligibility. We chose a composite endpoint to capture the total burden of both fatal and nonfatal adverse events that, cumulatively, could affect quality of life following LVAD implantation.¹⁵ We also observed a significant difference in 1-year all-cause mortality between the NTL and TL groups. Compared to the TL group, NTL patients suffered more adverse events including early acute right heart failure requiring temporary RVAD support. Even though we did not find an association between NTL patients and the individual nonfatal adverse events, NTL patients had an increased risk of adverse LVAD-related outcomes including both LVADI and HRAE.

Our findings were consistent with previous registry-based data reporting an increased risk of all-cause mortality among trial-ineligible patients on HM3⁸ and other LVAD devices.6, 7 This study is unique because of the detailed analysis of both post-implant morbidity and mortality including the use of temporary RVAD support post-HM3, hospital length of stay, and rehospitalization rates. Approximately 90% of TL and 77% of NTL patients in our cohort were alive without any stroke at 1 year following device implantation, while 84% of the HM3 cohort in MOMENTUM 3 met its primary outcome. Similarly, the clinical trial had an 86.6% overall survival at 1-year post-implant while our NTL cohort had an 84% overall survival.⁴

In our study, all-cause mortality was the primary driver of the composite primary outcome. This observation may be a cumulative effect from the burden of nonfatal adverse events. We found that even though having low prealbumin/albumin levels was prevalent in our NTL patient, this exclusion criterion was not associated with the primary composite outcome. Having multiple exclusion criteria did appear to identify a patient population at increased risk of adverse events. VA-ECMO support and the presence of other exclusion criteria in patients on tMCS were associated with a higher risk of adverse events. In patients without VA-ECMO support, the NTL group had a lower event-free survival, but there was a nonsignificant association between NTL and all-cause mortality.

Rather than a singular factor, the interplay between these risk factors (i.e., the exclusion criteria) might synergistically increase the perioperative complexity of NTL patients. The NTL group also consisted of a significantly higher proportion of INTERMACS profile 1 and 2 patients. These patients would have a greater degree of illness severity and thus, additional risk of post-operative mortality.¹⁶ These factors might contribute to greater perioperative risk and early acute right heart failure needing temporary RVAD support. Previous studies have shown that the use of temporary RVAD support is associated with greater short-term mortality.11, 17, 18 Because of their severity of illness, NTL patients expectedly might have a higher postoperative clinical acuity than the TL patients which resulted in a complicated index HM3 implant hospitalization. These findings collectively could explain the increased cumulative adverse event burden even in those without VA-ECMO support.

Approximately 70% of the NTL patients had an uncomplicated index hospitalization defined as having no post-HM3 adverse events during their implant hospitalization. In this subgroup of NTL patients, adverse outcomes were similar to that of TL patients. Furthermore, the incidence of initial adverse event following index hospitalization did not differ significantly between the 2 cohorts with uncomplicated hospital stay. The length of stay for implant hospitalization and the frequency of rehospitalizations were also similar between the 2 groups. These findings suggest that a subgroup of NTL patients might have adequate outcomes when compared to TL patients. Additional research is needed to characterize this patient population.

Clinical implications

Our study provides insights into the adverse outcomes in NTL patients who are considering HM3 therapy. This shared decision-making process is especially relevant for “destination therapy” patients who may remain on device support for a prolonged period.5, 19, 20 The HM3 risk score has been developed to estimate mortality risk, in which a score ≥2.97 designates a high-risk cohort.¹⁴ After adjusting for the risk score, the NTL group remains associated with the primary outcome. In addition, our results suggest that when NTL patients achieve clinical stability after index hospital discharge, the effects of the exclusion criteria (i.e., “increased risk profile”) on the health care burden of HM3 patients may be less relevant. Reducing the time spent in the hospital21, 22 and the idea of “thrivival” on LVAD support²³ are crucial aspects of improving health status following LVAD implantation. Our observations potentially reinforce the clinical success of patients who fall outside of clinical trial criteria, indicating that HM3 device therapy is viable for patients classified as NTL.

Limitations

Our study has several key limitations. As a small cohort from 2 high-volume institutions, our findings may not be generalizable because of the wide range of site-specific variability in patient selection and postdevice management.²⁴ Furthermore, our study may be underpowered to detect differences in nonfatal adverse events. Future research is needed to study NTL patients who did not require VA-ECMO support. We do not have social determinants of health data, patient-reported outcome measurements, and objective performance measures or markers of frailty and deconditioning. These outcomes may differentiate NTL from TL patients, and represent important factors for patients who are considering LVAD therapy.

Conclusions

In this multicenter analysis of patients on commercial HM3 support, patients outside of traditional clinical trial eligibility had a significantly lower 1-year event-free survival. However, NTL patients with an uncomplicated index hospitalization had a similar clinical course as TL patients. This subgroup may have nominally increased risk of post-LVAD adverse events. Shared decision-making between providers and patients should include HM3 therapy as a viable treatment option in those with an NTL profile.

Author contributions

L.C.: Study design, acquisition of data, interpreting data, drafting and revising manuscript. C.G., N.G., L.K., R.P., K.H., K.S.: Acquisition of data, interpreting data, drafting and revising manuscript. A.S., N.S.: Study design, acquisition of data, interpreting data, drafting and revising manuscript. G.P.M., S.P., J.M., M.D.C., W.G.C., P.S.P., A.J.T.: Interpreting data, drafting and revising manuscript. N.K., V.Q.C.: Study design, analyzing data, interpreting data, drafting and revising manuscript. All authors approved the final version of the manuscript and agreed to being accountable for its content.

Appendix A. Supplementary material

Supplementary material

.