Nationwide Outcomes and Predictors of Mortality After Left Ventricular Assist Device Implantation in Korea

Abstract

Background

Durable left ventricular assist device (LVAD) therapy was introduced in Korea in 2018 within a transplant environment characterized by frequent preoperative extracorporeal membrane oxygenation (ECMO) support. We evaluated nationwide outcomes and predictors of mortality after LVAD implantation.

Methods and Results

This registry-based cohort study included all durable LVAD implantations performed in Korea between September 2018 and December 2023 using the Korean Mechanically Assisted Circulatory Support registry. The primary outcome was all-cause mortality. Survival was assessed with Kaplan–Meier and competing risk analyses, and predictors were identified using multivariable Cox models. Among 465 patients (median age 62 years; 78% male), 367 underwent bridge-to-transplantation (BTT) and 98 destination therapy (DT). Preoperative ECMO was used in 27.5%. One-year survival was 83.1% overall and higher in BTT than DT (87.4% vs. 69.6%; P<0.001). At 1 year, 59.6% remained on LVAD support, 24.8% underwent transplantation, and 15.6% died. Arrhythmia, bleeding, and infection were the most common complications. In multivariable analysis, age 60–69 years (hazard ratio [HR] 3.80; 95% confidence interval [CI] 1.73–8.31) and preoperative ECMO (HR 2.85; 95% CI 1.76–4.60) independently predicted mortality, whereas treatment strategy did not.

Conclusions

In this high-acuity population with frequent preoperative ECMO support, nationwide LVAD outcomes were comparable to international benchmarks. Mortality was more closely associated with preoperative clinical status than with treatment strategy.

Central Figure.

Left ventricular assist devices (LVADs) are increasingly used as bridge-to-transplantation (BTT) or destination therapy (DT) in patients with advanced heart failure and have demonstrated improved survival and quality of life compared with medical therapy.^1,2 Contemporary international registries, including Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS), International registry for Mechanically Assisted Circulatory Support (IMACS), and Japanese registry for Mechanical Assisted Circulatory Support (J-MACS), report favorable outcomes with current-generation devices, although patient profiles and preoperative acuity vary substantially across regions.^3–7

In Korea, durable LVAD therapy has been available under the national insurance system since 2018 for patients with end-stage heart failure refractory to guideline-directed therapy or temporary mechanical support. Patient eligibility is reviewed through a national committee, and clinical data are prospectively collected in the Korean registry for Mechanically Assisted Circulatory Support (KMACS).

A notable characteristic of the Korean transplant environment is the high prevalence of preoperative extracorporeal membrane oxygenation (ECMO) support, reflecting the frequent use of temporary mechanical circulatory support in patients with severe cardiogenic shock or hemodynamic instability.⁸ Nationwide data indicate that 35% of heart transplant recipients were bridged with ECMO, with a median waiting time of 16 days among ECMO-supported patients.⁹ Prior reports also suggest a substantial proportion of LVAD implantations are performed in patients previously supported with ECMO.¹⁰ Such a high-acuity, ECMO-predominant setting may influence outcomes following durable LVAD implantation.

Using nationally representative data from KMACS, this study aimed to evaluate outcomes and predictors of mortality after durable LVAD implantation in Korea, providing insight into LVAD performance within a high-acuity allocation environment.

Methods

Study Design and Population

This was a nationwide, registry-based cohort study including all durable LVAD implantations performed in Korea between September 2018 and December 2023, as recorded in the KMACS (Figure 1). Health insurance claims and national death registration records from September 2018 to June 2024 were utilized to verify preoperative medical histories, heart transplantation status and date, as well as mortality status and date. All patients meeting national reimbursement criteria for durable LVAD implantation during the study period were included.¹¹

Figure 1.

Overview of the Korean Mechanically Assisted Circulatory Support (KMACS) registry data collection framework. The registry prospectively captures baseline, perioperative, and longitudinal follow-up data through linkage with insurance claims and death registries.

LVAD is reimbursed in Korea for patients with irreversible end-stage heart failure who remain severely symptomatic despite optimal medical therapy or temporary mechanical circulatory support (e.g., intra-aortic balloon pump [IABP] or ECMO). Eligibility criteria include severe left ventricular dysfunction with reduced exercise capacity, inotrope dependency, or progressive end-organ dysfunction attributable to heart failure.¹¹ In the Korean heart transplantation allocation system, patients receiving ECMO support, those hospitalized with LVAD support, or those experiencing severe LVAD-related complications are assigned higher urgency status for transplantation.¹² In this context, ECMO primarily refers to veno-arterial ECMO (VA-ECMO), which is commonly used as bridge therapy for severe cardiac or cardiopulmonary failure, including patients awaiting heart transplantation.

In Korea, LVAD implantation is permitted only at institutions that have performed at least 3 heart transplants within the preceding 24 months.¹² The study encompassed 3 durable LVAD types: axial-flow pumps (HeartMate 2; Abbott Laboratories), centrifugal hybrid levitation (HeartWare Ventricular Assist Device [HVAD]; Medtronic, Inc.), and centrifugal full magnetic levitation (HeartMate 3 [HM3]; Abbott Laboratories). For multivariable analyses, devices were categorized as HM3 vs. Others (HM 2 and HVAD).

The KMACS registry was established by the Health Insurance Review and Assessment Service (HIRA) to prospectively monitor patients with advanced heart failure receiving durable LVADs under the national insurance program. Patients with paracorporeal pulsatile VADs were excluded.

A total of 465 LVAD implantations (367 BTT, 98 DT) performed at 20 medical institutions between September 2018 and December 2023 were included in the final analysis. Mortality data were collected through to June 30, 2024. Time zero was defined as the date of LVAD implantation, and follow up continued until death or June 30, 2024, whichever occurred first. The mean follow-up duration was 767 days (standard deviation 553 days).

Data Source and Variables

The KMACS registry prospectively collects comprehensive preoperative clinical data, including demographic characteristics, primary diagnosis, New York Heart Association (NYHA) functional class, INTERMACS profile, prior cardiac interventions, laboratory findings, echocardiographic parameters, and hemodynamic data. In addition to these patient-level variables, hospital-level characteristics were incorporated into the multivariable analyses.

To ensure data accuracy and minimize underreporting, registry data were cross-matched with health insurance claims and national death records. In cases of discrepancies, additional validation was conducted using institutional medical records and clinical reports.

The primary outcome was all-cause mortality after LVAD implantation. Secondary outcomes included postoperative complications (e.g., arrhythmia, bleeding, and infection) assessed at discharge and during follow up. Complication data were based on registry-reported events and were not fully standardized according to IMACS criteria.

Statistical Analysis

Continuous variables are expressed as mean ± standard deviation or median (interquartile range) and compared using t-tests or Mann–Whitney U tests after assessment of normality with the Shapiro–Wilk test. Categorical variables are presented as frequencies (%) and compared using chi-square or Fisher’s exact tests.

Survival was estimated using the Kaplan–Meier method and compared with the log-rank test, censoring patients at the time of heart transplantation. However, because heart transplantation represents a competing event rather than simple censoring, and such competing risks may not be adequately accounted for in Kaplan–Meier analysis, additional competing risk analysis was performed using the Fine and Gray model. The Fine and Gray method estimates the cumulative incidence function while accounting for competing risks.¹³

Univariate Cox proportional hazards regression was first performed to identify candidate variables for multivariable analysis (P<0.2; Supplementary Table 1). Variables with P<0.2 in the univariate analysis were subsequently included in the multivariable Cox proportional hazards model. Changes in laboratory and echocardiographic parameters over follow up were assessed using ANOVA or the Kruskal–Wallis test, as appropriate.

All statistical analyses were performed using R (version 2024.9.1.394). This study was reviewed and approved by the HIRA Research Committee (HIRAIRB-2024051-001), and the procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation and the Declaration of Helsinki.

Results

Baseline Characteristics of Patients Undergoing LVAD Implantation

Between September 2018 and December 2023, 465 LVAD implantations were performed at 20 institutions (367 BTT, 98 DT), with mortality data collected through to June 2024 (mean follow up 767±553 days). Overall, 78.3% were male, and the median age was 62 years, with BTT patients significantly younger than DT patients (59 vs. 74 years; P<0.001). Nearly 90% of implantations occurred at tertiary centers in metropolitan regions.

DT patients had greater comorbidity (CCI ≥8: 36.7% vs. 20.7%; P<0.001; Table 1). All patients received intravenous inotropes preoperatively, and 27.5% required ECMO support, markedly higher than in international registries (~5–6%). Most were NYHA III–IV (99.5%), and none were INTERMACS 5–7; 9.0% were profile 1 and 25.4% profile 2, indicating advanced disease severity.

Table 1.

Baseline Characteristics of Patients Undergoing LVAD Implantation

Characteristic	Overall (n=465)		BTT (n=367)		DT (n=98)		P value
	Data (n)	Value	Data (n)	Value	Data (n)	Value
Demographic
Sex, male	465	364 (78.3)	367	298 (81.2)	98	66 (67.3)	0.003
Age (years)
<50	465	105 (22.6)	367	102 (27.8)	98	3 (3.1)	<0.001
50–59		97 (20.9)		95 (25.9)		2 (2.0)
60–69		150 (32.3)		145 (39.5)		5 (5.1)
≥70		113 (24.3)		25 (6.8)		88 (89.8)
Health insurance	465	426 (91.6)	367	333 (90.7)	98	93 (94.9)	0.187
Physical characteristic
BMI (kg/m2)
<18.5	465	47 (10.1)	367	32 (8.7)	98	15 (15.3)	0.015
18.5–22.9		226 (48.6)		171 (46.6)		55 (56.1)
23.0–24.9		78 (16.8)		64 (17.4)		14 (14.3)
≥25.0		114 (24.5)		100 (27.2)		14 (14.3)
Blood type
A	464	163 (35.1)	366	130 (35.4)	98	33 (33.7)	0.007
B		120 (25.8)		98 (26.7)		22 (22.4)
O		149 (32.0)		121 (33.0)		28 (28.6)
AB		32 (6.9)		17 (4.6)		15 (15.3)
Diagnosis
Dilated/congestive cardiomyopathy	465	238 (51.2)	367	186 (50.7)	98	52 (53.1)	0.098
Other cardiomyopathy		19 (4.1)		18 (4.9)		1 (1.0)
Chronic ischemic heart disease		134 (28.8)		99 (27.0)		35 (35.7)
Congestive heart failure		45 (9.7)		38 (10.4)		7 (7.1)
Other		29 (6.2)		26 (7.1)		3 (3.1)
Medical history
CCI
0–3	465	128 (27.5)	367	115 (31.3)	98	13 (13.3)	<0.001
4–5		127 (27.3)		102 (27.8)		25 (25.5)
6–7		98 (21.1)		74 (20.2)		24 (24.5)
≥8		112 (24.1)		76 (20.7)		36 (36.7)
IV inotropic agents	465	465 (100.0)	367	367 (100.0)	98	98 (100.0)	–
ECMO	465	128 (27.5)	367	105 (28.6)	98	23 (23.5)	0.311
IABP	465	6 (1.3)	367	4 (1.1)	98	2 (2.0)	0.611
Ventilator	465	68 (14.6)	367	57 (15.5)	98	11 (11.2)	0.284
Severity of illness
NYHA classification
Class I	465	0 (0.0)	367	0 (0.0)	98	0 (0.0)	0.188
Class II		2 (0.4)		2 (0.5)		0 (0.0)
Class III		109 (23.4)		92 (25.1)		17 (17.3)
Class IV		354 (76.1)		273 (74.4)		81 (82.7)
INTERMACS profile
Level 1	465	42 (9.0)	367	36 (9.8)	98	6 (6.1)	0.193
Level 2		118 (25.4)		99 (27.0)		19 (19.4)
Level 3		266 (57.2)		201 (54.8)		65 (66.3)
Level 4		39 (8.4)		31 (8.4)		8 (8.2)
Laboratory values
Creatinine (mg/dL)	465	1.1 [0.9–1.5]	367	1.1 [0.9–1.5]	98	1.3 [1.0–1.6]	0.013
AST (IU/L)	464	26 [20–40]	366	26 [20–39]	98	26 [19–43]	0.801
ALT (IU/L)	464	23 [15–38]	366	23 [15–38]	98	22 [14–36]	0.499
BNP (pg/mL)	121	1,864.9 [1,026.2–3,768.0]	99	1,814.0 [1,009.0–3,573.0]	22	2,806.0 [1,094.4–4,204.4]	0.586
NT-proBNP (pg/mL)	387	5,991.0 [2,753.0–13,130.0]	296	5,507.2 [2,431.0–11,927.7]	91	8,223.0 [4,824.0–18,460.0]	<0.001
Hemoglobin (g/dL)	465	11.2 [9.6–13.3]	367	11.4 [9.7–13.6]	98	10.6 [9.4–12.2]	0.013
Echocardiography
LVEF (%)	463	21.7 [17.0–25.1]	365	21.3 [17.0–26.0]	98	22.0 [18.0–25.0]	0.532
LVESD (cm)	444	5.9 [5.3–6.8]	350	6.0 [5.3–6.9]	94	5.8 [5.3–6.5]	0.087
LVEDD (cm)	451	6.8 [6.1–7.5]	357	6.8 [6.1–7.7]	94	6.7 [6.2–7.2]	0.169
Cardiac catheterization
PASP (mmHg)	391	50 [36–63]	307	51 [38–64]	84	44.5 [31.0–58.0]	0.018
PADP (mmHg)	392	24 [18–31]	308	25 [19–32]	84	21.5 [15.0–27.0]	<0.001
PCWP (mmHg)	388	26 [20–31]	306	26 [20–32]	82	24 [17–29]	0.030
PVR (Wood units)	383	3.0 [2.0–4.9]	300	3.1 [2.0–5.0]	83	2.9 [1.9–4.8]	0.475
RAP (mmHg)	388	10 [6–14]	305	10 [7–15]	83	9 [5–14]	0.156
Cardiac index (L/min/m2)	389	1.9 [1.6–2.2]	306	1.9 [1.6–2.2]	83	1.8 [1.5–2.3]	0.385
LVAD device
HM 3	465	326 (70.1)	367	259 (70.6)	98	67 (68.4)	0.672
Medical facility characteristics
General	465	48 (10.3)	367	38 (10.4)	98	10 (10.2)	0.965
Non-metropolitan region	465	48 (10.3)	367	40 (10.9)	98	8 (8.2)	0.429

Unless indicated otherwise, data are presented as n (%), or median [Q1–Q3]. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BNP, B-type natriuretic peptide; BTT, bridge to transplantation; CCI, Charlson Comorbidity Index; DT, destination therapy; ECMO, extracorporeal membrane oxygenation; HM 3, HeartMate 3; IABP, intra-aortic balloon pump; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; LVAD, left ventricular assist device; LVEDD, left ventricular diastolic dimension; LVEF, left ventricular ejection fraction; LVESD, left ventricular end systolic dimension; NT-proBNP, N-terminal pro B-type natriuretic peptide; NYHA, New York Heart Association; PADP, pulmonary artery diastolic pressure; PASP, pulmonary artery systolic pressure; PCWP, pulmonary capillary wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure.

Preoperative laboratories showed higher creatinine (1.3 vs. 1.1 mg/dL; P=0.013), lower hemoglobin (10.6 vs. 11.4 g/dL; P=0.013), and higher N-terminal pro B-type natriuretic peptide (NT-proBNP; 8,223 vs. 5,507 pg/mL; P<0.001) in DT vs. BTT. BTT patients had higher pulmonary artery systolic pressure (51 vs. 44.5 mmHg; P=0.018) and pulmonary capillary wedge pressure (26 vs. 24 mmHg; P=0.030). HM 3 accounted for 70.1% of implants, fully replacing HM 2 and HVAD after 2020. The AB blood type was more frequent in DT (15.3% vs. 4.6%; P=0.007), suggesting allocation differences.

LVAD Implantation Outcomes

Overall survival was 83.1% at 1 year, 77.2% at 2 years, and 69.9% at 3 years (Figure 2). One-year survival was higher in BTT than DT (87.4% vs. 69.6%; log-rank P<0.001). Competing risk analysis showed that at 1 year, 59.6% remained on LVAD, 24.8% were transplanted, and 15.6% died; among BTT patients, corresponding rates were 57.4%, 31.0%, and 11.6% (Figure 3). LVAD support led to marked improvements in laboratory and echocardiographic parameters, including decreases in NT-proBNP and LV dimensions and normalization of liver enzymes (all P<0.001; Supplementary Table 2).

Figure 2.

Kaplan–Meier survival estimates for patients who received left ventricular assist device (LVAD) between September 2018 and December 2023. (A) Overall survival for the entire cohort. (B) Stratified survival according to device strategy (bridge-to-transplantation [BTT] vs. destination therapy [DT]).

Figure 3.

Competing risk analysis for post-left ventricular assist device (LVAD) outcomes. (A) All patients. (B) Subgroup of patients receiving LVAD as bridge-to-transplantation. Events were classified as continued LVAD support, heart transplantation, or death.

Complications and Predictors of Mortality After LVAD Implantation

At discharge, arrhythmias were the most frequent complication (18.0%), followed by bleeding (15.4%) and infection (14.9%), with higher infection rates in DT (23.5% vs. 12.6%; P=0.008; Supplementary Table 2).

In univariate Cox proportional hazards analysis, LVAD strategy, sex, age, insurance type, cardiomyopathy diagnosis, CCI ≤3, ECMO, IABP, previous LVAD implantation type, INTERMACS profile 4, aspartate aminotransferase (AST), hemoglobin, and LVEF showed P<0.2 and were selected as candidate variables for multivariable analysis (Supplementary Table 1). These variables were subsequently included in the multivariable Cox proportional hazards model.

In multivariable analysis, age 60–69 years (hazard ratio [HR] 3.80; 95% confidence interval [CI] 1.73–8.31) and preoperative ECMO (HR 2.85; 95% CI 1.76–4.60) independently predicted mortality. HM3 was associated with lower mortality risk compared with other devices (HR 0.63; 95% CI 0.41–0.98), while DT strategy was not significant after adjustment (Table 2).

Table 2.

Multivariable Cox Proportional Hazards Model for Factors Associated With All-Cause Mortality After LVAD Implantation

Variable	HR	95% CI	P value
LVAD strategy
LVAD strategy (ref. BTT)	2.083	0.964–4.502	0.062
Demographic
Sex (ref. men)	1.346	0.833–2.175	0.226
Age (ref. <50 years)
50–59 years	1.379	0.509–3.740	0.528
60–69 years	3.795	1.733–8.310	0.001
≥70 years	2.345	0.872–6.308	0.091
Insurance (ref. medical benefit)	1.122	0.484–2.602	0.789
Physical characteristic
Dilated/congestive cardiomyopathy	1.135	0.723–1.780	0.582
Medical history
CCI (0–3 group)	0.713	0.417–1.221	0.218
ECMO	2.849	1.764–4.602	<0.001
IABP	1.149	0.275–4.809	0.849
LVAD device
HM 3	0.633	0.407–0.984	0.042
Severity of illness
INTERMACS profile 4	0.507	0.157–1.637	0.256
Laboratory value
AST (IU/L)	1.000	0.999–1.002	0.603
Hemoglobin (g/dL)	0.971	0.893–1.056	0.495
LVEF (%)	0.784	0.559–1.100	0.159

CI, confidence interval; HR, hazard ratio; HVAD, HeartWare ventricular assist devices. Other abbreviations as in Table 1.

To further characterize differences according to preoperative ECMO support, baseline characteristics stratified by ECMO status are presented in Supplementary Table 3. Patients requiring ECMO demonstrated greater clinical severity, including higher proportions of INTERMACS profile 1–2, greater ventilator dependence, and worse laboratory markers of end-organ dysfunction compared with non-ECMO patients.

Discussion

Key Findings

This nationwide, registry-based analysis using the KMACS database represents the first comprehensive evaluation of durable LVAD therapy in Korea, encompassing clinical characteristics, survival outcomes, and predictors of mortality. Previous studies in Korea have been limited to single-center experiences, while most regional data from Asia originate from Japan.^10,14,15 We found favorable survival outcomes, with survival rates of 83.1%, 77.2%, and 69.9% at 1, 2, and 3 years, respectively, despite a remarkably high prevalence of preoperative ECMO support (27.5%), which is substantially higher than the 5–6% reported in international registries.^16,17 Unadjusted survival was significantly higher in BTT than DT patients, but after adjustment, mortality risk was not significantly different between groups, suggesting that clinical outcomes are primarily driven by patients’ preoperative clinical status and comorbidities rather than treatment intent itself. This interpretation is consistent with recent Japanese experience showing that patients receiving LVAD as DT had slightly lower 1-year survival than those receiving BTT, but were approximately 10 years older on average, underscoring the importance of patient selection when comparing treatment strategies.¹⁸ Advanced age (60–69 years) and preoperative ECMO support were independent predictors of mortality, whereas LVAD implantation led to significant improvements in laboratory and echocardiographic markers of cardiac function.

Survival Outcomes and Clinical Context in Korea

Despite the relatively recent introduction and lower procedural volume of LVAD therapy in Korea, the 1-year survival rate of 83.1% was comparable with major international registries, including INTERMACS (82.7%), IMACS (75–86%), and the Multicenter Italian Study on Radial Mechanically Assisted Circulatory Support (MIRAMACS; 76.3%).^16,17,19 The higher survival reported in J-MACS (92%) may be related to differences in patient characteristics, treatment strategies, and health care systems.^14–16

A unique feature of the Korean system is its ECMO-oriented allocation and reimbursement structure. National data indicate that 35% of heart transplant recipients are bridged with ECMO, with a median waiting time of 16 days, and 27.5% of LVAD recipients were supported with ECMO prior to implantation.⁹ This rate far exceeds those in INTERMACS (5.5%) and IMACS (6%), reflecting both higher clinical acuity and systemic prioritization of ECMO.^14,17–22

Although overall outcomes in this ECMO-dominant environment were not inferior to international benchmarks, earlier LVAD implantation – before progression to cardiogenic shock or cardiac arrest – could further improve outcomes by reducing end-organ injury and perioperative risk. Such a shift would require adjustments in patient selection and allocation strategies. This consideration is supported by recent Japanese data showing that, among LVAD-supported heart transplant candidates, right heart failure, infection-related stroke, and sudden stroke were significant risk factors for waiting-list mortality, suggesting that high-risk patients may require more tailored prioritization and management during the bridging period.²³ Notably, catheter-based temporary LVAD systems (e.g., Impella) were not available during the study period; their future introduction may alter bridging patterns and reduce reliance on ECMO.

The 1-year survival rate of 83.1% is comparable with major international registries, such as INTERMACS (82.7%) and IMACS (75–86%), which is noteworthy given Korea’s ECMO-oriented allocation and reimbursement structure. In Korea, 35% of heart transplant recipients are bridged with ECMO, and national insurance criteria often necessitate high-acuity bridging. Our results demonstrate that the Korean medical system provides effective perioperative management for these high-acuity patients. However, preoperative ECMO remained a potent independent predictor of mortality (HR 2.85; 95% CI 1.76–4.60), likely reflecting a state of profound cardiogenic shock and subsequent end-organ vulnerability. Earlier LVAD implantation – before the need for ECMO – could further optimize survival by reducing perioperative risks. During the study period, catheter-based temporary LVADs like Impella were unavailable in Korea. Their recent introduction may reduce the heavy reliance on ECMO and shift bridging patterns toward more hemodynamically stable conditions before durable LVAD surgery.

Complications and Predictors of Mortality

LVAD implantation was associated with significant improvements in cardiac (BNP, NT-proBNP) and hepatic (AST, alanine aminotransferase) biomarkers, supporting its effectiveness in reversing congestion and improving cardiac output.^24,25 However, postoperative complications remained common, with arrhythmias, bleeding, and infections as the most frequent events. Notably, arrhythmias were the leading complication in KMACS, differing from international registries where infections and bleeding predominate, possibly reflecting regional variations in reporting, management, or patient characteristics.^26,27 Standardized definitions and detailed reporting would enable better international comparisons.

Multivariable analysis identified advanced age (60–69 years) and preoperative ECMO support as independent predictors of mortality, consistent with previous studies.^28–32 HM 3 implantation was associated with a lower mortality risk, in line with global evidence on the improved durability and hemocompatibility of fully magnetically levitated pumps.^7,33–36 Interestingly, patients aged ≥70 years did not exhibit higher adjusted mortality, likely reflecting strict selection and mandatory national review. In Korea, LVAD implantation in elderly patients is subject to careful candidate selection, potentially resulting in the inclusion of older patients with relatively favorable baseline clinical characteristics. The association between ECMO support and mortality underscores the prognostic impact of preoperative hemodynamic instability, emphasizing timely patient optimization before LVAD implantation.²⁷

Study Limitations

This study has several limitations. As with all registry-based analyses, underreporting and incomplete data capture are possible, particularly for cause-specific mortality and late complications. Reporting compliance declined over time, limiting the scope of long-term analyses. Institutional variability in surgical volume and experience may also have influenced outcomes, and despite adjustments for hospital characteristics, residual confounding may persist. Although durable LVAD implantation in Korea is restricted to experienced centers under national reimbursement criteria, inter-institutional differences in operative expertise and perioperative management may still have affected clinical outcomes. Due to the limited number of events and relatively small patient numbers in some centers, institution-level mixed-effects modeling could not be robustly implemented in the present analysis. In addition, although ECMO use was significantly associated with mortality, detailed ECMO-related clinical information, including ECMO duration, indications for initiation, and clinical course (e.g., successful weaning or bridge to LVAD), was not available in the registry. These factors may influence patient severity and prognosis, potentially resulting in residual confounding in the observed association between ECMO use and mortality. Future studies should explicitly incorporate procedural volume, operator experience, and detailed ECMO-related clinical variables to better understand these effects.

Conclusions

Durable LVAD therapy in Korea achieved survival outcomes comparable with international benchmarks despite a substantially higher prevalence of preoperative ECMO support. Mortality was primarily driven by patient severity rather than treatment strategy. Given the scarcity of donor organs, LVAD therapy remains a critical component of advanced heart failure management in Korea. Continued efforts to optimize patient selection, perioperative care, and longitudinal follow up may further improve outcomes.

Disclosures

The authors declare no conflicts of interest related to this study.

IRB Information

This study was reviewed and approved by the Institutional Review Board of the Health Insurance Review and Assessment Service (HIRAIRB-2024051-001). The requirement for informed consent was waived because this study used de-identified administrative data.

Supplementary Files

Supplementary File 1

Supplementary Table 1. Supplementary Table 2. Supplementary Table 3.

Funding Statement

Funding: The article publication charge was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2026-25488413).

Data Availability

The data that support the findings of this study were obtained from HIRA. Because of data use agreements and privacy regulations, the data are not publicly available. Access to the data may be granted by HIRA upon reasonable request and with appropriate approval.