Genetic Determinants of Early Mortality in Hypoplastic Left Heart Syndrome: A Prospective NC-DEFINE Study

Abstract

Background

Survivors of hypoplastic left heart syndrome (HLHS), the most severe form of congenital heart disease, are at high risk for heart failure (HF). HF in early life is a major contributor to mortality in this vulnerable population. However, reliable approaches to identify infants at highest risk for early HF are currently lacking.

Objective

To evaluate whether ultra-rare variants in cardiomyopathy-associated genes are associated with HF risk in HLHS.

Methods

Neonates with HLHS were prospectively enrolled within the first 21 days of life at Duke University Health System (DUHS). Children and adults with HLHS who were older than 21 days were enrolled from DUHS and the University of North Carolina into an ambispective cohort. External HLHS cohorts from Nationwide Children’s Hospital (NCH) and Vanderbilt University Medical Center (VUMC) were evaluated to assess for reproducibility across institutions. Participants underwent genome sequencing, and ultra-rare variants in dilated cardiomyopathy (DCM)-associated genes (minor allele frequency ≤0.01%) were evaluated. The primary outcome was HF, categorized as severe (ventricular assist device implantation, heart transplant, or death) or medically-managed (reduced systemic ventricular ejection fraction and/or HF diagnosis requiring initiation or escalation of HF therapy). Associations between variant status and HF risk were assessed using Cox regression.

Results

Among 35 neonates in the prospective cohort, 7(20.0%) developed severe HF, 10(28.6%) developed medically-managed HF, and 18(51.4%) remained HF free. The presence of a likely pathogenic/pathogenic (LP/P) variant was associated with a marked 9-fold increased risk of severe HF compared with genotype-negative individuals (P=0.02). Most severe HF events occurred within the first month of life (67%). Similar associations between LP/P variants and severe HF were observed in the NCH and VUMC cohorts (11- and 3-fold increased risk, respectively; all P<0.05). Associations were attenuated in the ambispective cohort (all P>0.05), which consisted of individuals significantly older than the prospective cohort (P<0.0001).

Conclusions

This study provides the first prospective evidence linking DCM-associated variants to early-onset HF in HLHS. These findings suggest that genetic variation may contribute to myocardial vulnerability in HLHS, highlighting the potential for genetic screening to enable early risk stratification and guide precision medicine approaches in this high-risk population.

Condensed Abstract

In this cohort of prospectively enrolled neonates with hypoplastic left heart syndrome, variants associated with cardiomyopathy were associated with a 9-fold increased risk of severe heart failure, with most events occurring within the first month of life. These findings suggest that genetic variation may contribute to myocardial vulnerability in HLHS and support further investigation of genetically informed risk stratification approaches in this high-risk population.

Introduction

Congenital heart disease (CHD) affects ~1% of live births and remains a major driver of childhood morbidity and mortality^1,2. Hypoplastic left heart syndrome (HLHS), the most common and severe form of single ventricle disease (SVD), occurs in ~3 per 10,000 live births³. Despite advances in surgical and medical management, first-year mortality remains high, largely due to early-onset heart failure (HF)^4,5. Current risk stratification relies primarily on anatomic and perioperative factors, yet substantial heterogeneity in outcomes persists among infants with apparently similar anatomy and clinical courses⁶. This unexplained variability suggests that important determinants of HF susceptibility in HLHS remain unidentified, including potential genetic modifiers of myocardial resilience.

Growing evidence suggests that variants in dilated cardiomyopathy (DCM)-associated genes contribute to inherited HF risk. A landmark Framingham study estimated that ~20% of HF is heritable⁷, with higher estimates among African American individuals (~33%)⁸. Truncating variants in TTN occur in up to 25% of HF cases compared with 1-2% of the general population⁹. These DCM-associated variants are increasingly recognized as stress-sensitive modifiers of cardiac function, contributing to HF in settings such as peripartum cardiomyopathy and anthracycline exposure^10,11, and may similarly exacerbate vulnerability in HLHS, where early myocardial stress arises from abnormal hemodynamics and cardiopulmonary bypass¹². Supporting this, rare variants in MYH6 have been associated with reduced transplant-free survival in patients with HLHS¹³. Collectively, these findings raise the possibility that DCM-associated variants contribute to HF susceptibility in HLHS and represent an underrecognized source of variability in clinical outcomes.

To investigate whether variants in DCM-associated genes contribute to HF risk in CHD, we established a prospectively enrolled cohort of individuals with CHD aimed at determining the factors that influence mortality in congenital heart disease in North Carolina (NC-DEFINE). Focusing on individuals with HLHS, we tested the hypothesis that this condition represents a physiologic stress state capable of unmasking latent myocardial vulnerability conferred by DCM-associated variants. We identified a robust association between these variants and early-onset HF, particularly within the first month of life, with supporting evidence from two independent cohorts. These findings represent the first prospective evaluation of genetic susceptibility and early HF and mortality in infants with HLHS and support a framework in which genetic risk contributes to the marked heterogeneity of outcomes observed in this vulnerable population.

Methods

Study Design

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Boards at Duke University Health System (DUHS), University of North Carolina (UNC), Nationwide Children’s Hospital (NCH), and Vanderbilt University Medical Center (VUMC). Written informed consent was obtained from all participants or their legal guardians. Eligibility required a diagnosis of HLHS, available genomic data, and absence of a known chromosomal or syndromic diagnosis. This study included 4 independent HLHS cohorts: 1.) A prospective neonatal cohort enrolled at DUHS (≤21 days old at enrollment), 2.) An ambispective cohort of older children and adults with HLHS enrolled at DUHS and UNC with retrospective outcome ascertainment and prospective follow-up (Figure 1), and external retrospective cohorts from 3.) NCH and 4.) VUMC. Additional cohort details are provided in Supplemental Materials.

Figure 1. NC-DEFINE study design.

Infants with hypoplastic left heart syndrome (HLHS) were enrolled within the first 21 days of life (prospective cohort), or thereafter (ambispective cohort), and underwent retrospective and/or prospective clinical follow-up. All participants underwent genome sequencing. Exonic variants in dilated cardiomyopathy (DCM)-associated genes were compared between individuals with severe or medically-managed heart failure (HF) and those who remained HF-free. Curves shown are illustrative schematics representing analytic workflow and are not intended to depict actual study data. Created in https://BioRender.com

Genomic Sequencing

Detailed sequencing protocols are provided in Supplemental Materials. Whole genome and exome sequencing were performed using Illumina platforms and reads were aligned to the GRCh38 reference genome.

Variant Filtering

Variant calling and annotation methods were performed using harmonized pipelines (Supplemental Materials). Analyses were restricted to ultra-rare variants (minor allele frequency <0.0001) in curated DCM-associated genes. Coding and canonical splice-site variants were included, and pathogenicity was assessed using ACMG/AMP criteria¹⁵.

Study Outcomes

Participants were prospectively monitored for HF outcomes, categorized as severe, medically managed, or HF free (Supplemental Figure 1). Severe HF was defined as heart transplant listing, ventricular assist device implantation, or HF-related death associated with ventricular dysfunction. Medically managed HF was defined by persistent ventricular dysfunction on serial echocardiography and/or cardiologist-adjudicated HF requiring escalation of HF-directed therapy. An additional overall HF endpoint was defined as the composite of either medically managed or severe HF. Detailed outcome definitions and adjudication procedures are provided in Supplemental Materials.

Estimated Genetic Ancestry

Genetic ancestry estimation is described in Supplemental Materials. Individuals were assigned to a predominant ancestry group based on ≥55% probability using gnomAD v4.1 reference categories.

Statistics

Categorical variables were compared using Fisher’s exact test. Kaplan-Meier analyses with log-rank testing and Cox proportional hazards models were used to evaluate HF outcomes. Univariable Cox models were used as the primary analysis in the prospective, ambispective and NCH cohorts. Given the exploratory and hypothesis-generating nature of these analyses, formal adjustment for multiple comparisons was not performed. In the larger VUMC cohort, multivariable Cox regression was used as the primary analytic approach. Statistical significance was set at P<0.05. Additional methods are described in Supplemental Materials.

Results

Prospective Observational Cohort of Neonates with HLHS

Detailed cohort demographics are summarized in Table 1. Thirty-five neonates diagnosed with HLHS were prospectively enrolled within the first 21 days of life at DUHS (mean age: 7.2 ± 7.7 days; range: 0-20) and were prospectively followed for 29.8 ± 16.2 months, on average. By the end of the study period, 18 (51.4%) remained HF free, 10 (28.6%) developed medically managed HF, and 7 (20.0%) met the composite severe HF endpoint. Thirty-one individuals (88.6%) underwent Norwood, 1 (2.9%) received hybrid palliation, and 3 did not undergo surgery due to early mortality or need for primary transplantation, all of whom were classified within the severe HF group. HF trajectories relative to stage I palliation are summarized in Supplemental Tables 1–2. Detailed cohort co-morbidities are summarized in Supplemental Table 3. No participants exhibited a consistent congenital extracardiac or syndromic phenotype, and non-cardiac comorbidities observed during follow-up were heterogeneous and not enriched among HF or variant-positive groups (Supplemental Table 4). These findings support the establishment of a robust cohort of prospectively enrolled neonates for identifying genetic determinants of early HF and mortality in HLHS.

Table 1.

Prospective Cohort Demographics

	Prospective Cohort; N=35	HF Free; N=18	MM HF; N=10	Severe HF; N=7
DCM-Associated Gene Variant-Positive;N (%)	12 (34.3)	2 (11.1)	6 (60.0)	4 (57.1)
Biological Sex; N (% Female)	13 (37.1)	7 (38.9)	3 (30.0)	3 (42.9)
Age at Enrollment (days; mean ± SD)	7.2 ± 7.7	11.3 ± 8.1	3.3 ± 4.4	2.1 ± 3.4
Follow-Up Time (months; mean ± SD)	29.8 ± 16.2	32.2 ± 16.1	23.2 ± 20.1	32.4 ± 5.6
Race; N (%)
Black or African American	4 (11.4)	3 (16.7)	0 (0.0)	1 (14.3)
Native Hawaiian or Other Pacific Islander	1 (2.9)	0 (0)	0 (0.0)	1 (14.3)
White or Caucasian	23 (65.7)	12 (66.7)	8 (80.0)	3 (42.9)
Multi-Racial/Other	4 (11.4)	2 (11.1)	1 (10.0)	1 (14.3)
Choose not to Identify	3 (8.6)	1 (5.6)	1 (10.0)	0 (0.0)
Ethnicity; N (%)
Hispanic or Latino	5 (14.3)	2 (11.1)	2 (20.0	1 (14.3)
Not Hispanic, Latino or Middle Eastern	29 (82.9)	16 (88.9)	7 (70.0)	6 (85.7)
Other	1 (2.9)	0 (0)	1 (10.0)	0 (0.0)
Inferred Genetic Ancestry; N (%)
Admixed American (AMR)	6 (17.1)	1 (5.6)	3 (30.0)	2 (28.6)
African/African American (AFR)	4 (11.4)	3 (16.7)	0 (0.0)	1 (14.3)
Non-Finnish European (NFE)	21 (60.0)	13 (72.2)	5 (50.0)	3 (42.9)
Other or Mixed	4 (11.4)	1 (5.6)	2 (20.0)	1 (14.3)
HLHS Anatomy; N (%)
MS/AS	8 (22.9)	2 (11.1)	4 (40.0)	2 (28.6)
MS/AA	9 (25.7)	6 (33.3)	2 (20.0)	1 (14.3)
MA/AA	15 (42.9)	8 (44.4)	3 (30.0)	4 (57.1)
Other	3 (8.5)	2 (11.1)	1 (10.0)	0 (0.0)
Stage I Palliation; N (%)
Norwood	31 (88.6)	17 (94.4)	10 (100.0)	4 (57.1)
Hybrid Stage I	1 (2.9)	1 (5.6)	0 (0.0)	0 (0.0)
None	3 (8.6)	0 (0.0)	0 (0.0)	3 (42.9)

Abbreviations: AA, aortic atresia; AS, aortic stenosis; HF, heart failure; MM, medically managed; MS, mitral stenosis.

Characterization of Variants in DCM-Associated Genes in the Prospective Cohort

We identified 40 genes with definitive, strong, moderate, or limited human genetic evidence of DCM association based on ClinGen curation¹⁶ (Supplemental Table 5) and additional supporting evidence of DCM pathogenicity¹⁷. Leveraging a customized variant filtering informatics pipeline, we identified a total of 17 ultra-rare variants in DCM-associated genes across 12 individuals in the prospective cohort (34.3%; Table 2). Multiple variants were identified in OBSCN (n=5), TTN (n=2), NEXN (n=2), and MYH7 (n=2) and single variants were found in RBM20, MYBPC3, LDB3, SCN5A, PRDM16 and MYH6. Manual evaluation of pathogenicity following ACMG guidelines classified variants as LP/Ps, (n=8; TTN x2, MYH7 x2, NEXN x2, MYH6, SCN5A) or VUSs (n=9). Nine (52.9%) variants were present in ClinVar¹⁸ (as of 02.2025), of which 5 (55.6%) demonstrated concordant classifications with ACMG criteria. All ClinVar variants classified as LP/P were also categorized as LP/P by ACMG. Moreover, ACMG criteria identified additional LP/P variants that were either absent from ClinVar (n=2), had conflicting classifications (n=1) or were classified as VUSs (n=2). Together, these results support a substantial burden of ultra-rare variants in DCM-associated genes among individuals with HLHS.

Table 2.

Ultra-Rare DCM-Associated Variants Identified in the Prospective Cohort

Subject	Dx	HF Group	Gene	Variant class	NM	Zygosity	DNA change(c.)	Protein change (p.)	ACMG Classiflcation	ClinVar Review Status
1	HLHS	Severe	TTN	Del	NM_001267550.2	Het	52524del	Lys17508AsnfsTer17	P (PS4, PVS1, PM2,6)	1* P
2	HLHS	Severe	MYH7	SNV	NM_000257.4	Het	1640C>T	Thr547Ile	P (PS4, PM1,2, PP2,3,5)	1* LP
3	HLHS	MM	RBM20	SNV	NM_001134363.3	Het	542G>A	Gly181Glu	VUS (PM2,6)	1* VUS
4	HLHS	Severe	MYH7	SNV	NM_000257.4	Het	3168G>C	Glu1056Asp	LP (PM1,2,6, PP2,3,5)	2* VUS
5	HLHS	Severe	MYBPC3	SNV	NM_000256.3	Het	1255C>T	Arg419Cys	VUS (PM2,6)	2* VUS
6	HLHS	HF Free	OBSCN	SNV	NM_001386125.1	Het	21217G>T	Ala7073Ser	VUS (PM2,6)	n/a
7	HLHS	MM	LDB3	SNV	NM_007078.3	Het	254G>A	Arg85His	VUS (PM2,6)	2* VUS
8	HLHS	HF Free	SCN5A	SNV	NM_000335.5	Het	659C>T	Thr220Ile	LP (PP3, PM2,6)	C
			PRDM16	SNV	NM_022114.4	Het	2777C>A	Pro926His	VUS (PM2,6)	n/a
			OBSCN	SNV	NM_001386125.1	Het	10978G>T	Gly3660Trp	VUS (PM2,6)	n/a
9	HLHS	MM	OBSCN	SNV	NM_001386125.1	Het	15254G>T	Gly5085Val	VUS (PM2,6)	n/a
			OBSCN	SNV	NM_001386125.1	Het	20995G>T	Gly6999Trp	VUS (PM2,6)	n/a
			MYH6	Del	NM_002471.4	Het	1510_1512del	Lys504del	LP (PM2,4,6)	n/a
10	HLHS	MM	TTN	SNV	NM_001267550.2	Het	51100G>A	Gly17034Ser	LP (PP3, PM2,6)	1* VUS
11	HLHS	MM	NEXN	Ins	NM_144573.4	Het	529_530insTT	Ser241Phefs*8	P (PVS1, PM2,6)	n/a
			NEXN	Ins	NM_144573.4	Het	531dup	Leu242Thrfs*11	P (PVS1, PM2,6)	n/a
12	HLHS	MM	OBSCN	SNV	NM_001386125.1	Het	13673G>A	Arg5515His	VUS (PM6)	C

Abbreviations: C, conflicting; Del, deletion; HF, heart failure; Het, heterozygous; HLHS, hypoplastic left heart syndrome; Ins, insertion; LP, likely pathogenic; MM, medically managed; P, pathogenic; SNV, single nucleotide variation; VUS, variant of uncertain significance.

^†n/a: variant not present in ClinVar (as of Feb. 1 2025).

LP/P Variants in DCM-Associated Genes Are Associated with Severe HF in Neonates with HLHS

We next evaluated whether the burden of DCM-associated variants differed by HF outcome in the prospective cohort. Compared with individuals who remained HF-free, those with medically managed HF showed a trend toward higher odds of carrying either a VUS or an LP/P variant in a DCM-associated gene, but this was not statistically significant (all ORs: 7.3, P=0.12; Figure 2). While individuals with severe HF demonstrated no clear enrichment of VUSs (OR: 2.8; P=0.49), the odds of hosting LP/P variants were significantly increased in this group (OR: 12.8, P=0.03; Figure 2C).

Figure 2. DCM-associated gene variant carriers by heart failure (HF) group.

A. Distribution of individuals in the prospective cohort and across HF groups carrying any rare variant in a dilated cardiomyopathy (DCM)-associated gene (either a likely pathogenic/pathogenic (LP/P) variant or a variant of uncertain significance (VUS)). DCM-associated gene variant positivity is further classified based on gene evidence for DCM association (ClinGen 02.2025). Numbers within bars represent number of positive individuals. B-C. Distribution of DCM-associated gene VUS-positive (B) or LP/P-positive (C) individuals in the prospective cohort and across HF groups. Numbers within bars represent the number of positive individuals/total individuals. OR=odds ratio, *=P<0.05 (Fisher’s exact test with Baptista-Pike method for calculating odds ratio).

Given these findings, we next explored whether LP/P variant status was associated with HF risk over time in the prospective cohort. LP/P variant carriers demonstrated a marked 8.7-fold increased hazard of developing severe HF compared with genotype-negative individuals (HR: 8.7 [95% CI, 1.7–34.6], P=0.02; Figure 3A, Table 3). Notably, two-thirds (66.7%) of LP/P carriers who developed severe HF met composite endpoints within the first 10 days of life, and all by 4 months (Figure 3A). LP/P variants did not exhibit a significantly increased risk of medically-managed HF (HR: 3.0 [95% CI, 0.6–14.2], P=0.17; Figure 3B). When evaluating overall HF, LP/P variant carriers had a 4.6-fold increased risk of meeting any severe or medically-managed HF endpoint (HR: 4.6 [95% CI, 1.6–12.9], P=0.004; Figure 3C). These associations persisted after adjusting for genetic ancestry (Supplemental Figures 2 and 3). To determine whether these associations were observed among older individuals, we next assessed an ambispective cohort, which included individuals who were significantly older than those in the prospective cohort (8.7±7.7 years versus 7.2±7.7 days, P<0.0001). In this older ambispective cohort, LP/P variants were not associated with severe, medically managed or overall HF outcomes (all P>0.05; Table 3). Overall, these findings support the hypothesis that LP/P variants in DCM-associated genes may contribute to early HF susceptibility in infants with HLHS, and that this effect diminishes in later childhood, highlighting the potential for survivor bias to obscure this association when assessed across the lifespan.

Figure 3. Variants in dilated cardiomyopathy (DCM)-associated genes are associated with heart failure (HF) in the prospective cohort.

Penetrance of severe (A), medically-managed (B), or overall HF (C), stratified by genotype: genotype-negative, variant of uncertain significance (VUS), and likely pathogenic/pathogenic (LP/P). Kaplan-Meier curves show time to HF onset. Shaded areas represent 95% confidence intervals. Odds ratios (OR) were calculated using the Baptista-Pike method with significance assessed by Fisher’s exact test. Survival curve P values represent comparisons versus the genotype-negative group (log-rank test). P<0.05 indicates statistical significance.

Table 3.

Association of DCM-Associated Variants with Heart Failure Outcomes Across Cohorts

Cohort (N)	Hazard Ratio† [95% CI]; P-value
	Severe HF		Medically-Managed HF		All HF
	VUS vs. GN	LP/P vs. GN	VUS vs. GN	LP/P vs. GN	VUS vs. GN	LP/P vs. GN
Prospective (35)	1.3 [0.2-8.1]; P=0.76	8.7 [1.7-34.6]; P=0.02*	5.2 [1.5-18.7]; P=0.01*	3.0 [0.6-14.2]; P=0.17	3.8 [1.2-12.0]; P=0.002*	4.6 [1.6-12.9]; P=0.004*
NCH (36)	3.4 [0.7-20.2]; P=0.13	11.2 [2.5-51.2]; P=0.005*	Insufficient Events	Insufficient Events	4.9 [1.2-20.5]; P=0.03*	9.2 [2.2-39.0]; P=0.005*
VUMC (157)	0.6 [0.2-1.6]; P=0.34	3.3 [1.2-7.1]; P=0.01*	1.2 [0.3-4.1]; P=0.81	1.1 [0.3-3.2]; P=0.93	0.8 [0.3-1.9]; P=0.63	1.94 [1.1-3.6]; P=0.04*
Ambispective (40)	0.8 [0.2-3.0]; P=0.74	0.9 [0.2-5.1]; P=0.89	0.5 [0.1-2.0]; P=0.3	0.3 [0.04-1.8]; P=0.18	0.7 [0.2-2.4]; P=0.56	0.8 [0.1-4.2]; P=0.77

Abbreviations: CI, confidence interval; GN, genotype-negative; LP/P, likely pathogenic/pathogenic; NCH, Nationwide Children’s Hospital; VUMC; Vanderbilt University Medical Center.

^†Hazard ratios were estimated using Cox proportional hazards models over cohort-specific follow-up periods. The prospective cohort and NCH cohort included all available follow-up. The VUMC cohort was restricted to events occurring within the first year of life to focus on early heart failure and reduce heterogeneity in later follow-up, which may reflect distinct clinical and biological processes.

VUSs in DCM-Associated Genes Are Associated with Medically Managed HF in Neonates with HLHS

We next assessed whether VUSs showed similar associations. This revealed an increased risk of developing medically managed HF over time among VUS carriers compared with genotype-negative individuals in the prospective cohort (HR: 5.2 [95% CI, 1.5–18.7], P=0.01; Figure 3B, Table 3). Although VUS carriers did not exhibit an increased risk of severe HF (HR: 1.3 [95% CI, 0.2–8.1], P=0.76; Figure 3A), they demonstrated an elevated risk of overall HF (HR: 3.8 [95% CI, 1.2–12.0], P=0.02), with most affected individuals meeting HF endpoints within the first month of life (Figure 3C). These associations remained consistent when adjusted for genetic ancestry (Supplemental Figure 3–4). As with LP/P variants, VUSs did not confer an increased risk of HF in the older ambispective cohort, again suggesting attenuation of these associations later in life (all P>0.05; Table 3). Overall, these findings support the hypothesis that VUSs in DCM-associated genes may be associated with early HF susceptibility in infants with HLHS.

Multi-Institutional Validation of Associations Between Variants in DCM-Associated Genes and Early HF in HLHS

To assess the reproducibility of associations between DCM-associated variants and HF outcomes, we evaluated independent HLHS cohorts from NCH and VUMC. The NCH cohort included 36 retrospectively enrolled individuals with HF outcomes ascertained by chart review through ~4 years of age (Supplemental Table 6). Notably, most individuals (94.4%) underwent hybrid palliation and 5.6% received a Norwood procedure, reflecting institutional preference and differing significantly from the prospective DUHS cohort, in which Norwood palliation predominated (all P<0.0001). Of the NCH cohort, 25 individuals (69.4%) remained HF free, 9 (25.0%) developed severe HF, and 2 (5.6%) experienced medically-managed HF. Consistent with findings from the prospective cohort, LP/P variants were associated with increased risk of severe HF in the NCH cohort (HR: 11.2 [95% CI, 2.5–51.2], P=0.005; Figure 4A, Table 3). Medically managed HF events were infrequent (2/36 individuals), limiting statistical power for additional analyses. When HF outcomes were combined, LP/P carriers demonstrated increased risk of overall HF (HR: 9.2 [95% CI, 2.2-39.0], P=0.005; Figure 4B). Although VUSs were not associated with a significant increase in severe HF risk (HR: 3.4 [95% CI, 0.7–20.2], P=0.13; Figure 4A), they were associated with an increased risk of overall HF in the NCH cohort (HR: 4.9 [95% CI, 1.2–20.5], P=0.03; Figure 4B).

Figure 4. Variants in dilated cardiomyopathy (DCM)-associated genes are associated with heart failure (HF) in the Nationwide Children’s Hospital (NCH) and Vanderbilt University Medical Center (VUMC) cohorts.

A-B. Penetrance of severe (A) and overall HF (B) in the NCH cohort, stratified by genotype: genotype-negative, variant of uncertain significance (VUS), or likely pathogenic/pathogenic (LP/P). C. Penetrance of severe HF within the first year of life in the VUMC cohort, stratified by genotype. Kaplan-Meier curves show time to HF onset, with follow-up extended to two years. Shaded areas represent 95% confidence intervals. Odds ratios (OR) were calculated using the Baptista-Pike method and compared using Fisher’s exact test. Survival curve P values represent comparisons versus the genotype-negative group (log-rank test). P<0.05 indicates statistical significance.

We then employed a second independent validation cohort to further test the rigor of the observed association. The VUMC cohort included 157 individuals with HLHS with a mean follow-up duration of 5.6±8.3 years (Supplemental Table 7). Similar to the prospective cohort, most individuals underwent Norwood palliation (94.0%) while 6.0% received hybrid procedures (all P>0.05). Within the cohort, 65.0% remained HF free, 9.5% developed medically managed HF, and 25.5% developed severe HF. Notably, LP/P variants in DCM-associated genes were associated with increased odds of severe HF by both 6 months (OR: 4.3 [95% CI 1.21–15.59]; P=0.03) and 1 year of age (OR: 3.5 [95% CI 1.31–9.33]; P=0.01; Figure 4C). In adjusted Cox analyses, LP/P variant carriers also demonstrated increased risk of both severe HF (HR: 3.3 [95% CI 1.2–7.1], P=0.01) and overall HF (HR: 1.9 [95% CI 1.1–3.6], P=0.04) within the first year of life (Figure 4C). In contrast, VUS carriers did not exhibit an increased risk of severe, medically managed, or overall HF (all P>0.05; Table 3). Together, these findings support reproducible associations between LP/P variants in DCM-associated genes and early HF across independent HLHS cohorts.

Discussion

Despite major surgical advances, HF remains a leading cause of early death in HLHS, with up to 40% of affected individuals dying in the first year of life¹⁹ and 6-year transplant-free survival remaining only ~60%²⁰. Current risk-stratification frameworks rely largely on post-Norwood clinical features²¹ and therefore provide limited insight into risk during the immediate postnatal period, when severe HF may emerge rapidly. In this study, we identified a distinct genetic signature of early HF in infants with HLHS, most pronounced within the first month of life. These findings suggest that genetic susceptibility contributes to early HF risk in HLHS, beyond that captured by current anatomic and perioperative models, and may help explain the heterogeneity in outcomes among infants with clinically similar anatomy and management.

LP/P variants in DCM-associated genes were strongly associated with severe HF, with most affected individuals reaching composite endpoints within the first 10 days of life. Similar findings were observed across the independent NCH and VUMC cohorts, supporting the reproducibility of these findings across distinct clinical settings and surgical strategies. In contrast, associations were attenuated in older individuals, consistent with survivor bias observed in prior largely retrospective studies of SVD, which often relied on ICD code-based outcome ascertainment, and may have limited inclusion of the highest-risk individuals and obscured genetic associations^22–27. By combining prospective enrollment with rigorous variant filtering, we identified a previously unrecognized association between variants in DCM-associated genes and early-onset HF. Importantly, associations persisted after adjustment for genetic ancestry, addressing the underrepresentation of diverse populations in genomic studies and strengthening the robustness of the findings. Given that post-transplant survival remains substantially lower in individuals with HLHS than in those with biventricular physiology²⁸, these findings underscore the importance of early risk stratification to support timely intervention and potentially reduce progression to advanced HF.

LP/P variants in DCM-associated genes are rare (<1%) but are associated with modest increases in HF and mortality (~1.5-2-fold) in population-based cohorts ²⁹. In contrast, LP/P variants in the prospective HLHS cohort were associated with a ~9-fold increased risk of early HF. This marked difference in effect size raises the possibility that HLHS functions as a unique physiologic stress state in which abnormal hemodynamics, perioperative stress, and early ventricular remodeling unmask latent myocardial vulnerability associated with DCM-associated variants. In this model, genetic susceptibility may not independently cause HF but may influence how an individual myocardium responds to the extraordinary physiologic demands of single-ventricle circulation. VUSs demonstrated weaker and less consistent associations across cohorts, although their enrichment among HF cases in the prospective cohort raises the possibility that a subset may influence HF susceptibility in specific physiologic contexts. These observations should be interpreted cautiously, as VUSs are not clinically actionable and should not independently guide decisions regarding palliation or treatment. Rather, they may help inform future studies of variant interpretation and disease mechanisms in HLHS, including iPSC-based modeling and targeted therapies such as myosin modulators^30,31.

Clinically, early identification of infants with increased genetic susceptibility may support more timely and tailored decision-making, including consideration of hybrid palliation, mechanical support strategies, or earlier optimization of medical therapy. Established predictors of outcomes in SVD include ventricular morphology, shunt type, early postoperative clinical course and genetic syndromes⁶; however, these variables incompletely account for differences in outcomes observed among infants with similar clinical profiles³². Our findings suggest that variants in DCM-associated genes may represent one biological contributor to this unexplained variability and support investigation of genetic susceptibility as a complement to existing clinical risk models. In adult cardiac surgery, the Society of Thoracic Surgeons Exceptional Risk Exclusion Committee provides a framework for adjudicating rare cases that fall outside standard risk models. Similarly, identification of pathogenic genetic variants in infants with HLHS could define a subgroup with exceptional biologic risk, improving risk adjustment and reducing bias in outcome comparisons.

As genetic screening becomes more widespread, defining the clinical utility of variants that may modify disease risk is increasingly urgent. While ACMG guidelines recommend return of LP/P variants in actionable genes^15,33, there is no consensus on reporting variants that modify disease risk. This uncertainty is especially relevant in CHD, where disclosure of uncertain results may heighten psychological burden³⁴. Genetic risk stratification must therefore be implemented with caution, ensuring equitable access and minimizing harm. Frameworks such as ClinGen and the NHGRI ELSI program offer guiding principles, including stakeholder engagement and respect for autonomy³⁵. Community engagement will be essential to apply these frameworks with cultural humility, prioritize patient needs, and ensure relevance to affected populations³⁶.

By leveraging a prospective neonatal cohort and rigorous variant filtering, we identified a distinct genetic signature of early-onset HF in HLHS. These findings suggest that genetic susceptibility may represent an additional layer of biologic risk, beyond traditional risk factors, that contributes to outcome variability in HLHS. Larger studies integrating genomic and clinical predictors will be needed to define the incremental prognostic value of genetic risk stratification. More broadly, these data support a model in which HLHS functions as a physiologic stress state that may unmask latent myocardial vulnerability, contributing to early HF risk in genetically susceptible individuals.

Limitations

This study should be interpreted in the context of several limitations. 1.) Despite prospective enrollment, cohort size remained modest, limiting power to detect gene-specific effects and contributing to imprecision in risk estimates. 2.) Although findings showed reproducible associations between LP/P variants and severe HF across neonatal and external validation cohorts, heterogeneity in surgical strategies and clinical management may have influenced cohort-specific results. Differences in institutional management practices may have influenced ascertainment of medically managed HF, particularly at NCH, where routine prophylactic HF-directed therapy during staged palliation may have reduced later clinical HF diagnoses. 3.) Inclusion of VUSs introduces uncertainty regarding pathogenicity, and observed associations may reflect aggregate or context-dependent effects rather than causal relationships at the individual variant level. 4.) Our outcome framework focused on clinically adjudicated systolic HF and may not capture other myocardial phenotypes, including subclinical or isolated diastolic dysfunction. Future studies incorporating dedicated diastolic assessment and myocardial tissue-level characterization will be important to define underlying mechanisms. 5.) Finally, longer longitudinal follow-up will be necessary to better define HF trajectories and determine whether genetic risk influences transition between medically-managed and severe HF phenotypes over time. Larger prospective studies are needed to validate these findings and define the clinical utility of genetic risk stratification in HLHS.

Clinical Perspectives.

Competency in Medical Knowledge

In neonates with hypoplastic left heart syndrome, rare variants in cardiomyopathy-associated genes are associated with early-onset heart failure, particularly within the first month of life. These findings suggest that genetic variation may identify a subset of patients with increased susceptibility to myocardial dysfunction not fully captured by current clinical risk models, which rely primarily on anatomic and perioperative factors.

Translational Outlook

Future studies should evaluate whether genetic information can be integrated into early clinical decision-making and whether genetically informed risk stratification can guide surgical planning, timing of intervention, and targeted therapies. The potential to identify high-risk variants in the immediate postnatal period may enable earlier and more personalized management strategies in infants with HLHS.

Abbreviations

CHD

Congenital Heart Defect

DCM

Dilated Cardiomyopathy

HF

Heart Failure

HLHS

Hypoplastic Left Heart Syndrome

NC-DEFINE

determining the factors that influence mortality in congenital heart disease in North Carolina

NCH

Nationwide Children’s Hospital

SVD

Single Ventricle Disease

VUMC

Vanderbilt University Medical Center

Data Availability

All informatic analysis files are available via GitHub¹⁴. Genomic sequencing data are deposited in the database of Genotypes and Phenotypes (dbGaP; accession number phs002010.v1.p1).