Epigenetic age acceleration in young adults with congenital heart disease

Abstract

Background

Adults with congenital heart disease (ACHD) having undergone palliative surgery experience chronic stress due to altered physiology and repeated surgical interventions since infancy.

Objectives

To investigate whether ACHD, who had experienced chronic physiological stress from their underlying condition and early-life cardiac surgeries, was associated with epigenetic age acceleration (EAA) and other DNA methylation (DNAm)-based biomarkers, and to assess the potential contribution of derived inflammatory markers to EAA.

Methods

A case–control study comparing ACHD patients and healthy adults. Whole blood DNAm profile was used to estimate DNAm-based blood cell type proportions, multiple epigenetic age measures, and interleukin-6 (IL-6) and C-reactive protein (CRP) scores. Two ACHD subgroups were recruited: one with multiple palliative surgeries since birth (Fontan group, n = 13), and another with a single corrective surgery as an infant (SS group, n = 5). Healthy controls (n = 20) had no chronic medical conditions. EAA was calculated using four epigenetic clocks (Horvath, Hannum, GrimAge, PhenoAge) and the pace of aging (DunedinPACE). Comparisons were made across groups using robust linear regression models, adjusting for age, sex, self-reported ethnicity, and estimated cell type proportions. Associations between DNAm-based IL-6 and CRP scores and surgery group were tested, and their potential contribution to differences in EAA was evaluated.

Results

Participants were 20–30 years (25.6 ± 2.7 years), predominantly non-Hispanic white. After controlling for age/sex/ethnicity/immune-cell-type proportions, the Fontan group had significantly higher GrimAge (Cohen’s f = 0.90, p < 0.001) and PhenoAge (Cohen’s f = 0.82, p < 0.001) and higher DunedinPACE (Cohen’s f = 0.69, p = 0.01). The Fontan group also had statistically higher predicted IL-6 (Cohen’s f = 0.84, p < 0.001) and CRP scores (Cohen’s f = 0.62, p < 0.001).

Conclusions

Young ACHD patients who undergo multiple childhood surgeries for Fontan palliation were associated with accelerated aging. These changes could reflect the long-term effects of underlying CHD condition, early-life physiological stress and other factors, potentially involving inflammatory pathways. Further research is needed to identify and validate the key factors contributing to EAA in this population and to clarify the role of chronic stress and physiological alterations over time.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-026-02049-5.

Background/Introduction

Congenital heart disease (CHD) is the most common birth defect in the US, affecting 1% of live births [1]. Advances in care have improved survival for adults with congenital heart disease (ACHD) [2, 3]. However, ACHD patients face significant comorbidities due to imperfect hemodynamics, repeated surgeries, and chronic inflammation, with higher mortality burden in more severe cases [4–7]. ACHD patients with Fontan circulation undergo three stages of surgical palliation, starting with the Norwood procedure as neonate, Glenn procedure at 4–6 months of age, with final completion of Fontan circulation by age 2–3 years, resulting in passive systemic venous return to the heart. As a result, ACHD with Fontan circulation suffer chronic stress due to altered physiology [8] and lower physical health status [2], and increased age-related comorbidities like dementia and coronary artery disease, with early signs of cardiac aging due to excessive inflammation and oxidative stress stress [9].

DNA methylation (DNAm), which involves adding a methyl group to the 5’ position of cytosine at cytosine-guanine dinucleotides (CpGs), plays a key role in several important biological processes, including development, cell differentiation, and imprinting [10]. The accumulation of DNAm data has led to the development of DNAm-based tools, such as cell type deconvolution algorithms [11, 12], epigenetic clocks [13–17], and DNAm-based biomarker predictors [18, 19]. which are widely used in health-related epigenetic research [20]. Among all these tools and biomarkers, research and development of epigenetic clocks is a prominent and rapidly advancing field. Epigenetic clocks use DNAm patterns at specific CpGs that are highly correlated with chronological age [21–23], and can be used as biomarkers to investigate premature aging in CHD [2, 7]. Epigenetic age acceleration (EAA), is a derived measure of rate of aging [24, 25], and higher EAA has been associated with negative outcomes of various health conditions and diseases, such as cardiovascular disease [14–16, 26–28] and metabolic syndrome [17, 29].

The biological embedding of early life experience is the process by which childhood exposures can alter developmental trajectories, leading to individual variations in health and wellbeing in adulthood [30]. DNAm acts as a bridge between environment and genome, embedding experiences like trauma into biology [30], and can be detected by DNAm-based biomarkers [31]. While epigenetic aging is studied in cardiovascular disease [32, 33], it is less explored in ACHD population, especially in young adults, when early intervention is possible. We hypothesize that the early life exposure of highly invasive surgical procedures, plus the cumulative metabolic stress and systemic inflammation imposed by altered circulatory physiology, can potentially become embedded in the epigenome of young ACHD patients. This impact may be detectable by DNAm-based tools, namely cell type proportion estimations, inflammation scores and epigenetic clocks, and with strongest impact in those who underwent multiple palliative interventions including Fontan surgery in early childhood. Here, we investigated whether ACHD, who had experienced chronic physiological stress from their underlying condition and early-life cardiac surgeries, exhibit differences in DNAm-based biomarkers compared to controls. We further explore a potential contribution of inflammation in EAA.

Materials and methods

Study participants description

This case–control study involved ACHD patients aged 20–30 years, and healthy controls of the same age range with no past medical or surgical history (Table 1). Two subgroups of ACHD patients were recruited during their routine outpatient visits: the Fontan group included patients who underwent multiple palliative surgical interventions during infancy and early childhood with Fontan surgery being the latest intervention, resulting in passive pulmonary circulation and increased venous congestion. The second subgroup consisted of patients with a single corrective surgery as an infant (SS group). The study was approved by the Baylor College of Medicine IRB and informed consent was obtained from all participants (cases and controls) before blood sampling. After obtaining informed consent, 5 mL of peripheral whole blood was collected via venipuncture into EDTA tubes from all patients during their routine outpatient visits with an ACHD cardiologist. All samples including controls were collected following the same protocol and were immediately stored at − 80 °C until further processing.

Table 1.

Cohort demographics

	Control	Fontan	Single surgery	p-value
No. patients (Total 38)	20	13	5
Sex				0.241
Male	5	6	3
Female	15	7	2
Ancestry				0.96
White non-Hispanic	13	7	4
White Hispanic	3	2	1
African American	2	4	0
Asian	2	0	0
Age at sample collection: Median (IQR)	27 years (25- 28)	23 years (21–26)	25 years (23 – 28)
Age at last Surgery: Median (IQR)	NA	3 years (2 – 4 years)	3 months (1 – 10 months)
Surgical repair	NA	Fontan (13)	TOF (3) ASD/VSD (1) ASO (1)

Number of patients or controls in each group, with sex, ancestry, age at sample collection and at last surgery, and type of surgical repair. TOF, Tetralogy of Fallot repair, ASD/VSD, Atrial Septal Defect/Ventricular Septal Defect; ASO, Arterial Switch Operation. The p-value was calculated using chi-square test

Microarray genome-wide DNA methylation profiling and data processing

Genomic DNA was extracted from whole blood using the Qiagen DNeasy Blood and Tissue Kit according to the manufacturer’s instructions [34]. The EZ DNA Methylation Kit (Zymo Research, Irvine, CA, USA) was used to bisulfite-convert 750 ng of genomic DNA, allowing for sequence-based differentiation of unmethylated and methylated cytosine nucleotides. An input of 160 ng of converted DNA was used for whole-genome amplification and enzyme fragmentation, then hybridized to Infinium MethylationEPIC BeadChip V1 (Illumina, CA, USA), and scanned with an Illumina HiScan. Raw IDAT files capturing the signal intensities were exported for analysis in R (version 4.2.0). Sample quality control (QC) was assessed as described previously [35, 36]. Briefly, all samples passed the QC metrics including Illumina quality control metrics [37], average methylated and unmethylated intensity [38], detection p-value [38], beadcount [38, 39], outlier detection [39], and predicted sex and reported sex matching [38]. Signals were background-subtracted and color-corrected, followed by Beta-Mixture Quantile (BMIQ) normalization [40], then technical batch correction was performed using ComBat from the R package sva (version 3.46.0) [41]. DNA methylation levels are reported as β values ranging from 0 (unmethylated) to 1 (fully methylated). Background-subtracted and color-corrected DNAm β values were used as input for cell type proportion estimation as recommended [11], while normalized and batch-corrected DNAm β values were used to calculate epigenetic clocks and DNAm-based inflammation markers since normalization had been shown to improve clock performance [42].

Cell type proportion estimation

Cell type composition is one of the key drivers of variation in DNAm profiles [43, 44]. In the absence of direct cell counts of the whole blood samples, we estimated the relative proportions of 12 blood cell types (neutrophils, eosinophils, basophils, monocytes, naïve and memory B cells, naïve and memory CD4 + and CD8 + T cells, natural killer, and T regulatory cells) from DNAm data using the IDOL method [15]. The proportional cell type estimations were adjusted by the isometric log-ratio (ilr) transformation, and then a robust principal component analysis (PCA) was applied for dimension reduction [45].The top 3 principal components (PCs) accounted for 93.1% of predicted cell type proportion variability and were thus included in subsequent statistical models.

To determine the extent of influence of surgery group on immune cell type proportions, we applied robust linear regression models (rlm) to test estimated cell type proportion differences between the control, Fontan and SS groups, while adjusting for age, biological sex, and self-reported ethnicity:

Each estimated cell type was tested independently using the same model. To correct for multiple testing, we applied the Bonferroni correction with threshold of adjusted p-value < 0.05.

Estimation of first- and second-generation epigenetic clocks and rate-based clock

Epigenetic age as measured by the Horvath pan-tissue clock [25] (Horvath), Hannum clock [46], GrimAge [47] and PhenoAge [48] were calculated using the DNA Methylation Age Calculator (https://dnamage.genetics.ucla.edu/new), while DunedinPACE was calculated using the DunedinPACE R package [49]. Additionally, DNAmPACKYRS, one of the DNAm-based biomarkers incorporated in GrimAge to estimate smoking exposure (“pack-years”), was extracted for downstream analysis. For the first-and second-generation clocks, epigenetic age acceleration (EAA) was calculated by extracting the residuals from a linear model of DNAm age regressed on chronological age for Horvath, Hannum, GrimAge and PhenoAge estimates. DunedinPACE measures rate of aging, therefore without the need for regression on chronological age.

The differences between the three groups were tested by rlm independently for the four EAA measures and DunedinPACE, all adjusted for age, biological sex, self-reported ethnicity and top three cell type PCs:

Bonferroni correction was applied. Effect size was calculated as Cohen’s f, using the effsize R package (version 0.8.1) [50].

DNAm-based inflammation marker predictions

DNAm-based scores of inflammation biomarkers, interleukin-6 (IL-6) and C-reactive protein (CRP), positively correlate with measured plasma levels, thus providing proxy measures of chronic inflammation [18, 19]. We calculated two DNAm-based markers of inflammation: IL-6 score [18] and CRP score [19], both transformed into z-scores (standardized to mean of 0 and standard deviation of 1). Using the same statistical model as described above, DNAm IL-6 and CRP score differences between groups were tested by the following rlm, and p-values were Bonferroni corrected:

Contribution analysis

To test whether the inflammation biomarkers and other predicted factors influenced EAA, we performed contribution analysis, as demonstrated previously [35, 51, 52]. We calculated the percent contribution to the regression β coefficient of surgery groups by including contributors: (1) DNAm IL-6 score, (2) DNAm CRP score, and (3) DNAmPACK in the adjusted models, and compared these to the base model.

Percent contribution was calculated as,

Results

We enrolled 13 ACHD patients in the Fontan group, 5 ACHD patients in SS group, and 20 healthy controls. The control group was predominantly White females. The Fontan and SS groups had almost equal male to female patients, and had predominantly White patients of non-Hispanic descent. Median ages at sampling were 27 years for controls, 23 years for the Fontan group, and 25 years for the SS group. The median age for the last surgery was 3 years for Fontan and 3 months for SS patients. Cohort demographics are summarized in Table 1, with cardiac anatomy and surgical details in Table S1. A power calculation accounting for the unequal sample sizes between groups (n = 13 for the Fontan group and n = 20 for controls) was performed after sample collection, as the sample size was not determined a priori using power analysis. We estimated the statistical power to detect small (Cohen’s d = 0.2), medium (d = 0.5), and large (d = 0.8) effect sizes using a two-tailed test with an alpha level of 0.05. The estimated power was 8%, 27%, and 59% for small, medium and large effect sizes, respectively, and a minimum sample size of n = 15 per group would be required to achieve 80% power to detect a large effect size.

Fontan group had higher GrimAge, PhenoAge EAA and DunedinPACE

We tested the correlation of epigenetic ages from four clocks with chronological age, expecting a high correlation. GrimAge and Horvath age estimates were moderately positively correlated with chronological age (r(36) = 0.57, p < 0.001; r(36) = 0.49, p = 0.002), while Hannum had the lowest correlation and highest mean absolute error (17.96 years) of the four clocks (r(36) = 0.05, p = 0.76) (Figure S1, Table S2).

To determine the difference in EAA between Fontan or SS groups and the controls, we applied the robust linear models as described in the Methods. The Fontan group had a significantly higher GrimAge EAA than the control group (t(27) = 4.78, Bonferroni-adjusted p < 0.001; Cohen’s f = 0.90). Similarly, the Fontan group’s PhenoAge EAA was significantly higher than controls (t(27) = 4.12, Bonferroni-adjusted p = 0.002; Cohen’s f = 0.82). Neither the Horvath nor the Hannum clock, showed significant differences between the Fontan group and controls, and no clocks showed significant differences between SS group and controls (Fig. 1 and Table 2).

Fig. 1.

Fontan patients show higher EAA on GrimAge and PhenoAge, and DunedinPACE. EAA/DunedinPACE were compared between the Fontan, single surgery (SS) and control groups, and no significant differences among groups in EAAs of Horvath pan-tissue and Hannum (***: Bonferroni-adjusted p < 0.001, **: Bonferroni-adjusted p < 0.01)

Table 2.

Test results for epigenetic age acceleration of first- and second-generation clocks, and DunedinPACE

Clock EAA	Comparison	β coefficient (SE)	β 95% CI	t(27)	p	Bonferroni-adjusted p	Cohen’s f	Contribution
								IL-6 score	CRP score	DNAmPACKYRS
GrimAge	Fontan group and controls	4.211 (0.881)	2.484–5.939	4.778	7.4E-05	< 0.001	0.895	55.95%	44.25%	82.12%
PhenoAge		8.895 (2.158)	4.666–13.124	4.123	3.46E-04	0.002	0.818	53.20%	49.71%	62.10%
Dunedin PACE		0.136 (0.038)	0.062–0.21	3.593	1.67E-03	0.01	0.685	39.67%	54.0%	47.56%
Horvath		0.009 (1.287)	-2.513–2.531	0.007	9.94E-01	1	0.047
Hannum		1.987 (0.975)	0.075–3.898	2.037	4.74E-02	0.284	0.441
GrimAge	SS group and controls	1.951 (0.974)	0.042–3.861	2.003	4.93E-02	0.296	0.388
PhenoAge		5.444 (2.385)	0.77–10.117	2.283	2.86E-02	0.171	0.483
Dunedin PACE		0.042 (0.042)	-0.04–0.124	1.001	3.09E-01	1	0.153
Horvath		2.215 (1.078)	0.102–4.328	2.055	6.09E-02	0.365	0.377
Hannum		2.163 (1.422)	-0.624–4.95	1.521	1.35E-01	0.81	0.354

For the five clocks, the Fontan and SS groups were compared to the control group. Contributions to EAA by DNAm-based IL-6, CRP scores and DNAmPACKYRS are shown for the clocks showing significant EAA differences

DunedinPACE is a metric estimating the rate of biological aging per chronological year and we found that the Fontan group had significantly higher DunedinPACE than controls (t(27) = 3.59, Bonferroni-adjusted p = 0.01; Cohen’s f = 0.69), while the SS group was not significantly different from controls (Fig. 1 and Table 2).

Fontan group had lower estimated memory B cell and naïve CD4 + T cell proportions

As health conditions may relate to immune cell type proportions among individuals, we tested predicted cell type proportion differences among groups for each cell type. Estimated memory B cells (Bmem) (t(31) = 3.80, Bonferroni-adjusted p = 0.007; Cohen’s f = 0.60) and naïve CD4 + T cells (CD4nv) (t(31) = 7.01, Bonferroni-adjusted p < 0.001; Cohen’s f = 1.26) were significantly lower in the Fontan group than in controls. The remaining cell types did not show significant differences between the Fontan group and controls (Supplementary Figure S2 and Table S3), and there were no significant differences between the SS and control groups (Supplementary Figure S2 and Table S4).

DNAm-based inflammation markers were higher in the Fontan group and contributed to GrimAge and PhenoAge EAA, and DunedinPACE

To further illuminate possible immune system alterations in our ACHD cohort, we used DNAm-based predictors to estimate inflammatory markers as surrogates for serum measures. DNAm-based IL-6 and CRP scores were normally distributed as tested by Shapiro–Wilk test (W = 0.95, p = 0.70 and W = 0.95, p = 0.12, respectively). Expectedly, a weak positive association was observed between predicted IL-6 and CRP scores (r(36) = 0.37, p = 0.02; Supplementary Figure S3). When comparing these predicted inflammatory profiles between groups, we observed that IL-6 and CRP scores of the Fontan group were significantly higher than controls (t(27) = 4.44, Bonferroni-adjusted p < 0.001 and t(27) = 4.69, Bonferroni-adjusted p < 0.001, respectively) with large Cohen’s effect size for IL-6 score (Cohen’s f = 0.84) and medium effect size for CRP score (Cohen’s f = 0.62) (Fig. 2 and Table S4). While IL-6 and CRP scores of the SS group were higher than controls, the trends did not reach statistical significance (t(27) = 2.27, Bonferroni-adjusted p = 0.054 and t(27) = 0.27, Bonferroni-adjusted p = 0.055, respectively) (Fig. 2 and Table S5).

Fig. 2.

Fontan patients have higher DNAm-based IL-6 and CRP scores. DNAm-based IL-6 and CRP scores are compared between the Fontan, single surgery (ss) and control groups (***: Bonferroni-adjusted p < 0.001)

Additionally, DNAmPACKYRS was moderately correlated with DNAm-based IL-6 and CRP scores (r(36) = 0.43, p = 0.007; r(36) = 0.42, p = 0.008, respectively) (Supplementary Figure S3). DNAmPACKYRS scores were significantly higher in Fontan patients compared to controls (t(27) = 4.19, Bonferroni-adjusted p = 0.001), with a large effect size (Cohen’s f = 0.92) (Supplementary Figure S4 and Table S5). No significant difference was observed between the SS group and controls.

Due to the observed inflammation score and differences between the groups and accelerated epigenetic age in some clocks, we sought to quantify the potential influence of DNAm-based predictors to the EAA and rate of aging differences between the Fontan group and controls. We calculated the contribution of IL-6, CRP scores, and DNAmPACKYRS separately to the GrimAge and PhenoAge EAA differences and to the DunedinPACE. In Fontan group GrimAge EAA, DNAm IL-6, CRP scores and DNAmPACKYRS contributed 55.95, 44.25, 82.12% respectively (Table 2). We observed similar contributions to the Fontan group PhenoAge EAA, with 53.20, 49.71, 62.10% respectively; and slightly lower IL-6 (39.67%), lower DNAmPACKYRS (47.56%) and higher CRP (54%) contributions to DunedinPACE (Table 2).

Discussion

Adverse childhood experiences can become biologically embedded, resulting in disrupted physiology associated with poor long-term health outcomes, and widened gap between chronological and biological age [31, 53]. ACHD patients with Fontan circulation experience early childhood adversity of multiple surgical stresses, and have altered hemodynamics and physiology resulting in chronic morbidity. Using DNA methylation profiles from 18 ACHD patients and 20 controls in their young adulthood, we observed that ACHD patients with Fontan circulation had significantly accelerated biological aging, as measured by EAA, compared to healthy controls. Furthermore, our findings suggest that inflammation may contribute to increased EAA in the ACHD population, although further validation in larger, well-characterized cohorts is warranted.

Our study indicates that patients with Fontan circulation had significantly higher GrimAge and PhenoAge EAA, and DunedinPACE. All three of these measures of epigenetic aging had shown to be associated with cardiometabolic risk factors and mortality associated with cardiovascular disease [54, 55]. With improved survival, the ACHD population is at additional risk for acquired cardiovascular disease, and this higher morbidity/mortality risk of Fontan patients may potentially be indicated by the higher EAA as determined by the second-generation epigenetic clocks and rate-based clock. The EAA observed in this study may reflect a complex interplay between the underlying CHD condition, early-life surgical exposures, and the lifelong burden of comorbidities and chronic stress. However, our cross-sectional design precludes causal inference, and without longitudinal sampling beginning in the neonatal period or before surgery, it is not possible to determine whether EAA originated at birth, potentially due to underlying congenital factors, or accumulated over time. Future longitudinal studies starting in early life would be important to disentangle the temporal and mechanistic contribution to accelerated aging in ACHD.

The single surgery (SS) group showed a trend towards increased EAA but did not reach statistical significance. Although not significant, the SS group had lower or similar EAA, CRP, and IL-6 scores compared to the Fontan group but higher than controls. This suggested a potential cumulative effect of early life stress, from cardiac surgery and hospitalization, along with other adverse factors associated with ACHD, which may be reflected in DNA methylation changes observed in young adults and may be indicative of biological embedding.

Inflammaging, which describes the role of chronic, low grade, sterile inflammation in accelerated ageing, is increasingly recognized as an important factor in premature ageing [56]. We found that the DNAm-based proxies of inflammatory biomarkers CRP and IL-6 were significantly elevated in the Fontan subgroup of the ACHD population, consistent with measured IL-6 and CRP levels in patients with Fontan physiology [56, 57]. High sensitivity CRP measures have been shown to be elevated in ACHD and are associated with heart failure, arrhythmia and mortality [57], and the correlation of inflammation with positive EAA has been described in cardiovascular disease [58]. Predicted inflammation scores contributed to the EAA differences in GrimAge, PhenoAge and DunedinPACE, as these three clocks incorporated inflammatory biomarker for the clock training (PhenoAge: CRP, GrimAge: predicted B2M, DunedinPACE: CRP). The contribution analysis also indicated that these three clocks were sensitive to DNAm-predicted smoking scores. However, due to the lack of self-reported smoking data in our cohort, it remains unclear whether these DNAm smoking scores accurately reflected actual smoking behavior or rather capture a broader inflammatory state that may produce similar DNA methylation signatures, given that previous studies have reported associations between smoking and inflammation [59]. Considering the relatively low prevalence of smoking in the ACHD population previously reported (~ 10% current smokers and ~ 4% former smokers) [60], we hypothesize that the elevated DNAmPACKYRS observed in Fontan patients may, at least in part, reflect chronic inflammation rather than smoking exposure alone. Overall, our results may suggest inflammation is potentially an important contributor to EAA differences in this cohort. These predicted inflammatory biomarkers and their potential contribution to EAA offer insights into the functional status and chronic physiological stress state of this vulnerable population. These findings may suggest that longitudinal monitoring of inflammatory markers could be useful for evaluating disease progression and assessing treatment effects.

This study found that Fontan subgroup had significantly lower estimated memory B cell and naïve CD4 + T cells compared to healthy controls. These results, based on DNAm-predicted cell type proportions, may link to previous observation that early thymectomy, often part of surgical intervention in ACHD patients, had been associated with T-cell lymphopenia predominantly affecting memory and naïve cells [61]. Thymectomy is associated with long term autoimmune diseases, cancer, infections and atopic diseases [62] and may result in changes in the immunological system akin to aging [63]. However, these potential alterations in memory and naïve cell proportions require validation through direct measurement of cell counts.

Our study provides insight into accelerated epigenetic age in ACHD patients, especially in the high-risk Fontan subgroup, with some limitations. The small sample size, as indicated by the power analysis, was underpowered and may limit the generalizability of our findings. Therefore, these results require validation in independent cohorts with adequately powered sample sizes. We also acknowledge that in lieu of measured cell type proportions and inflammation markers, we used DNAm derived estimations, and the estimations have not yet been directly assessed in the participants. Even though these DNA methylation-based markers demonstrated high concordance with plasma protein measurements in generalized populations [37], our results should be validated using direct measurements in future studies. Additionally, the absence of detailed clinical profiling, such as comorbidities, anthropometric measures, smoking status, medication use, physical activity, and heart failure status, limited our ability to attribute the observed differences to specific underlying factors or to adjust for them in the statistical models. While the use of blood as a surrogate tissue may not fully capture tissue-specific aging processes, it can still provide valuable insights into systemic biological aging, particularly in ACHD patients who had been exposed to significant physiological stress from early-life surgeries and lifelong cardiovascular burden. Furthermore, due to the invasive nature of cardiac biopsies, especially in healthy controls, blood remains one of the most feasible options for assessing biological age in this study setting. Another limitation is the predominantly white, non-Hispanic population in the patient group, which may limit generalization of the results to other ancestries.

Conclusions

ACHD patients with Fontan physiology had significant EAA as compared to their healthy counterparts. These epigenetic clocks can serve as a potential biomarker of the disease progression, which could be associated with chronic physiological stress resulting from immune and inflammation dysregulation. This study lays the groundwork for large-scale studies to investigate EAA, and the role of inflammation in EAA, in the broader ACHD population.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: P.N.J, B.C.Z, T.W.M, M.S.K., P.A.C.; Data curation: B.C.Z, J.W., P.N.J., J.L.M., K.D., M.G., P.E., P.A.C.; Formal analysis: B.C.Z., J.W.; Visualization: B.C.Z.; Methodology: B.C.Z, J.W., P.N.J., T.W.M., P.A.C., M.S.K.; Resources: P.A.C., M.S.K.; Supervision: Writing—original draft: P.N.J., B.C.Z., J.W., M.S.K.; Writing—review and editing: all authors.

Funding

None.

Data availability

DNA Methylation datasets available upon request with proper justification.

Declarations

Ethics approval and consent to participate

The study was approved by the IRB of Baylor College of Medicine, Houston TX USA.

Competing interest

The authors declare no competing interests.