Skip to main content
Download PDF

Dissecting the causal effects of smoking, alcohol consumption, and related DNA methylation markers on electrocardiographic indices

Clinical Epigenetics volume 17, Article number: 40 (2025) Cite this article

Abstract

Background

Tobacco and alcohol are recognized risk factors for heart disease, yet their causal effects on electrocardiogram (ECG) signaling and mechanisms remain unclear. Previous studies may be susceptible to confounding or bias, and this study dissected the genetic architecture linking tobacco and alcohol consumption with P-wave duration, PR interval, and QT interval.

Methods

Utilizing genetic instruments for tobacco and alcohol consumption, associated methylation quantitative trait locus (mQTL), and summary-level GWAS data for ECG indices, we assessed heritability and genetic causal associations using linkage disequilibrium score regression and Mendelian randomization (MR) analysis. Fine mapping was performed via colocalization analysis and summary-data-based MR (SMR) to identify potential shared genetic variants.

Results

A positive causal relationship was found between drinks per week (DrnkWk) and QT interval [β (95%CI): 1.06 (0.91, 5.05), P = 0.005], with causality substantiated through multiple robust MR models. Multivariable MR confirmed independence from smoking phenotypes. In epigenetic MR analyses, two alcohol-related CpG loci (cg03345232 and cg04605617) were causally associated with QT interval changes, with cg04605617 mapping to PLA2G2C gene significantly prolonging QT. The mQTL rs10916683 at cg04605617 is a strong eQTL for PLA2G2C. Additionally, cg03345232 shared a causal variant (rs12881206) with QT interval predisposition through colocalization analysis. SMR analysis did not identify shared putative functional genes passing the HEIDI test between DrnkWk and the QT interval.

Conclusions

There is a causal relationship between DrnkWk and QT interval prolongation, and targeting specific DNA methylation sites like cg04605617 mapped to PLA2G2C may provide novel targets for preventing QT interval prolongation.

Graphic abstract

Introduction

Non-invasive indices derived from the electrocardiogram (ECG), such as P-wave duration, PR interval, and QT interval, play a pivotal role in providing insights into cardiac structural and electrical conduction abnormalities [1,2,3]. These indices are closely associated with the risk of atrial and ventricular arrhythmias, with prolonged QT intervals linked to ventricular arrhythmias, particularly torsade de pointes (TdP) [4,5,6,7,8,9]. The clinical significance of these ECG parameters lies in their causal impact on conditions such as atrial fibrillation and sudden cardiac death (SCD) [10, 11].

Observational studies reveal that ECG signals are influenced by various pathophysiological factors, such as age, gender, genetics, medications, electrolytes, hormones, and the autonomic nervous system [12, 13]. Underlying genetic and environmental determinants, such as poor lifestyle choices, possess the potential to disturb normal cardiac rhythms by modifying these pivotal factors that contribute to alterations in ECG indices [14]. Large cohort studies suggest that the amalgamation of genetic composition and collective health behaviors and factors exhibits a log-additive impact on the susceptibility to cardiovascular disease [15, 16]. The investigation of the causal association between these underlying factors and ECG signals prompts exploration into the impact of two prevalent adverse lifestyles, smoking and alcohol consumption, on ECG indices.

Smoking and alcohol use, global modifiable risk factors, exhibit intricate connections with cardiovascular disease. Both tobacco use and alcohol consumption, whether acute or chronic, are associated with arrhythmias [17,18,19,20,21,22]. Alcohol overdose is frequently associated with "holiday heart syndrome," characterized by a heart block PR and a prolonged QT interval [23,24,25]. Notably, epigenome-wide association studies (EWAS) show the effects of smoking and alcohol on DNA methylation, a regulatory mechanism influencing gene expression and potentially regulating cardiac ion channel activity [26,27,28]. However, the extent to which such epigenetic changes mediate the response of electrocardiographic signaling molecules to smoking and alcohol consumption, thereby affecting the normal regulation of cardiac rhythm, remains unclear. Additionally, establishing causal associations in observational studies is inherently challenging due to confounding bias, necessitating causal inference, and genetic architecture investigation.

Here, using large-scale genome-wide association studies (GWAS) and EWAS summary statistics, we sought to explore genetic correlations, causality, and shared risk loci with putative functions between tobacco and alcohol consumption and ECG indices, unraveling the underlying genetic association.

Method

Study design

This study aimed to investigate the genetic architecture between smoking and drinking exposure and ECG parameters. Five different smoking and drinking phenotypes were used as exposure factors, encompassing three distinct types of alcohol consumption behavior: drinks per week (DrnkWk, indicating the average number of drinks reported by each participant per week), alcohol use disorders (AUD, a binary phenotype representing a clinical diagnosis of chronic alcohol dependence), and problematic alcohol use (PAU, a broader phenotype incorporating both clinical and subclinical patterns of alcohol misuse), and two different stages of tobacco use: cigarettes per day (CigDay, a continuous measure reflecting the average number of cigarettes smoked daily), smoking initiation (SmkInit, a binary phenotype denoting whether an individual has ever smoked, serving as a proxy for early-life smoking behavior and lifetime exposure risk). The ECG markers, specifically P-wave duration (PWD), PR, and QT interval durations, were utilized as outcome measures. PWD, PR, and QT intervals were chosen due to their established roles in arrhythmic risk stratification and their potential modulation by genetic and environmental factors. Prolonged QT intervals, in particular, are linked to ventricular arrhythmias and SCD. Other ECG parameters, such as QRS duration, were not included due to the limited availability of GWAS data with sufficient statistical power, specifically within European populations. We used linkage disequilibrium score regression (LDSC) to estimate the heritability and genetic correlations between exposures and outcomes [29]. LDSC provides an initial assessment of shared genetic background between smoking, alcohol consumption, and ECG traits. A genetic correlation would suggest common genetic influences but does not establish causality. Two-sample MR (TSMR) was used to provide genetically causal evidence in two independent, non-overlapping populations, and multivariable MR (MVMR) was applied to address crosstalk genetic effects between smoking and drinking phenotypes, given genetic correction between the two (|rg|= 0.16–0.27) [30,31,32]. To identify whether DNA methylation (mQTL) mediates the observed causal relationships, we conducted epigenetic methylation MR. Colocalization analysis was then used to determine whether the same genetic variant drives both methylation changes and ECG alterations, strengthening the mechanistic interpretation of the findings.[33]. Finally, we adopted summary-data-based MR (SMR) to investigate shared gene expression pathways using eQTL from whole blood and cardiac tissue [34].

GWAS summary statistics sources

We obtained GWAS summary statistics of smoking (CigDay and SmkInit) and drinking (DrnkWk) from the GWAS & Sequencing Consortium of Alcohol and Nicotine Use (GSCAN) database [35]. GSCAN Phase 2 integrated data from 60 GWAS involving up to 3.4 million participants from four major ancestries, focusing on nicotine and substance use [30]. To ensure genetic homogeneity, we extracted stratified GWAS data for European populations (n = 2,669,029) stratified without United Kingdom Biobank (UKB). Additionally, GWAS in the FinnGen consortium for AUD (diagnostic codes: ICD-10 F10.2 or F10.3, 376,477 individuals) and the Million Veteran Program (MVP) GWAS meta-analysis for PAU (435,563 individuals) were used [36]. Summary statistics for PR and QT intervals were obtained from the Cardiovascular Disease Knowledge Portal (CVDKP). The CVDKP database curates a multi-ancestry GWAS dataset of 40 studies for the PR interval, with 271,570 Europeans accounting for 92.6% of the total participants [37]. The QT interval GWAS is a large-scale meta-analysis of 84,630 participants from the UKB [38]. PWD GWAS was derived from a recently published GWAS study (GCST004826) involving 37,678 Europeans, available in the NHGRI-EBI GWAS Catalog [39] (Table 1). Measurements of these parameters in milliseconds (ms) were obtained from resting supine 12-lead or 3-lead ECGs taken in the first 15 s before exercise testing on a bicycle. Heart rate-corrected QT (QTc) intervals were calculated using the Bazett formula, defined as \(\text{QTc}=\frac{\text{QT}(\text{ms})}{\sqrt{\text{RR }}(s)}\)[38]. More detailed recordings of ECGs and covariate adjustment for reliable results analysis can be found in the original studies [37,38,39]. All studies, including the FinnGen consortium, have their study protocols and have been approved by local ethics committees.

Table 1 Information for GWAS summary statistics
Full size table

Genetic proxies for alcohol and tobacco use

Genetic variants used as instrumental variables (IVs) meet the core assumptions of MR (Fig. 1). To serve as genetic proxies for smoking and alcohol use, single nucleotide polymorphisms (SNPs) needed to exhibit genome-wide significance (GWS) of P < 5 × 10−8 from each exposure summary-level dataset. Extracted SNP data were subsequently subjected to linkage disequilibrium (LD) clumping (r2 = 0.001, kb = 10,000) to ensure genetic independence. Palindromic variants were excluded if they had intermediate allele frequencies that hindered the inference of positive-strand alleles. The F-statistic greater than 10 was used to exclude weak instrumental bias [40].

Fig. 1
figure 1

Core assumptions of MR using genetic proxies for alcohol and tobacco use. The relevance assumption ensures IVs are strongly associated with the exposures. The independence assumption indicates IVs are independent of confounders. The exclusion restriction assumption ensures IVs affect ECG indices only through alcohol and tobacco use, not other pathways

Full size image

Genetic instruments for alcohol‐related DNA methylation

We obtained whole-blood DNA methylation patterns associated with alcohol consumption from an epigenome-wide association study (EWAS) with a meta-analysis of 13 populations (13,317 individuals) in the MRC-IEU EWAS Catalog [41]. The Infinium HumanMethylation450 BeadChip was employed to measure DNA methylation levels in blood samples, and an EWAS model was constructed using DNA methylation β value (the ratio of methylated probe intensities divided by the sum of methylated and non-methylated probe intensities) as the outcome variable and for multiple covariates (e.g., age, sex, BMI, smoking, batch effect, and white cell composition). A total of 363 CpG loci in 9643 Europeans were identified as being associated with alcohol consumption within the Bonferroni-adjusted threshold of epigenomic significance at P < 1 × 10−7 (0.05/440,000). We then derived mQTL robustly associated with these CpG loci based on the Accessible Resource for Integrated Epigenomic Studies (ARIES) project database (mQTLdb). The ARIES initiative utilized the Illumina Infinium HumanMethylation450 (HM450) BeadChip for epigenetic data and the Illumina Infinium Human Hap550/660-w quad SNP genotyping platform for genetic data in 1018 mother–offspring pairs from the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort. Matrix eQTL software conducted preliminary SNP-CpG association analysis, followed by exact linear regression and genome-wide complex trait analysis (GCTA) for significantly independent mQTLs related to each CpG site. Identified mQTLs at middle‐age time points were further analyzed for subsequent methylation MR.

Statistical analyses

Heritability and genetic correlation

LDSC methods have proven effective in estimating genetic correlations across multiple traits or diseases [32]. We used the ldscr package to assess single-trait SNP heritability (h2) and cross-trait genetic correlations (rg), including all exposures and outcomes. The analysis began with preparing and cleaning GWAS summary statistics for each trait. LD scores were computed using reference panel data to quantify the extent of LD across the genome. SNP-based heritability was then calculated using a univariate LDSC model to estimate the proportion of phenotypic variance explained by the SNPs. For cross-trait genetic correlations, a bivariate LDSC model was employed, leaving the intercept unconstrained to account for potential confounders such as population stratification. The genetic correlations (rg) between different phenotypes were estimated by fitting the bivariate LDSC model to the GWAS summary statistics of the exposures and outcomes. The estimated rg values were interpreted to understand the genetic overlap between different traits.

Mendelian randomization

TwoSampleMR package (version 0.5.7) and MR pleiotropy residual sum and outlier (MR-PRESSO) package for TSMR and MVMR, respectively, allowed us to assess causality through a variety of models and to perform sensitivity analyses on the results [31]. The analysis relies on the inverse variance weighted (IVW) method for unbiased and effective causal estimates. This method assumes the validity of all SNPs as IVs and estimates the causal effect by regressing the SNP-ECG measure on the SNP-alcohol and tobacco effect sizes, weighted by the inverse of the SNP-ECG marker standard error. Robust analyses were implemented using the MR-Egger regression, weighted median models, and the radial MR framework [42]. Radial MR, particularly leveraging a Cochran’s Q-statistic-based weighted regression method, generates radial plots to transform data, providing immediate visual diagnostics and robust weighting that can detect subtle outlier effects often overlooked by traditional statistical methods [42]. Potential directional horizontal pleiotropy was addressed using the Egger intercept test and the MR-Robust Adjusted Profile Score (MR-RAPS) [43]. MR-RAPS generated a consistent and asymptotically normal estimate by adjusting for profile scores, with ANOVA results (P > 0.05), confirming that IVs weights were independent of regression residuals. This suggests effective control of potential pleiotropy, thereby enhancing the reliability of causal effect estimation. Furthermore, additional sensitivity analyses, including heterogeneity assessments and leave-one-out analysis, were performed to ensure the robustness of the findings. When evaluating the causal effects of DNA methylation at alcohol-related CpG sites, the method varies based on the number of associated mQTLs. For CpG sites linked to a single mQTL, only the Wald ratio and its corresponding SE are computed. In contrast, if a CpG site is associated with multiple independent mQTLs, the aforementioned IVW method is applied. To address multiple tests, we employed Bonferroni correction.

Fine mapping under a single causal variant assumption

For those individual CpG loci convincingly associated with ECG marker changes (Bonferroni corrected P < 0.05/number of CpG loci), we performed an approximate Bayes factor colocalization analysis to finely map each feature under a single causal variant assumption. All available mQTLs for the CpG loci of interest were obtained from the Genetics of DNA Methylation Consortium (GoDMC) [44], which includes a diverse range of 36 cohorts from various studies and populations, such as ALSPAC, Brisbane Systems Genetics Study (BSGS), and GSK, providing a valuable resource for examining genetic influences on DNA methylation. These mQTL data were then integrated with the GWAS summary data. Colocalization analyses using the "coloc" R package were designed to investigate whether the association of these CpG locus-associated mQTL with ECG marker change was driven by shared causal variation. This analysis tested the posterior probability of five hypotheses: H0 (no causally variable loci), H1 (only trait 1 has causally variable loci), H2 (only trait 2 has causally variable loci), H3 (each of the two traits has a different causally variable locus), and H4 (the two traits share a causally variable locus). The results provide the Bayes factor and posterior probability for each SNP, and we considered an H4 posterior probability of 95% or higher for a single mQTL as evidence of colocalization.

SMR

SMR is a robust methodology for investigating the shared genetic basis between traits using summary statistics from GWAS and QTL [34]. We used GWAS and eQTL data to detect associations between trait-associated SNPs and gene expression and applied the heterogeneity in dependent instrument (HEIDI) test to identify potential horizontal pleiotropy in causal associations. Whole-blood eQTL gene expression was obtained from cis-eQTL summary data from eQTLGen Consortium, while heart left ventricle eQTL data were from the GTEx project (v8, only SNPs with P < 1 × 10−5 are included). We utilized SMR software (version 1.3.1) with default parameters, configuring the cis window to 1 Mb, the MAF threshold to 0.01, and requiring at least one cis-eQTL at P < 5 × 10−8. The hg19 reference genome and Ensembl were used for gene annotation. Significant SMR probes were identified using Bonferroni-corrected thresholds (0.05/number of probes), and the HEIDI test P value thresholds > 0.05 were considered indicative of a lack of heterogeneity.

Result

Genetic risk effects of drinking, smoking, and ECG indices

We assessed the heritability and genetic associations of smoking, alcohol consumption, and ECG indices using univariate and bivariate LDSC. The phenotype of cigarette and alcohol consumption could be explained by 3.1–9.3% genetic background, whereas PWD, PR, and QT interval showed higher genetic effects from 10.5 to 19.3%. Regarding cross-trait genetic correlations, we observed significant genetic correlations between AUD and QT interval (rg = 0.126, P = 0.002), SmkInit and QT interval (rg = 0.096, P = 0.002), and CigDay and QT interval (rg = 0.184, P = 2.04E-6) (Table S1). Before performing the TSMR analysis, we evaluated bias and Type I error rates related to sample overlap using the Burgess et al. tool (https://sb452.shinyapps.io/overlap/) under the null hypothesis. By inputting exposure and outcome sample sizes, we estimated overlap proportions (0.0–1.0) with default settings. The Type I error rate remained stable at 0.05, with no inflation, and the bias was consistently 0.000, confirming the reliability of our estimation method. When MR analyses were performed, we identified 10 (F-statistic: 42.17–286.69), 10 (F-statistic: 30.68–71.09), 29 (F-statistic: 39.88–484), 16 (F-statistic: 31.38–790.45), and 46 (F-statistic: 34.34–103.69) genetic instruments for DrnkWk, AUD, PAU, CigDay, and SmkInit, respectively (Table S2). IVW results showed that genetically predicted PAU was associated with longer PWD [β (95%CI): 11.64 (8.25, 15.03), P = 1.76E-11], and SmkInit was associated with shorter PR intervals [β (95%CI): −1.23 (−2.12, −0.33), P = 0.007]. Notably, the QT interval was more significantly affected by the genetic predisposition to tobacco and alcohol consumption, with three of the five exposures significantly prolonging the QT interval [AUD, β (95%CI): 1.35 (0.24, 2.46), P = 0.017; DrnkWk, β (95%CI): 1.06 (0.91, 5.05), P = 0.005; SmkInit, β (95%CI): 2.06 (0.32, 3.80), P = 0.020] (Fig. 2).

Fig. 2
figure 2

Mendelian randomization (MR) estimates the causal effect between tobacco and alcohol consumption and cardiac ECG indexes. A MR estimates are evaluated by the inverse variance weighted (IVW) method. B Scatterplot for MR-estimated individual variant regression coefficients for tobacco and alcohol consumption on PWD. C Scatterplot for MR-estimated individual variant regression coefficients for tobacco and alcohol consumption on PR interval. D Scatterplot for MR-estimated individual variant regression coefficients for tobacco and alcohol consumption on QT interval. PWD, P-wave duration; AUD, alcohol use disorders; PAU, problematic alcohol use; DrnkWk, drinks per week; CigDay, cigarettes per day; SmkInit, smoking initiation; β, effect sizes; 95%CI, 95% confidence interval

Full size image

To gauge the stability of the TSMR results, multiple robust models were further used to provide causal estimates consistent with IVW. The casual correlation between PAU and longer PWD passed the test of weighted median models and radial IVW, but failed MR-Egger's assumptions [β (95%CI): −0.17 (−24.28, 23.93), P = 0.990]. In contrast, MR-Egger suggested a weak correlation between SmkInit and longer PR intervals, but weighted median models radial IVW did not provide significant causal inference (Fig. 3A and Table S3). The radial plot revealed that no influential outliers were affecting the causal relationship between PAU and PWD. However, 9 potential outlier points were detected in the association between SmkInit and longer PR intervals (Fig. 3B). The outlier-corrected analysis (radial Egger corrected) maintained the negative result consistent with other robust methods (Fig. 3A). For QT intervals, genetic predispositions to AUD and SmkInit were found to have potential outliers, with 1 and 5 outlier points, respectively. Although the causal relationship remained significant after outlier removal, this association between genetically determined AUD, SmkInit, and prolonged QT intervals was almost exclusively rejected in other sensitivity analyses due to inconsistent causal directions (Fig. 3A). Nevertheless, all robust analytical models, including radial MR outlier correction, supported a causal effect between DrnkWk and longer QT intervals [βMR-Egger (95% CI): 4.59 (0.85, 8.33), P = 0.018; βpenalized weighted median (95% CI): 5.05 (2.00, 8.10), P = 0.001; βRadial IVW (95%CI): 2.98 (0.91, 5.05), P = 0.005] (Fig. 3 and Table S3). Importantly, after adjusting for CigDay and SmkInit, the MVMR analysis substantiated that the DrnkWk is causally linked to prolonged QT intervals (Table 2). Additionally, examinations based on the MR-Egger intercept and MR-RAPS tests indicated the absence of directional horizontal pleiotropy (P > 0.05) for DrnkWk and QT interval, further suggesting adherence to the MR exclusive restriction assumption in the aforementioned analyses (Table 3). Leave-one-out analysis showed no single SNP significantly impacted results and no heterogeneity was observed (IVW: P = 0.061) (Fig. 4A and B). MR-RAPS diagnostics (ANOVA: P = 0.221) indicated IV weights were independent of residuals, ensuring unbiased estimates. Residual and Q-Q plots validated normality assumptions, supporting reliable MR findings (Fig. 4C).

Fig. 3
figure 3

Robust Mendelian randomization (MR) estimates the causal relationship between tobacco and alcohol consumption and ECG indices. A Forest plot of causal results assessed by applying robust MR methods. B Radial MR plots showing ratio estimates for each genetic variant as well as overall MR estimates, and variant outliers that affect causal estimates. PWD, P-wave duration; AUD, alcohol use disorders; PAU, problematic alcohol use; DrnkWk, drinks per week; SmkInit, smoking initiation; PWM, penalized weighted median; radial IVW, radial inverse variance weighted; β, effect sizes; 95%CI, 95% confidence interval. Radial Egger corrected represents outlier(s)-corrected analyses

Full size image
Table 2 Multivariable Mendelian randomization analysis adjusted for smoking phenotypes
Full size table
Table 3 Directional horizontal pleiotropy test
Full size table
Fig. 4
figure 4

MR sensitivity analysis of the causal relationship between DrnkWk and prolonged QT intervals. A Leave-one-out analysis identifying the effect of individual variables on the total effect. B Funnel plot showing the results of the heterogeneity analysis of the instrumental variable (IV) for causal estimation; C MR-RAPS diagnostic plots of the association between DrnkWk and QT interval prolongation, including residual plots showing the IV weights versus the standardized residuals and Q-Q plots assessing normality

Full size image

The causal effect of alcohol‐related DNA methylation CpG loci on QT interval predisposition

Given the observed causal relationship between DrnkWk and prolonged QT intervals, and the fact that EWAS suggests that drinking influences DNA methylation in blood and tissues, and that aberrant epigenetic mechanisms (patterns of DNA methylation) potentially modulate a wide range of cardiovascular disorders, including cardiac arrhythmias [26,27,28], using GCTA results, we identified 65 SNPs associated with 56 CpG sites from the mQTLdb database, all of them belonging to cis-mQTL (Table S4). These mQTL were integrated with the QT interval GWAS to be included in the methylation MR analyses. Eight CpG loci (cg03345232, cg04605617, cg07091481, cg07104958, cg07512517, cg07567724, cg12807764, and cg20374917) were identified as causally associated with QT interval susceptibility, achieving nominal significance at the P < 0.05 level; two of these loci, cg03345226 and cg04605617, in particular, survived Bonferroni correction (P < 0.05/56 = 8.93E-04) (Fig. 5A, Tables 4 and S5). The CpG site cg04605617 was most significantly associated with QT interval prolongation [β (95%CI): 0.175 (0.13, 0.22), P = 1.82E-13] and mapped to the promoter region of the protein-coding gene Phospholipase A2 Group IIC (PLA2G2C) (chr1: 20,501,558) (Fig. 5B). Although the function of PLA2G2C is poorly understood, it is predicted to enable calcium ion binding activity. To investigate whether the two mQTLs rs10916683 and rs2872909 of cg04605617 affect the expression of the PLA2G2C gene in blood and or heart, we queried it on eQTLGen and GTEx Portal and found rs10916683 to be a strong eQTL for the PLA2G2C gene in whole blood (P = 9.36E-16) (Table S6). Another CpG locus cg03345232 exhibited a significant causal relationship with a shortened QT interval [β (95%CI): −0.435 (−0.657, −0.213), P = 1.22E-04], and it was mapped to the protein-coding gene Ras and Rab interactor 3 (RIN3). Nevertheless, we did not find evidence supporting the classification of either of the two mQTLs, rs77826962, or rs12884739 at this locus, as eQTLs influence gene expression (Table S6).

Fig. 5
figure 5

Epigenetic MR estimates and colocalization analysis. A Epigenetic causal estimates of significant CpG loci for alcohol-related methylation on QT interval susceptibility. B Genomic localization of DNA methylation site: CpG island cg04605617.cg04605617 localizes to chromosome 1 at position 20,501,558 and maps to the promoter region of the gene PLA2G2C. C Regional map of evidence for colocalization of methylated CpG sites with QT interval susceptibility. 99.7% posterior probability supports that rs12881206 is a shared causal variant for cg03345232 methylation and QT interval susceptibility. IVW, radial inverse variance weighted; mQTL, methylation quantitative trait loci

Full size image
Table 4 Causal estimates of significant CpG loci for alcohol-related methylation on QT interval susceptibility by epigenetic Mendelian randomization
Full size table

Genetic basis shared between alcohol consumption and QT interval susceptibility

For the convincing causal relationship between the CpG locus and QT interval alteration described above, we performed an approximate Bayes factor colocalization analysis to replicate the MR findings and attempted to find individual causal variants mediating this relationship. We obtained a broader spectrum of mQTLs associated with cg03345232 and cg04605617 in the GoDMC dataset. Figure 5C shows that methylation at cg03345232 shared a genetic causal variant, rs12881206, with a 99.7% posterior probability of H4 (PPH4) with QT GWAS signaling. However, for the mQTLs at cg04605617, the colocalization analysis did not reveal a single shared variant site with a posterior probability exceeding 95.0%. Furthermore, we applied SMR integrating eQTL and GWAS summary data, to identify putative functional genes underlying the association between DrnkWk and QT interval. We identified 77 and 33 significant variants in whole-blood eQTL for DrnkWk and QT intervals, respectively (PDrnkWk < 0.05/15629 = 3.20E-06, PQT < 0.05/15617 = 3.20E-06), and 19 significant variants (P < 0.05/4662 = 1.07E-05) were found for QT intervals in heart left ventricle (Table S7). Among these variants, we observed that DrnkWk and QT intervals share a protein-coding gene, NPIPB6, and a long non-coding RNA gene, NONHSAG018995.2 (ENSG00000251417). Unfortunately, both of these variants did not pass the horizontal pleiotropy test by HEIDI (P < 0.05) (Table S7).

Discussion

This study presents evidence of a genetic association and causality between tobacco and alcohol consumption and ECG indices. It sheds light on the molecular genetic architecture underlying alcohol consumption's impact on QT interval prolongation and provides valuable insights into their regulatory mechanisms.

The evaluation of genetic associations across traits supports the hypothesis that genetic factors play a crucial role in the association between tobacco and alcohol consumption and ECG indices. Substantial genetic causal associations between tobacco and alcohol exposure and cardiac electrophysiological changes were highlighted by MR analysis. Using multiple robust MR hypothesis models, we provided compelling causal evidence in univariate MR analyses between weekly alcohol consumption and QT interval prolongation. To maintain the direction and significance of the causal effects, we employ stringent criteria, such as MR-Egger and weighted median models, to ensure the robustness of the MR estimates and to reduce the Type 1 error rate. MVMR further confirmed the independence of this genetic association from the smoking phenotype, eliminating the confounding effect of genetic correction between smoking and alcohol phenotypes. Epigenetic MR analyses revealed a significant causal relationship between alcohol-related DNA methylation and QT interval prolongation. Eight CpG loci were identified as potential mediators of this causality, with two surviving Bonferroni corrections, including cg04605617 mapping to the promoter region of PLA2G2C. The mQTL rs10916683 at cg04605617 is a robust eQTL for PLA2G2C. Despite SMR not identifying shared functional genes without pleiotropy for alcohol-induced QT interval prolongation, the colocalization analysis revealed an mQTL, rs12881206, shared with a high posterior probability between cg03345232 methylation and QT GWAS signals. They underscore the potential regulatory role of DNA methylation in modulating QT intervals.

These results align with observational studies indicating a positive correlation between alcohol consumption and QT interval prolongation, suggesting a potential increased arrhythmia risk by affecting ventricular repolarization [25, 38, 45, 46]. Importantly, the identification of the CpG locus cg04605617 in the promoter of the relatively obscure gene PLA2G2C provides an intriguing avenue for cardiovascular genetic exploration. PLA2G2C encodes a calcium-dependent phospholipase belonging to the phospholipase A2 (PLA2) family, contributing to signal transduction, membrane homeostasis, and immune regulation [47,48,49]. Although limited, polymorphisms in PLA2G2C sense changes in plasma triglycerides, and functional experiments have found that it produces a lysophospholipid antigen in mouse hepatocytes, leading to the propagation of an antiviral immune response via NKT cells [50,51,52]. Moreover, allergy and inflammation appear to induce the expression of PLA2, including PLA2G2C [53]. Other members of the PLA2 family, such as SPLA2-IIA, have also been associated with cardiovascular diseases, particularly atherosclerosis. Elevated levels of SPLA2-IIA are linked to an increased cardiovascular risk, likely due to its role in promoting inflammation within the arterial wall [54]. Given the structural similarities, PLA2G2C may also affect similar pathways; however, direct evidence supporting this hypothesis is currently lacking. Notably, EWAS reveals that methylation at the cg04605617 site within PLA2G2C is significantly linked to high arsenic exposure, hinting at a potential association with cardiovascular disease [55]. We also note that PLA2G2C is functionally annotated in Gene Ontology to enable calcium binding [56], and that the significant PLA2G2C mQTL and eQTL identified, rs10916683, are also notable polymorphic sites where its homolog, PLA2G2A, functions [57]. The observed causal link between cg04605617 methylation and QT interval prolongation suggests that alterations in DNA methylation at this locus may impact PLA2G2C expression and calcium ion kinetics in cardiac cells, which are integral to cardiac excitation–contraction coupling and have profound effects on cardiac electrophysiology [58, 59]. However, the precise mechanistic contributions remain incompletely understood. Determining how methylation at this specific locus regulates PLA2G2C expression could uncover novel epigenetic regulatory networks governing cardiac electrical activity. Functional validation studies addressing these gaps would significantly enhance the translational relevance of the current findings.

Furthermore, we identified rs12881206 between alcohol consumption and QT interval predisposition with a posteriori probability exceeding 0.95 to match the H4 hypothesis from colocalization analysis, implying that they share the same causal variant in this genomic region. Upon querying dbSNP, we identified this polymorphic locus as an intronic variant of RIN3 on chromosome 14, previously undocumented for clinical significance. However, RIN3 is a guanine nucleotide exchange factor that activates Ras and Rab5 proteins [60]. Given their roles in cell signaling and endocytosis regulation [60], further exploration of the potential mechanistic links between these findings has the potential to uncover new insights into cardiovascular risk factors. In contrast, this study did not identify a definitive genetic causal relationship between smoking and ECG alterations. Nevertheless, it is imperative to emphasize that smoking constitutes a well-established role as a cardiovascular risk factor through diverse mechanisms such as inflammation, oxidative stress, and endothelial dysfunction [61]. Consequently, the research findings should not be misconstrued as indicating that smoking is benign for cardiovascular health. Statistical significance should not be confused with biological relevance, and caution should be exercised in attributing causality based solely on observational data and statistical associations. Previous studies have reported associations between smoking and ECG indices, particularly QT interval prolongation [17, 20]. While CigDay showed a strong genetic correlation with QT interval, MR did not confirm causality, underscoring the need to distinguish correlation from causation. LDSC quantifies the overall genetic overlap between traits by assessing genome-wide correlations, which may arise from shared genetic influences rather than direct causal pathways. In contrast, MR leverages genetic variants as IVs to estimate the direct effect of an exposure on an outcome, minimizing confounding and reverse causation. By relying on independent genetic instruments, MR helps distinguish true causal relationships from associations driven by pleiotropy or indirect genetic linkages, providing stronger evidence for causal inference. The strengths of our study lie in leveraging large-scale, high-quality GWAS data, employing multiple robust MR models to ensure stable causal inference, and using mQTL datasets for genetic co-localization and SMR analyses to explore potential sharing mechanisms. However, limitations include the use of blood DNA samples, which may not fully represent tissue-specific methylation profiles. The reliance on mQTL effects from a specific demographic and age group (middle-aged time point), along with the European pedigree in MR analyses, raises concerns about generalizability. Our MR analyses, primarily in European ancestry individuals, enhance population homogeneity but limit generalizability to other populations. Additionally, the CpG loci associated with alcohol consumption in our study were derived from the latest EWAS in the EWAS Catalog, which linked alcohol use to specific methylation sites. The mQTL data for these loci were obtained from ALSPAC. However, due to differences in alcohol consumption rates across cohorts, while the mQTL-CpG associations are robust, the extent of behavioral influence across cohorts is difficult to estimate [62]. These limitations underscore the need for diversified populations and tissues in future research to enhance the validity and applicability of the findings. In conclusion, our study reveals a potential genetic mechanism for alcohol-induced QT interval prolongation, identifying the CpG locus cg04605617 in the promoter region of the little-known PLA2G2C gene. This suggests that alcohol-related DNA methylation may influence the QT interval through a novel pathway. The presence of rs10916683 as a significant eQTL variant implies broader systemic effects of alcohol on the cardiovascular system. However, the limited understanding of the biological function of the CpG locus and its associated shared variant (rs12881206) raises questions about its direct impact on cardiac physiology. Despite these limitations, our findings provide valuable insights into the complex genetic architecture linking alcohol consumption to specific cardiac electrophysiologic changes. The identification of a causal relationship between alcohol consumption and QT interval prolongation underscores the importance of lifestyle factors in cardiovascular health. Specifically, the study highlights the significant role that genetic and epigenetic mechanisms play in mediating the effects of alcohol on cardiac electrical activity. Understanding these pathways can guide clinical interventions aimed at reducing alcohol-related cardiac risks. Identifying specific DNA methylation sites like cg04605617 mapped to PLA2G2C offers potential biomarkers for risk stratification. Clinically, these findings support routine ECG monitoring in individuals with heavy alcohol use and genetic predisposition to prolonged QT intervals. Moreover, epigenetic interventions, such as targeted methylation modulators, could be explored as a novel therapeutic approach for arrhythmia prevention. To further validate and expand these findings, future studies should replicate the associations in larger, multi-ethnic cohorts with broader age ranges. Functional experiments in vitro or animal models could elucidate the underlying mechanisms, and prospective clinical studies evaluating changes in alcohol consumption on methylation profiles and ECG parameters would strengthen the causal inference. Importantly, although blood-based mQTLs provide useful insights into systemic regulatory mechanisms, their relevance to cardiac-specific methylation remains uncertain. Tissue-specific differences in DNA methylation patterns could influence the functional interpretation of our findings. Future studies integrating cardiac tissue-derived mQTL and single-cell epigenomic data will be crucial to validating whether methylation changes at PLA2G2C directly impact QT interval regulation.

Conclusion

This study reveals a causal link between alcohol consumption and QT interval prolongation through specific genetic and epigenetic mechanisms. The identification of cg04605617 within PLA2G2C as a candidate locus offers new insights into the regulatory mechanisms underlying cardiac electrophysiology. However, given the tissue-specific nature of epigenetic modifications, further validation in cardiac tissues is warranted. These findings contribute to the broader understanding of alcohol’s influence on cardiac function and may inform future research into epigenetic-based therapeutic strategies.

Availability of data and materials

No datasets were generated or analyzed during the current study.

References

  1. Joukar S. A comparative review on heart ion channels, action potentials and electrocardiogram in rodents and human: extrapolation of experimental insights to clinic. Lab Anim Res. 2021;37(1):25. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s42826-021-00102-3.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Narayanan K, Chugh SS. The 12-lead electrocardiogram and risk of sudden death: current utility and future prospects. Europace. 2015. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/europace/euv121.

    Article  PubMed  Google Scholar 

  3. Breijo-Márquez FR, Pardo Ríos M, Alcaraz BM. Association of short PR interval, long QT interval and sudden cardiac death in a young male. Rev Esp Cardiol. 2010;63(3):362–4. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s1885-5857(10)70071-2.

    Article  PubMed  Google Scholar 

  4. Shimizu W, Makimoto H, Yamagata K, et al. Association of genetic and clinical aspects of congenital long QT syndrome with life-threatening arrhythmias in Japanese patients. JAMA Cardiol. 2019;4(3):246–54. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jamacardio.2018.4925.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Varró A, Tomek J, Nagy N, et al. Cardiac transmembrane ion channels and action potentials: cellular physiology and arrhythmogenic behavior. Physiol Rev. 2021;101(3):1083–176. https://doiorg.publicaciones.saludcastillayleon.es/10.1152/physrev.00024.2019.

    Article  CAS  PubMed  Google Scholar 

  6. Marbán E. Cardiac channelopathies. Nature. 2002;415(6868):213–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/415213a.

    Article  PubMed  Google Scholar 

  7. Osadchii OE. Role of abnormal repolarization in the mechanism of cardiac arrhythmia. Acta Physiol (Oxf). 2017;220(Suppl 712):1–71. https://doiorg.publicaciones.saludcastillayleon.es/10.1111/apha.12902.

    Article  CAS  PubMed  Google Scholar 

  8. Viskin S. Long QT syndromes and torsade de pointes. Lancet. 1999;354(9190):1625–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/s0140-6736(99)02107-8.

    Article  CAS  PubMed  Google Scholar 

  9. Prifti E, Fall A, Davogustto G, et al. Deep learning analysis of electrocardiogram for risk prediction of drug-induced arrhythmias and diagnosis of long QT syndrome. Eur Heart J. 2021;42(38):3948–61. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/eurheartj/ehab588.

    Article  PubMed  Google Scholar 

  10. Gajendragadkar PR, Von Ende A, Ibrahim M, et al. Assessment of the causal relevance of ECG parameters for risk of atrial fibrillation: a mendelian randomisation study. PLoS Med. 2021;18(5):e1003572. https://doiorg.publicaciones.saludcastillayleon.es/10.1371/journal.pmed.1003572.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Young WJ, Lahrouchi N, Isaacs A, et al. Genetic analyses of the electrocardiographic QT interval and its components identify additional loci and pathways. Nat Commun. 2022;13(1):5144. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41467-022-32821-z.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Wojakowski A, Izbizky G, Carcano ME, et al. Fetal Doppler mechanical PR interval: correlation with fetal heart rate, gestational age and fetal sex. Ultrasound Obstet Gynecol. 2009;34(5):538–42. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/uog.7333.

    Article  CAS  PubMed  Google Scholar 

  13. Ahnve S. Correction of the QT interval for heart rate: review of different formulas and the use of Bazett’s formula in myocardial infarction. Am Heart J. 1985;109(3 Pt 1):568–74. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/0002-8703(85)90564-2.

    Article  CAS  PubMed  Google Scholar 

  14. Chung MK, Eckhardt LL, Chen LY, et al. Lifestyle and risk factor modification for reduction of atrial fibrillation: a scientific statement from the American heart association. Circulation. 2020;141(16):e750–72. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/cir.0000000000000748.

    Article  PubMed  Google Scholar 

  15. Said MA, Verweij N, van der Harst P. Associations of combined genetic and lifestyle risks with incident cardiovascular disease and diabetes in the UK biobank study. JAMA Cardiol. 2018;3(8):693–702. https://doiorg.publicaciones.saludcastillayleon.es/10.1001/jamacardio.2018.1717.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Ye Y, Chen X, Han J, et al. Interactions between enhanced polygenic risk scores and lifestyle for cardiovascular disease, diabetes, and lipid levels. Circ Genom Precis Med. 2021;14(1):e003128. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circgen.120.003128.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Ruedisueli I, Lakhani K, Nguyen R, et al. Electronic cigarettes prolong ventricular repolarization in people who smoke tobacco cigarettes: implications for harm reduction. Am J Physiol Heart Circ Physiol. 2023;324(6):H821–32. https://doiorg.publicaciones.saludcastillayleon.es/10.1152/ajpheart.00057.2023.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Brunner S, Herbel R, Drobesch C, et al. Alcohol consumption, sinus tachycardia, and cardiac arrhythmias at the Munich octoberfest: results from the Munich beer related electrocardiogram workup study (MunichBREW). Eur Heart J. 2017;38(27):2100–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/eurheartj/ehx156.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Brunner S, Drobesch C, Herbel R, et al. Effects of acute alcohol consumption on cardiac excitation, conduction, and repolarization: results from the Munich beer related electrocardiogram workup study (MunichBREW). Clin Res Cardiol. 2021;110(6):916–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00392-021-01839-6.

    Article  PubMed  PubMed Central  Google Scholar 

  20. D’Alessandro A, Boeckelmann I, Hammwhöner M, et al. Nicotine, cigarette smoking and cardiac arrhythmia: an overview. Eur J Prev Cardiol. 2012;19(3):297–305. https://doiorg.publicaciones.saludcastillayleon.es/10.1177/1741826711411738.

    Article  PubMed  Google Scholar 

  21. Gepner AD, Piper ME, Leal MA, et al. Electrocardiographic changes associated with smoking and smoking cessation: outcomes from a randomized controlled trial. PLoS ONE. 2013;8(4):e62311. https://doiorg.publicaciones.saludcastillayleon.es/10.1371/journal.pone.0062311.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Manolis TA, Apostolopoulos EJ, Manolis AA, et al. The proarrhythmic conundrum of alcohol intake. Trends Cardiovasc Med. 2022;32(4):237–45. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.tcm.2021.03.003.

    Article  PubMed  Google Scholar 

  23. Ghadri JR, Templin C, Duru F, et al. Holiday heart block: alcohol-induced PR prolongation. Am J Med. 2013;126(9):776–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.amjmed.2013.05.007.

    Article  CAS  PubMed  Google Scholar 

  24. Lekx AW, Lingius S, Barten DG. Second-degree atrioventricular block in an adolescent with an acute alcohol intoxication. Am J Emerg Med. 2020;38(2):407.e1-407.e3. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ajem.2019.158419.

    Article  PubMed  Google Scholar 

  25. de Veld L, van der Lely N, Hermans BJM, et al. QTc prolongation in adolescents with acute alcohol intoxication. Eur J Pediatr. 2022;181(7):2757–70. https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s00431-022-04471-2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Nascimento LV, Neto FL, Ribeiro Moreira DA, et al. Influence of antidepressant drugs on DNA methylation of ion channels genes in blood cells of psychiatric patients. Epigenomics. 2022;14(14):851–64. https://doiorg.publicaciones.saludcastillayleon.es/10.2217/epi-2022-0089.

    Article  CAS  PubMed  Google Scholar 

  27. Wang M, Tu X. The genetics and epigenetics of ventricular arrhythmias in patients without structural heart disease. Front Cardiovasc Med. 2022;9:891399. https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fcvm.2022.891399.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Zhong R, Zhang F, Yang Z, et al. Epigenetic mechanism of L-type calcium channel β-subunit downregulation in short QT human induced pluripotent stem cell-derived cardiomyocytes with CACNB2 mutation. Europace. 2022;24(12):2028–36. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/europace/euac091.

    Article  PubMed  Google Scholar 

  29. Ni G, Moser G, Wray NR, et al. Estimation of genetic correlation via linkage disequilibrium score regression and genomic restricted maximum likelihood. Am J Hum Genet. 2018;102(6):1185–94. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.ajhg.2018.03.021.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Saunders GRB, Wang X, Chen F, et al. Genetic diversity fuels gene discovery for tobacco and alcohol use. Nature. 2022;612(7941):720–4. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41586-022-05477-4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Rasooly D, Patel CJ. Conducting a reproducible Mendelian randomization analysis using the R analytic statistical environment. Curr Protoc Hum Genet. 2019;101(1):e82. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/cphg.82.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Grotzinger AD, Rhemtulla M, de Vlaming R, et al. Genomic structural equation modelling provides insights into the multivariate genetic architecture of complex traits. Nat Hum Behav. 2019;3(5):513–25. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41562-019-0566-x.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Giambartolomei C, Vukcevic D, Schadt EE, et al. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet. 2014;10(5):e1004383. https://doiorg.publicaciones.saludcastillayleon.es/10.1371/journal.pgen.1004383.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Zhu Z, Zhang F, Hu H, et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat Genet. 2016;48(5):481–7. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/ng.3538.

    Article  CAS  PubMed  Google Scholar 

  35. Liu M, Jiang Y, Wedow R, et al. Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nat Genet. 2019;51(2):237–44. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41588-018-0307-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Zhou H, Sealock JM, Sanchez-Roige S, et al. Genome-wide meta-analysis of problematic alcohol use in 435,563 individuals yields insights into biology and relationships with other traits. Nat Neurosci. 2020;23(7):809–18. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41593-020-0643-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Ntalla I, Weng LC, Cartwright JH, et al. Multi-ancestry GWAS of the electrocardiographic PR interval identifies 202 loci underlying cardiac conduction. Nat Commun. 2020;11(1):2542. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41467-020-15706-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Nauffal V, Morrill VN, Jurgens SJ, et al. Monogenic and polygenic contributions to QTc Prolongation in the population. Circulation. 2022;145(20):1524–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circulationaha.121.057261.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Christophersen IE, Magnani JW, Yin X, et al. Fifteen genetic loci associated with the electrocardiographic P wave. Circ Cardiovasc Genet. 2017. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circgenetics.116.001667.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–64. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyr036.

    Article  PubMed  Google Scholar 

  41. Liu C, Marioni RE, Hedman ÅK, et al. A DNA methylation biomarker of alcohol consumption. Mol Psychiatry. 2018;23(2):422–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/mp.2016.192.

    Article  CAS  PubMed  Google Scholar 

  42. Bowden J, Spiller W, Del Greco MF, et al. Improving the visualization, interpretation and analysis of two-sample summary data Mendelian randomization via the Radial plot and radial regression. Int J Epidemiol. 2018;47(4):1264–78. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyy101.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Zhao Q, Wang J, Hemani G, et al. Statistical inference in two-sample summary-data Mendelian randomization using robust adjusted profile score. Ann Statist. 2020;48:1742–69. https://doiorg.publicaciones.saludcastillayleon.es/10.1214/19-AOS1866.

    Article  Google Scholar 

  44. Min JL, Hemani G, Hannon E, et al. Genomic and phenotypic insights from an atlas of genetic effects on DNA methylation. Nat Genet. 2021;53(9):1311–21. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s41588-021-00923-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Platiša MM, Gal V, Nestorović Z, et al. Quantification of the acute effect of a low dose of red wine by nonlinear measures of RR and QT interval series in healthy subjects. Comput Biol Med. 2014;53:291–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.compbiomed.2014.08.015.

    Article  PubMed  Google Scholar 

  46. Lee AS, Sung YL, Pan SH, et al. A common East Asian aldehyde dehydrogenase 2*2 variant promotes ventricular arrhythmia with chronic light-to-moderate alcohol use in mice. Commun Biol. 2023;6(1):610. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/s42003-023-04985-x.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Murakami M. Novel functions of phospholipase A(2)s: overview. Biochim Biophys Acta Mol Cell Biol Lipids. 2019;1864(6):763–5. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.bbalip.2019.02.005.

    Article  CAS  PubMed  Google Scholar 

  48. Murakami M, Sato H, Taketomi Y. Modulation of immunity by the secreted phospholipase A(2) family. Immunol Rev. 2023;317(1):42–70. https://doiorg.publicaciones.saludcastillayleon.es/10.1111/imr.13205.

    Article  CAS  PubMed  Google Scholar 

  49. Murakami M, Taketomi Y, Miki Y, et al. Recent progress in phospholipase A2 research: From cells to animals to humans. Prog Lipid Res. 2011;50(2):152–92. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.plipres.2010.12.001.

    Article  CAS  PubMed  Google Scholar 

  50. Tischfield JA, Xia YR, Shih DM, et al. Low-molecular-weight, calcium-dependent phospholipase A2 genes are linked and map to homologous chromosome regions in mouse and human. Genomics. 1996;32(3):328–33. https://doiorg.publicaciones.saludcastillayleon.es/10.1006/geno.1996.0126.

    Article  CAS  PubMed  Google Scholar 

  51. Tremblay BL, Cormier H, Rudkowska I, et al. Association between polymorphisms in phospholipase A2 genes and the plasma triglyceride response to an n-3 PUFA supplementation: a clinical trial. Lipids Health Dis. 2015;14:12. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-015-0009-2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Zeissig S, Murata K, Sweet L, et al. Hepatitis B virus-induced lipid alterations contribute to natural killer T cell-dependent protective immunity. Nat Med. 2012;18(7):1060–8. https://doiorg.publicaciones.saludcastillayleon.es/10.1038/nm.2811.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Bickford JS, Mueller C, Newsom KJ, et al. Effect of allergy and inflammation on eicosanoid gene expression in CFTR deficiency. J Cyst Fibros. 2013;12(3):258–65. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jcf.2012.08.014.

    Article  CAS  PubMed  Google Scholar 

  54. Holmes MV, Simon T, Exeter HJ, et al. Secretory phospholipase A(2)-IIA and cardiovascular disease: a mendelian randomization study. J Am Coll Cardiol. 2013;62(21):1966–76. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.jacc.2013.06.044.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  55. Argos M, Chen L, Jasmine F, et al. Gene-specific differential DNA methylation and chronic arsenic exposure in an epigenome-wide association study of adults in Bangladesh. Environ Health Perspect. 2015;123(1):64–71. https://doiorg.publicaciones.saludcastillayleon.es/10.1289/ehp.1307884.

    Article  PubMed  Google Scholar 

  56. Gaudet P, Livstone MS, Lewis SE, et al. Phylogenetic-based propagation of functional annotations within the Gene ontology consortium. Brief Bioinform. 2011;12(5):449–62. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/bib/bbr042.

    Article  PubMed  PubMed Central  Google Scholar 

  57. Hoeft B, Linseisen J, Beckmann L, et al. Polymorphisms in fatty-acid-metabolism-related genes are associated with colorectal cancer risk. Carcinogenesis. 2010;31(3):466–72. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/carcin/bgp325.

    Article  CAS  PubMed  Google Scholar 

  58. Gardner RT, Ripplinger CM, Myles RC, et al. Molecular mechanisms of sympathetic Remodeling and arrhythmias. Circ Arrhythm Electrophysiol. 2016;9(2):e001359. https://doiorg.publicaciones.saludcastillayleon.es/10.1161/circep.115.001359.

    Article  PubMed  PubMed Central  Google Scholar 

  59. Varodayan FP, de Guglielmo G, Logrip ML, et al. Alcohol dependence disrupts Amygdalar L-type voltage-gated calcium channel mechanisms. J Neurosci. 2017;37(17):4593–603. https://doiorg.publicaciones.saludcastillayleon.es/10.1523/jneurosci.3721-16.2017.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Kajiho H, Sakurai K, Minoda T, et al. Characterization of RIN3 as a guanine nucleotide exchange factor for the Rab5 subfamily GTPase Rab31. J Biol Chem. 2011;286(27):24364–73. https://doiorg.publicaciones.saludcastillayleon.es/10.1074/jbc.M110.172445.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  61. de Faria HCF, Garcez A, da Costa JSD, et al. Overweight and obesity among Brazilian healthcare university students: prevalence and associated factors. Arch Endocrinol Metab. 2023;67(3):416–26. https://doiorg.publicaciones.saludcastillayleon.es/10.20945/2359-3997000000602.

    Article  PubMed  Google Scholar 

  62. Zhou X, Wang L, Xiao J, et al. Alcohol consumption, DNA methylation and colorectal cancer risk: Results from pooled cohort studies and Mendelian randomization analysis. Int J Cancer. 2022;151(1):83–94. https://doiorg.publicaciones.saludcastillayleon.es/10.1002/ijc.33945.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We appreciate the contributions of participants and investigators from GSCAN, FinnGen studies, CVDKP, NHGRI-EBI GWAS Catalog, MRC-IEU EWAS Catalog, mQTLdb, GoDMC, eQTLGen, and GTEx projects for sharing their data publicly. Their collaboration is critical to the success of this research and the advancement of scientific understanding.

Funding

This work was supported by the National Natural Science Foundation of China (NSFC) [Grant Number 82073659 and 82300347].

Author information

Author notes

  1. Zequn Zheng and Yongfei Song have contributed equally to this work.

Authors and Affiliations

  1. Department of Cardiology, First Affiliated Hospital of Shantou University Medical College, Shantou, Guangdong, China

    Zequn Zheng & Xuerui Tan

  2. Clinical Research Center, First Affiliated Hospital of Shantou University Medical College, Shantou, Guangdong, China

    Zequn Zheng & Xuerui Tan

  3. Institute of Innovation for Combined Medicine and Engineering, Ningbo Medical Center Lihuili Hospital, Ningbo University, Ningbo, Zhejiang, China

    Zequn Zheng & Yongfei Song

  4. Center for Molecular Medicine, Hangzhou Medical College, Hangzhou, Zhejiang, China

    Yongfei Song

  5. Department of Cardiology, Taizhou Hospital of Zhejiang Province affiliated to Wenzhou Medical University, Taizhou, Zhejiang, China

    Xinhan Li

  6. Center of Stem Cell and Regenerative Medicine, and Bone Marrow Transplantation Center of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, Zhejiang, China

    Tao Luo

  7. Human Phenome institute of SUMC, Guangdong Engineering Research Center of Human Phenome, Chemistry and Chemical Engineering Guangdong Laboratory, Shantou, Guangdong, China

    Zequn Zheng & Xuerui Tan

Authors
  1. Zequn Zheng
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yongfei Song
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Xinhan Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Tao Luo
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Xuerui Tan
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

ZZ and YS helped in conceptualization, methodology, data curation, formal analysis, resources, software, visualization, writing—original draft, funding acquisition. XL and TL contributed to methodology, investigation, formal analysis, supervision. XT helped in conceptualization, writing—review & editing, project administration, supervision, funding acquisition.

Corresponding author

Correspondence to Xuerui Tan.

Ethics declarations

Ethics approval and consent to participate

The data used in this study were sourced from publicly available GWAS and DNA methylation datasets, all of which received ethical approval in their original studies with informed consent from participants. As the data were anonymized and de-identified, no additional ethical approval was required for our reanalysis. We adhered to the ethical guidelines set by the original studies.

Consent for publication

All authors approved the final version and agreed to be responsible for the study.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zheng, Z., Song, Y., Li, X. et al. Dissecting the causal effects of smoking, alcohol consumption, and related DNA methylation markers on electrocardiographic indices. Clin Epigenet 17, 40 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-025-01851-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-025-01851-x

Keywords

Clinical Epigenetics

ISSN: 1868-7083

Contact us