Background Studies on DNA methylation have the potential to discover mechanisms of cardiovascular disease (CVD) risk. However, the role of DNA methylation in CVD etiology remains unclear. Results We performed an epigenome-wide association study (EWAS) on CVD in a longitudinal sample of Swedish twins (535 individuals). We selected CpGs reaching the Bonferroni-corrected significance level (2 $$\times$$ ×  10–7) or the top-ranked 20 CpGs with the lowest P values if they did not reach this significance level in EWAS analysis associated with non-stroke CVD, overall stroke, and ischemic stroke, respectively. We further applied a bivariate autoregressive latent trajectory model with structured residuals (ALT-SR) to evaluate the cross-lagged effect between DNA methylation of these CpGs and cardiometabolic traits (blood lipids, blood pressure, and body mass index). Furthermore, mediation analysis was performed to evaluate whether the cross-lagged effects had causal impacts on CVD. In the EWAS models, none of the CpGs we selected reached the Bonferroni-corrected significance level. The ALT-SR model showed that DNA methylation levels were more likely to predict the subsequent level of cardiometabolic traits rather than the other way around (numbers of significant cross-lagged paths of methylation → trait/trait → methylation were 84/4, 45/6, 66/1 for the identified three CpG sets, respectively). Finally, we demonstrated significant indirect effects from DNA methylation on CVD mediated by cardiometabolic traits. Conclusions We present evidence for a directional association from DNA methylation on cardiometabolic traits and CVD, rather than the opposite, highlighting the role of epigenetics in CVD development.

Aim: Alcohol intake alters DNA methylation profiles and methylation might mediate the association between alcohol and disease, but limited number of positive CpG sites repeatedly replicated. Materials & methods: In total, 57 monozygotic (MZ) twin pairs discordant for alcohol drinking from the Chinese National Twin Registry and 158 MZ and dizygotic twin pairs in the Swedish Adoption/Twin Study of Aging were evaluated. DNA methylation was detected using the Infinium HumanMethylation450 BeadChip. Results: Among candidate CpG sites, cg07326074 was significantly correlated with drinking after adjusting for covariates in MZ twins in both datasets but not in the entire sample or dizygotic twins. Conclusion: The hypermethylation of cg07326074, located in the tumor-promoting gene C16orf59, was associated with alcohol consumption.

Introduction Statins are lipid-lowering drugs and starting treatment has been associated with DNA methylation changes at genes related to lipid metabolism. However, the longitudinal pattern of how statins affect DNA methylation in relation to lipid levels has not been well investigated. Methods We conducted an epigenetic association study in a longitudinal Swedish twin sample in previously reported lipid-related CpGs (cg10177197, cg17901584 and cg27243685). First, we applied a mixed-effect model to assess the association between blood lipids (total cholesterol (TC), low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), total triglyceride (TG)) and DNA methylation. Then, we performed a piecewise latent linear–linear growth curve model (LGCM) to explore the long-term changing pattern of lipids and methylation in response to statin treatment. Finally, we used a bivariate autoregressive latent trajectory model with structured residuals (ALT-SR) to analyze the cross-lagged effects in different lipid-CpG pairs in statin users and non-users. Results We replicated the associations between TC, LDL, HDL and DNA methylation level in cg17901584 and cg27243685 (P values ranged from 4.70E−12 to 1.84E−04). From the piecewise LGCM, we showed that TC and LDL significantly decreased in statin users before treatment started and then remained stable. For non-statin users, we only found a slightly significant decreasing trend for TC and TG. We observed a similar dynamic pattern for methylation levels at cg27243685 and cg17901584. Before statin initiation, cg27243685 showed a significantly increasing trend and cg17901584 a decreasing trend, but post-treatment, there were no additional changes. From the ALT-SR model, we found TG levels to be significantly associated with the DNA methylation level of cg27243685 at the next measurement in statin users (estimate = 0.383, 95% CI: 0.173, 0.594, P value < 0.001). Conclusions Longitudinal blood lipid and DNA methylation levels change after statin treatment initiation, where the latter is mostly a response to alterations in lipid levels and not vice versa.