Congenital heart disease (CHD) remains one of the most common structural birth defects, and improving prenatal classification beyond imaging-based screening continues to be a clinical priority. A case-control sequencing study published in Clinical Epigenetics evaluated whether amniotic fluid DNA methylation patterns could distinguish fetuses with CHD from controls across three independent cohorts.

In this case-control study, 135 second-trimester amniotic fluid samples (80 CHD cases and 55 controls) were divided into three independent cohorts. Cohort I underwent whole-genome bisulfite sequencing (WGBS) to identify differentially methylated regions (DMRs). Candidate DMRs were evaluated by target bisulfite sequencing (TBS) in cohort II. Final biomarker selection was based on adjusted P-values, sequencing depth, group differences, and biological relevance. A predictive model using individual biomarker cut-offs was validated in cohort III.

WGBS identified 52 DMRs (adjusted P<0.01). TBS in cohort II detected 25 differentially methylated sites. A four-marker predictive model was constructed from the most discriminatory hypermethylated sites located within a single intron of the PCNT gene. Validation in cohort III identified 11 differentially methylated sites, with 8 overlapping cohort II; 7 were associated with PCNT. Model performance in cohort III showed specificity 95.65% and sensitivity 70.45%. In cohort II, specificity was 73.08% and sensitivity 83.33%. Performance was similar in simple CHD (specificity 83.67%, sensitivity 76.09%) and complex CHD (specificity 83.67%, sensitivity 75.00%).

Amniotic fluid methylation profiles differed between CHD and controls. The four-marker PCNT methylation panel showed consistent classification performance across cohorts.