HDL cholesterol as a mediator of the relationship between breastfeeding and coronary atherosclerosis from a two-step Mendelian randomization analysis.

The association between breastfeeding and coronary atherosclerosis (CA) is still controversial. In this study, we employed Mendelian randomization (MR) analysis to inverstigate the association between breastfeeding and CA, and to identify potential mediators underlying this link. Breastfeeding status was determined via a recall-based questionnaire assessing whether individuals had been breastfed in infancy. Two independent datasets were selected, a discovery dataset from GWAS catalog (GCST90041823, 247,160 cases and 99,661 controls), and a replication dataset from Neale Lab (193,838 cases and 273,743 controls). CA was defined based on related diseases phenotypes, with summary statistics obtained from the FinnGen, including 56,685 cases and 378,019 controls. Nine variables were selected as candidate mediators. All data were obtained from European adult cohorts. MR was employed to estimate the causal effect of breastfeeding on CA. To explore potential pathways, a two-step MR analysis was used to assess the mediating roles of candidate variables. MR analysis supported a causal association between breastfeeding and reduced risk of CA, both in discovery (OR: 0.826, 95%CI: 0.724-0.941, p = 0.004) and replication dataset (OR: 0.439, 95%CI: 0.274-0.703, p = 0.0006). High-density lipoprotein cholesterol (HDL-C) was identified as a potential mediator, accounting for 10.26% of the total effect, 95% CI 1.4-19.13%. Breastfeeding exerted a protective effect against CA, in part via increased HDL-C levels, highlighting a potential early-life intervention target for cardiovascular health.