Abstract
BACKGROUND: In women, both earlier and later age at menarche (AAM) are associated with increased risk of coronary artery disease (CAD). This study examined if the relationship of AAM with CAD and CAD risk factors differs for different underlying sources of variation in AAM-specifically, variation that is attributable to common genetic variants, as represented by a polygenic score (PGS), vs variation in AAM that is independent of the PGS (eg, from environment, rare genetic variants).</p>
METHODS: Primary analyses were conducted on data from 201 037 women in the UK Biobank and validation studies on data from 23 268 women in the Women's Genome Health Study. For each individual, a PGS for AAM was calculated; then, 2 variables were estimated from linear regression models: genetically predicted AAM (the estimated AAM for each woman solely due to the effects of common genetic variants) and PGS-adjusted AAM (the estimated AAM for each woman solely due to factors other than the PGS). Logistic and linear regression with linear splines were then used to study the relationships of these variables with CAD and CAD risk factors.</p>
RESULTS: Genetically predicted AAM demonstrated linear or roughly linear relationships with CAD and CAD risk factors. In contrast, PGS-adjusted AAM demonstrated a U-shaped relationship with CAD, hemoglobin A1c, triglycerides, high-density lipoprotein cholesterol, and waist-hip ratio.</p>
CONCLUSION: These results are consistent with earlier AAM causally increasing risk of CAD but suggest that later AAM itself does not cause increased risk of CAD; rather, sources of variation in AAM other than common genetic variants can cause both later AAM and increased risk of CAD. Dysglycemia, dyslipidemia, and central adiposity are candidate mediators of the association of later AAM with increased risk of CAD.</p>