Later Age at Menarche is Associated With Metabolite Signatures for Cardiovascular Diseases: A Metabolome-Wide Association Study of Racially and Ethnically Diverse Women From the Boston Birth Cohort

Introduction: Later menarche onset portends lower risk of cardiometabolic outcomes in adults (e.g., obesity, diabetes, hypertension, and myocardial infarction) and in pregnancy (e.g., preterm birth). However, the underlying mechanisms underlying these associations are unclear. Metabolome-wide association studies afford an opportunity to identify mechanistic pathways. Objective: To identify metabolome signatures of age at menarche in a diverse sample of women. Methods: This analysis includes 1301 women from the Boston Birth Cohort who were enrolled 24-72 hrs. after delivery, when we collected maternal plasma sample for metabolome and information on maternal age at menarche (range 7-18y) via a standard questionnaire interview. We used linear regression and restricted cubic splines to examine the association between age at menarche and metabolites. We adjusted for maternal year of birth, place of birth, education, smoking status, Mediterranean diet score, pre-pregnancy BMI, hypertensive disorders, and diabetes. Results: Of the 1301 study women, 63.7% identified as Black and 21.3% as Hispanic. Maternal age at delivery was 28.3y ±6.6y and age at menarche was 12.87y ±1.95y. Of the 389 metabolites we examined, 63 were associated with late menarche (>15y vs. menarche 12-13y; Fig., A ). A 1-yr increment in menarche age was associated with 7 metabolites ( Fig., B ); of these, imidazole propionate and C18:2 SM were (inverse) linear dose-response ( Fig., C ). Other metabolites, including some implicated in cardiovascular disease (e.g., dimethylguanidino valeric acid), showed a threshold effect starting at menarche >14 years (Fig., C). Conclusion: In this racially and ethnically diverse cohort of women from Boston, later age at menarche is associated with many metabolites, including inverse associations with metabolites that have been previously implicated in cardiovascular disease. Our findings help inform mechanistic pathways and that, if replicated, could be targeted for cardiometabolic prediction and prevention.