Stuart, Jennifer

Abstract Context Menstrual irregularities have been associated with cardiovascular disease (CVD) risk, but underlying etiologies may confer distinct risk profiles. Functional hypothalamic amenorrhea (FHA) has not been evaluated as a separate risk phenotype. Objective To assess the association between FHA phenotype and incident CVD and cardiometabolic risk factors. Design Prospective cohort study with biennial follow-up from 1993 to 2019. Setting Nurses’ Health Study II cohort. Participants The sample included 52,655 premenopausal women free of CVD in 1993. FHA phenotype was defined as irregular or absent cycles without PCOS traits combined with low body mass index and/or high physical activity. Intervention None Main Outcome Measures Incident CVD (coronary heart disease [CHD; myocardial infarction or coronary revascularization] and stroke) and incident elevated cholesterol, hypertension, and type 2 diabetes (T2DM). Results During up to 26 years of follow-up, 1,007 CVD events occurred. Among women who developed CVD, the median time to the first CVD event was 17.0 years (IQR, 11.3-21.0). FHA phenotype in mid-adulthood was associated with a 62% higher risk of CVD (95% confidence interval [CI]: 1.01–2.59), driven by an almost two-fold higher risk of CHD (hazard ratio [HR] = 1.91; 95% CI: 1.11–3.27), compared with women with regular cycles. FHA was also associated with T2DM (HR = 1.86; 95% CI: 1.41–2.47), but not hypertension or elevated cholesterol. Non-FHA menstrual irregularity was associated with elevated cholesterol, hypertension, and T2DM, but not CVD. FHA in early adulthood was not associated with outcomes. Conclusions FHA phenotype in mid-adulthood was associated with increased CHD risk independent of traditional risk factors. Menstrual history may serve as an early marker of cardiometabolic risk.

Introduction: Hypothalamic amenorrhea (HA) is a neuroendocrine disorder characterized by ovulatory dysfunction, anovulation, and infertility, which can be prolonged for months to years. Menstrual cycle irregularity and amenorrhea identify individuals with higher CVD risk, but existing evidence has not examined HA separately from polycystic ovary syndrome (PCOS). It remains unknown whether HA is associated with CVD events across the lifecourse. Hypothesis: We hypothesized that the HA phenotype (oligo/amenorrhea without PCOS traits) would identify individuals at increased risk for CVD. Methods: Nurses’ Health Study II participants with available information on menstrual cycle characteristics who were premenopausal and free of CVD at baseline comprised the analytic sample (n=83,281). Menstrual cycle regularity was retrospectively reported for ages 18-22y on the baseline questionnaire in 1989 while the 1993 biennial questionnaire captured current menstrual cycle regularity (at ages 29-49y); PCOS traits (severe acne, hirsutism) were additionally ascertained in 1993. Participants were followed for confirmed incident CVD (MI, stroke, and fatal CHD) from 1993 through 2017. Cox proportional hazards models estimated hazard ratios (HR) and 95% confidence intervals (CI) for the relationship between HA and CVD, adjusted for age, race/ethnicity, parental education, parental history of CVD <60y, and behavioral risk factors (smoking, physical activity, diet, alcohol intake, body mass index). Results: Thirteen percent of participants (n=11,071) ever had HA: 8.5% of participants (n=7,081) had HA at ages 18-22y while 6.5% (n=5,428) had HA at ages 29-46y. Persistent HA (HA at both 18-22y and 29-46 y) was identified in 1.7% of participants (n=1,438). CVD events occurred in 253 individuals who ever had HA (HA at either 18-22y or 29-46y) and 1,228 individuals without the HA phenotype. Individuals who ever had HA had a 20% higher rate of CVD (CI: 1.05-1.38). HA phenotype at ages 18-22y was not associated with CVD risk (HR=1.14; CI: 0.96-1.35) while HA phenotype at ages 29-46y was associated with a 30% higher rate of CVD (CI: 1.10-1.54) compared to individuals without HA phenotype. Individuals with persistent HA across both age ranges had a 69% higher rate of CVD (CI: 1.24-2.28). Underlying associations were strongest between ever HA and stroke (HR=1.25; CI: 1.03-1.51) and between persistent HA and CHD (HR=1.90; CI: 1.27-2.86), although case counts were small. Conclusions: HA phenotype in mid-adulthood identified individuals at increased risk of CVD compared to those without HA. As there are multiple HA subtypes, defined by varying combinations of psychosocial stress, anxiety, physical activity and weight loss, future studies should examine which HA subtypes are associated with CVD risk.