The relationship of longitudinal change in reproductive hormones and vasomotor symptoms during the menopausal transition

Context: The relationship of reproductive hormones to vasomotor symptoms (VMS) has been incompletely explored, although an increase in such symptoms at midlife and their reduction with hormone therapy suggest a strong and direct relationship. Vasomotor symptoms are reported by 65-76% of women traversing the menopausal transition and are a primary reason for medical intervention during this life stage. Objective: The purpose of this report was to relate longitudinal serum concentrations of the reproductive hormones estradiol (E2), FSH, testosterone ( T), dehydroepiandrosterone sulfate (DHEAS), and SHBG and the free hormone indices free E2 index (FEI) and free T index (FTI) with the occurrence of VMS in women traversing the menopausal transition. Participants and Main Outcome Measures: At baseline, 3302 menstruating women who belonged to one of five ethnic/racial groups were recruited and followed up with annual visits. Frequencies of symptoms ( hot flashes, night sweats) for the prior 2 wk and measures of other covariates as well as potentially confounding variables were self-reported in the annual interview. Serum was obtained annually, on d 2-5 of a spontaneous cycle in cycling women or within 90 d of the anniversary of the baseline study visit in noncycling women and assayed for FSH, E2, T, SHBG, and DHEAS. FTI and FEI were calculated. This analysis incorporated available longitudinal data from 3293 women, excluding information collected at or after first report of hormone therapy use or hysterectomy. Data were analyzed using longitudinal marginal logistic regression models and a partial proportional odds model. Results: After adjusting for age, body mass index, and other related covariates, VMS prevalence increased with higher logFSH concentrations, and the increase was greater when blood was drawn more than 5 d after menses began. FSH concentrations were positively associated with the frequency of either hot flashes or night sweats, and higher FSH concentrations were associated with greater odds of reporting more frequent symptoms. Vasomotor symptom prevalence decreased with higher (log)E2, (sqrt)SHBG, and logFEI but only when these hormone values were modeled independently of logFSH values and the specimens were obtained outside the d 2-5 window. When modeled simultaneously with logFSH, (log)E2, (sqrt)SHBG, and logFEI were no longer significantly associated with symptom prevalence. (Cubic) T-root and (sqrt)DHEAS concentrations and logFTI were not associated with the prevalence of VMS. Annual serum FSH concentrations, but not E2, T, DHEAS, FTI, or FEI when collectively modeled longitudinally, are associated with both the prevalence and frequency of VMS in women at midlife.