The risk for depression markedly rises during the 5–6 years leading up to the cessation of menstruation, known as the menopause transition. Exposure to extreme estradiol levels may help explain this increase but few studies have examined individual sensitivity to estradiol in predicting perimenopausal depression.

The current study recruited 101 perimenopausal women. During Phase 1, we quantified each woman's sensitivity to changes in estradiol using 12 weekly measures of estrone-3-glucuronide (E1G), a urinary metabolite of estradiol, and concurrent depressive symptoms. The weekly cortisol awakening response was measured to examine the hypothalamic-pituitary-adrenal (HPA) axis in mediating mood sensitivity to estradiol. In Phase 2, depressive symptoms and major depression diagnoses were assessed monthly for 9 months. The relationship between Phase 1 E1G sensitivity and Phase 2 depressive symptoms and major depressive episodes was examined. Several baseline characteristics were examined as potential moderators of this relationship.

The within-person correlation between weekly E1G and mood varied greatly from woman to woman, both in strength and direction. Phase 1 E1G mood sensitivity predicted the occurrence of clinically significant depressive symptoms in Phase 2 among certain subsets of women: those without a prior history of depression, reporting a low number of baseline stressful life events, and reporting fewer months since their last menstrual period. HPA axis sensitivity to estradiol fluctuation did not predict Phase 2 outcomes.

Mood sensitivity to estradiol predicts risk for perimenopausal depression, particularly among women who are otherwise at low risk and among those who are early in the transition.