Context: The constellation of endocrine patterns accompanying menopausal depression remains incompletely characterized.

Objective: To test the hypothesis that the amplitude or phase (timing) of melatonin circadian rhythms differs in menopausal depressed patients (DP) vs. Normal Controls (NC) women.

Design: We measured plasma melatonin every 30 minutes from 18:00–10:00 h in dim light (< 30 lux) or dark, serum gonadotropins and steroids (18:00, 06:00 h) and mood (Hamilton and Beck depression ratings).

Setting: University hospital

Participants: 29 (18 NC, 11 DP) peri- or post-menopausal women

Intervention: None

Main Outcome Measures: Plasma melatonin (onset, offset, synthesis offset, duration, peak concentration, area under the curve) and mood.

Results: Multi- and univariate analyses of covariance (MANCOVA, ANCOVA) showed melatonin offset time was delayed (P = .045) and plasma melatonin was elevated in DP compared with NC (P = .044) across time intervals. Multiple regression analyses showed that years past menopause predicted melatonin duration, and that melatonin duration, body mass index (BMI), years past menopause, Follicle Stimulating Hormone (FSH) level and sleep end time were significant predictors of baseline Hamilton (P = .0003) and Beck (P = .00004) depression scores.

Conclusions: Increased melatonin secretion that is phase-delayed into the morning characterized menopausal DP vs. NC. Years past menopause, FSH, sleep end time and BMI may modulate effects of altered melatonin secretion in menopausal depression.