Sleep plays an important role in energy homeostasis. The present study tests the hypothesis that circulating levels of leptin, a hormone that signals energy balance to the brain, are influenced by sleep duration. We also analyzed associations between leptin and sympathovagal balance, cortisol, TSH, glucose, and insulin under different bedtime conditions. Twenty-four-hour hormonal and glucose profiles were sampled at frequent intervals, and sympathovagal balance was estimated from heart rate variability in 11 subjects studied after 6 d of 4-h bedtimes (mean +/- sem of sleep duration during last 2 d: 3 h and 49 +/- 2 min) and after 6 d of 12-h bedtimes (sleep: 9 h and 03 +/- 15 min). A study with 8-h bedtimes was performed 1 yr later (sleep: 6 h and 52 +/- 10 min). Caloric intake and activity levels were carefully controlled in all studies. Mean levels, maximal levels, and rhythm amplitude of leptin were decreased (-19%, -26%, and -20%, respectively) during sleep restriction compared with sleep extension. The decrease in leptin levels was concomitant with an elevation of sympathovagal balance. The effects of sleep duration on leptin were quantitatively associated with alterations of the cortisol and TSH profiles and were accompanied by an elevation of postbreakfast homeostasis model assessment values. Measures of perceived stress were not increased during sleep restriction. During the study with 8-h bedtimes, hormonal and metabolic parameters were intermediate between those observed with 4-h and 12-h bedtimes. In conclusion, sleep modulates a major component of the neuroendocrine control of appetite.

The 24-h profile of plasma ACTH and cortisol levels was determined in 18 men suffering from major depressive illness (8 with unipolar depression and 10 with bipolar depression) as well as in 7 age-matched normal men. Blood was sampled every 15 min. The circadian variation and episodic fluctuations were analyzed for each individual profile. Both unipolar and bipolar depressed patients had higher 24-h mean cortisol levels (P less than 0.01) than normal men, but no significant difference in 24-h mean ACTH level was found. The nadir of cortisol secretion occurred almost 3 h earlier in older normal subjects and patients with unipolar depression, regardless of age, than in younger normal subjects. This shift paralleled a similar advance of the ACTH nadir. Early timing of the quiescent period of ACTH-cortisol secretion was also found in several patients with bipolar depression, but did not reach significance at the group level. The hypercortisolism in the depressed patients was associated with an increase in the magnitude, but not the number, of cortisol secretory episodes. About 90% of the cortisol pulses could be related to a concomitant ACTH pulse in normal subjects as well as in both groups of depressed patients. However, concomitant ACTH and cortisol pulses were less correlated in magnitude in depressed patients than in normal subjects. These results indicate that major depressive illness is associated with disturbances of pituitary-adrenal function. The early timing of the nadir of ACTH-cortisol secretion suggests that disorders of circadian time keeping may characterize major endogenous depression.

Throughout gestation, maternal insulin-like growth factor I (IGF-I) increases progressively despite suppressed pituitary growth hormone (GH) secretion. We have previously shown that in normal pregnancy, a specific placental GH variant, rather than human placental lactogen (hPL), substitutes for pituitary GH in the regulation of maternal IGF-I. We studied the maternal IGF-I secretion in a cohort of 286 normal and abnormal pregnancies (617 blood samples). Regardless of pathology and gestational age, IGF-I values correlated with corresponding placental GH but not with hPL values. Similar correlations were evidenced for each 2-wk gestational period between 32 and 39 wk. In pathological pregnancies, when only those hormonal results that are obtained before any treatment are considered and diabetes is excluded, IGF-I levels were closely related to corresponding placental GH, but not to hPL. In women with a fetoplacental unit disorder, low placental GH levels resulted in low IGF-I and in a secondary pituitary GH increase, whereas in patients without detectable impairment of the fetoplacental unit normal placental GH corresponded to normal IGF-I. These results suggest that in pathological as well as in normal pregnancy, placental GH, and not hPL, substitutes for pituitary GH to regulate the maternal IGF-I secretion.

The aim of this study was to investigate, in normal young men, whether gamma-hydroxybutyrate (GHB), a reliable stimulant of slow-wave (SW) sleep in normal subjects, would simultaneously enhance sleep related growth hormone (GH) secretion. Eight healthy young men participated each in four experiments involving bedtime oral administration of placebo, 2.5, 3.0, and 3.5 g of GHB. Polygraphic sleep recordings were performed every night, and blood samples were obtained at 15-min intervals from 2000 to 0800. GHB effects were mainly observed during the first 2 h after sleep onset. There was a doubling of GH secretion, resulting from an increase of the amplitude and the duration of the first GH pulse after sleep onset. This stimulation of GH secretion was significantly correlated to a simultaneous increase in the amount of sleep stage IV. Abrupt but transient elevations of prolactin and cortisol were also observed, but did not appear to be associated with the concomitant stimulation of SW sleep. Thyrotropin and melatonin profiles were not altered by GHB administration. These data suggest that pharmacological agents that reliably stimulate SW sleep, such as GHB, may represent a novel class of powerful GH secretagogues. ( J. Clin. Invest. 1997. 100:745-753.)