When applied centrally to animals, growth hormone-releasing hormone (GHRH) stimulates slow-wave sleep (SWS), whereas somatostatin (SRIF) increases REM sleep. We investigated whether these peptides also affect the sleep EEG in humans when given intravenously by comparing polysomnographically the effects of four boluses of (1) placebo, (2) 50 micrograms GHRH or (3) 50 micrograms SRIF administered at 22.00, 23.00, 24.00 and 1.00 h to 7 male controls. In addition, we collected blood samples through a long catheter every 20 min from 22.00 to 7.00 h and measured plasma cortisol and growth hormone (GH) levels. In comparison with SRIF and placebo, GHRH produced a significant increase in plasma GH concentration throughout the night (mean +/- SD: 10.8 +/- 2.0 ng/ml after GHRH; 3.0 +/- 1.7 ng/ml after SRIF and 3.2 +/- 2.0 ng/ml after placebo). SRIF failed to substantially attenuate the nocturnal GH release. Nocturnal cortisol secretion was blunted after GHRH but remained unaffected by SRIF (61.4 +/- 12.9 ng/ml after placebo; 46.6 +/- 19.7 ng/ml after GHRH and 70.8 +/- 12.6 ng/ml after SRIF). Quantitative sleep EEG staging showed a significant increase in SWS after GHRH administration but no change after SRIF (percent spent in SWS per night: 14.0 +/- 5.6 after placebo, 20.2 +/- 6.6 after GHRH and 15.1 +/- 8.2 after SRIF). Application of SRIF was accompanied by a trend toward increased REM density. The effects of episodic GHRH administration upon SWS, GH and cortisol secretion were opposite to those previously reported for corticotropin-releasing hormone, which supports the view that neuroregulation of human sleep involves an interaction of central GHRH and corticotropin-releasing hormone.