To assess the relative roles of growth hormone-releasing hormone (GHRH) pulse and somatostatin withdrawal as potential generators of pulsatile growth hormone (GH) release in humans, we studied GH responses to iv bolus GHRH (1 μg/kg) and to termination of a 4 h iv somatostatin infusion (7.2 μg·kg−1·h−1) in five normal young men, and in five men with previously diagnosed isolated GH deficiency. The patients were diagnosed 8–15 years previously on the basis of typical auxological and hormonal criteria, were treated with exogenous GH and were off GH therapy for 1.5–8.9 years prior to this study. Growth hormone rises to a bolus GHRH were similar between the controls and the patients (maximum GH 27.3±15.3 vs 8.0±4.0 μg/l). The controls exhibited only a small GH rise to somatostatin withdrawal (maximum GH 2.9±1.2 μg/l), while the patients did not (maximum GH 0.7±0.1 μg/l; p<0.05). We conclude that somatostatin withdrawal by itself is an ineffective promoter of GH pulsatility. Periodic quiescence of somatostatinergic neurons must be associated with a concomitant GHRH pulse in order to result in a robust GH pulse.