The episodic and pulsatile nature of GH secretion in normal man is well established. Studies in hypophysectomized rats have indicated that pulsatile administration of GH is superior to continuous infusion in promoting growth, but similar studies have not yet been conducted in human subjects. We compared three different iv GH administration schedules in six GH-deficient patients. They were hospitalized three times for 44 h on three occasions, separated by at least 4 weeks without GH treatment. On each occasion they received 2 IU GH, administered iv as either 1) two boluses (at 2000 and 0200 h), 2) eight boluses (at 3-h intervals starting at 2000 h), or 3) a continuous (2000-0200 h) infusion. Serum insulin-like growth factor-I (IGF-I) after eight boluses and that after continuous infusion were almost identical, with a steep increase reaching a peak at 2000-2400 h, followed by a steady decline. The total areas under the curve, expressed as mean levels (micrograms per L), were 147.6 +/- 11.8 (eight boluses) and 151.2 +/- 8.9 (infusion; P = NS). The change with time in IGF-I after the two-bolus regimen differed significantly from that in the other studies (P less than 0.001), displaying only a modest increase, as also reflected in a smaller area under the curve of serum IGF-I (125.3 +/- 8.7 micrograms/L; P less than 0.05). No differences in blood glucose, serum insulin, or plasma glucagon were observed when comparing the three studies. Both blood glucose and serum insulin tended to be elevated during the second night of each study. Almost identical fluctuations were recorded in lipid intermediates in the three studies, with nightly elevations being more pronounced on the first night. Alanine and lactate exhibited nearly identical patterns in the three studies and were characterized by low nocturnal levels. These data indicate that small but frequent iv boluses and continuous infusion of GH are equally effective in generating an increase in IGF-I in GH-deficient patients, whereas the same amount of GH given as two large boluses results in a significantly smaller increase in IGF-I. This could mean that a prolongation of the period during which serum GH is above zero in GH-treated subjects is just as essential as pulsatility for the growth-promoting effects of the hormone.