We have recently presented experimental evidence indicating that insulin has a physiologic inhibitory effect on growth hormone (GH) release in healthy humans. The aim of the present study was to determine whether in obesity, which is characterized by hyperinsulinemia and blunted GH release, insulin contributes to the GH defect. To this aim, we used a simplified experimental protocol previously used in healthy humans to isolate the effect of insulin by removing the interference of free fatty acids (FFAs), which are known to block GH release. Six obese subjects (four men and two women; age, 30.8 +/- 5.2 years; body mass index, 36.8 +/- 2.8 kg/m2 [mean +/- SE]) and six normal subjects (four men and two women; age, 25.8 +/- 1.9 years; body mass index, 22.7 +/- 1.1 kg/m2) received intravenous (i.v.) GH-releasing hormone (GHRH) 0.6 microg/kg under three experimental conditions: (1) i.v. 0.9% NaCl infusion and oral placebo, (2) i.v. 0.9% NaCl infusion and oral acipimox, an antilipolytic agent able to reduce FFA levels (250 mg at 6 and 2 hours before GHRH), and (3) euglycemic-hyperinsulinemic clamp (insulin infusion rate, 0.4 mU x kg(-1) x min(-1)). As expected, after placebo, the GH response to GHRH was lower for obese subjects versus normals (488 +/- 139 v 1,755 +/- 412 microg/L x 120 min, P < .05). Acipimox markedly reduced FFA levels and produced a mild reduction of insulin levels; under these conditions, the GH response to GHRH was increased in both groups, remaining lower in obese versus normal subjects (1,842 +/- 360 v 4,871 +/- 1,286 microg/L x 120 min, P < .05). In both groups, insulin infusion yielded insulin levels usually observed under postprandial conditions and reduced circulating FFA to the levels observed after acipimox administration. Again, the GH response to GHRH was lower for obese subjects versus normals (380 +/- 40 v 1,075 +/- 206 microg/L x 120 min, P < .05), and in both groups, it was significantly lower than the corresponding response after acipimox. In obese subjects, as previously reported in normals, the GH response to GHRH was inversely correlated with the mean serum insulin (r = -.70, P < .01). In conclusion, our data indicate that in the obese, as in normal subjects, the GH response to GHRH is a function of insulin levels. The finding that after both the acipimox treatment and the insulin clamp the obese still show higher insulin levels and a lower GH response to GHRH than normal subjects suggests that hyperinsulinemia is a major determinant of the reduced GH release associated with obesity.