OBJECTIVE: To verify the hypothesis of an increased sensitivity to GH in obesity (OB) and Cushing's syndrome (CS). DESIGN: We studied the effects of short-term administration of low-dose rhGH on circulating IGF-I levels in patients with simple OB or CS and in normal subjects (NS). METHODS: Nineteen women with abdominal OB aged (mean +/- s.e.m.) 38.2+/-3.1 years, body mass index 40.7+/-2.5 kg/m(2), waist to hip ratio 0.86+/-0.02, ten with CS (50.4+/-4.2 years, 29.7 +/- 3.3 kg/m(2)) and 11 NS (35.0+/-3.6 years, 20.5+/-0.5 kg/m(2)) underwent s.c. administration of 5 microg/kg per day rhGH at 2200 h for four days. Serum IGF-I, IGF-binding protein-3 (IGFBP-3), GH-binding protein (GHBP), insulin and glucose levels were determined at baseline and 12 h after the first and the last rhGH administration. RESULTS: Basal IGF-I levels in NS (239.3+/-22.9 microg/l) were similar to those in OB (181.5+/-13.7 microg/l) and CS (229.0+/-29.1 microg/l). Basal IGFBP-3, GHBP and glucose levels in NS, OB and CS were similar while insulin levels in NS were lower (P<0.01) than those in OB and CS. In NS, the low rhGH dose induced a sustained rise of IGF-I levels (279.0+/-19.5 microg/l, P<0.001), a non-significant IGFBP-3 increase and no change in GHBP, insulin and glucose levels. In OB and CS, the IGF-I response to rhGH showed progressive increase (246.2+/-17.2 and 311.0+/-30.4 microg/l respectively, P<0.01 vs baseline). Adjusting by ANCOVA for basal values, rhGH-induced IGF-I levels in CS (299.4 microg/l) were higher than in OB (279.1 microg/l, P<0.01), which, in turn, were higher (P<0.05) than in NS (257.7 microg/l). In OB, but not in CS, IGFBP-3 and insulin levels showed slight but significant (P<0.05) increases during rhGH treatment, which did not modify glucose levels in any group; thus, in the OB patient group a significant fall in glucose/insulin ratio was observed. CONCLUSIONS: Short-term treatment with low-dose rhGH has enhanced stimulatory effect on IGF-I levels in OB and, particularly, in hypercortisolemic patients. These findings support the hypothesis that hyperinsulinism and hypercortisolism enhance the sensitivity to GH in humans.