To determine if diet composition influences responses to GH, we fed 11 obese women diets containing 12 Cal/kg ideal BW (IBW) for 2 5-week study intervals. Nonprotein calories were supplied to 6 subjects as 72% carbohydrate (high carbohydrate diet), and 5 subjects received 80% of their nonprotein calories as lipid (high lipid diet). Protein intake was constant (1.0 g/kg IBW) in both groups. During 1 study interval we gave injections of GH (0.1 mg/kg IBW) every other day for 3 weeks (weeks 2-4), and in the other interval injections of vehicle were given. Subjects ingesting the high carbohydrate diet excreted significantly less urinary nitrogen [660.2 +/- 124.1 mmol/day (mean +/- SD)] than those who received high lipid (794.2 +/- 198.5 mmol/day; P less than 0.001), and GH injections reduced nitrogen excretion in the high carbohydrate subjects (532.8 +/- 123.8 mmol/day), but not in the high lipid subjects (743.7 +/- 126.6 mmol/day). The subjects receiving the high carbohydrate diet had a significant increase in serum insulin-like growth factor-I (IGF-I; from 36.2 +/- 9.7 to 55.9 +/- 6.6 nmol/L) and urinary C-peptide excretion (from 43.9 +/- 25.6 to 60.8 +/- 29.4 nmol/day) in response to GH. The IGF-I response attenuated slowly over the 3-week treatment interval. On the other hand, the high lipid group had lesser increases in IGF-I (from 31.0 +/- 6.5 to 41.7 +/- 8.8 nmol/L) and C-peptide excretion (from 24.3 +/- 28.9 to 29.8 +/- 32.8 nmol/day), and IGF-I concentrations declined to control values after only 5 days of GH injection. We believe that this initial IGF-I response was due to an antecedent 35-Cal balanced diet. The mean increment in serum FFA 4 h after GH injection was significantly less in subjects fed the high lipid diet (0.53 +/- 0.40 meq/L) than in those fed the high carbohydrate diet (0.83 +/- 0.43 meq/L). GH injections produced more body fat loss than injections of vehicle whether expressed as percent body fat lost (GH, 3.7 +/- 1.0%; vehicle, 2.8 +/- 0.7%, P less than 0.05) or as the fraction of weight lost as fat (fat loss/weight loss; GH, 0.81 +/- 0.13; vehicle, 0.64 +/- 0.08; P less than 0.005). There was an inverse correlation between the percentage of body fat lost and mean urinary C-peptide excretion during GH injections (r = -0.70; P less than 0.05), suggesting that stimulation of insulin secretion by GH might antagonize the tendency of GH to promote fat loss.(ABSTRACT TRUNCATED AT 400 WORDS)