Lower growth hormone and higher cortisol are associated with greater visceral adiposity, intramyocellular lipids, and insulin resistance in overweight girls

Abstract

Although body composition, insulin sensitivity, and lipids are markedly altered in overweight adolescents, hormonal associations with these parameters have not been well characterized. Growth hormone (GH) deficiency and hypercortisolemia predispose to abdominal adiposity and insulin resistance, and GH secretion is decreased in obese adults. We hypothesized that low-peak GH on the GH-releasing hormone (GHRH)-arginine stimulation test and high cortisol in overweight adolescents would be associated with higher regional fat, insulin resistance, and lipids. We examined the following parameters in 15 overweight and 15 bone age-matched control 12- to 18-yr-old girls: 1) body composition using dual-energy X-ray absorptiometry and MR [visceral and subcutaneous adipose tissue at L(4)-L(5) and soleus intramyocellular lipid ((1)H-MR spectroscopy)], 2) peak GH on the GHRH-arginine stimulation test, 3) mean overnight GH and cortisol, 4) 24-h urinary free cortisol (UFC), 5) fasting lipids, and 6) an oral glucose tolerance test. Stepwise regression was the major tool employed to determine relationships between measured parameters. Log peak GH on the GHRH-arginine test was lower (P = 0.03) and log UFC was higher (P = 0.02) in overweight girls. Log mean cortisol (overnight sampling) was associated positively with subcutaneous adipose tissue and, with body mass index standard deviation score, accounted for 92% of its variability, whereas log peak GH and body mass index standard deviation score accounted for 88% of visceral adipose tissue variability and log peak GH for 34% of the intramyocellular lipid variability. Log mean cortisol was independently associated with log homeostasis model assessment of insulin resistance, LDL, and HDL and explained 49-59% of the variability. Our data indicate that lower peak GH and higher UFC in overweight girls are associated with visceral adiposity, insulin resistance, and lipids.

Fig. 1.

Insulin and glucose levels at different times during the oral glucose tolerance test in adolescent overweight (black line) and normal-weight (gray line) girls. Insulin and glucose levels were higher in overweight than in normal-weight girls. Significances for insulin data are from comparison of logarithmically converted data: *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 2.

Growth hormone (GH)-releasing hormone (GHRH)-arginine stimulation test in adolescent overweight and normal-weight girls. A: GH levels at different points of the GHRH-arginine stimulation test were lower in adolescent overweight (black line) than in normal-weight (gray line) girls. Significances are from comparison of logarithmically converted data: *P < 0.05; **P < 0.01. B: log peak GH levels were lower in adolescent overweight (black bar) than in normal-weight (gray bar) girls. *P < 0.05.

Fig. 3.

Relationship of log peak GH (GHRH-arginine test) with visceral adipose tissue in overweight girls and all girls. Log peak GH levels correlated strongly and inversely with visceral adipose tissue.