Hormonal and body composition predictors of soluble leptin receptor, leptin, and free leptin index in adolescent girls with anorexia nervosa and controls and relation to insulin sensitivity.

Anorexia nervosa (AN) is associated with very low levels of leptin, a cytokine secreted by adipose tissue and known to suppress appetite. Leptin may play a permissive role in onset of puberty and in resumption of gonadal function in conditions of undernutrition. The soluble leptin receptor (sOB-R) is the main leptin binding protein, and the ratio of serum leptin to sOB-R provides a measure of the free leptin index (FLI), which may be a more accurate determinant of leptin function. Determinants of sOB-R and FLI have not been examined in an adolescent population. We examined levels of sOB-R, leptin, and FLI, and body composition and hormonal determinants of these variables in 23 adolescent girls with AN and 21 healthy adolescent girls of comparable maturity prospectively over 1 yr. Measures of insulin resistance and adiponectin were also examined. We determined changes in levels of sOB-R, leptin, and FLI with weight recovery (defined as an increase in body mass index of >/=10%, n = 11), and with resumption of menstrual cycles (n = 13). Girls with AN had significantly higher levels of sOB-R (P = 0.0008) and significantly lower levels of leptin and FLI (P < 0.0001 for both) than healthy controls, and levels of FLI were reduced more than levels of leptin in girls with AN compared with controls. An inverse correlation was noted between levels of leptin and sOB-R for the group as a whole (r = -0.64, P < 0.0001) but not in girls with AN considered alone. The most important predictor of levels of sOB-R was cortisol in the group as a whole (r = 0.61, P < 0.0001) and in girls with AN considered alone (r = 0.66, P = 0.0008). Other independent predictors of sOB-R levels for the entire group were percent body fat (r = -0.44, P = 0.003) and levels of IGF-I (r = -0.37, P = 0.01). The most important predictors of leptin and FLI were body mass index and percent body fat. An inverse relationship was noted between measures of insulin resistance and sOB-R levels, whereas a positive association was noted between these measures and leptin and FLI. Adiponectin values did not differ in girls with AN compared with healthy controls and did not correlate with sOB-R, leptin, or FLI. Weight recovery resulted in significant decreases in levels of the sOB-R (24.7 +/- 1.7 to 17.6 +/- 1.2 U/ml, P = 0.004), and increases in levels of leptin (4.4 +/- 1.0 to 13.7 +/- 2.9 microg/liter, P = 0.02). Resumption of menstrual function, but not weight recovery alone, was associated with significant increases in FLI (0.19 +/- 0.04 to 0.50 +/- 0.09 microg/U x 10(-3), P = 0.02).We demonstrate an increase in levels of sOB-R and a decrease in the FLI in adolescent girls with AN, and also demonstrate that cortisol is the most important predictor of levels of sOB-R in this condition. Levels of leptin and FLI, conversely, are primarily predicted by body composition. Weight recovery is associated with a decrease in sOB-R and an increase in leptin. Resumption of menses is associated with significant increases in the FLI, suggesting that free leptin may be an important determinant of menstrual recovery.