Preservation of neuroendocrine control of reproductive function despite severe undernutrition

Anorexia nervosa (AN) is characterized by low weight and self-imposed caloric restriction and leads to severe bone loss. Although amenorrhea due to acquired GnRH deficiency is nearly universal in AN, a subset of patients maintains menses despite low weight. The mechanisms underlying continued GnRH secretion despite low weight in these patients and the impact of gonadal hormone secretion on bone mineral density (BMD) in such eumenorrheic, low-weight patients remain unknown. We hypothesized that 1) eumenorrheic women with AN would have higher body fat and levels of nutritionally dependent hormones, including leptin and IGF-I, than amenorrheic women with AN and comparable body mass index; and 2) BMD would be higher in these women. We also investigated whether the severity of eating disorder symptomatology differed between the groups. We studied 116 women: 1) 42 low-weight women who fulfilled all Diagnostic and Statistical Manual of Mental Disorders (fourth edition) diagnostic criteria for AN, except for amenorrhea; and 2) 74 women with AN and amenorrhea for at least 3 months. The two groups were similar in body mass index (17.1 +/- 0.2 vs. 16.8 +/- 0.2 kg/m(2)), percent ideal body weight (78.2 +/- 0.8% vs. 76.7 +/- 0.8%), duration of eating disorder (70 +/- 13 vs. 59 +/- 9 months), age of menarche (13.2 +/- 0.3 vs. 13.5 +/- 0.2 yr), and exercise (4.5 +/- 1.0 vs. 4.2 +/- 0.5 h/wk). As expected, eumenorrheic patients had a higher mean estradiol level (186.6 +/- 19.0 vs. 59.4 +/- 2.5 nmol/liter; P < 0.0001) than amenorrheic subjects. Mean percent body fat, total body fat mass, and truncal fat were higher in eumenorrheic than amenorrheic patients [20.9 +/- 0.9% vs. 16.7 +/- 0.6% (P = 0.0001); 9.8 +/- 0.5 vs. 7.8 +/- 0.3 kg (P = 0.0009); 3.4 +/- 0.2 vs. 2.7 +/- 0.1 kg (P = 0.006)]. The mean leptin level was higher in the eumenorrheic compared with the amenorrheic group (3.7 +/- 0.3 vs. 2.8 +/- 0.2 ng/ml; P = 0.04). Serum IGF-I levels were also higher in the eumenorrheic than in the amenorrheic group (41.8 +/- 3.7 vs. 30.8 +/- 2.3 nmol/liter; P = 0.02). There were only minor differences in severity of eating disorder symptomatology, as measured by the Eating Disorders Inventory, and where differences were observed, eumenorrheic subjects manifested more severe symptomatology than amenorrheic subjects. Mean BMD at the posterior-anterior and lateral spine were low in both groups, but were higher in patients with eumenorrhea than in those with amenorrhea [posterior-anterior spine T-score, -0.9 +/- 0.1 vs. -1.9 +/- 0.1 (P < 0.0001); lateral spine T-score, -1.2 +/- 0.1 vs. -2.3 +/- 0.2 (P < 0.0001)]. In contrast, preservation of menstrual function was not protective at the total hip (total hip T-score, -0.9 +/- 0.1 vs. -1.1 +/- 0.1; P = 0.27), trochanter, or femoral neck. In summary, patients with eumenorrhea had more body fat and higher serum leptin levels than their amenorrheic counterparts of similar weight. Moreover, reduced bone density was observed in both groups, but was less severe at the spine, but not the hip, in women with undernutrition and preserved menstrual function than in amenorrheic women of similar weight. Therefore, fat mass may be important for preservation of normal menstrual function in severely undernourished women, and this may be in part mediated through leptin secretion. In addition, nutritional intake and normal hormonal function may be independent contributors to maintenance of trabecular bone mass in low-weight women.