This study investigated the hypothalamus–pituitary–adrenal (HPA) axis activity in children with anorexia nervosa (AN) before and after inpatient treatment. Salivary cortisol levels were measured to ascertain whether changes in the HPA axis activity following therapeutic intervention could be applicable as a prognostic predictor. This study comprised 21 females with AN and 22 control subjects. Saliva was collected at 2-hour intervals from 9 a.m. to 7 p.m. before and after inpatient treatment. The concentrations for areas under the curve (AUC) were compared with physical parameters, eating attitude score, profile of mood states (POMS), and prognostic factors. Mean salivary cortisol levels at all points and mean AUC cortisol levels in subjects with AN before therapy were significantly higher than those in controls, but returned to control levels after inpatient treatment. Higher AUC cortisol levels were associated with lower standard deviation for weight in AN. A significant positive correlation between the AUC cortisol level and POMS subscale of “Fatigue” was apparent in the control group, but not in the AN group. The increased change values of AUC cortisol level before and after inpatient treatment correlated with increased body weight gain ratio just after treatment, but not with the ratio after one year. The present study indicated that HPA axis activity could reflect severity of illness, but did not show an accurate neuroendocrine response for mood states. Changes in HPA axis activity following treatment could therefore be used to predict prognosis and particularly in the short term.
Anorexia nervosa (AN) is a disturbance of eating behavior or weight-control behavior, which commonly occurs during adolescence in girls and may result in a clinically significant impairment of physical and psychosocial function (Anderluh et al., 2003 and Fairburn and Harrison, 2003). Loss of appetite, distortion of body image, overevaluation of shape and weight, obsessive fears of being fat, and amenorrhea characterize this disorder. The recent increase in childhood AN has underlined the important adverse consequences that such a trend may have on the individual's physical and mental development, and on their risk of chronic diseases in adulthood.
Increased hypothalamus–pituitary–adrenal (HPA) axis activity has been postulated to influence the onset and course of AN in adulthood (Licinio et al., 1996). The endpoint of HPA axis activation is the release of the glucocorticosteroid cortisol from the adrenal glands. Several studies showed increased salivary or serum cortisol levels in the acute phase of adult AN (dos Santos et al., 2007, Gold et al., 1986 and Putignano et al., 2001), and suggested that this condition may occur as a consequence of chronic food restriction, as a biological adaptation to starvation. High cortisol levels decrease gonadotropin-releasing hormone (GnRH) secretion and luteinizing hormone (LH) pulse frequency (Barbarino et al., 1989). The secretion of GnRH from hypothalamus controls the release and synthesis of the pituitary gonadotropins including follicle stimulating hormone (FSH) and LH, which are both important in regulating the menstrual cycle. FSH grows the ovarian follicles and LH induces follicular rupture and sustains the corpus luteum, while both FSH and LH induce proliferation and differentiation of the uterine endometrium. An inverse association has been noted between cortisol level and menstrual frequency (Laughlin et al., 1998). Cortisol excess also inhibits osteoblasts and stimulates osteoclasts, impairs calcium absorption from the gut (Misra and Klibanski, 2009), and is associated with low bone mineral density in AN (Misra et al., 2008). Consequently, measuring baseline levels of neuroendocrine modulators/hormones including ghrelin, leptin, peptideYY, and cortisol was proposed as a way to predict overall recovery in AN (Lawson et al., 2010 and Misra et al., 2006a).
Salivary cortisol is also routinely used as a biomarker of psychological stress and related mental or physical diseases (Hellhammer et al., 2009 and Sjörs et al., 2010). Depressed adults show increased total cortisol secretion and a flattened diurnal rhythm is characteristic of HPA axis hyperactivity (Goodyer et al., 2001). Young people at familial risk of depression, but with no personal history of mood disorder, show higher cortisol secretion compared to controls with no familial history of depression, indicating that elevated cortisol secretion is a prospective predictor of major depression and may serve as a vulnerability marker (Mannie et al., 2007). Lawson et al. (2009) reported a positive association between baseline cortisol levels and anxiety or depression scales in AN, suggesting that hypercortisolemia is a potential mediator of mood disturbance in these patients.
The relationship between HPA axis activity and physical or psychological condition in childhood AN (Zonnevylle-Bender et al., 2005) has not been investigated. It also remains unclear whether longitudinal variations in such activity would predict prognosis in childhood AN. To address these research questions, we measured changes in salivary cortisol level from baseline before and after inpatient treatment to determine the prognostic significance of those levels as predictors for recovery in childhood AN.
