Is hypercortisolism in anorexia nervosa detectable using hair samples?

Anorexia nervosa (AN) is a severe mental disorder accompanied by extensive metabolic and endocrine abnormalities. It has been associated with hypercortisolism using short-term measurement methods such as 24 h-urine, saliva, and blood. The aim of this study was to examine whether the proposed hypercortisolism in acutely underweight AN (acAN) is also reflected in a long-term measure: hair cortisol (HCC). To gain further insight, we compared hair cortisol to a well-established classical cortisol measure (24 h-urine; UCC) longitudinally in acAN. Hair samples were collected and analyzed using a LC-MS/MS-based method to provide a monthly cortisol value. We compared HCC in samples of 40 acAN with 40 pairwise age-matched healthy controls (HC) as well as 23 long-term recovered AN participants (recAN) with 23 pairwise age-matched HC (cross-sectional design). In the second part, UCC collected weekly during 14 weeks of weight-restoration therapy in 16 acAN was compared with the (time-corresponding) HCC using linear mixed models and bivariate correlations (longitudinal design). No group differences in HCC occurred comparing acAN and recAN to HC (cross-sectional study). The longitudinal analysis revealed a decrease of UCC but not HCC with weight gain. Furthermore, there was no overall significant association between UCC and HCC. Only in the last four weeks of weight-restoration therapy we found a significant moderate correlation between UCC and HCC. HCC did not reflect the expected hypercortisolism in acAN and did not decrease during short-term weight-restoration. Time-corresponding measurements of UCC and HCC were not consistently associated in our longitudinal analysis of acAN undergoing inpatient treatment. Given the drastic metabolic disturbances in acutely underweight AN our findings could be interpreted as disturbed cortisol incorporation or altered activity of 11-β-HSD-enzymes in the hair follicle.