Salivary cortisol measurement in normal-weight, obese and anorexic women: comparison with plasma cortisol

Objective: To compare salivary, plasma and urinary free cortisol (UFC) measurements in patients with anorexia nervosa. in whom an overdrive of the hypothalamic-pituitary-adrenal (HPA) axis is well established but information on salivary cortisol is lacking, in viscerally obese patients in whom subtle abnormalities of cortisol secretion and metabolism are postulated, and in normal-weight healthy women. Participants and experimental design: Measurement of salivary cortisol offers a convenient way to assess the concentrations of free, biologically active cortisol in plasma in different physiopathological settings. Forty-seven drug-free, newly diagnosed women with active restrictive anorexia nervosa, 30 restrictive anorexic women undergoing chronic psychopharmacological treatment, 47 women with mild-to-moderate visceral obesity, 103 women with severe central obesity and 63 normal-weight healthy women entered the study. Salivary and blood samples were collected at 0800 h, 1700 h and 2400 h, together with three consecutive 24-h urine specimens for UFC determination. In controls and patients with anorexia nervosa (n = 83), salivary and plasma cortisol were also measured after a 1-mg overnight dexamethasone suppression test (DST). In patients with anorexia nervosa, mood was rated by the Hamilton scale for anxiety and depression. Results: Untreated patients with anorexia nervosa showed increased plasma and salivary cortisol and UFC concentrations (all P < 0.001 compared with controls), and decreased cortisol suppression after DST in plasma and saliva (P < 0.0001 and P < 0.005 respectively compared with controls). These alterations were less pronounced, although still statistically significant, in treated patients with anorexia nervosa. Salivary cortisol was highly correlated with paired plasma cortisol in the whole population and after splitting the participants by group (P < 0.0001). However. for plasma cortisol values greater than 500 nmol/l (the corticosteroid-binding globulin saturation point), this parallelism was lost. Taking plasma cortisol as a reference, the level of agreement for post-dexamethasone salivary and plasma cortisol was 58.9% among suppressors and 77.8% among non-suppressors (chi (2) test: P < 0.01). Decreased 0800 h/2400 h cortisol ratios were observed in plasma and saliva in drug-free patients with anorexia nervosa (P < 0.005 and P < 0.05 respectively compared with controls), and in saliva in severely obese patients (P < 0.05 compared with controls). Depression and anxiety scores were unrelated to cortisol concentrations in any compartment. Conclusions: Salivary cortisol measurement is a valuable and convenient alternative to plasma cortisol measurement. It enables demonstration of the overdrive of the HPA axis in anorexia nervosa and subtle perturbations of the cortisol diurnal rhythm in women with visceral obesity. With the establishment of more specific and widely acceptable cut-off values for dynamic testing, measurement of salivary cortisol could largely replace plasma cortisol measurement.