Objectives: The insulin tolerance test (ITT) is considered the gold standard to evaluate corticotropic function in patients with suspected pituitary disease, but is limited by several contraindications. As an alternative, we evaluated the diagnostic value of basal cortisol in serum and saliva.
Methods: volunteers (V) and 61 patients with suspected pituitary disease (P) were enrolled into this study. Basal serum and saliva samples were collected simultaneously between 8 and 9 am.
Results: A significant correlation was found between age and SeC (r=−0.22, P<0.005) as well as SaC (r=−0.16, P<0.05), but not for sex or BMI to either parameter. By using a serum cortisol cut point (CP) of 500 nmol/l during ITT, 35 P were considered adrenal insufficient (AI), whereas 26 P were adrenal sufficient (AS). Applying ROC analysis to SeC, an optimal threshold of 260 nmol/l was found (Sens 74%, Spec 73%, AUC 0.81). By using an upper CP of 382 nmol/l (Sens >95%) and a lower CP of 103 nmol/l (Spec >95%), 22 of 61 P (36%) were correctly identified. Regarding SaC, ROC analysis led to an optimal threshold of 7.6 nmol/l (Sens 54%, Spec 85%, AUC 0.76). With an upper CP of 17.5 nmol/l (Sens >95%) and a lower CP of 5.0 nmol/l (Spec >95%), 21 of 61 P (34%) were correctly identified.
Conclusions: Measurement of basal cortisol in either serum or saliva does not require sex- or BMI-dependent reference ranges. Although a weak correlation with age was observed, age-dependent reference ranges calculated by X±2S.D. were nearly identical. By consideration of lower and upper CPs with high Sens and Spec, respectively, measurement of SeC or SaC classified about one third of P correctly, thereby reducing the necessity for ITT to 65% of subjects. Saliva samples taken at 8 am may be especially useful in an outpatient setting.
01 - 05 Apr 2006
European Society of Endocrinology