Endocrine Abstracts

Introduction: The insulin tolerance test (ITT) is accepted as the gold-standard test in the evaluation of adrenal and GH axis in patients with pituitary disorders. Diagnostic criteria that requires a minimum increment in serum cortisol is considered invalid although individuals who have a lower basal serum cortisol concentration because of recent ACTH deficiency may be maximally stimulated by ITT and thus able to further increase cortisol secretion without reaching the cut-off. Analyzing the relationship between cortisol Increment and basal cortisol could lead to precious information, and perform a prediction of adrenal insufficiency (AI). We therefore decided to investigate the relationship between the peak and basal cortisol values after the ITT.

Patients and methods: This was a prospective study in which ITT was performed in 81 patients with pituitary disorders. Serum cortisol was measured. We divided our population in Group 1 (G1): Adrenal Insufficiency (Peak cortisol<200ng/mL) and Group 2 (G2): normal response (Peak cortisol>200 ng/ml). Sampling took place at 0, 10, 20, 30, 45, and 60 min. The cortisol increment was plotted against basal cortisol. Receiver-operating characteristic (ROC) analysis was performed to identify a cortisol increment peak with the best sensitivity and specificity for AI prediction.

Results: The mean baseline cortisol levels was 95.40±47.08 ng/ml with a peak level of 179.75±79.005 ng/ml (60^th min, P<10⁻³). In ITT, 44/81 (54.3%) subjects had a peak of cortisol response <200 ng/ml and were classified as Group 1 (G1). Basal cortisol was significantly lower in the group 1 with 72.68±33.13 ng/ml than in the group 2 with 122.43±47.26 ng/ml (P<10⁻³). The mean cortisol increment peak in group 1 subjects was 50.61±30.34 ng/ml significantly lower compared to a mean increment peak of 122.02±51.40ng/ml in Group 2 patients (P<10⁻³). The highest proportion of correctly classified patients (84.04%) evaluated by ROC curve analysis was obtained for ITT-induced cortisol increment peak cut-off of 87 ng/ml (sensitivity 85%; specificity 84%; AUC=0.885; 95% confidence interval 0.80–0.96).

Conclusion: Our study showed a marked interdependence of the basal cortisol concentration, peak cortisol concentration, and increases in serum cortisol concentration. Our finding indicates that, considering the induced cortisol increment peak cut-off of 87 ng/ml, we can identify, with a statistical concordance, 84.04% of adrenal insufficient under ITT.