Revisiting peak serum cortisol response to insulin-induced hypoglycemia in children

Purpose To evaluate factors that could potentially affect the hypothalamic-pituitary adrenal (HPA) axis response to insulin-induced hypoglycemia in children without history or symptoms of adrenal insufficiency and to propose a cut-off value to define a normal response in this population. Methods Exploratory single-center study involving 78 children that prospectively underwent insulin tolerance test (ITT) for suspected growth hormone (GH) deficiency. Methods Glucose, cortisol, GH, adrenocorticotrophic hormone (ACTH), epinephrine and norepinephrine levels were measured at baseline and after insulin-induced hypoglycemia. Serum cortisol was measured using Access automated immunoassay. Results Mean (range) basal morning serum cortisol of 8 (2.2-19.5) mu g/dL/222 (61-542) nmol/L increased after hypoglycemia to 20.5 (14.6-29.5) mu g/dL/570 nmol/L (405-819) nmol/L. Peak serum cortisol levels of 14.6 mu g/dL (405 nmol/L) and 15.4 mu g/dL (428 nmol/L) corresponded to the 2.5th and 5th percentiles, respectively. Peak serum cortisol correlated with peak plasma epinephrine (r = 0.367;P = 0.0014) but did not correlate with age, BMI-SD or peak serum GH. Children with intact and abnormal GH responses presented similar mean peak serum cortisol levels (20.0 vs. 20.6 mu g/dL/555 vs. 572 nmol/L;P = 0.21). Conclusion Our data indicate that the current cut-off to define normal HPA axis response in children after insulin-induced hypoglycemia warrants reevaluation to avoid over-diagnosis of adrenal insufficiency. Our results suggest that peak serum cortisol levels >= 15.4 mu g/dL (428 nmol/L) in children undergoing ITT might represent a normal cortisol response to stress, regardless of age, BMI or GH secretory capacity.