Background Previous studies of glucagon stimulation test (GST) in children have shown variable results regarding the utility and reliability of the cortisol response to this test and its correlation with clinical parameters. The aim of this study was to assess cortisol levels at GST and to evaluate how clinical parameters, such as age, sex, pubertal status and Body Mass Index (BMI), correlate to cortisol levels in children. Methods A retrospective study of children evaluated for short stature with the GST. Cortisol, glucose and growth hormone (GH) levels at GST, as well as clinical parameters (age, sex, pubertal status, BMI), were collected from medical records. A peak cortisol of >= 450 nmol/L was used as a cut-off indicative of a sufficient response. Non-parametric tests were applied in the statistical analysis, and linear regression was used to examine factors affecting cortisol max at the GST. Results In total, 171 children were included; median age 7.8 years (1.0-18.0), 60 (35.1%) female, 23 (13.5%) pubertal. Of all children, 145 (84.8%) achieved a peak cortisol >= 450 nmol/L. There was a negative correlation between peak cortisol levels and age (Spearman's rho - 0.26, p = < 0.001). Peak cortisol levels were higher in females vs. males: 667.5 nmol/L (range 400-995) vs. 602 nmol/L (range 202-1008), p = 0.005. A higher number of boys than girls did not reach the cortisol cut-off value of 450 nmol/L (p = 0.022). The difference in maximum stimulated cortisol levels between the sexes remained after adjusting for age with a linear regression model (beta (95% CI) 65.3 (15.9-114.6), p = 0.01). Conclusion GST is a reliable test of the hypothalamic-pituitary-adrenal (HPA) axis in children. Girls and younger children had higher peak cortisol at GST. The results support a need for sex- and age-dependent reference values for cortisol.
