The impact of cranial irradiation (CIR) and chemotherapy on the hypothalamo-pituitary (HP)-adrenal (HPA) axis was assessed in a population-based follow-up study of patients treated for childhood brain tumor not directly involving the HP axis. HPA function was evaluated and compared with that in healthy controls (n = 17), measuring basal cortisol and the peak cortisol response to an insulin tolerance test (ITT) and an ACTH test(.) The cortisol cut-off level was 500 nmol/liter. The biological effective dose (BED) of radiotherapy was determined for the HP region and spine and was expressed in Gray units, as BED gives a means of expressing the biological effects of different dosage schedules in a uniform way. Seventy-three children (46 males and 27 females), less than 15 yr of age when diagnosed during 1970-1997 in the Eastern part of Denmark, were included. The median age at time of radiotherapy was 8.4 yr (range, 0.8-14.9). The median length of follow-up was 15 yr (range, 2-29). Fourteen patients (19%) had basal cortisol levels below 500 nmol/liter and did not respond with a peak cortisol above the cut-off level to either an ACTH test (30 or 60 min) or an ITT, and thus, they had insufficiency of the HPA axis. Even though a peak cortisol above 500 nmol/liter was reached in the rest of the cohort (n = 59) after either an ACTH test (30 or 60 min) or an ITT, they had significantly lower peak cortisol levels compared with controls (P = 0.0099). Thirteen patients failed the ACTH test (30 min), but passed the ACTH test (60 min), implying a risk of misinterpreting the cortisol capacity of the patient if only the ACTH test (30 min) is obtained. The basal cortisol levels and the cortisol levels in the ACTH test (30 min) and the ACTH test (60 min) were significantly lower in the patient group compared with controls. There was a significant correlation between the peak cortisol after the ITT compared with the peak cortisol after the ACTH test (30 or 60 min; r(s) = 0.56; P = 0.0006), but 48% failed the ITT, and there was discordance in 10 of 33 (30%) patients who passed the ACTH but failed the ITT, indicating the recommendation of continuous use of the ITT as the gold standard for evaluation of the HPA axis. Stepwise backward multiple linear regression analysis showed that the best-fit model to predict the peak cortisol level after an ITT included BED (P = 0.04) and length of follow-up (P = 0.06). In contrast, age at RT, chemotherapy, BED to the spine, and gender were not included in the model. In conclusion, these data suggest that CIR for a childhood brain tumor may affect the HPA axis, resulting in secondary adrenal insufficiency, whereas adjuvant chemotherapy does not seem to add to the deleterious effect of CIR. We recommend life-long surveillance of the HPA axis and performing regular ITTs.