The effect of craniospinal irradiation (CSI) vs. cranial irradiation (CIR) only with or without chemotherapy (CT) on the hypothalamus/pituitary (HP) thyroid axis was assessed in a population-based study of patients treated for a childhood brain tumor not directly involving the HP axis. Thyroid function was evaluated and compared with that in healthy controls (n = 27), measuring TSH, free T4, total T4, total T3, and TRH. The biological effective dose (BED) of radiotherapy, determined for the HP region and spine and expressed in grays (Gy) as BED, gives a means of expressing the biological effects of different dosage schedules in a uniform way. Seventy-one children (45 males and 26 females), less than 15 yr of age when diagnosed between 1970-1997 in the eastern part of Denmark, were included. Twenty-nine had received CSI, and 42 had received CIR only. The median age at time of radiotherapy was 8.4 yr (range, 0.8-14.9). The median length of follow-up was 12.0 yr (range, 2.0-28.0). There was no significant difference between CSI and the CIR only patients with respect to median BED to the HP region. Primary hypothyroidism was found in 24%, of whom 71% had been treated with CSI and 29% with CIR only; 73% had compensated hypothyroidism, and 27% had overt primary hypothyroidism. Central hypothyroidism was found in 6%. Free T4 and total T3 were significantly lower in the CSI and CIR only groups compared with controls. As the CIR only group had significantly higher median basal TSH levels compared with controls and as the CSI compared with the CIR only group and controls had significantly higher median basal TSH levels, we speculate that this was probably due to scattered irradiation from both cranial and spinal fields to the thyroid gland. There was a significant relation between basal TSH and time of follow-up (r(s) = -0.39; P = 0.001). Stepwise backward multiple linear regression analysis showed that the best-fit model to predict basal TSH was free T4 (P < 0.0001), the length of follow-up (P = 0.02), and total T3 (P = 0.06). In contrast, age at radiotherapy, BED to the HP region and spine, and whether the patient had been treated with CT were not included in the model. The TRH test showed significantly exaggerated and prolonged TSH responses for the CSI and CIR only groups compared with controls, indicating HP dysfunction. In conclusion, these data suggest that both CSI and CIR for childhood brain tumor may affect the HP-thyroid axis, resulting in hypothyroidism. CT had no significant influence on HP-thyroid function. We recommend prolonged surveillance of pituitary-thyroid function in long-term survivors of childhood brain tumor and institution of thyroid hormone replacement if the levels of TSH and free T4 are above and below the normal range, respectively, to ensure normal growth and metabolism.