Because some patients with growth hormone (GH) deficiency are found to be hypothyroid after initiation of treatment with GH, we assessed the predictive value of the nocturnal thyrotropin surge (a sensitive test for central hypothyroidism) in 56 untreated GH-deficient children and adolescents. Eighteen patients had a subnormal thyrotropin surge (mean 18% (range -30% to 46%)), significantly less than that of 96 normal control subjects (mean 124%; 95% confidence limits, 47% to 300%; p less than 0.01); 13 of the 18 had a subnormal total thyroxine (T4) level or a subnormal free T4 level, or both. These 18 patients were given thyroid hormone replacement therapy; GH deficiency was confirmed during treatment with thyroxine. Of the remaining 38 patients, who had no initial evidence of dysfunction of the hypothalamic-pituitary-thyroid axis, 23 were re-examined while they were receiving GH treatment. Hypothyroidism developed in none of those 23 children during GH therapy. The nocturnal thyrotropin surge test and determination of iodothyronine levels were repeated in 14 of these euthyroid patients. There was no significant change in mean thyrotropin surge (129% (range +49% to +300) vs 125% (range +51% to +222%)), mean serum level of total T4 (111 +/- 4 vs 103 +/- 3 nmol/L), mean serum level of free T4 (19 +/- 0.7 vs 18 +/- 0.8 pmol/L), mean serum level of triiodothyronine (2.5 +/- 0.1 vs 2.5 +/- 0.1 nmol/L), or mean serum level of thyrotropin (2.9 +/- 0.3 vs 2.9 +/- 0.5 mU/L (mean +/- SEM)). We conclude that GH treatment does not appreciably alter thyroid function in GH-deficient patients who have no evidence of thyroid axis dysfunction before GH treatment.