OBJECTIVE: The effects of GH therapy on thyroid function among previous reports have shown remarkable discrepancies, probably due to differences in hormone assay methods, degree of purification of former pituitary-derived GH preparations, dosage schedules, diagnostic criteria, patient selection, duration of treatment and study design. These considerations motivated us to investigate whether and how GH replacement therapy changes serum thyroid hormone levels, including the much less studied rT3 levels, in a group of unequivocally GH-deficient children receiving long-term recombinant human GH therapy. PATIENTS AND DESIGN: Twenty clinically and biochemically euthyroid children were studied in two therapeutic conditions: on GH replacement therapy for at least 6 months and without GH replacement, either before GH was started or after GH was withdrawn for 30-60 days. Eight patients were on thyroxine replacement treatment and thyroxine doses were kept constant during the study. Blood was collected before and after 15, 20 and 60 minutes of TRH administration in both therapeutic conditions (with GH and without GH). MEASUREMENTS: Concentrations of thyroid hormone levels were determined only in sera obtained before TRH administration. FT4, T3 and TSH were measured by immunoflourimetric assays and rT2 was measured by immunoradioassay. RESULTS: Patients were classified into two groups, according to basal TSH levels: group I (TSH > 0.4 mU/l, n = 12) and group II (on thyroxine and TSH < 0.05 mU/l, n = 8). In both groups, serum FT4 levels decreased (17. 0 +/- 1.1 vs. 14.3 +/- 0.9 mU/l, P < 0.001, and 18.0 +/- 1.7 vs. 14. 2 +/- 1.7 mU/l, P < 0.01, respectively), serum T3 levels increased (1.8 +/- 0.1 vs. 2.4 +/- 0.2 nmol/l, P < 0.001, and 1.9 +/- 0.3 vs. 2.4 +/- 0.2 nmol/l, P < 0.05, respectively), and serum rT3 levels decreased (0.35 +/- 0.03 vs. 0.25 +/- 0.03 nmol/l, P < 0.01, and 0. 48 +/- 0.06 vs. 0.34 +/- 0.06 nmol/l, P < 0.01, respectively). Basal (3.2 +/- 0.50 vs. 2.6 +/- 0.72 mU/l, P = 0.28, paired t-test), TRH-stimulated peak TSH levels (13.9 +/- 5.3 vs. 15.9 +/- 8.0 mU/l, P = 0.35, paired t-test) and TRH-stimulated TSH secretion, expressed as area under the curve (609 +/- 97 vs. 499 +/- 53 mU/l.minutes-1, P = 0.15, paired t-test), remained unchanged during GH replacement in group I patients. Low serum FT4 and high serum T3 levels were observed in only one patient each, but low serum rT3 levels were found in six patients (four in group I and two in group II) during GH replacement. CONCLUSIONS: These results show that long-term GH replacement therapy in children with unequivocal GHD significantly decreases serum FT4 and rT3 levels and increases serum T3 levels; that these changes are independent of TSH and result from increased peripheral conversion of T4 to T3 and that GH replacement therapy in GH deficient children does not induce hypothyroidism, but simply reveals previously unrecognized cases whose serum FT4 values fall in the low range during GH replacement.