The metabolic consequences of thyroxine replacement in patients with central hypothyroidism (CH) need to be evaluated. The aim was to examine the outcome of thyroxine replacement in CH. Adult hypopituitary patients (n = 1595) with and without CH from KIMS (Pfizer International Metabolic Database) were studied before and after 2 years of GH replacement. CH patients (CH, n = 1080) were compared with TSH sufficient patients (TSHsuff n = 515) as one group and divided by thyroxine dose/kg/day into tertiles (CHlow-mid-high). Anthropometry, fasting glucose, glycosylated haemoglobin (HbA1c), blood pressure, lipids, IGF-I SDS, quality of life and morbidity were studied. Analyses were standardized for gender, age, number and types of pituitary insufficiencies, stimulated GH peak, age at GH deficiency onset, aetiologies and, when appropriate, for weight and GH dose. At baseline, TSHsuff patients did not differ from CH or CHmid in any outcome. CHlow (≤ 1.18 μg thyroxine/kg/day) had increased weight, BMI and larger waist circumference (WC), CHhigh (≥ 1.58 μg thyroxine/kg/day) had lower weight, BMI, WC and IGF-I than TSHsuff and compared to their predicted weights, BMIs and WCs. For every 0.1 μg/kg/day increase of thyroxine dose, body weight decreased 1.0 kg, BMI 0.3 kg/m(2), and WC 0.65 cm. The GH sensitivity of the CH group was higher (0.76 ± 0.56 SDS/mg GH) than that of TSHsuff patients (0.58 ± 0.64 SDS/mg GH), P <0.001. The middle thyroxine dose (1.19-1.57 μg/kg/day) seems to be the most physiological. This is equivalent to 70, 100, 125 μg thyroxine/day for hypopituitary patients of 50, 70 or 90 kg weight, respectively.