To evaluate the effect of different initial levothyroxine (LT4) replacement doses on growth and intellectual outcome in patients with congenital hypothyroidism (CH) detected by neonatal screening program, the longitudinal growth and intelligence quotient (IQ) were assessed and compared at 4 years of age in 83 patients with CH. The patients were divided into three groups according to the initial LT4 dose used: (1) group 1 (n = 42) received the previously recommended dose of 6.0-8.0 microg/kg per day; (2) group 2 (n = 21) received a dose of 8.1-10.0 microg/kg per day; (3) Group 3 (n = 20) a dose of 10.1-15.0 microg/kg per day. The IQ, evaluated by the Wechsler Preschool and Primary Scale of Intelligence test at 4 years of age, was significantly higher in group 3 (IQ 98 +/- 9) compared to group 1 (IQ 88 +/- 13; p < 0.05) but not compared to group 2 (IQ 94 +/- 13). However, the IQs were below the normal range (< 85) in six patients from group 2 (28%), but in none of the patients from group 3 (p = 0.03). Patients from group 3, with severe CH at diagnosis, had an IQ (97 +/- 9) at 4 years of age, which was not different from that of patients from the same group with moderate CH at diagnosis (IQ 99 +/- 9). Similar results were also observed in patients from group 2 however, mean IQ scores in these patients (93 +/- 12) were several points lower than those observed in patients from group 3 (95 +/- 15). After the first month of treatment, optimal serum levels of thyroxine (T4) and free thyroxine (FT4) were achieved in all groups, however, only patients from group 3 were able to normalize thyrotropin (TSH) (group 1, 16.0 +/- 12.0; group 2, 9.2 +/- 10.0; and group 3, 2.4 +/- 3.3 mU/L; p < 0.0001). Twelve patients from group 2 treated with an initial LT4 dose above 9 microg/kg per day were able to normalize TSH levels within the first 3 months of life and this resulted in a better IQ (97 +/- 16) compared to the remaining patients from the same group (IQ 90 +/- 9). In the whole group of 83 patients the IQ at 4 years of age was positively correlated to both initial LT4 dosage (r = 0.27, p < 0.02) and FT4 concentrations after the first month of treatment (r = 0.29, p < 0.02), and negatively correlated to TSH concentrations after the first month of treatment (r = -0.27, p < 0.02). No significant differences were observed in height, weight, head circumference, and bone age maturation among the three groups of patients. No clinical signs or symptoms of overtreatment were observed during follow-up in patients receiving the higher LT4 dosage. Our results indicate that high LT4 starting doses rapidly normalize serum TSH concentrations resulting in an improvement of the IQ at 4 years of age, even in patients with severe CH at diagnosis. Growth and bone age maturation are not affected by such a high dose.