We report an epi-analysis of 6-yr growth responses obtained with GH treatment in short children born small for gestational age (SGA). Four randomized, multicenter studies explored the effects of continuous and discontinuous regimens of GH treatment in short, non-GH-deficient SGA children. A total of 49 untreated and 139 treated children were followed over 2 and 6 yr, respectively. At the start of the study, the age of these 188 children averaged 5.2 yr (range, 2-8 yr), height was -3.4 SD score, and height adjusted for parental height was -2.4 SD score. Onset of puberty was observed in 46% of the GH-treated cohort, on the average, at 10.7 yr in girls and 11.7 yr in boys. Two studies essentially investigated the effects of continuous GH treatment at a dose of 33 or 67 microg/kg, day, and two studies focused on the growth characteristics during an initial GH treatment for 2-3 yr (dose range, 33-100 microg/kg x day), followed by a withdrawal phase of 1-2 yr, and then by either no or 1 or more episodes of further GH treatment (33 or 67 microg/kg x day). Continuous GH treatment for 6 yr resulted in height increments of 2.0 +/- 0.2 SD (33 microg/kg x day; n = 35) and 2.7 +/- 0.2 SD (67 microg/kg x day; n = 27). Discontinuous GH treatment was given to 77 children, most of them experiencing only 1 (n = 47) or 2 (n = 26) treatment phases with an average duration of 2.0 yr. All these children received GH during the first 2 yr; the dose was only 32 microg/kg x day when averaged over 6 yr. Some individualization of treatment schedules was allowed, and the majority of investigators seemed to aim for a low normal height level, adjusted for parental height. After 2 yr, the mean adjusted height SD score had increased to -0.4 +/- 0.1 and stabilized thereafter. Bone maturation progressed similarly in all treatment subgroups, and after 6 yr of study, bone age remained slightly delayed compared to chronological age. Multivariate analysis identified the average GH dose over 6 yr, parental-adjusted height SD score, and age at start as prime predictors of the growth response. GH treatment was well tolerated. In conclusion, this epi-analysis of growth responses over 6 yr confirms the administration of GH as an effective approach to normalize the stature of short, non-GH-deficient SGA children, at least during childhood and early puberty. In addition, it is now increasingly apparent that a relatively broad spectrum of GH regimens is effective, and this experience should facilitate the design of more individualized treatment schedules in the future, in particular for young children.