Within the Kabi Pharmacia International Growth Study (KIGS) database, there is information on 1017 (700 male/317 female) patients with idiopathic short stature (ISS). These patients were started on recombinant human growth hormone (GH) at a median age of 10.8 years, a bone age of -1.8 SDS, a height of -2.6 SDS and a predicted adult height (PAH) (Bailey-Pinneau method) of -2.5 SDS. The median dose of GH was 0.6 IU/kg body weight/week and the frequency of injections was six/week. According to the relationship with target height the patients were classified into 'familial short stature (FSS)' (height SDS > target height SDS -1.28) and into 'non-FSS' (height SDS < target height SDS -1.28). During the first year of GH treatment there was an overall increment in the median height velocity from 4.4 to 7.4 cm/year. Over 3 years of GH treatment, cross-sectional analysis demonstrated an overall increment in median PAH of 1.2 SDS. There was a positive correlation between gain in PAH and the GH dose (n = 202, r = 0.18, p < 0.01) during the first year. Longitudinal analysis in 84 patients showed an overall increment of PAH of 0.7 SDS over 2 years of treatment. When applying the KIGS first-year prediction model for patients with idiopathic GH deficiency on cohorts of prepubertal children with FSS and non-FSS, a lower responsiveness to GH in the non-FSS group was observed. It is concluded that higher than substitutive doses of GH are required for the long-term improvement of growth in ISS.