Context: GH treatment significantly increased adult height (AH) in a dose-dependent manner in short non-GH-deficient children in a randomized, controlled, clinical trial; the mean gain in height SD score (height(SDS)) was 1.3 (range 0-3), compared with 0.2 in the untreated group. Objective: The objective of the study was to analyze the relationship between IGF-1(SDS), IGF binding protein-3 SDS (IGFBP3(SDS)), and their ratio(SDS) with a gain in the height(SDS) until AH in non-GH-deficient short children. Design and Setting: This was a randomized, controlled, multicenter clinical trial. Intervention: The intervention included GH treatment: 33 or 67 mu g/kg.d plus untreated controls. Subjects: One hundred fifty-one non-GH-deficient short children were included in the intent-to-treat (ITT) population and 108 in the per-protocol (PP) population; 112 children in the ITT and 68 children in the PP populations had idiopathic short stature (ISS). Main Outcome Measures: Increments from baseline to on-treatment study mean IGF-1(SDS) (Delta IGF-1(SDS)), IGFBP3(SDS), and IGF-1 to IGFBP3 ratio(SDS) were assessed in relationship to the gain in height(SDS). Results: Sixty-two percent of the variance in the gain in height(SDS) in children on GH treatment could be explained by four variables: Delta IGF-1(SDS) (explaining 28%), bone age delay, birth length (the taller the better), and GH dose (the higher the better). The lower IGF-1(SDS) was at baseline, the higher was its increment during treatment. For both the All(PP)- and the ISSPP-treated groups, the attained IGF-1(SDS) study level did not correlate with height gain. Conclusion: In short non-GH-deficient children, the GH dose-related increment in IGF-1(SDS) from baseline to mean study level was the most important explanatory variable for long-term growth response from the peripubertal period until AH, when IGF-1(SDS), IGFBP3(SDS), and their ratio(SDS) were compared concurrently.