Analyses to predict the growth response to recombinant human growth hormone (GH) in prepubertal children during the first year of treatment were performed on data from 472 patients with idiopathic GH deficiency (IGHD), 202 children with Turner's syndrome, 327 children with idiopathic short stature (ISS) and 135 children with intrauterine growth retardation (IUGR). In IGHD, 56% of the variability of the response could be predicted from a model based on six variables. These variables could be ranked in order of importance as follows: target height SDS minus height SDS, chronological age, frequency of GH injections, dose of GH, weight-for-height index, and birth weight SDS. When the model for IGHD was applied to Turner's syndrome, ISS and IUGR, there was a high degree of similarity between the predicted and achieved growth response in ISS and IUGR. However, an uneven distribution within the plot of Studentized residuals in ISS and IUGR suggested heterogeneity within these populations. Prediction of growth in Turner's syndrome was greatly exaggerated by the model for IGHD, suggesting a different pathogenesis as the basis of the growth disorder. Specific prediction models were therefore developed for Turner's syndrome, ISS and IUGR. In all three disorders, the dose of GH was found to be the most important predictor, suggesting that, in contrast to IGHD, first-year growth is governed less by the difference between height and the presumed genetically determined target height. Again, in contrast to IGHD, this suggests that catch-up phenomena are not involved.(ABSTRACT TRUNCATED AT 250 WORDS)