The objective of this study was to investigate whether short-term changes in metabolism, as a result of GH therapy, could be used to predict its growth effect after 1 yr. Twenty-eight children (8.7 +/- 2.8 yr) were selected, based on anthropometric criteria characterizing GH-deficient patients. In addition, 21 healthy, age- and sex-matched controls (8.9 +/- 3.1 yr) were included. Total body water (TBW) and height were measured before and at 6 wk and 1 yr after the start of treatment. After 1 yr of treatment, patients were divided into good and poor responders, based on a change in height of at least 0.7 SD. Because individuals of different heights were compared, changes in TBW after 6 wk were corrected for height(2), in accordance with the body mass index. Eighty percent of the children who showed a good response to GH therapy had a change in TBW divided by height(2) exceeding the 2 SD reference line of the controls. In contrast, poor responders did not differ from controls. Maximum GH concentrations found during endocrine tests were not significantly different between good and poor responders. Changes in body composition data, after 6 wk, proved valuable in identifying good responders to GH therapy.