Fifteen prepubertal short stature children (10 girls, 5 boys), mean age 9.6 years (range 5.2-12.7 years), with normal response to growth hormone stimulation tests (group A) or partial growth hormone deficiency (GHD) of idiopathic nature (group B) were included in a controlled longitudinal study for evaluation of predictive parameters for the long-term growth response after administration of biosynthetic human growth hormone (B-hGH). The average knee-heel length velocity for the first 3 months was significantly correlated to total body height velocity during the following 9 months (p less than 0.0008). By contrast, this association could not be found for height velocity during the same period. The increase in serum values of alkaline phosphatase and insulin-like growth factor I (IGF-1) during the first month of treatment was not significantly correlated to height velocity during the first year. During one year of treatment with B-hGH the mean height velocity for groups A and B increased from 4.4 cm/year (range 2.5-6.5) to 7.6 cm/year (range 4.7-10.6). Bone age advanced by 1.08 +/- 0.60 per chronological year. The ratio between total height and knee-heel length prior to treatment was 3.34 +/- 0.10 and after one year 3.33 +/- 0.10, suggesting a proportional linear growth. An inverse relationship was observed between the ratio and chronological age. In conclusion, early knee-heel measurement may be a useful non-invasive predictor of long-term linear growth in children during treatment with growth hormone, and the ratio of total height to lower leg length may be of importance in detecting dysproportional growth.