Growth hormone (GH) responses to growth hormone-releasing factor (GRF) were evaluated in 55 children with growth failure. The study groups consisted of group 1, severe GH deficiency; group 2, partial GH deficiency; group 3, patients with prior cranial radiation for nonpituitary brain tumors; and group 4, children with idiopathic growth failure. Children in group 1 were unresponsive to GRF (mean GH peak +/- SEM, 1.6 +/- 0.5 ng/ml). Higher GH responses to GRF were observed in both groups 2 (17.2 +/- 4.1 ng/ml) and 3 (10.4 +/- 2.8 ng/ml). The highest GH responses to GRF were observed in group 4 (35.9 +/- 4.3 ng/ml). ANOVA revealed a significant difference between groups (F = 12.9; df = 3; p less than 0.01), and further analysis by the Scheffe and Student-Newman-Keuls tests revealed that group 4 was significantly higher than groups 1, 2, or 3 (p less than 0.05). These data suggest that GRF unresponsiveness is a reliable predictor of severe GH deficiency. In patients with partial GH deficiency or idiopathic growth failure, the GRF gives semiquantitative information about somatotrope responsivity to exogenous stimulation.