Growth-hormone-releasing hormone (GHRH) tests were performed once [GHRH(1-29)NH2, 1 microgram/kg] or on 2 consecutive days [GHRH(1-44)NH2, 1 and 2 microgram/kg administered in random order] in 27 children with idiopathic, isolated growth hormone (GH)-deficiency and in 49 short normal children, all clinically prepubertal. No differences in GH release were found between the tests performed on the 1st and 2nd day or according to GHRH dose or sex, both in GH-deficient and control children. 80% of GH-deficient and 87% of control children responded (GH peak greater than 10 ng/ml) to GHRH(1-29)NH2, and 65% of GH-deficient and all control children to GHRH(1-44)NH2. No differences in GH release were found between GH-deficient GHRH responders and control children. 17% of GH-deficient and 10% of control children responded only to one of the two tests performed on 2 consecutive days; the lack of responsiveness was unrelated to GHRH dose and sequence of GHRH administration (1st or 2nd day). The GHRH test does not seem to be a reproducible test for the evaluation of GH release, nor is it useful to differentiate GH-deficient GHRH responders from short normal children.