Diagnostic significance of urinary growth hormone measurements in children with growth failure: correlation between serum and urine growth hormone.

Twelve-h overnight urine and serum samples obtained simultaneously at 20-min intervals were assayed for growth hormone (GH). Ninety-one children, 5 to 16 y (Tanner stage 1 to 3) participated; group 1 were healthy children, group 2 were children with organic GH deficiency, and group 3 had idiopathic growth failure and normal GH stimulation tests. Serum pool GH concentrations in group 1 were similar to those in group 3 (3.3 +/- 0.3 versus 3.4 +/- 0.2 micrograms/L); group 2 had significantly lower GH concentrations (1.6 +/- 0.2 micrograms/L). Plasma IGF-I levels were significantly greater in groups 1 (14.2 +/- 2.6 nmol/L, p less than 0.001) than in groups 2 and 3 (2.6 +/- 0.5 and 5.5 +/- 0.7 nmol/L, respectively). Urinary GH (mean +/- SEM) standardized for body weight (micrograms/kg) in group 1 (0.31 +/- 0.02) was significantly greater than in group 2 (0.14 +/- 0.01) and group 3 (0.20 +/- 0.01). However, when expressed as microgram/mol creatinine, the output of GH was similar in group 1 (4.0 +/- 0.3) and group 3 (3.4 +/- 0.3); both groups had significantly greater output compared to group 2 (1.3 +/- 0.2). Urinary IGF-I (nmol/kg) in group 1 (0.22 +/- 0.02) was significantly greater than in group 2 (0.12 +/- 0.01) or group 3 (0.07 +/- 0.01). Urinary GH correlated with serum pool GH concentration (r = 0.64, p less than 0.001). Although urinary GH output reflects endogenous GH secretion, the overlap between groups 1 and 3 precludes using urinary GH measurements as a diagnostic test for GH deficiency in children with idiopathic growth failure.