VARIABILITY IN THE URINARY-EXCRETION OF GROWTH-HORMONE IN CHILDREN - A COMPARISON WITH OTHER URINARY PROTEINS

As a basis for assessment of the clinical validity of urinary GH (uGH) measurements in children, the day-to-day variability in renal handling of GH has been compared with that of albumin, N-acetylglucosaminidase (NAG) and creatinine. Five overnight urine specimens were collected over a 2-week period from 78 healthy children (aged 5-16 years), 20 of normal stature and 58 with growth disorders; ten children were classified as GH-deficient (GHD) and 48 were designated short normal (SN). The variability of excretion of each substance was expressed as a coefficient of variation (C.V.) which was not influenced by expressing the urine results as total mass excreted, concentration, excretion rate or as a ratio to creatinine. There was considerable night-to-night variability in the excretion of all substances (mean C.V. values for all groups: 56% for albumin, 41% for GH, 33% for NAG and 27% for creatinine). No differences were found in the variability of GH excretion between males and females, nor between prepubertal and pubertal subjects. The mean C.V. for uGH excretion ranged from 37% in normal and 35% in SN children to 52% in those with GHD (P<0.05). Assay variation rather than a change in renal protein handling accounted for the large variations in uGH concentrations of < 5 pg/ml, thus contributing to the high uGH C.V. of the GHD group. Increasing the number of samples collected (up to five) decreased the expected sample variation (error) for uGH but not significantly and only improved efficiency in the diagnosis of growth disorders from 91 to 95%, while reducing the convenience and practicality of the test. These results indicate that variation in urinary protein excretion over a 2-week period is considerable (albumin>GH>NAG) in both normal children and those with growth disorders. To apply this test to routine clinical management, we recommend comparison of a single overnight uGH measurement with normal ranges derived from age-, sex- and pubertal status-matched children.