RESUMEN
Nocturnal urinary growth hormone (U-hGH) levels measured by a sensitive immunoenzymometric assay were compared with hGH levels in serum before and after a clonidine test in healthy children and in children with short stature to determine whether U-hGH measurement is useful for the screening of hGH deficiency. The study was carried out on 19 healthy children (10 prepubertal and 9 pubertal subjects) and on 20 children with short stature, 10 with growth hormone deficiency (hGHD) and 10 with constitutional growth retardation. The diagnosis of hGHD was based on a blunted response to two provocative hGH tests in the appropriate clinical setting. Overnight urinary hGH secretion (mean of 3 collections) was measured by an immunoenzymometric assay. The best discrimination was obtained when the results were expressed as ng/h. Only one individual in the prepubertal group (U-hGH, 0.05 ng/h) and one patient in the growth retardation group (U-hGH, 0.08 ng/h) had a urinary hGH value below the highest value (0.17 ng/h) observed in the growth hormone deficiency group. The coefficient of correlation between urinary hGH in ng/h and post-clonidine peak was 0.50 (P = 0.0015), between urinary hGH in ng/l and post-clonidine peak was 0.48 (P = 0.0025), between urinary hGH in ng/l per hour and post-clonidine peak was 0.47 (P = 0.0027). The highest specificity (0.93), sensitivity (0.90), false negative rate (0.96) and false positive rate (0.82) were obtained when U-hGH was expressed as ng/h per night. Measurement of urinary nocturnal hGH excretion is a useful, simple, noninvasive method for the diagnosis of hGH deficiency. However, the day-to-day variability and wide normal range limit its usefulness in mild forms of hGH insufficiency