BACKGROUND: Pharmacological stress tests for the diagnosis of GH deficiency are unpleasant, labor intensive and potentially dangerous. Reports on urinary GH measurement for the assessment of GH have been published after highly sensitive immunoassaies were developed. The aim of this study was to determine whether a 24- hour urine GH as an alternative method for GH assessment were reliable in predicting GH deficiency defined by L-dopa stimulation test. METHODS: Thirty women, ages 45 to 67, were studied. L-dopa stimulation tests were performed with an ingestion of 500 mg of L-dopa. Serum GH and IGF-1 were measured by a radioimmunoassay using commercially available reagents and uGH was estimated from the 24-hour urine. Then, the mean and its distribution of uGH values were compared according to the GH status defined by L-dopa stimulation test. RESULTS: The correlation between the uGH and the peak values after L-dopa stimulation test was significant (r=0.46; p<0.01). The mean value of uGH in the GH deficiency group was significantly lower than that of the normal group P<0.05). But because of the broad overlap of uGH in the two groups uGH showed no good separation GH deficiency from normal group although uGH reflects GH status significantly. CONCLUSION: uGH reflects GH status significantly, but because of broad overalp uGH couldn't separate GH deficiency form normal group defined by L-dopa stimulation test.