It is not known how patients who acquire GH deficiency (GHD) in adulthood differ in measures of GH secretion from normal adults. To characterize measures of GH secretion in such patients compared to those in normal subjects, we studied 23 men (median age, 51 yr; range, 32-62 yr) with adult-onset pituitary disease, defined as GH-deficient based on having no detectable GH response to two pharmacological agents, and 17 normal men. Patients less than 50 yr old received insulin (0.1 U/kg, i.v.) and clonidine (0.15 mg, orally), whereas those 50 yr of age or older as well as normal controls received arginine (30 g, i.v.) and clonidine. Patients were compared to normal men by investigating GH sampling every 10 min for 24 h and serum levels of insulin-like growth factor I (IGF-I), IGF-binding protein 2 (IGFBP-2), IGFBP-3, and GH-binding protein. Frequent venous sampling of GH was analyzed in terms of mean 24-h levels, pooled 24-h GH, mean levels over the 12 h between 2000-0800 h (mean overnight GH level), and pulse analysis (pulses per 24 h and pulse amplitude) by the Pulsar computer program. Although there were significant differences between the two groups for almost all measures of GH secretion, overlap between the groups was always present. GH levels measured using a highly sensitive chemiluminescence assay on 24-h pools derived from frequent sampling displayed the least overlap between the two groups, as only 2 of 17 normal controls overlapped with the GHD patients. The pooled 24-h GH level using this technique was significantly lower in patients with GHD than in controls (0.117 +/- 0.021 vs. 0.861 +/- 0.098 micrograms/L; P < 0.0001). In the analysis of frequent GH sampling using a standard immunoradiometric assay, mean overnight GH levels provided the best separation between the two groups, as all 23 patients had values of 0.6 microgram/L or less, and 13 of 17 normal controls had values greater than 0.6 microgram/L. The mean overnight GH level in patients was 0.6 +/- 0.0 microgram/L compared to 1.0 +/- 0.1 microgram/L in controls (P < 0.0001). The mean 24-h GH level in patients was 0.5 +/- 0.0 microgram/L compared to 0.8 +/- 0.1 microgram/L in normal controls (P < 0.0001). GH pulse frequency and pulse amplitude were also reduced in patients with GHD compared to those in normal controls [1.7 +/- 0.5 vs. 5.1 +/- 0.5 pulses/24 h (P < 0.0001) and 0.6 +/- 0.1 vs. 2.8 +/- 0.4 microgram/L (P < 0.0001), respectively]. The mean serum IGF-I level was significantly lower in patients with GHD than in normal controls (106.7 +/- 8.0 vs. 218.7 +/- 16.7 microgram/L; P < 0.0001). Three of 23 patients overlapped with control values. Mean serum levels of IGFBP-3 and the serum IGF-I/IGFBP-2 ratio were also significantly lower in patients than in controls, but values overlapped substantially. We conclude that overlap occurs on measures of GH secretion between normal men and men identified as GH deficient despite a stringent definition of GHD. The best separation was obtained using pooled 24-h GH levels determined by a highly sensitive chemiluminescence assay.