Modifications were made to a commercially available human (h) GH chemiluminescence assay (Nichols Luma Tag hGH assay), which improved its sensitivity to 0.002 micrograms/L. The results of this assay had a high correlation with those of the Nichols hGH immunoradiometric assay (IRMA; r = 0.91; P < 0.001). The addition of recombinant hGH-binding protein (0.1-10 nmol/L) to standards and serum samples caused a dose-responsive reduction in measured GH in both the chemiluminescence assay and the IRMA; at physiological concentrations of hGH-binding protein, a 10-20% reduction was observed. Fifteen normal young adults (nine men and six women) underwent a standard 100-g oral glucose tolerance test, and plasma GH was measured from 30 min before until 5 h after glucose ingestion. GH was measurable in all samples with the chemiluminescence assay, but fell below the sensitivity of the IRMA in 59% of the samples. There was no difference between baseline or peak glucose levels in male and female subjects, but serum GH concentrations (mean +/- SD) measured by the enhanced sensitivity chemiluminescence assay were lower in male than female subjects at both baseline (0.12 +/- 0.08 vs. 2.3 +/- 2.3 micrograms/L; P < 0.01) and the postglucose GH nadir (0.029 +/- 0.014 vs. 0.25 +/- 0.23 micrograms/L; P < 0.01). The high correlation between baseline and nadir GH (r = 0.82; P < 0.001) and the equivalent fractional decline in mean GH levels in men and women after glucose administration (67 +/- 17% vs. 84 +/- 8%; P = 0.06) suggest that the lower GH levels in men after glucose treatment are due to lower baseline values and not to a greater suppressive effect of glucose.