The quantitation of human GH in a serum sample is not consistent among various commercially available immunoassays. We measured serum GH concentrations with four RIAs [Cambridge, Kallestad, National Hormone and Pituitary Program, and Radioassay Systems Laboratories (RSL)] and two immunoradiometric assays (IRMAs; Hybritech and Nichols). Serum GH concentrations measured by the RIAs were between 1.9 and 2.8 times higher than those determined by the Hybritech IRMA, whereas the concentrations measured by the Nichols IRMA were approximately 3.0 times higher than the Hybritech values. We evaluated the effects of differences in standards, assay diluents, and antibody specificity on GH measurement in the various assays. When GH standards from each of the assays were measured in the Hybritech IRMA, only the RSL standard was less immunoreactive than the other assay standards. Different assay diluents also resulted in varying GH values. In the RIAs, GH diluted in serum was more immunoreactive than GH diluted in phosphate-buffered saline-0.5% BSA. This enhanced immunoreactivity appeared to be due to a nonspecific effect generated by serum. The Nichols and Hybritech IRMAs provide standards diluted in horse serum. In the Nichols assay, GH diluted in human serum was more immunoreactive than GH diluted in horse serum, whereas the immunoreactivity of GH diluted in either serum was equal in the Hybritech IRMA. These IRMAs also differ in that the Nichols assay detected the 20K variant of GH, whereas the Hybritech assay did not. Considering these discrepancies, comparison of data obtained using different assays should be made carefully.