Immunoassays for quantifying choriogonadotropin compared for assay performance and clinical application

We examined calibration and accuracy, precision, sensitivity, specificity, and "hook" effects for recently revised automated choriogonadotropin (hCG) immunoassay systems (Baxter-Dade Stratus II, Abbott IMx intact hCG and total beta hCG) and compared them with a widely used immunoradiometric assay (Hybritech). We estimated hCG in pregnant women, women with trophoblastic disease, nonpregnant young and menopausal women, normal men, and men with testicular tumors. We found clinically unimportant differences in calibration (all calibrated to the 3rd International Standard). Detection of hCG by all four assays was limited by their responses in serum from nonpregnant women and men. Precision within-run was best for the automated instruments, but all four assays had similar between-run precision. The Hybritech, Stratus, and IMx intact assays are specific for intact hCG. The IMx total beta assay quantifies both free beta subunit and beta subunit present in intact hCG. There is a clinically important hook effect in the Hybritech assay but not the Stratus or IMx assays (to 1.2 x 10(6) int. units/L). Results for pregnant women were similar by all four assays. We measured "hCG" to 8 int. units/L in menopausal women, which weakly correlated with concentrations of lutropin and follitropin and was, in part, explained by crossreactivity. There was no sample-probe carryover in either instrument. We found the IMx diluting module as well as results at the extremes of the IMx calibration curves (less than 10, 800-1200 int. units/L) unreliable but encountered no such problems with the Stratus system. Both automated systems involve batch analyzers with limited throughput but provide hCG concentration estimates much more quickly than the Hybritech assay can.