Utility of commonly used commercial human chorionic gonadotropin immunoassays in the diagnosis and management of trophoblastic diseases

Background: Patients with trophoblastic diseases produce ordinary and irregular forms of human chorionic gonadotropin (hCG; e.g., nicked hCG, hCG missing the β-subunit C-terminal segment, hyperglycosylated hCG, and free β subunit) that are recognized to differing extents by automated immunometric hCG (or hCGβ) assays. This has led to low or false-negative results and misdiagnosis of persistent disease. False-positive hCG immunoreactivity has also been detected, leading to needless therapy for trophoblastic diseases. Here we compare seven commonly used hCG assays. Methods: Standards for five irregular forms hCG produced in trophoblastic diseases, serum samples from 59 patients with confirmed trophoblastic diseases, and serum samples from 12 women with previous false-positive hCG results (primarily in the Abbott AxSYM assay) were blindly tested by commercial laboratories in the Beckman Access hCGβ, the Abbott AxSYM hCG/β, the Chiton ACS:180 hCGβ, the Baxter Stratus hCG test, the DPC Immulite hCG test, the Serono MAIAclone hCGβ tests, and in the hCGβ RIA. Results: Only the RIA and the DPC appropriately detected the five irregular hCG standards. Only the Beckman, DPC, and Abbott assays gave results similar to the RIA in the patients with confirmed trophoblastic diseases (values within 25% of RIA in 49, 49, and 54 of 59 patients, respectively). For samples that were previously found to produce false-positive hCG results, no false-positive results were detected with the DPC and Chiron tests (5 samples, median <2 IU/L), but up to one-third of samples were false positive (>10 IU/L) in the Beckman (1 of 5), Serono (2 of 9), and Baxter assays (1 of 5), and the hCGβ RIA (3 of 9; median for all assays, <5 IU/L). These samples, which produced false-positive results earlier in the Abbott AxSYM assay, continued to produce high values upon reassessment (median, 81 IU/L). Conclusions: Of six frequently used hCG immunometric assays, only the DPC detected the five irregular forms of βhCG, agreed with the RIA, and avoided false-positive results in the samples tested. This assay, and similarly designed assays not tested here, seem appropriate for hCG testing in the diagnosis and management of trophoblastic diseases.