Selecting human chorionic gonadotropin immunoassays: consideration of cross-reacting molecules in first-trimester pregnancy serum and urine.

Abstract

We investigated the variation in human chorionic gonadotropin results found with different commercial kits. Levels of human chorionic gonadotropin and related molecules were determined in pregnancy serum and urine and compared with the specificities of different laboratory, office, and home test kits. Total human chorionic gonadotropin (nicked+nonnicked), nonnicked human chorionic gonadotropin, free beta subunit, and beta core fragment were measured in 242 serum samples and 125 urine samples from early pregnancies. In serum, in the 2 weeks after the missed period when most pregnancy tests are performed, median levels of total, nonnicked, and beta human chorionic gonadotropin (total+free beta+beta core) were similar (< or = 12% difference). Individual values, however, varied significantly. For nonnicked human chorionic gonadotropin, values ranged from 41% to 145% and for beta from 101% to 145% of the total human chorionic gonadotropin level. In urine individual nonnicked values varied from < 1% to 148% and beta values from 102% to 547% of the total human chorionic gonadotropin level. A survey of 29 kits revealed that 10 were types detecting total human chorionic gonadotropin, five detecting nonnicked only, and 14 were beta assays. Results from total, nonnicked, and beta human chorionic gonadotropin kits are not necessarily interconvertible. Individual variations in levels of nicked human chorionic gonadotropin, free beta and beta core, and differences in their recognition by immunoassays causes discordant results.