Hyperglycosylated human chorionic gonadotropin as a predictor of ongoing pregnancy

BACKGROUND: Hyperglycosylated human chorionic gonadotropin, the predominant human chorionic gonadotropin variant secreted following implantation, is associated with trophoblast invasion. OBJECTIVE: To determine whether the initial serum hyperglycosylated human chorionic gonadotropin differs between ongoing and failed pregnancies, and to compare it to total serum human chorionic gonadotropin as a predictor of ongoing pregnancy. MATERIALS AND METHODS: Women undergoing fresh/frozen in vitro fertilization cycles at a university-based infertility clinic with an autologous day 5 single embryo transfer resulting in serum human chorionic gonadotropin >3 mIU/mL (n = 115) were included. Human chorionic gonadotropin was measured 11 days after embryo transfer in a single laboratory (coefficient of variation <6%). Surplus frozen serum (80 degrees C) was shipped to Quest Laboratories for measurement of hyperglycosylated human chorionic gonadotropin (coefficient of variation <9.1%). Linear regression analyses adjusted for oocyte age a priori were used to compare human chorionic gonadotropin and hyperglycosylated human chorionic gonadotropin in ongoing pregnancies (>8 weeks of gestation) and failed pregnancies (clinical pregnancy loss, biochemical and ectopic pregnancies). RESULTS: A total of 85 pregnancies (73.9%) were ongoing. Hyperglycosylated human chorionic gonadotropin and human chorionic gonadotropin values were highly correlated (Pearson correlation coeffi-cient 92.14, P <.0001), and mean values of both were positively correlated with blastocyst expansion score (P value test for trend <.0004). Mean human chorionic gonadotropin and hyperglycosylated human chorionic gonadotropin were significantly higher in ongoing vs failed pregnancies. Among ongoing pregnancies vs clinical losses, mean hyperglycosylated human chorionic gonadotropin, but not human chorionic gonadotropin, was significantly higher (19.0 vs 12.2 ng/mL, beta = 8.1, 95% confidence interval beta = 13.0 to beta -3.2), and hyperglycosylated human chorionic gonadotropin comprised a higher proportion of total human chorionic gonadotropin (4.6% vs 4.1%; risk ratio, 0.79; 95% confidence interval, 0.66e0.94). CONCLUSION: Measured 11 days after single blastocyst transfer, hyperglycosylated human chorionic gonadotropin and human chorionic gonadotropin values were highly correlated, but only mean hyperglycosylated human chorionic gonadotropin and its ratio to total human chorionic gonadotropin were significantly higher in ongoing pregnancies vs clinical pregnancy losses. Further evaluation of hyperglycosylated human chorionic gonadotropin, including in multiple embryo transfers and multiple pregnancy, and using serial measurements, is required.