Hyperglycosylated human chorionic gonadotropin (HCG-H) as a novel predictor of ongoing pregnancy

Abstract

HCG-H, the predominant hCG variant secreted immediately following implantation, is a novel marker of trophoblast invasion (1). Our objective was to determine whether the initial serum hCG-H differs between ongoing and failed pregnancies, and to compare its performance to total serum hCG as a predictor of ongoing pregnancy. Prospective cohort study. Fresh/frozen autologous IVF cycles with a day 5 single embryo transfer (ET) during 2017 with positive serum hCG (>3 IU/L) were included (n=115). Exclusion criteria were body mass index >35 kg/m2, recurrent pregnancy loss, donor egg, gestational carrier, preimplantation genetic testing and multiple gestation. HCG levels were checked exactly 11 days after ET at a single laboratory (CV<6%). Surplus frozen serum (-80oC) was batched and shipped to Quest Laboratories, where hCG-H was measured using an electrochemiluminescence assay (CV<9.1%). Linear regression analyses with robust “sandwich” standard errors adjusted for oocyte age a priori were used. Differences in hCG and hCG-H for ongoing pregnancies (>8 weeks of gestation) and failed pregnancies (clinical pregnancy loss [CPL], biochemical and ectopic pregnancies [EP]) were reported as linear regression βs with 95% confidence intervals (CI), representing differences in mean analyte values between outcomes. EP (n=1) was included in failed pregnancies but not evaluated separately. Receiver operator curves (ROC) were generated for each analyte for the prediction of ongoing pregnancy. Eighty-five patients (73.9%) had ongoing pregnancies. Both mean hCG-H and hCG were significantly higher in ongoing versus failed pregnancies (hCG-H: 19.0 vs. 7.5 mcg/L, hCG: 422.1 vs. 187.5 IU/L, p<0.001), as well as versus biochemical pregnancies (Table). No significant differences were observed between hCG values for ongoing vs. CPL. In contrast, hCG-H values were significantly higher in ongoing vs. CPL (46% of which had fetal cardiac activity) (p=0.01, Table). The areas under the ROC curves (AUC) for HCG-H and HCG were 0.825 and 0.798.Tabled 1HCG-H and HCG among pregnancy outcome groupsOutcomeOngoing pregnancy (n=85) ReferentFailed pregnancy (n=30)Clinical pregnancy loss (n=11)Biochemical pregnancy (n=18)AnalyteMean (SD) min-maxMean (SD) min-maxLinear regression β (95% CI)Mean (SD) min-maxLinear regression β (95% CI)Mean (SD) min-maxLinear regression β (95% CI)HCG-H(mcg/L)19.0 (10.3) 3.3-47.57.5 (7.9) 0.0-31.6-11.5 (-15.0,-8.0)12.2 (8.4) 1.0-31.6-6.8 (-12.0,-1.6)4.1 (5.7) 0.0-20.3-14.9 (-18.3,-11.5)HCG(IU/L)422.1 (222.5) 56.0-1193.0187.5 (201.9) 4.0-805.0-234.6 (-319.8,-149.4)338.1 (216.3) 16.0-805.0-84.0 (-214.6,46.6)84.7 (118.0) 4.0-442.0-337.4 (-408.3,-266.6) Open table in a new tab HCG-H has exciting potential as a very early predictor for ongoing pregnancy in IVF. The AUC for hCG-H was only slightly higher than that for hCG; however, in contrast to hCG, values for hCG-H were significantly different between ongoing pregnancies and CPL, just 11 days post-ET. This observation suggests that hCG-H may be a better predictor of viable pregnancy than hCG. Further evaluation of this possibility in a broader IVF population, including those with multiple embryo transfer and multiple pregnancy, is required.