Hyperglycosylated human chorionic gonadotropin(h-hcg) asanearlypredictor of biochemical and clinical pregnancy loss: a prospective analysis of single, euploid, frozen embryo transfers.

Hyperglycosylated hCG (H-hCG) is a glycoprotein with the same polypeptide structure as ß-hCG, and much larger N- and O-linked oligosaccharides. H-hCG is produced by extravillous cytotrophoblast cells, while ß-hCG originates from syncytiotrophoblast cells. During the first 2-3 weeks of pregnancy (when invasive trophoblast activity is high), H-hCG comprises up to 90% of total hCG measurable in serum and urine, dropping to less than 5% of total hCG at the end of the first trimester. In this study, we aimed to determine whether H-hCG early in the pregnancy could help predict pregnancy loss after a single, euploid, frozen embryo transfer (FET). H-hCG was measured by enzyme-linked immunosorbent assay (ELISA) on day 9 and 11 after FET, once ß-hCG was confirmed positive. Samples were assessed in duplicate, and the average measurement was obtained. Patients were categorized as having an ongoing pregnancy (discharged pregnant at 8 weeks gestational age), or having a biochemical or clinical pregnancy loss. Receiver operator characteristics (ROC) curves were created for both H-hCG and ß-hCG, as well as for the ratio between them and the trend from day 9 to day 11, and area under the curve (AUC) of the different curves were compared. Sensitivity and specificity of H-hCG for the prediction of pregnancy loss was compared to that of ß-hCG. Blood samples were obtained prospectively from 274 patients on days 9 and 11 after FET, where 200/274 (73%) resulted in ongoing pregnancy, 34/274 (12%) in a clinical loss, 37/274 (14%) in a biochemical loss, and 3/274 (1%) in pregnancy of unknown location. In comparing the AUC of H-hCG and ß-hCG measurements for the detection of any pregnancy loss, H-hCG levels on days 9 and 11, the ratio of H-hCG to ß-hCG on either day, and the change in either absolute value or percentage from day 9 to 11 provided AUC that were inferior to those of ß-hCG measurements on the same days. The highest AUC using H-hCG for the prediction of any loss was 0.78 in the case of the level of H-hCG on day 11, compared to an AUC of 0.84 for ß-hCG on the same day. H-hCG measurements were more sensitive and specific for detection of biochemical loss, with the most predictive measurement being the change in H-hCG level between days 9 and 11 (AUC 0.92). However, similarly to the outcome for pregnancy loss of any type, every H-hCG measurement was outperformed by its ß-hCG counterpart (maximum AUC 0.94 for level of ß-hCG on day 11). For all studied parameters, the specificity of ß-hCG outperformed that of H-hCG for any given sensitivity. ß-hCG is superior to H-hCG as an early predictor of pregnancy loss and biochemical loss after FET. However, H-hCG does have high sensitivity and specificity for biochemical pregnancy loss, suggesting that perhaps a weighted model with both measurements and their change from day 9 to day 11 could provide an improved predictive value.