Cd. Chen et al., PAIRED HUMAN CHORIONIC-GONADOTROPIN DETERMINATIONS FOR THE PREDICTIONOF PREGNANCY OUTCOME IN ASSISTED REPRODUCTION, Human reproduction, 12(11), 1997, pp. 2538-2541
The aim of this study was to determine the prognostic value of single
and paired measurements of serum concentrations of human chorionic gon
adotrophin (HCG) for successful pregnancy following in-vitro fertiliza
tion (IVF) and tubal embryo transfer (TET). We analysed serum HCG conc
entrations 15 and 22 days after IVF or TET in 198 conception cycles. C
ut-off values of serum HCG were determined by a receiver operating cha
racteristic (ROC) curve. On the basis of single HCG samples on day 15
(HCG(15)) after transfer, using a cut-off value of HCG(15) = 150 mIU/m
l, the sensitivity was 71% and the specificity was 77%. The positive p
redictive value (HCG(15) greater than or equal to 150 mIU/ml indicatin
g a normal pregnancy) was 89%, while the negative predictive rate (HCG
(15) <150 mIU/ml indicating an abnormal pregnancy) was 51%. Patients w
ith HCG(15) <150 mIU/ml but HCG(22)/HCG(15) ratio greater than or equa
l to 15, still had a 90% chance of normal pregnancy. However, in patie
nts with HCG(15) <150 mIU/ml and an HCG(22)/HCG(15) ratio <15, there w
as an 84% chance of an abnormal pregnancy. We conclude that a single H
CG(15) determination combined with the ratio of HCG(22) to HCG(15) has
a higher diagnostic accuracy for prediction of pregnancy outcome than
either analysis alone.