Early hormonal markers of pregnancy outcome after in vitro fertilization and embryo transfer.

Abstract

To investigate the relative power of HCG, estradiol, and progesterone determinations in the prediction of pregnancy outcome after IVF. These prognostic hormonal factors were studied as single and combined predictors. Serum concentrations of beta-HCG, progesterone, and estradiol were measured 12-13 days after embryo transfer (study point 1) and 7 days later (study point 2) in a series of 20 consecutive infertile patients having a first-trimester spontaneous clinical abortion after an IVF-embryo transfer cycle. As a control group (n = 60), the next three IVF-embryo transfer cycles resulting in an ongoing pregnancy after each miscarried IVF cycle in our assisted reproduction program was used. The discrimination attained between the two study groups (ongoing pregnancies and miscarriages) was evaluated by logistic regression and receiver operating characteristic (ROC) curve analysis. Mean hormone concentrations at study points 1 and 2 were higher in the ongoing pregnancy than in the abortion group. Regarding pregnancy outcome the percentage increment of HCG serum levels (> or = 1321%), with an accuracy (predictive value of pregnancy outcome) of 81.2% (sensitivity 98%, specificity 50%), had the best prognostic reliability but no significant differences were found when this parameter was compared with the predictive value of HCG concentration (> or = 72 IU/l) at study point 1 (diagnostic accuracy 80.5%; sensitivity 70%; specificity 80%). When ROC analysis was used, the best predictor of ongoing pregnancy according to the AUC(ROC) was HCG concentration at study point 2 but again no significant differences were found when this parameter was compared with the predictive value of HCG serum levels at study point 1. A multiple marker strategy did not help distinguish viable from nonviable pregnancies. A single, early (days 12-13 after embryo transfer) HCG quantitative serum measurement in IVF cycles not only is diagnostic but also has good predictive value for pregnancy outcome.