ObjectiveTo determine whether maternal midtrimester serum N-terminal peptide of proatrial natriuretic peptide, free beta subunit of human chorionic gonadotropin (hCGβ), or alpha-fetoprotein (AFP) levels can predict preeclampsia.MethodsA population-based cohort included 1037 nulliparous women, of whom 637 (61%) participated in a maternal serum Down syndrome screening program. Measurements of hCGβ, AFP, and N-terminal peptide of proatrial natriuretic peptide were made from maternal serum collected at 15-19 weeks' gestation. Sensitivity, specificity, and predictive values were calculated for elevated AFP (at least 2.0 multiples of the median [MoM]) and hCGβ (at least 2.0 MoM) values.ResultsNo difference was found in the concentrations of the N-terminal peptide of proatrial natriuretic peptide among the 30 women in whom preeclampsia developed later (median 270 [range 142-604] pmol/L) compared with 536 women who remained normotensive (274 [51-2626] pmol/L). The sensitivity and specificity of elevated AFP in predicting preeclampsia were 3% and 98% and those of elevated hCGβ were 20% and 84%, respectively. When a stepwise multiple logistic regression model was used, only mean arterial pressure was an independent risk factor in predicting preeclampsia.ConclusionDeterminations of the proposed new marker N-terminal peptide of proatrial natriuretic peptide, as well as serum hCGβ or AFP, are not helpful in predicting preeclampsia.