Prediction of pregnancy outcome with single versus serial maternal serum alpha-fetoprotein tests.

Abstract

The purpose of our study was to determine whether the trend of three maternal serum alpha-fetoprotein samples was more predictive of pregnancy outcome than the initial sample in the evaluation of patients with unexplained alpha-fetoprotein elevations. A total of 432 patients with unexplained elevation of their first two maternal serum alpha-fetoprotein samples had a third sample drawn. Pregnancy outcomes were determined. Patients were grouped for analysis according to the level of the initial sample, the final sample, and the trend of three samples. Statistical analysis was by chi 2 and logistic regression, with p < 0.05 considered significant. The initial maternal serum alpha-fetoprotein was most predictive of preterm delivery (p < 0.001), size small for gestational age (p < 0.001), and intrauterine fetal death (p = 0.009). Neither the final value nor the trend of three values was as prognostic. The first maternal serum alpha-fetoprotein is the best predictor of pregnancy outcome. Obtaining a second sample to confirm the elevation is appropriate, but additional samples provide minimal information.