Analysis of maternal alpha-fetoprotein: a comparison of three radioimmunoassays

Increased levels of alpha-fetoprotein in maternal serum or amniotic fluid may be indicative of a fetal neural tube defect. The present investigation compares, prospectively and retrospectively, the analysis of serum or amniotic fluid alpha-fetoprotein from humans using three radioimmunoassays. A total of 46 sera and 137 prospective and 202 retrospective amniotic fluid samples collected at 12-26 weeks' gestation were assayed using our in-house Farr technique assay, a kit Farr technique assay, and a kit double-antibody assay. Analysis of quality control data in all three assays showed that within- and between-assay variations were less with the double-antibody technique. Additionally, the double-antibody assay offers the advantage of considerably lower nonspecific binding (1.7% vs 16-18% for the Farr technique assays). Of the 375 pregnancies sampled during this study, 340 (91%) resulted in normal singleton births, whereas three (1%) pregnancies were found to have elevated serum or amniotic fluid by all three assay methods and resulted in confirmed neural tube defects. Of the remaining, non-normal outcome without neural tube defect or elevated alpha-fetoprotein was determined in 27 (7%) pregnancies, and false-positive elevated alpha-fetoprotein levels were found for 5 (1%). Analysis of maternal alpha-fetoprotein by radioimmunoassay is a sensitive and accurate adjunct in the prenatal diagnosis of neural tube defects.