Improved reliability of a serum alpha-fetoprotein enzyme immunoassay compared with a radioimmunoassay

During the 6-month period between September 1982 and February 1983, serum samples were obtained from 850 pregnant women attending the municipal family clinics in Tel Aviv. All serum samples were tested by alpha-fetoprotein (AFP) enzyme immunoassay (EIA), and 650 samples were also tested by AFP radioimmunoassay (RIA). Comparisons were made between the absolute values, the medians, and the multiples of the median (MOM) obtained by each method. The ratios of the 90th percentile to the median and of the median to the 10th percentile were also compared. Inter- and intraassay variation were determined for both methods. Correlating the test results with the birth results showed that both methods detected the four fetuses with a neural tube defect (NTD) and three of nine twin pairs, when using a cutoff point of greater than or equal to 2.5 MOM. Elevated maternal serum AFP (MSAFP) levels were also noted in 32 normal singletons (3.8%). There were no false negative results during this phase of the study. Although the median values were similar for both methods, the mean of the ratios of the median to the 10th percentile and the 90th percentile to the median, showed a reduced spread for the EIA method.