Abstract
An enzyme-linked immunosorbent assay (ELISA) was evaluated for serum alpha-fetoprotein determination in the antenatal screening for fetal open neural tube defects. The ELISA was used concurrently with an existing radioimmunoassay (RIA) method until serum specimens from 5000 pregnant women, between 15 and 20 weeks gestation, had been tested. The accuracy of the ELISA was similar to that of the RIA; the median AFP values by gestational week obtained with the ELISA were, on average, 2 KIU/L higher than the corresponding RIA values; the 10th and 90th percentiles, in terms of multiples of the median (MoM), were very similar. The precision of the two methods was also similar. The ELISA method yielded 1·8% results from unaffected pregnancies above 2·5 MoM compared with 1·4% by RIA, a small but statistically significant difference (P=0·03). Both methods detected the same affected pregnancies identified during this period; five open neural tube defects, three with exomphalos and three intra-uterine deaths. The ELISA method was simple, required about one quarter less operator time than the RIA and enabled results to be generated in one day rather than the two days required by RIA. The ELISA method is a suitable alternative to RIA for routine use in screening for fetal neural tube defects.