ConcanavalinA Bindingof a-Fetoprotein in the Diagnosisof NeuralTube Defects-a Wordof Caution

We were most interested to see the report of Buamah et al. (1) on the use of a-fetoprotein (AFP) concanavalin A (Con A) binding studies as an adjunct to total AFP assay in the diagnosis of neural tube defects. In a preliminary study of 20 abnormal pregnancies (complicated by anencephaly, “open” spina bifida, or exomphalos) and 21 normal pregnancies we also found that this test appeared to be valuable (2). We have now completed a larger study of 121 pregnancies (49 normal, 25 anencephalic, 21 “open” spina bifida, and 20 abnormal pregnancies other than neural tube defects). Our results from this larger study now lead us to question the value of % Con A non-reactive AFP measurements (manuscript in preparation). Firstly, we found that the % Con A non-reactive AFP in amniotic fluid falls off markedly with gestational age over the period 16-20 weeks. A similar gestational age-related decrease has been shown by Toftager-Larsen et al. (3), using an electrophoretic rather than a column technique. In our study, the overlap between normal and abnormal pregnancies was considerable before 19 weeks of gestation, and extensive after 19 weeks. The data of Buamah et al. for normal and abnormal pregnancies are not matched in terms of gestational ages. For example, 68% of their normal pregnancies were of 17 weeks gestational age or less, as compared with 12.5% of abnormal pregnancies. The derivation of a mean and range for % Con A nonreactive AFP for all normal pregnancies taken as a group therefore results in considerable overestimation of the separation between normal and abnormal groups. The second point concerns the relative merits of additional tests in cases of “borderline” amniotic fluid total AFP values. To be of value, a test must aid in the correct classification of “false-positive” or “false-negative” total AFP results. We found that values for % Con A non-reactive AFP incorrectly classified as abnormal two of four cases with false-positive total AFP results (pregnancies with normal outcome and amniotic fluid total AFP results greater than four SDs above the mean for gestational age). Similarly, of six cases of exomphalos, only two could be definitely identified as abnormal by use of the Con A-Sepharose column method. We found polyacrylamide gel electrophoresis of amniotic fluid cholinesterase enzymes, estimation of total and acetylcholinesterase activities, or a2-macroglobulin measurements to be far superior to % Con A non-reactive AFP in the reduction of misclassification errors associated with results for amniotic fluid total AFP.