Evaluation of standard parameters to predict exchange transfusions in the erythroblastotic newborn.

During the time period 1983-90, 91,300 consecutive pregnancies were monitored for red cell alloimmunization. Once revealed, the immunizations were followed by means of repeated maternal antibody titers, maternal anti-D quantitation in D-immunized women, amniotic fluid bilirubin levels and fetal hemoglobin concentrations. High dose intravenous immunoglobulin and/or intrauterine intravascular transfusion was given to prevent or treat fetal anemia. Delivery was induced for all before term when antibody titers were > or = 16. Nevertheless, exchange transfusions were performed in 41 newborns with mothers alloimmunized to Rh(D), Rh(c), Rh(E) and Kell antigens. Eight of the mothers were Rh(D) positive. Phototherapy alone was given to 35 newborns. Both maternal antibody titers and amniotic fluid bilirubin levels were found to be unreliable to predict the need of exchange transfusions in the newborns. Quantitation of maternal anti-D concentration was found to be significantly better predicting 62% at a cut-off level of 0.7 microgram/mL. Analysis of fetal hemoglobin concentration by cordocentesis is the only direct method to evaluate the degree of fetal affection, and should probably be performed when maternal antibody titers are > or = 64, anti-D concentration is > or = 0.7 microgram/mL and data indicate an aggravation of the immunization.