Maternal thrombocytopenia. Predicting neonatal thrombocytopenia with cordocentesis

To evaluate the efficacy of cordocentesis for predicting fetal thrombocytopenia in the presence of maternal thrombocytopenia.

We studied platelet counts obtained by cordocentesis from 42 consecutive immune thrombocytopenia purpura patients. Platelet counts were obtained on 36 neonates, and the statistical analysis included only these infants. Presence of maternal antiplatelet antibodies, interval from fetal sampling to delivery, neonatal platelet counts and outcome were evaluated. Thrombocytopenia was defined as a platelet count < or = 150,000/microL, with < or = 50,000 microL considered severe.

No procedure-related complications occurred. A moderate correlation existed between fetal and neonatal platelet counts (r = .48, P = .003), unrelated to the interval between sampling and delivery. Eight of 36 fetuses had thrombocytopenia, and 4 were confirmed at delivery. Two neonates had thrombocytopenia at birth but not at cordocentesis. Two neonatal thrombocytopenia cases were severe. Neither was categorized as severe antenatally. The sensitivity, specificity, and positive and negative value for predicting severe neonatal thrombocytopenia were 0%, 100%, 0%, and 94%, respectively. Grade 1 intraventricular hemorrhages occurred in two neonates delivered at 35 weeks' with normal platelet counts.

Cordocentesis was not reliable in predicting severe neonatal thrombocytopenia; however, the high negative predictive value was reassuring. The clinical utility of the technique and the population in which it should be used remain to be defined.