Risk factors for cordocentesis and fetal intravascular transfusion

There is little information on the impact of technical aspects or patient characteristics on the risks of accessing the fetal circulation. We performed 594 diagnostic cordocenteses and 156 intravascular transfusions over 6 years. Pancuronium was administered during 52% of procedures. The number of needle punctures per successful procedure was unrelated to the placental location. However, the number of punctures required was lower if the placental cord origin rather than a midsegment was targeted (p less than 0.0001). Bleeding from either the uterine or umbilical cord puncture site was not believed to be clinically significant, although the duration of bleeding was greater after arterial puncture than after venous puncture (p = 0.01) and after intravascular transfusion than after diagnostic cordocentesis (p less than 0.0001). Amnionitis (suspected plus verified) complicated 0.5% of procedures. Preterm premature rupture of membranes (with or without amnionitis) followed 0.4% of procedures. Fetal bradycardia occurred in 6.6% (6.6 +/- 0.8 minutes; range, 0.1 to 35 minutes). There were five perinatal losses after a diagnostic procedure, yielding an uncorrected loss rate of 0.8% (5/594). Each was associated with a prolonged bradycardia; each fetus was ultimately demonstrated to have been unsalvageable. Two independent risk factors for bradycardia were identified--arterial puncture and severe, early onset intrauterine growth retardation. The administration of pancuronium reduced the incidence of bradycardia in appropriately grown fetuses (6% to 1.5%; p less than 0.05), but did not alter the incidence in growth-retarded fetuses. We conclude that cordocentesis performed with a needle guide is a safe procedure but that its risk varies with both the indication and the vessel punctured.