Perinatal Mortality Despite Prenatal Diagnosis of Vasa Previa: A Systematic Review.

Abstract

To determine the causes and potential preventability of perinatal deaths in prenatally identified cases of vasa previa. Reports of prenatally identified cases of vasa previa published in the English language literature since 2000 were identified in Medline and ClinicalTrials.gov with the search terms "vasa previa," "abnormal cord insertion," "velamentous cord," "marginal cord," "bilobed placenta," and "succenturiate lobe." All cases from the above search with an antenatally diagnosed vasa previa present at delivery in singleton or twin gestations with perinatal mortality information were included. Cases meeting inclusion criteria were manually abstracted, and multiple antenatal, intrapartum, and outcome variables were recorded. Deaths and cases requiring neonatal transfusion were analyzed in relation to plurality, routine hospitalization, and cervical length monitoring. A total of 1,109 prenatally diagnosed cases (1,000 singletons, 109 twins) were identified with a perinatal mortality rate attributable to vasa previa of 1.1% (95% CI 0.6-1.9%). All perinatal deaths occurred with unscheduled deliveries. The perinatal mortality rate in twin pregnancies was markedly higher than that in singleton pregnancies (9.2% vs 0.2%, P<.001), accounting for 80% of overall mortality despite encompassing only 9.8% of births. Compared with individuals with singleton pregnancies, those with twin pregnancies are more likely to undergo unscheduled delivery (56.4% vs 35.1%, P=.01) despite delivering 2 weeks earlier (33.2 weeks vs 35.1 weeks, P=.006). An institutional policy of routine hospitalization is associated with a reduced need for neonatal transfusion (0.9% vs 6.0%, P<.001) and a reduction in the perinatal mortality rate in twin pregnancies (0% vs 25%, P=.002) but not in singleton pregnancies (0% vs 0.5%, P=.31). Routine hospitalization and earlier delivery of twins may result in a reduction in the perinatal mortality rate. A smaller benefit from routine admission of individuals with singleton pregnancies cannot be excluded. There is currently insufficient evidence to recommend the routine use of cervical length measurements to guide clinical management.