Utilizing Fetal Echocardiography to Risk Stratify and Predict Neonatal Outcomes in Fetuses Diagnosed with Congenital Heart Disease

Abstract

Objective Risk stratification of fetuses diagnosed with congenital heart disease (CHD) helps provide a delivery plan and prepare families and medical teams on expected course in the delivery room. Our aim was to assess the accuracy of echocardiographically determined risk-stratification assignments in predicting postnatal cardiac outcomes beyond the delivery room. Study Design This was a retrospective study at a single center evaluating all fetuses with CHD who were risk-stratified by echocardiographically determined level of care (LOC) assignment, ranging from 1a (lowest risk) to 4 (highest risk). All data were collected from January 1, 2017, to November 1, 2021. Outcomes included any unexpected cardiac interventions and neonatal clinical outcomes including in-hospital mortality, the need for prostaglandins or inotropes, and defined critical illness. These outcomes were assessed for each LOC assignment by Fisher's exact test. Results Out of 817 patients assigned a LOC, a total of 747 fetuses were included in our final cohort with a separate subanalysis of 70 fetuses diagnosed with coarctation of the aorta. The sensitivity and specificity were high for all LOC levels in predicting delivery room needs (93–100%). Higher LOC levels (3–4) had a lower positive predictive value (66–67%) indicating a high false-positive rate. Subjects with higher LOC assignments had a greater frequency of critical illness, hospital mortality, need for inotropes, need for neonatal surgical or catheterization interventions, and need for prostaglandins (p < 0.001 for all outcomes). A post-hoc analysis reviewing LOC assignments revealed a greater tendency to over-assign LOC at higher assignments (19% for LOC 3 and 4) compared to lower assignments (4% for LOC 1 and 2). Conclusion Risk stratification based on fetal echocardiography can predict neonatal clinical outcomes and acuity of postnatal management needs. However, there is greater variability in expected clinical events and an expected degree of false positives for those with higher LOC assignments. Key Points