213 - Predictive Modeling and Postnatal Management of Neonates with Concern for Coarctation of the Aorta

Poster Number: 213
Publication Number: 213.220

Zach Anderson, Vanderbilt University Medical Center, Nashville, TN, United States; Leon D. Hatch, Vanderbilt University Medical Center, Nashville, TN, United States; David Parra, Monroe Carell Jr. Children's Hospital at Vanderbilt, Nashville, TN, United States; Jonathan H. Soslow, Vanderbilt University Medical Center, Nashville, TN, United States

Background: Coarctation of the aorta (CoA) is difficult to diagnose pre- and postnatally due to the presence of a patent ductus arteriosus (PDA). With this uncertainty, many infants with prenatal concern for CoA are admitted to the ICU for observation while awaiting PDA closure. These admissions incur undue parental stress and medical resources. Our prior work created and validated a coarctation prediction model (CPM) to differentiate neonates with and without CoA in the presence of a PDA, with the CPM calibrated for maximum sensitivity.  This resulted in a postnatal management algorithm for these neonates (Figure 1).

Objective: The objective of this study was to validate the CPM in a larger cohort of neonates. We hypothesized that the CPM and the postnatal management algorithm would identify neonates with CoA without adverse cardiovascular outcomes.

Design/Methods: This is a retrospective study of 133 infants with prenatal concern for CoA from 2016-2020. Neonates < 35 weeks post-menstrual age and < 2 kg were excluded from the study. Infants with simple congenital heart disease (atrial septal defect, ventricular septal defect) were included but patients with all other types of congenital heart disease were excluded. CPM scores were calculated using standardized arch measurements (Figure 2) and compared using Wilcoxon rank sum analysis. The postnatal management plan specified continued observation for a score of >0.15 and transfer to the newborn nursery (and discharge when appropriate) for a score £ 0.15. CoA was diagnosed with echocardiographic or clinical criteria, not CPM results.

Results: 39.1% of the infants were female, 60.9% were male. Average gestational age was 37.9 weeks +/- 1.3. The average birth weight was 3.1 kg +/- 0.5kg. Of 133 infants, 33 required CoA repair. The model yielded a sensitivity of 96.6% and a specificity of 56.4%. Those with CoA had a higher CPM score than those without (0.95 IQR (0.79, 0.99) vs 0.09 IQR (0.02,0.53), p< 0.001) (Figure 3). Sixteen infants represented to the ED post-discharge, all unrelated to cardiac disease. Two deaths occurred in the cohort during the first year of life but none due to cardiac disease. 42 infants were able to be discharged from the newborn nursery or general pediatrics floor as opposed to the NICU. Conclusion(s): This study demonstrates that the CPM is sensitive for detection of CoA. Over a 6-year period, we found that use of a postnatal management plan incorporating CPM resulted in no adverse cardiac outcomes. Further work will be done to enhance the sensitivity and specificity of the model and increase the accuracy with serial CPM scores.
Clinical Guideline Flowsheet
Figure 2: Coarctation Prediction Model Measures and EquationEquation and examples of measurements used to calculate the CPM