O1008: Impact of Sociodemographic Factors on Timing of Arterial Switch Operation and on Length of Stay in Infants Prenatally Diagnosed with d-Transposition of the Great Arteries

Background: Later timing of the arterial switch operation (ASO) has been associated with lack of prenatal diagnosis and increased morbidity. Limited data is available among infants prenatally diagnosed with d-transposition of great arteries (dTGA) with intact ventricular septum (IVS). Our objective was to determine associations between maternal sociodemographic factors, ASO timing, and length of stay (LOS) in infants prenatally diagnosed with dTGA IVS.

Methods: We conducted a multicenter retrospective cohort study of prenatally diagnosed infants with dTGA IVS from 21 centers in the Fetal Heart Society Research Collaborative. Infants with large ventricular septal defect (VSD), coarctation, or significant extracardiac anomalies were excluded. Our primary outcome variable was neonatal LOS after ASO (in days). Our primary predictor variable was age at ASO surgery categorized into four groups (referent: 0-3 days, ,4-6 days, 7-14 days, and >14 days). Covariates included maternal sociodemographics (age at delivery, race/ethnicity, geographic location, neighborhood poverty level) and infant characteristics (prematurity, birth weight, cardiac center delivery, having a balloon atrial septostomy (BAS)). We fit a multilevel linear regression model to evaluate the relationship between age at ASO and LOS, using Akaike Information Criterion to select model covariates. Random intercepts for each FHS center were used to account for hospital level variability. Additionally, we used unadjusted liner and logistic regression models stratified by having a BAS to evaluate the relationship of age at ASO, maternal poverty, and LOS.

Results: Among 283 infants prenatally diagnosed with dTGA IVS, 68.2% were male, 90.1% were full term, and 71.4% had BAS prior to ASO. Regarding maternal characteristics, 28.6% were publicly insured, 67.5% were non-Hispanic white, and 18.4% lived in neighborhoods with ≥ 20% poverty. ASO was performed at 4-6 days in 41.3% of cases, between 7-14 days in 30% and >14 days in 4.9%. In our multilevel linear model, percent poverty ≥ 20% and age at ASO were independent predictors of postoperative LOS (Figure 1), with those undergoing ASO > 14 days having an average 17 day longer LOS and those with >20% poverty having an average 7 day longer LOS. We did not find evidence of confounding by other maternal sociodemographic or infant characteristic covariates, including undergoing a BAS. Among those who did not undergo BAS, maternal poverty was associated with 4x the odds of later age at ASO (≥7 days of age) compared to infants with early ASO (OR 3.93; p=0.042). Additionally, maternal poverty was associated with an average 9.7 day longer postoperative LOS among infants with BAS (p=0.0016).

Conclusion: There is an association with later ASO, neighborhood poverty >20% and longer LOS. Further research is needed to expose how ASO timing is institutionally determined, how maternal poverty results in longer postoperative LOS, and the impact these factors have on maternal-infant bonding and neurodevelopment.