Routine postnatal chest x-ray and intensive care admission are unnecessary for a majority of infants with congenital lung malformations

Background: Postnatal evaluation of prenatally identified congenital lung malformations (CLMs) often includes a chest x-ray (CXR) and neonatal intensive care unit (NICU) admission for observation. With current efforts aimed at prioritizing value and resource utilization, we sought to assess the utility of this practice in infants with known CLMs. We hypothesized that CXR and NICU admission are overused and could be deferred in the majority of cases. Methods: Clinical and radiographic data for infants with CLM from 2007 to 2016 were reviewed with IRB approval. Regression models were developed for respiratory support (RS), symptoms within 30 days of discharge (Sx30), and abnormal CXR. Predictors included initial symptoms (IS), birth weight (BW), gestational age (GA), cyst-volume-ratio (CVR) and abnormal CXR. Odds ratios (ORs) and ROC curves were generated for significant predictors (p < 0.05). Results: Fifty-eight infants were identified. Eight were excluded because birth or surgery occurred outside of our institution. Another four were excluded for requiring immediate surgery, leaving forty-six for full analysis. All infants underwent initial CXR and NICU admission, and 22 (47.8%) had an abnormal CXR. Higher CVR (OR = 6.69, p = 0.024) and lower BW (OR = 0.27, p = 0.028) both increased the odds of an abnormal CXR. Applying optimal ROC cutoffs for CVR and BW would have safely eliminated 21 of 46 CXRs, increasing CXR sensitivity from 48% to 68%. For RS and Sx30, no variable, including abnormal CXR, significantly predicted outcomes. Twenty-seven infants (59%) had a NICU stay of < 24 h and only three patients (6.8%) developed Sx30. Conclusions: Both CXR and NICU admission appear to be overused in infants with CLM. CXR result did not predict need for respiratory support or symptoms following discharge, and thus may not aid in the initial evaluation or in the prediction of future care needs. Using CVR and birth weight can guide CXR use and optimize its sensitivity. Need for NICU admission could not be predicted, but a majority of infants spent < 24 h in the NICU without intervention, suggesting that NICU admission was likely not needed for all infants in this setting. Level of evidence: Study of diagnostic test, Level II evidence.