P0375 - Association Between Enteral Nutrition and Necrotizing Enterocolitis in Neonates with Congenital Heart Disease

Introduction and Objective
Enteral nutrition (EN) is often withheld from neonates with congenital heart disease due to concerns of necrotizing enterocolitis (NEC), despite limited evidence showing no association between the two in this population. This study sought to confirm the association between EN and NEC and determine if this relationship was influenced by feeding practices.
Methods
This single-center retrospective study included all neonates with congenital heart disease admitted to the Cardiac Critical Care Unit (CCCU) at the Hospital for Sick Children, Toronto between July 1, 2016 and May 30, 2018. Subjects were identified from a comprehensive nutrition database of all CCCU admissions; those meeting inclusion for whom EN was initiated and subsequently held for at least 48 hours were reviewed for NEC. Neonates born prematurely (less than 37 weeks gestational age), had a previous history of NEC, or diagnosed with gastrointestinal tract anomalies were excluded. NEC severity was graded retrospectively according to Bell’s criteria. Continuous data were analyzed as median with interquartile range and frequency with proportions were used for categorical variables. Univariable logistic regression was utilized to evaluate the direction and strength of association between NEC and subject demographics, cardiac physiology, surgical complexity, and enteral feeding practices. Multivariable models included predictors with possible associations but were limited due to a low frequency of NEC in the cohort.
Results
A total of 241 neonates were admitted and initiated on EN in the CCCU. The overall prevalence of NEC in this cohort was 5.4% (13/241). Subjects diagnosed with NEC had similar birthweight, gestational age, CCCU admission weight, prostaglandin dependence and surgical complexity as those without. Patients with NEC had longer duration of mechanical ventilation (23 vs 5 days, p=0.001), were slower to return to birthweight (25 vs 8 days, p=0.003), and had longer CCCU (59 vs 8 days, p=0.001) and hospital length of stay (91 vs 8 days, p=0.001). Both groups had similar timing of EN initiation, EN advancement, and maximum EN volume, but NEC was associated with higher EN caloric density (0.80 vs 0.67 kcal/mL, p=0.002) and older age at which maximum EN volume was obtained (26 vs 12 days, p=0.001). In logistic regression models, NEC was also associated with single ventricle physiology (p= 0.006) and presence of diastolic flow reversal (p=0.009). Multivariable analysis demonstrated that single ventricle physiology was no longer associated with NEC after controlling for caloric density (p=0.70).
Conclusion
Delivery of higher caloric density EN was the only feeding practice associated with NEC in neonates with congenital heart disease. Although NEC was associated with single ventricle physiology and diastolic flow reversal, the association with single ventricle physiology was lost when controlling for caloric density.