96 - Predicting surgical necrotizing enterocolitis (NEC) and mortality among preterm infants suspected of NEC

Poster Number: 96
Publication Number: 96.237

Sujir Pritha Nayak, UTSW -Dallas, Dallas, TX, United States; Mariela sanchez-Rosado, Joe DiMaggio Children's Hospital at Memorial Regional Hospital, Hollywood, FL, United States; Jordan D. Reis, Mednax, Baylor Scott & White Dallas, Dallas, TX, United States; Priya Sharma, University of Texas Southwestern Medical School, Dallas, TX, United States; Larry S. Brown, Parkland Health, Dallas, TX, United States; Kate Louise Mangona, University of Texas Southwestern Medical School, Dallas, TX, United States; David B.. Nelson, UT Southwestern Medical Center, Dallas, TX, United States; Myra H. Wyckoff, University of Texas Southwestern Medical School, Dallas, TX, United States; Patti J. Burchfield, University of Texas Southwestern Medical School, Dallas, TX, United States; Pollieanna M. Sepulveda, UTSouthwestern, Dallas, TX, United States; Imran N. Mir, University of Texas Southwestern Medical School, Plano, TX, United States; Luc P. Brion, UT Southwestern Medical Center, Dallas, TX, United States; Anita Thomas, University of Texas Southwestern Medical School, McKinney, TX, United States

Background: Necrotizing enterocolitis can be a devastating disease especially if not diagnosed and treated on time. Serial assessments of Neonatal Sequential Organ Failure Assessment (NSFOA) has been shown to predict surgical NEC and mortality, however that score may be low at the time of initial evaluation and increase progressively with clinical deterioration.

Objective: To develop a score which will accurately predict NEC stage III/ mortality at the time of initial sepsis evaluation in babies with a suspicion of NEC.

Design/Methods: This is a retrospective cohort study of preterm infants < 33 weeks or < 1500 grams at birth with a suspicion or diagnosis of NEC born at Parkland Hospital between 2009-2021. Infants were divided into stage I, II, III according to modified Bell's classification. Various variables obtained at the time of sepsis evaluation for suspicion of NEC were entered into logistic regression models and assessed using the area under the curve (AUC). We combined NSOFA score with Hyperglycemia, Hyperkalemia, Acidosis along with pneumoperitoneum and abdominal compartment syndrome into HASOFA score.

Results: In our Cohort (n=271), 122 infants had NEC stage I, 72 stage II, and 77 Stage III NEC. Maternal and infant characteristics at birth were similar in all groups except gestational age (GA) which was higher in NEC II(P=0.006). Several variables not included in NSOFA but, collected at the time of initial sepsis evaluation reached statistical significance in bivariate comparisons such as acidosis, hyperglycemia and hyperkalemia (Table 1&2). Stepwise logistic regression showed that HASOFA was a better predictor of NEC III/ death than NSOFA. The final models, including either NSOFA/HASOFA, GA and use of inotropes for hypotension within 1 week, had an AUC =0.85 with NSOFA vs AUC =0.93 with HASOFA. (Table 3). When the cut off probability of NECIII/Death based on the HASOFA model was set at 0.2 it resulted in a sensitivity of 91% and specificity of 88%, at the time of initial evaluation, at 0.4 it resulted in 82 % and 90% respectively, and at of 0.7 it was 74% and 96% respectively. Respective values for the NSOFA model are at 84% and 75% for the cut off 0.2, 65% and 91% for0.4 and 50%and 97% for 0.7.

Conclusion(s): This data suggests using HASOFA instead of NSOFA evaluation can enhance the accuracy in predicting NECIII/ death at the time of initial evaluation in this population. After further validation, we speculate this score can be used in decision making for babies with suspicion of NEC especially in hospitals that do not have surgical resources in house.