Advancing necrotizing enterocolitis prediction through iterative monitoring

Abstract

Background: Necrotizing enterocolitis (NEC) is a severe inflammatory intestinal disease in preterm infants, marked by heightened morbidity and mortality. Timely prediction of NEC is significant in the management of critical neonates. However, it is difficult to predict NEC accurately because of the multi-factorial pathogenesis. This study aimed to develop a prediction model through repeated measurement data to further improve the accuracy of prediction in NEC.

Methods: We retrospectively collected clinical data of premature infants admitted to the Neonatology Department of the First Affiliated Hospital of Anhui Medical University from January 2016 to December 2023. The infants were categorized into the NEC group (Bell's stage ≥ II) (n=150) and the non-NEC group (n=150). The clinical baseline data of the NEC and non-NEC groups were matched. Laboratory examination indicators were collected on the 1st day, the 7th day after birth, and the day of NEC onset. Univariate and multivariate logistic regression analyses were conducted to identify independent factors influencing NEC. A nomogram was constructed based on these factors to predict NEC. The concordance index and calibration plot were used to assess the efficiency of the nomogram in the training and validation cohorts.

Results: This study demonstrated that antenatal steroids, antenatal antibiotics, probiotics treatment before NEC, anion gap (AG, day 7), and mean corpuscular volume (MCV, day 7) were independent risk factors which combined to accurately predict NEC. A nomogram of NEC was created utilizing these five predictors. With an area under the receiver operator characteristic (ROC) curve of 0.835 [95% confidence interval (CI): 0.785-0.884]. Concordance index for the training and validation groups were 0.835 and 0.848, respectively. As the calibration plots indicate, the predicted probability of NEC is highly consistent with the actual observation.

Conclusions: The risk estimation nomogram for NEC offers clinical value by guiding early prediction, targeted prevention, and early intervention strategies for NEC.

Keywords: Preterm infant; anion gap (AG); necrotizing enterocolitis (NEC); prediction.

Figure 1

Comparison of hematologic parameters between the NEC and non-NEC groups on the 1st day, the 7th day after birth, and the day of NEC onset. (A) Anion gap; (B) lymphocyte count; (C) mean corpuscular volume; (D) red cell distribution width; (E) red cell distribution width to platelet ratio; (F) eosinophil count. *, P<0.05; **, P<0.01; ***, P<0.001. AG, anion gap; NEC, necrotizing enterocolitis; MCV, mean corpuscular volume; RDW-SD, red cell distribution width standard deviation; RPR, RDW to platelet ratio.

Figure 2

NEC risk nomogram. The NEC nomogram was developed in the cohort, with antenatal steroids, antenatal antibiotics, probiotics, AG day 7, and the MCV day 7. To estimate the probability of NEC, mark infant values at each axis, draw a straight line perpendicular to the point axis, and sum the points for all variables. Next, mark the sum on the total point axis and draw a straight line perpendicular to the probability axis. AG, anion gap; MCV, mean corpuscular volume; NEC, necrotizing enterocolitis.

Figure 3

Receiver operating characteristic curves for training cohort and validation cohort prediction model. (A) Area under the curve of the training cohort was 0.835 (95% confidence interval: 0.785–0.884); (B) area under the curve of the validation cohort was 0.848 (95% confidence interval: 0.743–0.952). AUC, area under the curve.

Figure 4

Calibration curves of the training (A) and validation (B) cohorts, corrected using 1,000 bootstrap samples to reduce over-fitting bias. NEC, necrotizing enterocolitis.