Risk factor analysis and nomogram prediction model construction for NEC complicated by intestinal perforation

Abstract

Objective: To investigate the clinical characteristics of neonatal necrotizing enterocolitis (NEC) complicated by intestinal perforation and predict the incidence of intestinal perforation in NEC. Methods: Neonates diagnosed with NEC at the Affiliated Hospital of Zunyi Medical University from January 2012 to May 2022 were enrolled, and the clinical data were collected and analyzed retrospectively. The patients were divided into two groups based on intestinal perforation occurrence or not. Mann-Whitney U tests, t-tests, chi-square tests, and fisher’s exact tests were performed between-group comparisons. Logistic and lasso regressions were applied to screen independent risk factors for concomitant bowel perforation, and R software (RMS package) was used to formulate the nomogram prediction model. In addition, the receiver operating curve (ROC) and the calibration curve were drawn to verify the predictive power, while decision curve analysis (DCA) was constructed to evaluate the clinical applicability of the nomogram model. Results: One hundred eighty neonates with NEC were included, of which 48 had intestinal perforations, and 132 did not; the overall incidence of intestinal perforation was 26.67% (48/180). Bloody stool (OR = 5.60), APTT ≥ 50 s (OR = 3.22), thrombocytopenia (OR = 4.74), and hypoalbuminemia (OR = 5.56) were identified as independent risk variables for NEC intestinal perforation (P < 0.05) through multivariate logistic regression analysis. These factors were then applied to develop a nomogram prediction model (C-index = 0.838) by using the R software. The area under the curve (AUC) for the nomogram in the training and validation cohorts were 0.838 (95% Cl: 0.768, 0.908) and 0.802 (95% CI: 0.659, 0.944), respectively. The calibration curve shown that the nomogram has a good predictive ability for predicting the risk of intestinal perforation occurrence. And the decision curve and clinical impact curve analyses demonstrated good clinical utility of the nomogram model. Conclusion: We found that Bloody stool, APTT ≥ 50 s, Thrombocytopenia, and hypoalbuminemia could be used as independent risk factors for predicting intestinal perforation in neonates with NEC. The nomogram model based on these variables had high predictive values to identify NEC patients with intestinal perforation.