Bedside predictors of in-hospital mortality in neonatal gastric perforation: Development of a clinical scoring system.

BACKGROUND

Neonatal gastric perforation has high mortality, and bedside tools for risk stratification at diagnosis remain limited.

OBJECTIVE

To develop and internally validate a clinical prediction model for mortality using variables available at diagnosis, and to characterize post-operative vasoactive inotropic score (VIS) trajectories.

METHODS

In this retrospective cohort study (2015-2025), neonates with surgically confirmed gastric perforation were included. A multivariable logistic regression model was developed and internally validated using bootstrap resampling (1000 iterations). Clinical utility was evaluated using decision curve analysis. A simplified bedside score was derived, and survival analyzed using Kaplan-Meier methods. VIS trajectories were examined in a secondary analysis.

RESULTS

Mortality was 41.5% (27/65). The final model included delivery room intubation (OR 3.26, 95% CI 1.04-10.94), onset-to-diagnosis interval (OR 1.03 per hour, 95% CI 1.00-1.05), and shock at diagnosis (OR 4.18, 95% CI 1.11-20.48). Discrimination was moderate (AUC 0.747; optimism-corrected 0.719), with acceptable calibration. The simplified score stratified patients into low- and high-risk groups (22.9% vs 63.3%, p = 0.001). Decision curve analysis demonstrated positive net benefit across clinically relevant threshold probabilities (approximately 25-60%). In secondary analyses, non-survivors showed marked escalation in post-operative vasoactive support, with higher VIS1 (110 [IQR 58-165] vs 9 [0-30]) and ΔVIS (65 [40-125] vs 0 [-2 to 9]).

CONCLUSION

A three-variable bedside model enables early risk stratification in neonatal gastric perforation. Post-operative VIS dynamics describe hemodynamic deterioration and may inform clinical reassessment.