Nomogram model for predicting post-intubation cardiac arrest in the emergency department: a retrospective study.

Objective: Cardiac arrest is the most serious complication of endotracheal intubation in the emergency department (ED). The aim of this study was to develop and validate a nomogram model for predicting post-intubation cardiac arrest (PICA) in ED setting. Methods: We conducted a retrospective study of patients who underwent endotracheal intubation in the ED between October 2022 and March 2024. Data collected including patient demographics, diagnosis, pre-induction, and post-intubation clinical parameters. PICA was defined as cardiac arrest occurred within 60 min of endotracheal intubation. Least absolute shrinkage and selection operator (LASSO) regression was utilized to identify potential predictor variables. Multivariable logistic regression was used to develop a nomogram risk prediction model. Internal validation was performed by bootstrap method. Receiver operating characteristic (ROC) curves, calibration curves, and decision curves were used to assess the performance of the nomogram. Results: A total of 241,840 patients visited the ED during this period, of whom 1591 underwent tracheal intubation, corresponding to an intubation rate in the ED of 6.8 per 1,000 patient visits. Of the 1167 cases included in the study, 32 (2.7 %) experienced cardiac arrest within 60 min after endotracheal intubation. The LASSO identified five non-zero coefficient variables (systolic blood pressure, heart rate, percutaneous arterial oxygen saturation <90 %, intubation within 5 min of ED arrival, and absence of induction). These variables were used to build a predictive nomogram model. The area under the curve (AUC) of nomogram was 0.834 (95 %CI: 0.738-0.931), it had a sensitivity of 0.781 and specificity of 0.850. The C-index of the model was 0.835, and internal validation showed a corrected C-index of 0.819. Decision curve analysis demonstrated the clinical utility of the model. Conclusions: Our nomogram model, based on systolic blood pressure, heart rate, percutaneous arterial oxygen saturation <90 %, intubation within 5 min of ED arrival, and absence of induction, effectively predicted PICA in ED. This model may serve as a valuable tool for clinicians to identify high-risk emergency patients and optimize airway management strategies.