Construction and validation of a risk prediction model for emergency endotracheal intubation in patients with spontaneous intracerebral hemorrhage

FENG Li; LIAO Limei

doi:10.12102/j.issn.1009-6493.2025.12.004

Abstract：Objective:To analyze the risk factors of emergency endotracheal intubation in patients with spontaneous intracerebral hemorrhage,and to construct a risk prediction model,and verify it.Methods:A total of 389 patients with spontaneous intracerebral hemorrhage who admitted to the emergency department in a tertiary grade A hospital from September 1,2018 to January 31,2022 were selected as subjects.According to the chronological order,they were divided into a modeling group(290 cases)and a validation group(99 cases).The LASSO regression was used to screen the influencing factors of emergency endotracheal intubation in patients with spontaneous intracerebral hemorrhage.And the Logistic risk prediction model was constructed.The area under the receiver operating characteristic curve(ROC),calibration curve,Hosmer-Lemeshow test,and decision curve(DCA)was used to verify the model.Results:Emergency endotracheal intubation occurred in 54 patients(18.6%)with spontaneous intracerebral hemorrhage in modeling group.Multivariate Logistic regression analysis results showed that the age-adjusted Charlson Comorbidity Index(aCCI),vomiting,onset-to-arrival time,heart rate,oxygen saturation,Modified Early Warning Scale(MEWS),diameter of the left pupil,light reflex of the left pupil,and light reflex of the right pupil were the influencing factors for emergency endotracheal intubation in patients with spontaneous intracerebral hemorrhage.The area under the ROC curve in the modeling group was 0.906(95%CI 0.867-0.946).The optimal cut-off value of the model prediction was 0.122,with a sensitivity of 94.4%,a specificity of 74.6%.The Youden index was 0.690.Hosmer-Lemeshow test results showed that χ²=8.91,P=0.350.The calibration curve had a high degree of coincidence with the reference line.The clinical decision curve indicated that the model had good clinical decision-making performance.Conclusions:The risk prediction model constructed in this study could better predict the occurrence of emergency endotracheal intubation in patients with spontaneous intracerebral hemorrhage,providing reference for clinical medical staff to identify the high-risk population requiring emergency endotracheal intubation as early as possible and formulate targeted interventions.