Risk of ICU aspiration and construction of a prediction model in postoperative patients with brain glioma

Shi Huan; Liu Yue; Yuan Yuan

doi:10.3969/j.issn.1009-0126.2024.10.018

Abstract：Objective To explore the risk factors for aspiration in intensive care unit(ICU)patients with brain glioma after surgery and construct a risk prediction model for aspiration.Methods Clinical data of 379 glioma patients undergoing surgical treatment in our hospital between March 2021 and March 2024 were collected and retrospectively analyzed.In a ratio of 2∶1,they were assigned into a modeling group(252 cases)and a verification group(127 cases).According to presence or absence of postoperative ICU aspiration,the patients in the modeling group were di-vided into aspiration subgroup(49 cases)and non-aspiration subgroup(203 cases).Multivariate logistic regression analysis was applied to identify the risk factors for ICU aspiration.Based on the regression coefficients(β)of these risk factors,a risk prediction model was constructed.ROC curve was employed to evaluate the discrimination efficiency of the risk prediction model.Results There was no significant difference in incidence of ICU aspiration between the modeling group and the verification group(x2=0.594,P=0.441).Multivariate logistic regression analysis showed that COPD,mechanical ventilation,enteral nutrition and postoperative coma time ≥24 h were risk fac-tors for ICU aspiration in brain glioma after surgery(P＜0.01).The incidence of ICU aspiration was 7.69％,39.13％and 64.29％,respectively in the low-,medium-and high-risk patients of the modeling group,with an AUC value of 0.771(95％CI:0.700-0.842,P＜0.01).And the incidence was 8.05％,51.61％and 66.67％,respectively in the low-,medium-and high-risk patients of the validation group,with an AUC value of 0.789(95％CI:0.698-0.881,P＜0.01).Conclusion COPD,mechanical ventilation,enteral nutrition and continuous postoperative coma can increase the risk of ICU aspiration in postoperative patients with brain glioma.Our constructed prediction model based on above risk factors has good predictive value.