Analysis Of Influencing Factors Of Hypoxemia During One-lung Ventilation Based On Logistic Regression Model And CHAID Decision Tree Model

Objective:The influencing factors and their relationship of hypoxemia during one-lung ventilation were investigated by using Logistic regression model andχ² automated interactive detection（CHAID）decision tree model.Methods:A total of 142 patients who underwent one-lung ventilation thoracic surgery in the Second Hospital of Shanxi Medical University from October 2021 to June 2022 were selected,and their gender,body mass index（BMI）,age,hemoglobin level,smoking history,operation side,forced expiratory volume in one second（FEV₁）,forced expiratory volume in one second to forced vital capacity ratio（FEV₁/FVC%）,Pa O₂ during spontaneous ventilation,Pa O₂ at 10 minutes in the lateral position,and the difference in end-expiratory partial pressure of carbon dioxide（d-Pet CO₂）between the dependent lung and the non-dependent lung at 10 minutes in the lateral position were collected.Observe and record the pulse oxygen saturation（Sp O₂）changes of all patients during one-lung ventilation,hypoxemia is considered to occur when Sp O₂ is less than 90%,according to whether hypoxemia occurred in 142 patients during one-lung ventilation,they were divided into hypoxemia group and non-occurrence group.Univariate correlation analysis was performed on the variable data of the two groups of patients by t test orχ² test,and the variables with P<0.05 were used to construct the Logistic regression model and the CHAID decision tree model to explore the occurrence of hypochondriacs during one-lung ventilation in patients undergoing thoracic surgery.Results:Among 142 patients undergoing one-lung ventilation thoracic surgery,the incidence of hypoxemia was 16.20%（23/142）,after univariate analysis excluded confounding factors,Logistic regression analysis showed that the d-Pet CO₂ value had a statistically significant effect on the occurrence of hypoxemia（OR=0.52,95%CI0.36-0.75,P<0.001）,the FEV₁/FVC%value had a statistically significant effect on the occurrence of hypoxemia（OR=0.82,95%CI0.70-0.97,P=0.016）,the 10min Pa O₂ value in the lateral position had a statistically significant effect on the occurrence of hypoxemia（OR=0.98,95%CI,0.96-1.00,P=0.018）;The CHAID decision tree suggests that d-Pet CO₂,FEV₁/FVC%and BMI are risk factors for hypoxemia during one lung ventilation,the first layer of the tree is divided according to the patient’s d-Pet CO₂ level,and the second layer of the tree is divided according to BMI and FEV₁/FVC%levels for patients whose d-Pet CO₂ is between(1,4]or>4,there were interactions between d-Pet CO₂ and FEV₁/FVC%,d-Pet CO₂ and BMI,respectively.Conclusion:D-Pet CO₂ and FEV₁/FVC%have important predictive value for the occurrence of hypoxemia during one-lung ventilation,and there is a certain relationship between them.The Logistic regression model and the CHAID decision tree model are mutually verified,which can more efficiently identify high-risk groups of hypoxemia during one-lung ventilation,so that anesthesiologists can take relevant intervention and treatment as soon as possible.