肥胖状态对呼吸机相关肺炎患者1年死亡率的影响及其列线图预测模型构建

The Impact of Obesity on the 1-Year Mortality Risk of Patients with Ventilator-Associated Pneumonia and the Construction of Its Nomogram Prediction Model

摘要: 背景：本研究主要探索肥胖对呼吸机相关肺炎(Ventilator-Associated Pneumonia, VAP)患者1年全因死亡率的影响，构建呼吸机相关肺炎的肥胖患者长期死亡风险模型，进一步探究“肥胖悖论”对呼吸机相关肺炎的影响。方法：我们使用大型公共数据库提取进入重症监护室(Intensive Care Unit, ICU)内的合适的患者，将患者根据BMI分为肥胖和非肥胖两类。首先，调节协变量后进行倾向性匹配评分(Propensity Score Matching, PSM)减少选择偏差，探究肥胖状态与VAP不良预后的具体关系。其次，为了进一步研究VAP肥胖患者1年全因死亡率的风险因素，我们通过套索回归分析(Least Absolute Shrinkage and Selection Operator, LASSO)联合多因素逻辑回归分析选择特征变量来构建VAP肥胖患者长期死亡率的列线图预测模型。最后，通过绘制校准曲线以评估模型的准确性和可靠性。结果：最终，我们提取到资料齐全的1506名VAP患者，其中肥胖患者有665名。调整一切混杂变量后，进行PSM后发现肥胖状态并不影响VAP患者短期死亡率，无明显统计学差异(P > 0.05)；而在1年死亡率中，肥胖患者的死亡风险显著低于非肥胖患者(P < 0.05)。于是在通过LASSO回归分析联合多因素逻辑回归分析选择特征变量后，我们构建了肥胖患者1年全因死亡率的列线图预测模型，发现年龄、CCI、是否气管插管及RRT是影响死亡率的最重要的三大因素。结论：我们的研究发现，在呼吸机相关肺炎中肥胖悖论仍然存在。较高的年龄、CCI以及住院期间进行RRT的患者，可能具有较高的长期死亡风险。

Abstract: Background: This study primarily explores the impact of obesity on the 1-year all-cause mortality rate in patients with Ventilator-Associated Pneumonia (VAP). It aims to construct a long-term mortality risk model for obese patients with VAP and further investigate the effect of “obesity paradox” on VAP. Methods: We extracted data from a large public database for patients admitted to the Intensive Care Unit (ICU) and categorized them into obese and non-obese groups based on BMI. Initially, Propensity Score Matching (PSM) was used after adjusting for covariates to reduce selection bias and explore the relationship between obesity and adverse outcomes in VAP. Subsequently, to further investigate the risk factors for 1-year all-cause mortality in obese VAP patients, we applied Least Absolute Shrinkage and Selection Operator (LASSO) regression combined with multivariate logistic regression analysis to select feature variables and constructed a nomogram prediction model for predicting long-term mortality in obese patients with VAP. Finally, we assessed the accuracy and reliability of the model through calibration curves. Results: A total of 1506 VAP patients with complete data were included, of whom 665 were obese. After adjusting for all confounding variables and performing PSM, it was found that obesity status did not affect the short-term mortality of VAP patients, with no statistically significant difference (P > 0.05). However, in terms of 1-year mortality, obese patients exhibited a significantly lower risk of death compared to non-obese patients (P < 0.05). Thus, after selecting characteristic variables by LASSO regression combined with multivariate logistic regression, we constructed a nomogram prediction model to predict the 1-year all-cause mortality in obese VAP patients, and found that the most important factors influencing mortality were age, Charlson Comorbidity Index (CCI), whether or not tracheal intubation, and renal replacement therapy (RRT). Conclusion: Our study found that the obesity paradox still exists in VAP. Older age, higher CCI, and the need for RRT during hospitalization may be associated with a higher long-term mortality risk in these patients.