【摘要】 目的 探讨重症肺炎患者死亡的危险因素，并构建预测模型。方法 回顾性分析2022年1月至2025年12月于武汉大学人民医院重症医学科住院的398例重症肺炎患者的临床资料。按照7∶3的比例将患者随机分为训练集283例和验证集115例，训练集根据患者入ICU后28天是否死亡分为死亡组129例和生存组154例。收集患者一般临床资料，记录入ICU 24 h内最差的急性生理学与慢性健康状况评分Ⅱ(acute physiology and chronic health evaluationⅡ, APACHEⅡ)、序贯器官衰竭评分(sequential organ failure assessment, SOFA),记录患者住院期间其他相关并发症发生情况、实验室指标及预后相关指标等，采用单因素分析和多因素Logistic回归分析结合临床实际情况筛选危险因素，构建重症肺炎患者死亡的预测模型，并利用列线图进行可视化。在训练集和验证集中通过受试者工作特征(ROC)曲线、校准曲线、决策曲线分析进行模型预测能力评估。结果 共纳入398例重症肺炎患者，221例生存(55.53%)。>80岁(OR=3.284, 95%CI 1.582～6.815,P=0.001),既往糖尿病史(OR=2.376,95%CI 1.288～4.386,P=0.006),合并脓毒症(OR=2.205,95%CI 1.265～3.845,P=0.005),APACHEⅡ评分(OR=1.091,95%CI 1.040～1.146,P<0.001),脑钠肽(OR=1.000, 95%CI 1.000～1.000,P=0.047),D-二聚体(OR=1.026,95%CI 1.004～1.049,P=0.022)是重症肺炎患者死亡的独立危险因素。构建的列线图在推导队列[ROC曲线下面积(AUC)=0.797]和验证集(AUC=0.719)中均显示出中等预测能力。结论 本研究识别了重症肺炎患者死亡的危险因素，并构建了有效的预测模型。该模型可辅助临床早期识别高风险患者，优化治疗策略。更多还原

【Abstract】 Objective To explore the risk factors for death in patients with severe pneumonia and construct a predictive model. Methods Clinical data from patients with severe pneumonia admitted to the Department of Critical Care Medicine at Renmin Hospital of Wuhan University from January 2022 to December 2025 were retrospectively analyzed. The patients were randomly divided into a training set of 283 cases and a validation set of 115 cases at a ratio of 7∶3. The training set was further divided into a death group of 129 cases and a survival group of 154 cases based on whether the patients died within 28 days of admission to the intensive care unit. General clinical data of the patients were collected, and the worst acute physiology and chronic health evaluationⅡ(APACHEⅡ)and sequential organ failure assessment scores within 24 hours of intensive care unit admission were recorded. Other related complications, laboratory indicators, and prognosis-related indicators during hospitalization were also recorded. Univariate analysis and multivariate logistic regression were used to screen for risk factors, and a predictive model for mortality in patients with severe pneumonia was constructed and visualized using a nomogram. The predictive performance of the model was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis in both the training and validation sets. Results A total of 398 patients with severe pneumonia were included, of whom 221 survived(55.53%). Age over 80 years [odds ratio(OR)=3.284, 95% confidence interval(CI) 1.582-6.815, P=0.001], a history of diabetes(OR=2.376, 95%CI 1.288-4.386, P=0.006), concurrent sepsis(OR=2.205, 95%CI 1.265-3.845, P=0.005), APACHEⅡ score(OR=1.091, 95%CI 1.040-1.146, P<0.001), BNP(OR=1.000, 95%CI 1.000-1.000, P=0.047), and D-dimer(OR=1.026, 95%CI 1.004-1.049, P=0.022) were independent risk factors for death in patients with severe pneumonia. The constructed nomogram demonstrated moderate predictive ability in both the derivation cohort [area under the receiver operating characteristic curve(AUC)=0.797] and the validation set(AUC=0.719). Conclusions This study identified risk factors for mortality in patients with severe pneumonia and developed an effective predictive model. This model can assist in the early identification of high-risk patients in clinical practice and optimize treatment strategies.更多还原