【摘要】 目的 鉴于脓毒症相关肝损伤(SALI)危害严重，本研究旨在运用Lasso-logistic回归构建SALI预测模型，并制定针对性的防控策略。方法 选取2024年4月1日—2025年4月1日安徽医科大学第一附属医院收治的192例脓毒症患者为研究对象，根据是否发生肝损伤分为SALI组(40例)和NSALI组(152例)。收集2组患者临床相关资料，采取单因素分析研究不同因素对脓毒症患者SALI发生的影响。采用Lasso-logistic回归分析研究SALI发生的危险因素，建立SALI风险预测模型并使用ROC曲线进行验证分析。结果 2组合并糖尿病、机械通气、合并休克、低氧血症、尿素、血小板计数、血乳酸(LAC)、序贯器官衰竭评分(SOFA评分)比较差异均有统计学意义(P<0.05)。Lasso-logistic回归分析显示，合并糖尿病、合并休克、LAC、SOFA评分是发生SALI的独立危险因素(P<0.05)。ROC曲线分析显示，合并糖尿病、合并休克、LAC、SOFA评分及列线图模型均可预测SALI,曲线下面积分别为0.615、0.620、0.843、0.837、0.948,其中列线图风险预测模型的AUC最高。结论 合并糖尿病、合并休克、LAC、SOFA评分是SALI发生独立危险因素，基于此构建的列线图模型预测与实用价值高，利于为SALI患者制定精准防控策略。更多还原

【Abstract】 Objective In view of the serious harm of sepsis-related liver injury(SALI), this study aims to use lasso logistic regression to build a SALI prediction model and to develop targeted prevention and control strategies. Methods A total of 192 sepsis patients admitted to the First Affiliated Hospital of Anhui Medical University from April 1, 2024 to April 1, 2025 were selected as the research subjects. They were divided into the SALI group(40 cases) and the non SALI group(152 cases) based on whether liver injury occurred. Clinical data from two groups of patients were collected, and univariate analysis was performed to identify factors associated with the occurrence of SALI in sepsis patients. The risk factors of SALI were analyzed by lasso logistic regression, and the SALI risk prediction model was established and verified by ROC curve. Results Significant differences were observed between the two groups in the scores of diabetes, mechanical ventilation, shock, hypoxemia, blood urea, platelet count, blood lactic acid, and sequential organ failure assessment(SOFA) scores(P<0.05). Lasso-logistic regression analysis showed that diabetes mellitus, combined shock, blood lactic acid, and SOFA scores were independent risk factors for SALI(P<0.05). ROC curve analysis results showed that SALI can be predicted by combining diabetes, combined shock, blood lactic acid, SOFA scores, and the nomogram model were predictive of SALI, with area under the curve values of 0.615, 0.620, 0.843, 0.937, 0.948, respectively. Among these, the nomogram-based prediction model exhibited the highest predictive performance. Conclusion Diabetes mellitus, combined shock, blood lactic acid levels, and SOFA score are independent risk factors for SALI. The nomogram model constructed based on these factors shows high predictive accuracy and clinical applicability, and may facilitate the development of accurate prevention and control strategies for SALI patients.更多还原