本文已被:浏览 216次 下载 162次
投稿时间:2025-04-17 网络发布日期:2025-11-26
投稿时间:2025-04-17 网络发布日期:2025-11-26
中文摘要: 目的 探讨老年脓毒症患者发生急性肾损伤(AKI)的危险因素,构建老年脓毒症患者发生AKI的预测模型并验证模型的预测价值。方法 以美国重症监护医学信息数据库-Ⅳ 2.2(MIMIC-Ⅳ 2.2)中收录的住院老年脓毒症患者数据进行回顾性队列研究。收集患者人口学、临床和实验室数据。按7∶3随机分为训练集和验证集。在训练集中使用10折交叉验证的最小绝对收缩和选择算法(LASSO)回归进行特征选择,然后进行logistic回归分析建立老年脓毒症并发AKI的预测模型并绘制列线图,并在验证集中验证。通过受试者工作特征曲线(ROC)、临床影响曲线(CIC)评价预测模型的预测值。结果 共纳入5 792例老年脓毒症患者,其中4 888例发生AKI(发生率84.4%)。多因素logistic 回归分析显示,机械通气(OR=2.115,95%CI:1.722~2.598)、充血性心力衰竭(OR=2.237,95%CI:1.771~2.824)、身体质量指数(BMI)(OR=1.108,95%CI:1.09~1.13)、活化部分凝血活酶时间(APTT)(OR=1.010,95%CI:1.004~1.017)、乳酸水平(OR=1.114,95%CI:1.018~1.225)、急性生理学评分(APS)(OR=1.025,95%CI:1.020~1.031)、动脉血氧分压(PaO2)(OR=1.003,95%CI:1.002~1.004)、尿量(OR=0.942,95%CI:0.932~0.951)是AKI发生的独立影响因素(P<0.05)。结合以上8个变量绘制老年脓毒症患者发生AKI预测模型的静态列线图。训练集中列线图预测脓毒症患者发生AKI 的ROC 曲线下面积(AUC)为0.803(95%CI:0.786~0.821),灵敏度为0.733,特异度为0.726,最佳截断值为0.829,提示该模型区分度尚可;Hosmer-Lemeshow检验显示,预测模型有较好的校准能力(P=0.976)。CIC曲线亦证明该模型具有良好的临床效用。结论 尿量、充血性心力衰竭、BMI、APTT、乳酸、PaO2和APS以及机械通气是老年脓毒症患者发生AKI的主要影响因素,基于上述因素构建的老年脓毒症并发AKI的预测模型能够帮助临床医生尽早识别高危患者,并及时干预。
中文关键词: 老年 脓毒症 急性肾损伤 美国重症监护医学信息数据库⁃Ⅳ 列线图 急性生理学评分 活化部分凝血活酶时间 充血性心力衰竭 机械通气
Abstract:Objective To investigate the risk factors for acute kidney injury (AKI)in elderly septic patients,construct a predictive model for AKI in elderly septic patients,and validate the predictive value of the model.Methods A retrospective cohort study was conducted using the data from hospitalized elderly septic patients collected in the Medical Information Mart for Intensive Care database - Ⅳ 2.2(MIMIC - Ⅳ 2.2). Demographic,clinical and laboratory data were collected. The patients were randomly divided into a training set and a validation set at a 7:3 ratio.Feature selection was performed in the training set using the least absolute shrinkage and selection operator(LASSO)regression with 10-fold cross-validation,followed by logistic regression analysis to establish a predictive model for AKI in elderly septic patients. A nomogram was draw and the model was then validated in the validation set. The predictive value of the model was evaluated using the receiver operating characteristic(ROC)curve and the clinical impact curve(CIC).Results A total of 5 792 elderly septic patients was included,of which 4 888 developed AKI(incidence rate:84.4%).Multivariate logistic regression analysis revealed that mechanical ventilation(OR=2.115,95%CI:1.722-2.598),congestive heart failure(OR=2.237,95%CI:1.771-2.824),body mass index(BMI)(OR=1.108,95%CI:1.09-1.13),activated partial thromboplastin time(APTT)(OR=1.010,95%CI:1.004-1.017),lactic acid level(OR=1.114,95%CI:1.018-1.225),Acute Physiology Score(APS)(OR=1.025,95%CI:1.020-1.031),arterial partial pressure of oxygen(PaO2)(OR=1.003,95%CI:1.002-1.004),and urine output(OR=0.942,95%CI:0.932-0.951)were independent influencing factors for the development of AKI(P<0.05). A static nomogram for predicting AKI in elderly septic patients was constructed based on these eight variables. In the training set,the area under the ROC curve(AUC)of the nomogram for predicting the occurrence of AKI in sepsis patients was 0.803(95%CI:0.786-0.821),with a sensitivity of 0.733 and specificity of 0.726,and an optimal cut-off value of 0.829,indicating that the model had moderate discriminatory ability.The Hosmer-Lemeshow test showed good calibration of the predictive model(P=0.976). The CIC also demonstrated that the model had good clinical utility. Conclusion Urine output,congestive heart failure,BMI,APTT,lactate,PaO2,APS,and mechanical ventilation are the main influencing factors for AKI in elderly septic patients. The predictive model for AKI in elderly septic patients,based on these factors,can help clinicians identify high-risk patients early and provide timely interventions.
keywords: Elderly Sepsis Acute kidney injury Medical Information Mart for Intensive Care database ⁃ Ⅳ Nomogram Acute Physiology Score Activated partial thromboplastin time Congestive heart failure Mechanical ventilation
文章编号: 中图分类号:R631 R692.5 文献标志码:A
基金项目:甘肃省自然科学基金(21JR11RA005);兰州市科技计划项目(2023?ZD?180);中央高校基本科研业务费专项资金资助(31920240061)
附件
| 作者 | 单位 |
| 方春天 | 1. 中国人民解放军联勤保障部队第九四〇医院肾内科, 甘肃 兰州 730050 |
| 刘冬梅 | 2. 中国人民解放军联勤保障部队第九四〇医院重症医学科, 甘肃 兰州 730050;3. 西北民族大学医学部, 甘肃 兰州 730050 |
| Author Name | Affiliation |
| FANG Chuntian*,LIU Dongmei | *Department of Nephrology,the 940th Hospital of Joint Logistics Support Force of PLA,Lanzhou,Gansu 730050,China |
引用文本:
方春天,刘冬梅.基于MIMIC⁃Ⅳ数据库的老年脓毒症患者急性肾损伤风险预测模型[J].中国临床研究,2025,38(11):1648-1653.
方春天,刘冬梅.基于MIMIC⁃Ⅳ数据库的老年脓毒症患者急性肾损伤风险预测模型[J].中国临床研究,2025,38(11):1648-1653.
