严重创伤患者脓毒症风险的数据挖掘  

Data mining of sepsis risk in patients with severe trauma

作　　者：许静[1] 焦小琴 关红 王鹏森 杜俊凯[1] XU Jing;JIAO Xiaoqin;GUAN Hong;WANG Pengsen;DU Junkai(Department of Emergency,the First Affiliated Hospital of Xi'an Jiaotong University,Xi'an,710061,China;Intensive Care Unit,Shaanxi Provincial People's Hospital;Department of Emergency,Xi'an Gao Xin Hospital;Department of Emergency,People's Hospital of Tong Chuan)

机构地区：[1]西安交通大学第一附属医院急诊科,西安710061 [2]陕西省人民医院重症医学科 [3]西安高新医院急诊科 [4]铜川市人民医院急诊科

出　　处：《临床急诊杂志》2022年第12期847-853,共7页Journal of Clinical Emergency

基　　金：陕西省自然科学基金(No:2018JM7152);睿E(睿意)急诊医学研究专项基金(No:R2019010)。

摘　　要：目的:探讨严重创伤后发生脓毒症的高危预警因素,以期建立创伤后脓毒症发生的预测体系。方法:回顾性分析2017年1月1日—2020年12月31日期间西安交通大学第一附属医院收治的严重创伤住院患者的临床资料,联合lasso-logistic回归分析、ROC分析及列线图预警模型,探讨影响其发生创伤后脓毒症的相关因素,并研究其预测效能。结果:①纳入234例严重创伤患者,其中合并脓毒症117例(病例组),未合并脓毒症117例(对照组)。Lasso-logistic回归分析结果显示年龄(OR=3.03,P=0.004,95%CI:1.48~6.87)、伤后入院时间(OR=1.11,P=0.021,95%CI:1.02~1.21)、ISS评分(OR=3.35,P=0.001,95%CI:1.59~7.97)、平均体温(OR=1.35,P=0.001,95%CI:5.38~44.3)、中性粒细胞计数(OR=6.10,P=0.007,95%CI:1.76~25.3)、PCT(OR=4.16,P<0.001,95%CI:2.12~9.23)和是否24 h内手术(OR=7.93,P=0.001,95%CI:2.40~30.6)是创伤后脓毒症发生的高危预警因素。②列线图预警模型经内部验证后有很好的预测价值。结论:年龄>40岁、伤后入院时间较长、ISS评分>20分、低热、中性粒细胞计数升高、PCT>0.5 ng/mL、24 h内手术是创伤后脓毒症发生的高危预警因素。列线图可以使预警模型量化、可视化、可读化,方便、简洁明了,临床应用价值较高,值得推广及更进一步的研究。Objective: To explore the high risk early warning factors of sepsis after severe trauma, so as to establish a prediction system for the occurrence of sepsis after trauma. Methods: The clinical data of severe trauma inpatients in our hospital from January 1, 2017 to December 31, 2020 were retrospectively included, and combined with lasso-logistic regression analysis, ROC analysis and nomograph early warning model, to explore the related factors affecting the occurrence of post-traumatic sepsis, and study its predictive efficiency. Results:①234 patients with severe trauma were included, including 117 patients with sepsis(case group) and 117 patients without sepsis(control group). Lasso logistic regression analysis showed that age(OR=3.03, P=0.004, 95%CI: 1.48-6.87), admission time after injury(OR=1.11, P=0.021, 95%CI: 1.02-1.21), ISS score(OR=3.35, P=0.001, 95%CI: 1.59-7.97), mean body temperature(OR=1.35, P=0.001, 95%CI: 5.38-44.3), neutrophil count(OR=6.10, P=0.007, 95%CI: 1.76-25.3), PCT(OR=4.16, P<0.001, 95%CI: 2.12-9.23) and whether to operate within 24 h(OR=7.93, P=0.001, 95%CI: 2.40-30.6) were the high-risk early warning factors for post-traumatic sepsis.②The nomograph early warning model has good prediction value after internal verification. Conclusion: Age >40 years old, long admission time after injury, ISS score >20, low fever, increased neutrophil count, PCT >0.5 ng/mL, and operation within 24 h are the high-risk early warning factors for post-traumatic sepsis. Nomogram can make the early warning model quantitative, visual, readable, convenient, concise and clear. It has high clinical application value and is worthy of promotion and further research.

关 键 词：严重创伤 脓毒症 预警因素 数据挖掘 

分 类 号：R631.2[医药卫生—外科学]