Logistic回归模型构建联合预测因子在细菌性血流感染中的早期诊断价值  

Early diagnostic value of Logistic regression model constructing joint predictors in bacterial bloodstream infection

作　　者：冷清清 张慧[1] 叶英[1] Leng Qingqing;Zhang Hui;Ye Ying(The First Affiliated Hospital of Anhui Medical University,Hefei,Anhui 230022,China)

机构地区：[1]安徽医科大学第一附属医院,安徽合肥230022

出　　处：《齐齐哈尔医学院学报》2025年第5期406-411,共6页Journal of Qiqihar Medical University

基　　金：国家自然科学青年科学基金项目(编号:82304209)。

摘　　要：目的探究中性粒细胞与淋巴细胞比值(NLR)与降钙素原(PCT)、C反应蛋白(CRP)对血流感染的诊疗价值及建立一个简单易行的联合预测因子计算公式用于血流感染早期诊断。方法回顾性分析2022年1月-2023年6月本院感染科收治的入院时疑似血流感染且同时抽取血培养、血常规、PCT、CRP等实验室指标的246例患者病例资料。根据血培养结果分为血培养阴性组(116例)和血培养阳性组(130例),比较两组NLR、PCT、CRP差异,相关性因素纳入二元logistic回归分析,构建联合预测因子并绘制ROC曲线,计算曲线下面积、灵敏度、特异度、阳性预测值、阴性预测值。结果血培养阳性组血清NLR、PCT、CRP水平高于血培养阴性组,差异具有统计学意义(P<0.05)。NLR、PCT、CRP的曲线下面积分别为0.807、0.833、0.746,PCT的ROC曲线下面积在NLR、PCT、CRP中最大。血清PCT诊断细菌性血流感染的最佳临界值为>1.9μmol/L,灵敏度、特异度分别为74.60%、81.00%,阳性预测值、阴性预测值分别为81.5%、74.0%。NLR联合PCT、CRP构建联合预测因子(L),联合预测因子(L)敏感度、特异度、阳性预测率和阴性预测率分别为77.7%、81.9%、82.8%、76.6%,均高于单独炎症指标。革兰阴性菌PCT、CRP水平比革兰阳性菌高,差异具有统计学意义。PCT联合CRP构建联合预测因子(L'),联合预测因子(L')在鉴别细菌类型的ROC曲线下面积最大,为0.750。结论基于NLR、PCT、CRP的logistic回归模型构建的联合预测因子对细菌性血流感染的早期判断准确性优于单一指标。基于PCT、CRP的联合预测因子可提高鉴别细菌类型的效能,进一步指导临床诊疗。Objective To explore the diagnostic and therapeutic value of neutrophil-lymphocyte ratio(NLR),procalcitonin(PCT)and C-reactive protein(CRP)in bloodstream infection,and to establish a simple and easy calculation formula of joint predictors for early diagnosis of bloodstream infection.Methods Retrospective analysis was conducted on the case data of patients with suspected bloodstream infection who were admitted to the Department of Infectious Diseases from January 2022 to June 2023,and laboratory indicators such as blood culture,blood routine,PCT and CRP were taken at the same time.Patients were divided into negative blood culture group(116 cases)and positive blood culture group(130 cases)based on the results of blood culture.The differences of NLR,PCT and CRP between the two groups were compared,and the related factors were included in logistic regression analysis.The joint predictor was constructed and the ROC curve was drawn,and the area,sensitivity,specificity,positive predictive value and negative predictive value under the curve were calculated.Results The levels of serum NLR,PCT and CRP in blood culture positive group were higher than those in blood culture negative group,and the difference was statistically significant(P<0.05).The areas under the curve of NLR,PCT and CRP were 0.807,0.833 and 0.746 respectively,and the area under the ROC curve of PCT was the largest among NLR,PCT and CRP.The optimal critical value of serum PCT in the diagnosis of bacterial bloodstream infection was>1.9μmol/L,with sensitivity and specificity of 74.60%and 81.00%,positive predictive value and negative predictive value of 81.5%and 74.0%respectively.A joint predictor(L)was constructed by NLR combining with PCT and CRP.The sensitivity,specificity,positive predictive rate and negative predictive rate of the joint predictor(L)were 77.7%,81.9%,82.8%and 76.6%,respectively,which were higher than those of the single inflammatory index.The levels of PCT and CRP in gram-negative bacteria were higher than those in gram-positive bacteria,a

关 键 词：血流感染 中性粒细胞与淋巴细胞比值 降钙素原 C反应蛋白 

分 类 号：R515[医药卫生—内科学]