儿童大环内酯类药物无反应性肺炎支原体肺炎临床预测模型的建立与评估  

Development and evaluation of a clinical prediction model for macrolide-unresponsive Mycoplasma pneumoniae pneumonia in children

作　　者：刘洁 孙杰 刘业琼 徐炜新 LIU Jie;SUN Jie;LIU Yeqiong;XU Weixin(Department of Clinical Laboratory,Central Hospital of Jiading District,Shanghai 201899,China)

机构地区：[1]上海市嘉定区中心医院检验科,上海201899

出　　处：《国际检验医学杂志》2024年第19期2386-2391,共6页International Journal of Laboratory Medicine

基　　金：上海市嘉定区自然科学研究课题(JDKW-2022-0014);嘉定区中心医院双馨培养计划(SX202206)。

摘　　要：目的 探讨大环内酯类药物无反应性肺炎支原体肺炎(MUMPP)的早期预测指标,并构建列线图预测模型。方法 回顾性收集2023年1月至2024年2月就诊于该院的159例肺炎支原体肺炎(MPP)患儿的临床资料,按照就诊时间分为建模组(112例)和验证组(47例)。建模组里再根据药物反应分为MUMPP组(51例)和MPP组(61例),比较各组临床资料及实验室指标,通过单因素和多因素分析MUMPP的独立预测因子,构建列线图预测模型,并通过受试者工作特征(ROC)曲线及曲线下面积(AUC)、决策曲线、校准曲线、Hosmer-Lemeshow拟合优度检验评估模型的区分度、临床实用性和校准度。结果 系统免疫炎症指数(SII)、C反应蛋白(CRP)/白蛋白(ALB)、D-二聚体是MUMPP的独立影响因素(P<0.05)。建模组的AUC为0.938(95%CI 0.890~0.986),验证组的AUC为0.912(95%CI 0.832~0.992)。Hosmer-Lemeshow拟合优度检验的χ^(2)=3.768,P=0.877,阈值概率在5%~99%有较高的临床净获益。结论 用SII、CRP/ALB、D-二聚体建立的列线图预测模型对MUMPP患儿的预测准确性较好,临床应用价值高。Objective To explore the early predictors of macrolide-unresponsive Mycoplasma pneumoniae pneumonia(MUMPP)and construct a nomogram prediction model.Methods The clinical data of 159 children with Mycoplasma pneumoniae pneumonia(MPP)admitted to the hospital from January 2023 to February 2024 were retrospectively collected.According to the time of admission,they were divided into modeling group(112 cases)and validation group(47 cases).The modeling group was further divided into MUMPP group(51 cases)and MPP group(61 cases)according to the drug response.The clinical data and laboratory indexes of each group were compared.The independent predictors of MUMPP were analyzed by univariate and multivariate analysis,and a nomogram prediction model was constructed.Receiver operating characteristic(ROC)curve,area under the curve(AUC),decision curve,calibration curve and Hosmer-Lemeshow goodness of fit test were used to evaluate the discrimination,clinical practicability and calibration of the model.Results The systemic immune inflammation index(SII),C-reactive protein(CRP)/albumin(ALB)and Ddimer were independent influencing factors of MUMPP(P<0.05).The AUC was 0.938(95%CI 0.890-0.986)in the modeling group and 0.912(95%CI 0.832-0.992)in the validation group.χ^(2) was 3.768 and P was 0.877 in Hosmer-Lemeshow goodness-of-fit test,threshold probability between 5%and 99%had high net clinical benefit.Conclusion The nomogram prediction model established by SII,CRP/ALB and D-dimer has good prediction accuracy and high clinical application value for children with MUMPP.

关 键 词：大环内酯类药物无反应性肺炎支原体肺炎 新型炎症指标 列线图 预测模型 

分 类 号：R446.1[医药卫生—诊断学]