儿童肺炎支原体合并EB病毒感染的列线图诊断模型构建与验证

Construction and validation of nomogram diagnosis model for EBV co-infection in children with Mycoplasma pneumoniae pneumonia

  • 摘要:
    目的 构建儿童肺炎支原体肺炎(MPP)合并EB病毒(EBV)感染列线图诊断模型。
    方法 回顾南京医科大学附属儿童医院2020年7月-2024年7月收治的427例MPP儿童的临床资料, 分为建模组(n=299)和验证组(n=128), 另根据建模组EBV感染情况分为MPP组(n=235)和MPP合并感染组(n=64)。多因素logistic回归分析MPP儿童合并EBV感染的危险因素, 并构建MPP儿童合并EBV感染的列线图诊断模型, 采用受试者工作特征(ROC)曲线、校准曲线、临床决策(DCA)曲线评估模型诊断价值和临床应用价值。
    结果 MPP合并感染组白细胞计数(WBC)为(12.37±2.32)×109/L, 高于MPP组(P<0.05), PLT和血红蛋白(Hb)分别为(197.95±32.85)×109/L和(102.58±13.74)g/L, 低于MPP组(P<0.05), 且MPP合并感染组发热时间≥10 d、呼吸困难以及胸腔积液的占比高于MPP组(P<0.05)。多因素logistic回归结果显示, WBC(OR=1.514)、PLT(OR=0.970)、Hb(OR=0.959)、发热时间(OR=4.790)、呼吸困难(OR=3.777)及胸腔积液(OR=4.795)与MPP儿童EBV感染有关(P<0.05)。列线图显示, MPP患儿模型总得分为219分时, 感染EBV概率达0.9。建模组和验证组ROC曲线下面积分别为0.882(95%CI:0.836~0.927)和0.943(95%CI:0.902~0.984), 敏感度为76.56%和91.30%, 特异度为82.55%和85.71%。H-L拟合度检验显示, 建模组χ2=4.124, P=0.846, 验证组χ2=4.203, P=0.838。DCA曲线显示模型临床应用价值较高。
    结论 WBC、PLT、Hb水平、发热时间、呼吸困难和胸腔积液诊断MPP患儿合并EBV感染有一定价值, 基于上述六个因素构建的列线图模型诊断效能较佳。

     

    Abstract:
    OBJECTIVE To construct a nomogram diagnosis model for Epstein-Barr virus (EBV) co-infection in children with Mycoplasma pneumoniae pneumonia (MPP).
    METHODS Clinical data of 427 children with MPP admitted to the Children′s Hospital of Nanjing Medical University from Jul. 2020 to Jul. 2024 were retrospectively analyzed. The children were divided into a modeling group (n=299) and a validation group (n=128). The modeling group was further categorized into an MPP group (n=235) and an MPP co-infection group (n=64) based on EBV infection status. Multivariate logistic regression was used to identify risk factors for EBV co-infection in children with MPP, and a nomogram diagnosis model was constructed. The diagnostic value and clinical application value of the model were evaluated by receiver operating characteristic (ROC) curves, calibration curves and decision curve analysis (DCA).
    RESULTS The white blood cell count (WBC) in the MPP co-infection group was (12.37±2.32)×109/L, significantly higher than that in the MPP group (P < 0.05). Platelet count (PLT) and hemoglobin (Hb) levels were (197.95±32.85)×109/L and (102.58±13.74) g/L, respectively, lower than those in the MPP group (P < 0.05). Additionally, the MPP co-infection group exhibited higher proportions of fever duration ≥10 days, dyspnea and pleural effusion compared to the MPP group (P < 0.05). Multivariate logistic regression analysis revealed that WBC (OR=1.514), PLT (OR=0.970), Hb (OR=0.959), fever duration (OR=4.790), dyspnea (OR=3.777) and pleural effusion (OR=4.795) were significantly associated with EBV infection in children with MPP (P < 0.05). The nomogram demonstrated that when the total model score reached 219 points, the probability of EBV infection in children with MMP was 0.9. The areas under the ROC curve for the modeling group and validation group were 0.882 (95%CI: 0.836-0.927) and 0.943 (95%CI: 0.902-0.984), respectively, with sensitivities of 76.56% and 91.30%, respectively, and specificities of 82.55% and 85.71%, respectively. The Hosmer-Lemeshow goodness-of-fit test showed χ2=4.124, P=0.846 for the modeling group and χ2=4.203, P=0.838 for the validation group. DCA curve indicated high clinical application value of the model.
    CONCLUSIONS WBC, PLT, Hb levels, fever duration, dyspnea and pleural effusion have diagnostic values for EBV co-infection in children with MPP. The nomogram model constructed based on these six factors demonstrates excellent diagnostic performance.

     

/

返回文章
返回