文章摘要
李丹丹,石艳玺.儿童难治性肺炎支原体肺炎的危险因素分析及列线图模型构建[J].安徽医药,2026,30(7):1344-1348.
儿童难治性肺炎支原体肺炎的危险因素分析及列线图模型构建
Risk factors for refractory mycoplasma pneumoniae pneumonia in children and construction of a nomogram model
  
DOI:10.3969/j.issn.1009-6469.2026.07.014
中文关键词: 肺炎,支原体  儿童  难治性肺炎  列线图  预测模型
英文关键词: Pneumonia, mycoplasma  Children  Refractory pneumonia  Nomogram  Prediction modeling
基金项目:河北省医学科学研究课题项目( 20211652)
作者单位E-mail
李丹丹 河北省儿童医院、河北省儿童健康与疾病临床医学研究中心呼吸五科,河北石家庄 050031  
石艳玺 河北省儿童医院、河北省儿童健康与疾病临床医学研究中心呼吸五科,河北石家庄 050031 784583822@qq.com 
摘要点击次数: 314
全文下载次数: 124
中文摘要:
      目的探索儿童难治性肺炎支原体肺炎( RMPP)的危险因素,构建 RMPP列线图风险预测模型。方法回顾性分析 2021年 1月至 2024年 1月在河北省儿童医院收治的 262例肺炎支原体肺炎( MPP)病例资料。将数据集按照 7∶3分为训练集( 182例)和验证集( 80例)。对训练集病例资料进行单因素、相关性、受试者操作特征曲线( ROC曲线)分析和多因素 logistic回归分析,筛选出 RMPP的独立影响因素,采用 R软件制作列线图,用验证集中样本进行验证,模型区分度评估采用一致性指数( C.index),校准度评估采用校准图。结果 262例 MPP病儿中 RMPP病儿 65例( 24.8%),RMPP组病儿的 发热时长、咳嗽时长、红细胞沉降率( ESR)、中性粒细胞计数( NEU)分别为( 9.1±4.0)d、(11.2±5.2)d、(58.6±32.5)mm/h、(7.2±4.0)×109/L,均高于非 RMPP组病儿( 8.0±3.8)d、(9.2±4.7)d、(43.3±23.0)mm/h、(5.2±4.6)×109/L,RMPP组病儿的白蛋白( ALB)、乳酸脱氢酶( LDH)、 C反应蛋白( CRP)、降钙素原( PCT)分别为( 34.6±5.3)g/L、(356.5±99.1)U/L、(25.4±16.3)mg/ L、(0.29±0.16)μg/L,均低于非 RMPP组病儿( 40.6±6.0)g/L、(407.4±123.6)U/L、(45.6±20.2)mg/L、(0.46±0.21)μg/L,均差异有统计学意义( P<0.05)。单因素分析、相关性分析、 ROC曲线及多因素 logistic回归分析结果显示 NEU[OR 95%CI:3.03(1.43,6.41)]、 ESR[OR 95%CI:3.53(1.69,7.38)]、 LDH水平[ OR 95%CI:4.29(2.05,8.98)]、 CRP水平[ OR 95%CI:5.47(2.28,13.13)]和 PCT水平[ OR 95%CI:4.68(2.26,9.68)]是 MPP病儿中 RMPP发生的最佳预测因子。预测模型的 C-index为 0.85[95%CI:(0.80,0.90)],特异度和灵敏度分别为 65.4%和 87.6%,验证集的 C-index为 0.86。校准图显示在训练集和验证集的校准曲线与理想曲线重合度较高。结论该研究构建的难治性肺炎支原体肺炎的列线图风险预测模型区分度与校准度较好,可为临床提供参考价值。
英文摘要:
      Objective To investigate the risk factors associated with refractory Mycoplasma pneumoniae pneumonia (RMPP) in chil-dren and to develop a nomogram-based risk prediction model.Methods A retrospective analysis was conducted on the medical re-cords of 262 Mycoplasma pneumoniae pneumonia (MPP) cases admitted to Hebei Children's Hospital from January 2021 to January2024. These cases were randomly divided into a training set (n=182) and a validation set (n=80) in a 7:3 ratio. Univariate analysis, cor-relation analysis, receiver operating characteristic curve (ROC curve) analysis, and multivariate logistic regression analysis were per-formed on the training set data to identify the independent predictors of RMPP. Subsequently, nomograms were designed using R soft-ware and validated with the validation set data. The model's discriminative ability was evaluated using the concordance index (C-in-dex), and calibration was assessed through calibration plots.Results Among the 262 children with MPP, 65 (24.8%) cases were identi-fied as RMPP. The duration of fever, cough, erythrocyte sedimentation rate (ESR), and neutrophil count (NEU) in the RMPP group were(9.1±4.0) d, (11.2±5.2) d, (58.6± 32.5) mm/h, and (7.2±4.0) × 109/L respectively, which were all higher than those in the non-RMPP group [(8.0±3.8) d, (9.2±4.7) d, (43.3± 23.0) mm/h, and (5.2±4.6) ×109/L, respectively]. Albumin (ALB), lactate dehydrogenase (LDH), C-reactive protein (CRP), procalcitonin (PCT) in the RMPP group were (34.6±5.3) g/L, (356.5±99.1) U/L, (25.4±16.3) mg/L and (0.29±0.16) μg/L respectively, all of which were lower than those in the non-RMPP group [(40.6±6.0) g/L, (407.4±123.6) U/L, (45.6±20.2) mg/ L and (0.46±0.21) μg/L], with statistically significant differences (P<0.05). The results of univariate analysis, correlation analysis, ROCanalysis, and multivariate logistic regression analysis indicated that NEU [OR 95%CI:3.03 (1.43, 6.41)], ESR [OR 95%CI: 3.53 (1.69, 7.38)], LDH level [OR 95%CI: 4.29 (2.05, 8.98)], CRP level [OR 95%CI: 5.47 (2.28, 13.13)], and PCT level [OR 95%CI: 4.68 (2.26, 9.68)] are significant predictors. The prediction model demonstrated a C-index of 0.85 [95%CI: (0.80, 0.90)], with a specificity of 65.4% and a sensitivity of 87.6%. The validation set exhibited a C-index of 0.86. Calibration plots for both the training and validationsets were closely aligned with the ideal curves, indicating good calibration.Conclusion The constructed nomogram-based risk predic-tion model for refractory Mycoplasma pneumoniae pneumonia in children exhibits excellent discrimination and calibration capabilities,providing valuable clinical insights.
查看全文   查看/发表评论  下载PDF阅读器
关闭