基于血清炎症标志物构建儿童肺炎支原体肺炎合并胸腔积液的预警模型

Construction of early warning model for mycoplasma pneumoniae pneumonia with pleural effusion in children based on serum inflammatory markers

摘要:

目的 基于血清炎症标志物构建儿童肺炎支原体肺炎（MPP）合并胸腔积液的预警模型。方法 回顾性分析2021年7月—2023年7月中国人民解放军北部战区总医院收治的171例MPP合并胸腔积液患儿的临床资料，根据8∶2随机分成训练集137例与验证集34例。根据患儿预后分为预后不良组与预后良好组。比较训练集预后不良组与预后良好组患儿的临床资料及血清炎症标志物［白细胞介素-17A（IL-17A）、高迁移率族蛋白B1（HMGB1）、降钙素原（PCT）、淀粉样蛋白A（SAA）］；采用多因素逐步Logistic回归模型分析MPP合并胸腔积液患儿预后不良的影响因素；基于血清炎症标志物构建并验证MPP合并胸腔积液患儿预后不良的预警模型。结果 训练集137例患儿中，预后不良发生率为35.77%（49/137）；验证集34例患儿中，预后不良发生率为35.29%（12/34）。训练集预后不良组患儿持续发热时间、住院时间、D-二聚体（D-D）均高于预后良好组（P <0.05）。训练集预后不良组患儿血清IL-17A、HMGB1、PCT、SAA水平均高于预后良好组（P <0.05）。多因素逐步Logistic回归分析结果显示，IL-17A水平［O^R=3.951（95% CI：1.737，8.988）］、HMGB1水平［O^R =5.339（95% CI：2.347，12.145）］、PCT水平［O^R =4.084（95% CI：1.795，9.290）］和SAA水平［O^R =4.749（95% CI：2.088，10.804）］为MPP合并胸腔积液患儿预后不良的危险因素（P <0.05）。训练集模型预测MPP合并胸腔积液患儿预后不良的敏感性为89.80%（95% CI：0.770，0.962），特异性为92.05%（95% CI：0.838，0.965），曲线下面积为0.907（95% CI：0.762，0.969）。验证集模型预测MPP合并胸腔积液患儿预后不良的敏感性为83.33%（95% CI：0.509，0.971），特异性为90.91%（95% CI：0.694，0.984），曲线下面积为0.865（95% CI：0.717，0.935）。结论 血清IL-17A、HMGB1、PCT、SAA与MPP合并胸腔积液患儿预后不良有关，基于血清IL-17A、HMGB1、PCT、SAA构建预警模型有助于早期甄别MPP合并胸腔积液患儿预后不良的风险。

Abstract:

Objective To construct an early warning model for mycoplasma pneumoniae pneumonia (MPP) with pleural effusion in children based on serum inflammatory markers.Methods A retrospective analysis of the clinical data of 171 children with MPP and pleural effusion admitted to the hospital from July 2021 to July 2023 was conducted. The data were randomly divided into a training set (n = 137) and a validation set (n = 34) at an 8:2 ratio. During hospitalization, the children were categorized into poor prognosis and good prognosis groups. Clinical data and serum inflammatory markers [interleukin-17A (IL-17A), high mobility group box 1 (HMGB1), procalcitonin (PCT), serum amyloid A (SAA)] were compared between the poor prognosis and good prognosis groups in the training set, and the factors influencing poor prognosis in children with MPP and pleural effusion were analyzed. An early warning model for poor prognosis in children with MPP and pleural effusion was constructed and validated based on serum inflammatory markers.Results Among the 137 children in the training set, the incidence of poor prognosis was 35.77% (49/137), while among the 34 children in the validation set, it was 35.29% (12/34). In the training set,the duration of fever, hospital stay and D-dimer (D-D) in the poor prognosis group were higher than those in the good prognosis group (P < 0.05). The serum levels of IL-17A, HMGB1, PCT, and SAA were also higher in the poor prognosis group than in the good prognosis group (P < 0.05). Logistic regression analysis showed that IL-17A level [O^R = 3.951 (95% CI: 1.737, 8.988)], HMGB1 level [O^R = 5.339 (95% CI: 2.347, 12.145)], PCT level [O^R = 4.084 (95% CI: 1.795, 9.290)] and SAA level [O^R = 4.749 (95% CI: 2.088, 10.804)] were risk factors for poor prognosis in MPP patients with pleural effusion (P < 0.05).The model in the training set predicted poor prognosis in children with MPP and pleural effusion with a sensitivity of 89.80% (95% CI: 0.770, 0.962), specificity of 92.05% (95% CI: 0.838, 0.965), and an area under the curve of 0.907 (95% CI: 0.762, 0.969). The model in the validation set predicted poor prognosis with a sensitivity of 83.33% (95% CI: 0.509, 0.971), specificity of 90.91% (95%CI: 0.694, 0.984), and an area under the curve of 0.865 (95%CI: 0.717, 0.935).Conclusion Serum IL-17A, HMGB1, PCT, and SAA are related to poor prognosis in children with MPP and pleural effusion. Constructing an early warning model based on serum IL-17A, HMGB1, PCT, and SAA is helpful for early identification of the risk of poor prognosis in children with MPP and pleural effusion.