2022 年12 期 第30 卷
论著![查看全文 查看全文](/Areas/201911120101/Content/img/icon_html.png)
![PDF下载 PDF下载](/Areas/201911120101/Content/img/icon_pdf.png)
肺炎支原体肺炎患儿发生慢性咳嗽的影响因素及其列线图模型构建
Influencing Factors and Construction of Nomogram Model of Chronic Cough in Children with Mycoplasma Pneumoniae Pneumonia
作者:申永旺,张钰,张丽珍,田君平,梁海联,陈欣
- 单位:
- 河北省定州市人民医院普儿一科
- 单位(英文):
- Department of General Paediatrics, Dingzhou People’s Hospital;
- 关键词:
- 肺炎,支原体; 慢性咳嗽; Lasso回归; 风险预测模型; 列线图;
- 关键词(英文):
- Pneumonia,mycoplasma;Chronic cough;Lasso regression;Risk prediction model;Nomogram;
- 中图分类号:
- DOI:
- 10.12114/j.issn.1008-5971.2022.00.306
- 基金项目:
- 河北省医学科学研究课题(20191752);
摘要:
目的 探讨肺炎支原体肺炎(MPP)患儿发生慢性咳嗽的影响因素,并构建其列线图模型。方法 选取2018年1月至2022年1月定州市人民医院收治的MPP患儿245例。根据MPP患儿慢性咳嗽发生情况将其分为慢性咳嗽组和对照组。收集患儿的一般资料(性别、年龄、慢性咳嗽病史、过敏性疾病史、父母吸烟史)、肺炎患病情况(患儿发热时间、体温、肺内干/湿啰音情况、开始治疗时间)、胸部影像学表现(胸腔积液、肺部大片实变影)、血清学指标[CRP、IgE、降钙素原(PCT)、中性粒细胞与淋巴细胞比值(NLR)、血小板与淋巴细胞比值(PLR)]、呼出气一氧化氮(FeNO)。采用LASSO回归和多因素Logistic回归模型分析MPP患儿发生慢性咳嗽的影响因素,构建MPP患儿发生慢性咳嗽的列线图模型,并评估其预测效能。结果 245例MPP患儿中有61例发生慢性咳嗽,发生率为24.9%。慢性咳嗽组有慢性咳嗽病史、有父母吸烟史、发热时间>7 d、体温≥38.5℃、开始治疗时间>3 d者占比及CRP、IgE、PCT、NLR、FeNO高于对照组(P<0.05)。多因素Logistic回归分析结果显示,发热时间>7 d、体温≥38.5℃、开始治疗时间>3 d、IgE、NLR、FeNO是MPP患儿发生慢性咳嗽的独立影响因素(P<0.05)。基于上述独立影响因素构建MPP患儿发生慢性咳嗽的列线图模型。采用Bootstrap法重复抽样1 000次,结果显示,列线图模型预测MPP患儿发生慢性咳嗽的一致性指数为0.847。ROC曲线分析结果显示,列线图模型预测MPP患儿发生慢性咳嗽的AUC为0.817[95%CI(0.743,0.985)]。列线图模型预测MPP患儿发生慢性咳嗽的校准曲线与实际曲线基本接近。决策曲线分析结果显示,当阈值概率为0.10~0.65时,列线图模型预测MPP患儿发生慢性咳嗽的净获益率>0。结论 发热时间>7 d、体温≥38.5℃、开始治疗时间>3 d、IgE、NLR、FeNO是MPP患儿发生慢性咳嗽的独立影响因素,本研究基于上述影响因素构建的列线图模型可以在一定程度上预测MPP患儿慢性咳嗽的发生风险,可用于筛选慢性咳嗽高风险的MPP患儿,这对针对性的预防和治疗具有重要指导价值。
英文摘要:
Objective To investigate the risk factors of chronic cough in children with mycoplasma pneumoniae pneumonia(MPP), and construct its nomogram model. Methods A total of 245 children with MPP admitted to Dingzhou People’s Hospital from January 2018 to January 2022 were selected. Children with MPP were divided into chronic cough group and control group according to the occurrence of chronic cough. The general information(gender, age, history of chronic cough, history of allergic diseases, parental smoking history), prevalence of pneumonia(duration of fever, body temperature, dry and wet rales in the lungs, time of starting treatment), chest imaging conditions(hydrothorax, massive consolidation of chest), serological indicators [CRP, IgE, procalcitonin(PCT), neutrophil-to-lymphocyte ratio(NLR), platelet-to-lymphocyte ratio(PLR) ], fractional exhaled nitric oxide(FeNO) were collected. LASSO regression and multivariate Logistic regression model were used to analysis the influencing factors of chronic cough in children with MPP, and the nomogram model of chronic cough in children with MPP was constructed and tested. Results Chronic cough occurred in 61 of 245 children with MPP, with an incidence rate of 24.9%. In the chronic cough group, the proportion of patients with chronic cough history, the proportion of patients with parental smoking history, the proportion of patients with fever time > 7 d, the proportion of patients with body temperature ≥ 38.5 ℃, the proportion of patients with treatment start time > 3 d, CRP, IgE, PCT, NLR, and FeNO were higher than those in the control group(P < 0.05). The results of multivariate Logistic regression analysis showed that fever time > 7 d, body temperature ≥ 38.5 ℃, treatment start time > 3 d, IgE, NLR and FeNO were the indepent influencing factors of chronic cough in children with MPP(P < 0.05). The nomogram model of chronic cough in children with MPP was constructed based on the above indepent influencing factors. The Bootstrap method was used to repeat the sample for 1 000 times. The results showed that the consistency index of the nomogram model in predicting the occurrence of chronic cough in MPP children was 0.847. The results of ROC analysis showed that the AUC of the nomogram model for predicting chronic cough in children with MPP was 0.817 [95%CI(0.743, 0.985) ]. The calibration curve of the nomogram model for predicting chronic cough in children with MPP was close to the actual curve. Decision curve results show that when the threshold probability was 0.10-0.65, the nomogram model predicted that the net benefit rate of chronic cough in MPP children was > 0. Conclusion Fever time > 7 d, body temperature ≥ 38.5 ℃, time to start treatment > 3 d, IgE, NLR, FeNO are independent factors affecting the occurrence of chronic cough in MPP children. The nomograph model constructed in this study based on the above influencing factors can predict the risk of chronic cough in children with MPP to a certain extent, can be used to screen MPP children with high risk of chronic cough, and provide important guidance value for targeted prevention and treatment of children with MPP.
参考文献: