个体化预测学龄前儿童哮喘急性发作风险的列线图模型的建立

Establishment of a nomogram prediction of acute asthma attacks in preschool children

目的通过对学龄前儿童哮喘急性发作的相关危险因素进行分析,建立个体化预测学龄前儿童哮喘急性发作风险的列线图模型。方法纳入2014年9月至2017年9月学龄前哮喘患儿226例,分为哮喘急性发作组98例(观察组)和非急性发作组128例(对照组),收集临床资料,采用多因素Logistic回归模型分析哮喘急性发作危险因素;应用R软件建立预测学龄前儿童哮喘急性发作风险的列线图模型,采用Bootstrap法进行模型验证,并作ROC曲线探索列线图模型对哮喘急性发作风险的预测效率。结果特应性体质(OR=3.079,95%CI:1.25-7.58)、近1年呼吸道感染次数(OR=2.343,95%CI:1.84-2.99)、哮喘家族史(OR=6.099,95%CI:2.01-18.52)、被动吸烟(OR=2.722,95%CI:1.01-7.34)及未规范用药(OR=4.135,95%CI:1.58-10.85)是学龄前儿童哮喘急性发作的独立危险因素(P<0.05);列线图模型预测哮喘急性发作风险的一致性指数为0.856;ROC曲线示列线图模型预测哮喘急性发作风险的曲线下面积为0.889(95%CI:0.821-0.955)。结论本研究基于特应性体质、近1年呼吸道感染次数、哮喘家族史、被动吸烟及未规范用药这5项独立危险因素,建立的列线图模型对学龄前儿童哮喘急性发作风险的预测作用。

Objective To analyze the risk factors for acute asthma attacks in preschool children by establishing a nomogram model to predict the risk of acute asthma attacks.Methods From September 2014 to September2017,226 preschool pediatric asthma patients were enrolled and divided into the acute asthma attacks group（ the observational group,n = 98） and the non-acute asthma attacks group（ the control group,n = 128）.The clinical characteristics were collected and multivariate logistic regression analysis model was used to analyze the risk factors for acute asthma attacks.R software was adopted to establish a predictive nomogram model of acute asthma attacks.Bootstrap method was used to validate the nomogram model and ROC curve was used to determine the predictive efficacy of the model in predicting acute asthma attacks.Results Atopic body（ OR = 3.079,95% CI：1.25-7.58）,recurrent respiratory infections in one year（ OR = 2.343,95% CI：1.84-2.99）,family history of asthma（ OR =6.099,95% CI： 2.01-18.52）,passive smoking（ OR = 2.722,95% CI：1.01-7.34） and non-standard medications（ OR = 4.135,95% CI： 1.58-10.85） were all independent risk factors for acute asthma attacks in preschool children（ P〈0.05）.The nomogram model showed a C-index of 0.856 with good discrimination and accuracy,while ROC curve demonstrated that the area under the curve for the predictive model was 0.889（ 95% CI： 0.821-0.955）.Conclusion The study establishes the nomogram model based on atopic body,recurrent respiratory infections in one year,family history of asthma,passive smoking and non-standard medications,which could be helpful for identifying the pediatric patients with high risk and has potentially high clinical application value.