新生儿呼吸窘迫综合征患儿发生支气管肺发育不良的Nomogram预测模型的建立与评估

Establishment and evaluation of a nomogram prediction model for bronchopulmonary dysplasia in infants with neonatal respiratory distress syndrome

摘要

目的 构建新生儿呼吸窘迫综合征（neonatal respiratory distress syndrome，NRDS）患儿发生支气管肺发育不良（bronchopulmonary dysplasia，BPD）的Nomogram 预测模型，并评估其应用价值。 方法 收集2019年9月至2023年6月赣南医科大学第一附属医院收治的NRDS患儿378例，依据是否发生BPD分成单纯NRDS组（n=271）与NRDS合并BPD组（n=107），选取与NRDS患儿BPD发病风险有关的高危因素，通过套索回归分析（least absolute shrinkage and selection operator，LASSO）的方法优化NRDS患儿发生BPD 风险预测模型的指标，最终缩减为4个与NRDS患儿BPD发病高度相关的指标，分别是：胎龄、出生体质量、持续气道正压通气（continuous positive airway pressure，CPAP）时间、有创机械通气（invasive mechanical ventilation，IMV）时间；对这4个指标列入多因素logistic回归分析，采用R软件构建NRDS患儿发生BPD的Nomogram预测模型。对该列线图模型的内部验证采用Bootstrap方法，使用受试者工作特征曲线（receiver operating characteristic curve，ROC）评估该列线图的区分度、校准曲线评估其预测的准确度。 结果 logistic回归分析结果显示，出生体质量［比值比（odds ratio，OR）=0.998，95%置信区间（confidence interval，CI）=0.997~0.999］、CPAP时间（OR=1.128，95%CI=1.093~1.164）、IMV时间（OR=1.121，95%CI=1.056~1.090）是NRDS患儿发生BPD的独立危险因素（P<0.05）。绘制ROC曲线与校准曲线：ROC曲线下面积（area under curve，AUC）=0.906（95%CI=0.877~0.938）；通过Bootstrap方法对样本进行1 000次重抽样进行内部验证AUC=0.904（95%CI=0.807~1.000），本研究在校准曲线中显示具有良好的一致性。 结论 通过多因素logistic回归分析筛选出出生体质量、CPAP时间、IMV时间是NRDS患儿发生BPD的独立危险因素，并成功构建NRDS患儿发生BPD的Nomogram预测模型。

Abstract

Objective To establish a nomogram prediction model for bronchopulmonary dysplasia（BPD） in infants with neonatal respiratory distress syndrome（NRDS），and to investigate its application value. Methods A total of 378 infants with NRDS who were admitted to The First Affiliated Hospital of Gannan Medical University from September 2019 to June 2023 were enrolled，and according to the presence or absence of BPD，they were divided into NRDS group with 271 infants and NRDS+BPD group with 107 infants. The high-risk factors associated with BPD in infants with NRDS were selected，and a least absolute shrinkage and selection operator regression analysis was used to optimize the indicators in the risk prediction model for BPD. Finally four indicators highly associated with BPD in infants with NRDS were obtained，i.e.，gestational age，birth weight，duration of continuous positive airway pressure（CPAP），and duration of invasive mechanical ventilation（IMV），which were included in the multivariate logistic regression analysis to establish a nomogram model for predicting the risk of BPD in infants with NRDS using R software. The Bootstrap method was used for internal validation of the nomogram model； the receiver operating characteristic curve（ROC） curve was used to evaluate the discriminatory ability of the model，and the calibration curve was used to assess its accuracy in prediction. Results The logistic regression analysis showed that birth weight（odds ratio［OR］=0.998，95%CI=0.997-0.999，P<0.05），duration of CPAP（OR=1.128，95%CI=1.093-1.164，P<0.05），and duration of IMV（OR=1.121，95%CI=1.056-1.090，P<0.05） were independent risk factors for BPD in infants with NRDS. The ROC curve and the calibration curve were plotted，and the results showed that the model had an area under the ROC curve（AUC） of 0.906（95%CI=0.877-0.938）. The Bootstrap method was used to perform 1000 times of resampling for internal validation，and the results showed an AUC of 0.904（95%CI=0.807-1.000）. The calibration curve showed good consistency in this study. Conclusion Birth weight，duration of CPAP，and duration of IMV are independent risk factors for BPD in infants with NRDS based on the multivariate logistic regression analysis，and a nomogram model is successfully established for predicting the risk of BPD in infants with NRDS.

关键词

新生儿呼吸窘迫综合征 / 支气管肺发育不良 / 危险因素 / Nomogram预测模型

Key words

neonatal respiratory distress syndrome / bronchopulmonary dysplasia / risk factor / Nomogram prediction model

基金资助

江西省自然科学基金资助项目(20224BAB206009)

江西省教育厅科学技术研究资助项目(GJJ211528)