早产儿支气管肺发育不良早期风险预测模型的构建

Construction of early risk prediction models for bronchopulmonary dysplasia in preterm infants

摘要:

目的 构建早产儿生后3、7、14 d的支气管肺发育不良（bronchopulmonary dysplasia，BPD）风险预测模型。 方法 回顾性分析2019年7月至2021年4月入住新生儿重症监护室，胎龄<32周且出生体重（birth weight，BW）<1 500 g的414例早产儿的临床资料，根据2018年BPD诊断标准分为BPD组（n=98）和非BPD组（n=316），比较两组早产儿的一般情况、实验室检查、治疗及并发症情况，应用logistic回归模型选择与BPD相关的变量构建风险预测模型，应用受试者工作特征曲线评价该模型的预测价值。 结果 经logistic回归分析发现，BW、窒息、Ⅲ~Ⅳ级呼吸窘迫综合征（respiratory distress syndrome，RDS）、急性绒毛膜羊膜炎、间质性肺炎、吸入氧浓度（fraction of inspired oxygen，FiO₂）和呼吸支持模式是BPD发生的主要风险因素（P<0.05）。生后7 d（P₇）、14 d（P₁₄）预测模型表达式分别为logit（P₇）=-2.049-0.004×BW（g）+0.686×窒息（无=0，有=1）+1.842×Ⅲ~Ⅳ级RDS（无=0，有=1）+0.906×急性绒毛膜羊膜炎（无=0，有=1）+0.506×间质性肺炎（无=0，有=1）+0.116×FiO₂（%）+0.816×呼吸支持模式（无=0，鼻导管=1，经鼻持续气道正压通气=2，常频机械通气=3，高频机械通气=4），logit（P₁₄）=-1.200-0.004×BW+0.723×窒息+2.081×Ⅲ~Ⅳ级RDS+0.799×急性绒毛膜羊膜炎+0.601×间质性肺炎+0.074×FiO₂+0.800×呼吸支持模式，两个模型的曲线下面积分别为0.876，0.880，高于生后3 d预测模型（P<0.05）。 结论 BW、窒息、Ⅲ~Ⅳ级RDS、急性绒毛膜羊膜炎、间质性肺炎、FiO₂和呼吸支持模式是BPD发生的主要风险因素，可用于风险预测模型的构建。生后7 d、14 d预测模型可有效预测BPD的发生。

Abstract:

Objective To construct risk prediction models for bronchopulmonary dysplasia (BPD) in preterm infants on postnatal days 3, 7, and 14. Methods A retrospective analysis was performed on the medical data of 414 preterm infants, with a gestational age of <32 weeks and a birth weight (BW) of <1 500 g, who were admitted to the neonatal intensive care unit from July 2019 to April 2021. According to the diagnostic criteria for BPD revised in 2018, they were divided into a BPD group with 98 infants and a non-BPD group with 316 infants. The two groups were compared in terms of general status, laboratory examination results, treatment, and complications. The logistic regression model was used to identify the variables associated with BPD. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of models. Results The logistic regression analysis showed that BW, asphyxia, grade III-IV respiratory distress syndrome (RDS), acute chorioamnionitis, interstitial pneumonia, fraction of inspired oxygen (FiO2), and respiratory support mode were the main risk factors for BPD (P<0.05). The prediction models on postnatal days 7 and 14 were established as logit (P7) =-2.049-0.004×BW (g) +0.686×asphyxia (no=0, yes=1) +1.842×grade III-IV RDS (no=0, yes=1) +0.906×acute chorioamnionitis (no=0, yes=1) +0.506×interstitial pneumonia (no=0, yes=1) +0.116×FiO2 (%) +0.816×respiratory support mode (no=0, nasal tube=1, nasal continuous positive airway pressure=2, conventional mechanical ventilation=3, high-frequency mechanical ventilation=4) and logit (P14) =-1.200-0.004×BW (g) +0.723×asphyxia+2.081×grade III-IV RDS+0.799×acute chorioamnionitis+0.601×interstitial pneumonia+0.074×FiO2 (%) +0.800×respiratory support mode, with an area under the ROC curve (AUC) of 0.876 and 0.880, respectively, which was significantly larger than the AUC of the prediction model on postnatal day 3 (P<0.05). Conclusions BW, asphyxia, grade III-IV RDS, acute chorioamnionitis, interstitial pneumonia, FiO2, and respiratory support mode are the main risk factors for BPD and can be used to construct risk prediction models. The prediction models on postnatal days 7 and 14 can effectively predict BPD.