天津医药 ›› 2026, Vol. 54 ›› Issue (9): 982-986.doi: 10.11958/20253697

• 临床研究 • 上一篇    下一篇

极早/超早产儿支气管肺发育不良的危险因素预测模型构建

许海娟(), 张迪, 姚文秀, 邱锐琴, 乔木   

  1. 秦皇岛市第一医院新生儿科(邮编066000)
  • 收稿日期:2025-12-19 修回日期:2026-03-25 出版日期:2026-09-15 发布日期:2026-09-14
  • 作者简介:许海娟(1985),女,主管护师,主要从事新生儿疾病相关研究。E-mail:xuhaijuan0414@163.com
  • 基金资助:
    秦皇岛市科学技术研究与发展计划(202401A052)

Risk factors for bronchopulmonary dysplasia and construction of a prediction model in very/extremely preterm infants

XU Haijuan(), ZHANG Di, YAO Wenxiu, QIU Ruiqin, QIAO Mu   

  1. Department of Neonatology, the First Hospital of Qinhuangdao, Qinhuangdao 066000, China
  • Received:2025-12-19 Revised:2026-03-25 Published:2026-09-15 Online:2026-09-14

摘要:

目的 探讨极早/超早产儿发生支气管肺发育不良(BPD)的危险因素,并建立危险预测模型。方法 选取206例极早/超早产儿,分为BPD组(101例)和非BPD组(105例)。比较2组孕妇围产期和新生儿临床资料的差异。采用Logistic回归分析BPD发生的影响因素并构建预测模型,采用受试者工作特征(ROC)曲线评估模型区分度,采用Bootstrap法(重复抽样1 000次)进行模型内部验证。进一步通过ROC曲线评估模型对BPD严重程度的预测价值。结果 BPD组母亲孕期宫内感染比例高于非BPD组(P<0.05),新生儿相关因素比较显示,BPD组胎龄<28周、新生儿呼吸窘迫综合征、动脉导管未闭、感染性肺炎以及有创机械通气时间>7 d的比例均显著高于非BPD组(P<0.05)。多因素Logistic回归分析显示,孕期宫内感染(OR=6.846,95%CI:1.757~26.667)、胎龄<28周(OR=4.241,95%CI:2.072~8.683)、新生儿呼吸窘迫综合征(OR=6.346,95%CI:3.003~13.411)、动脉导管未闭(OR=4.935,95%CI:1.919~12.695)、感染性肺炎(OR=3.323,95%CI:1.465~7.540)、有创机械通气时间>7 d(OR=4.954,95%CI:2.226~11.025)是极早/超早产儿发生BPD的独立危险因素(P<0.01)。根据上述影响因素构建联合预测模型,Hosmer-Lemeshow检验拟合优度良好(χ²=8.433,P=0.296);该模型预测BPD发生的曲线下面积(AUC)为0.837(95%CI:0.783~0.891),内部验证的AUC为0.814(95%CI:0.754~0.875),预测重度BPD的AUC为0.874(95%CI:0.816~0.933)。结论 基于极早/超早产儿产前及产后资料构建的BPD早期风险预测模型具有良好的区分度和校准度,有助于临床进行早期干预。

关键词: 早产儿, 支气管肺发育不良, 危险因素, 预测

Abstract:

Objective To explore risk factors of bronchopulmonary dysplasia (BPD) in very/extremely preterm infants and construct the prediction model. Methods A total of 206 very/extremely preterm infants were selected and divided into the BPD group (101 cases) and the non-BPD group (105 cases). The differences in perinatal and neonatal clinical data were compared between the two groups. Logistic regression analysis was used to analyze influencing factors of BPD occurrence and construct a prediction model. The receiver operating characteristic (ROC) curve was used to evaluate the discrimination of the model. Internal validation of the model was conducted by Bootstrap method (repeated sampling for 1000 times). The predictive value of the model for the severity of BPD was further evaluated by ROC curve. Results The proportion of intrauterine infection during pregnancy was higher in the BPD group than that in the non-BPD group (P<0.05). The comparison of early neonatal factors showed that the proportion of gestational age < 28 weeks, neonatal respiratory distress syndrome, patent ductus arteriosus, infectious pneumonia, and invasive mechanical ventilation time >7 d were significantly higher in the BPD group than those in the non-BPD group (P<0.05). Multivariate Logistic regression analysis showed that intrauterine infection during pregnancy (OR=6.846, 95%CI: 1.757-26.667), gestational age <28 weeks (OR=4.241, 95%CI: 2.072-8.683), neonatal respiratory distress syndrome (OR=6.346, 95%CI: 3.003-13.411), patent ductus arteriosus (OR=4.935, 95%CI: 1.919-12.695), infectious pneumonia (OR=3.323, 95%CI: 1.465-7.540) and invasive mechanical ventilation time >7 d (OR=4.954, 95%CI: 2.226-11.025) were independent risk factors for BPD in very/extremely premature infants (P<0.01). The combined prediction model was constructed based on the above influencing factors, and its goodness of fit was detected by Hosmer-Lemeshow test (χ2=8.433, P=0.296). The area under the curve (AUC) of the model for predicting BPD was 0.837 (95%CI: 0.783-0.891), AUC of internal validation was 0.814 (95%CI: 0.754-0.875), and AUC of the model for predicting severe BPD was 0.874 (95%CI: 0.816-0.933). Conclusion The risk prediction model of BPD constructed based on prenatal and postnatal data has good discrimination and calibration in very/extremely preterm infants, which is conducive to early clinical intervention.

Key words: premature infant, bronchopulmonary dysplasia, risk factor, prediction

中图分类号: