超早产儿支气管肺发育不良发生风险的列线图预测模型构建及应用  

Construction and application of nomogram model for predicting the risk of bronchopulmonary dysplasia in extremely preterm infants

作　　者：夏磊[1] 吕荣吉 赵家文 乔青 武慧杰 武田博 XIA Lei;LYU Rongji;ZHAO Jiawen;QIAO Qing;WU Huijie;WU Tianbo(Department of Neonatology,the Third Af filiated Hospital of Zhengzhou University,Zhengzhou,Henan 450052,China)

机构地区：[1]郑州大学第三附属医院新生儿科,河南郑州450052

出　　处：《中华实用诊断与治疗杂志》2025年第3期249-256,共8页Journal of Chinese Practical Diagnosis and Therapy

基　　金：河南省医学科技攻关计划项目(LHGJ20220540)。

摘　　要：目的 分析超早产儿出生28 d内发生支气管肺发育不良(BPD)的影响因素,构建列线图预测模型,探讨其预测效能。方法 回顾性分析2016年1月—2023年12月郑州大学第三附属医院诊治的344例超早产儿的临床资料,统计BPD发生率。以8∶2的比例将344例超早产儿分为训练队列275例和验证队列69例,比较训练队列、验证队列分娩方式、妊娠期高血压比率等母亲资料,胎龄、出生体质量等患儿资料,有创通气时间、应用肺表面活性物质等住院期间治疗相关资料。将训练队列患儿根据BPD发生情况分为BPD组211例和非BPD组64例,比较2组胎龄、身长、出生体质量、早产儿视网膜病发生率、输血次数、有创通气时间、总氧疗时间、住院时间等临床资料,选择P<0.20的指标纳入多因素logistic回归分析模型,分析超早产儿出生28 d内发生BPD的影响因素,根据回归系数,应用R软件rms包构建列线图预测模型。绘制ROC曲线,评估列线图模型预测训练队列、验证队列超早产儿发生BPD的效能;绘制校准曲线,采用Hosmer-Lemeshow拟合优度检验评价列线图模型的校准度;采用临床决策曲线评估列线图模型的净收益。结果 344例超早产儿中265例发生BPD,BPD发生率为77.0%,其中训练队列发生211例,验证队列发生54例。训练队列母亲妊娠期高血压比率(12.4%)低于验证队列(23.2%)(χ^(2)=4.368,P=0.037),分娩方式、胎龄、出生体质量、有创通气时间、肺表面活性物质应用等其余母亲、患儿及治疗相关资料与验证队列比较差异均无统计学意义(P>0.05)。BPD组胎龄[27.0(26.3,27.4)周]小于非BPD组[27.4(27.0,27.7)周](U=-0.519,P=0.001)、身长[35(34,37)cm]短于非BPD组[37(35,38)cm](U=-1.510,P=0.025),出生体质量[950(840,1 040)g]低于非BPD组[1 010(940,1 100)g](U=-0.278,P<0.001),输血次数[5(3,7)次]多于非BPD组[4(2,5)次](U=-6.724,P=0.003),有创通气时间[7(3,12)d]、总氧疗时间[61(49,75)d]、住院时�Objective To analyze the influencing factors of bronchopulmonary dysplasia(BPD)in extremely preterm infants within 28 d after birth,to construct a nomogram model,and to explore its predictive efficiency.Methods Totally 344 extremely preterm infants were diagnosed and treated in the Third Affiliated Hospital of Zhengzhou University from January 2016 to December 2023,and their clinical data were retrospectively analyzed.The incidence of BPD was calculated.The 344 extremely preterm infants were divided into the training cohort(n=275)and the validation cohort(n=69)with a ratio of 8∶2.The maternal data such as delivery mode and gestational hypertension rate,the infant data such as gestational age and birth body mass,and the therapeutic data during hospitalization such as invasive ventilation time and pulmonary surfactant application were compared between the training cohort and the validation cohort.The infants in the training cohort were divided into the BPD group(n=211)and non-BPD group(n=64).The gestational age,body length,birth body mass,incidence of retinopathy of prematurity,number of blood transfusions,invasive ventilation time,total oxygen therapy time,length of hospital stay,and other clinical data were compared between two groups.Variables with P<0.20 were selected and incorporated into a multivariate logistic regression analysis model to identify influencing factors for BPD occurring within 28 d of birth in extremely preterm infants.Based on the regression coefficients,a nomogram prediction model was constructed using the rms package in R software.ROC curves were plotted to evaluate the efficiency of the nomogram model on predicting BPD in extremely preterm infants in both training and validation cohorts.The calibration curves were drawn to evaluate the nomogram model calibration using Hosmer-Lemeshow goodness-of-fit test.The clinical decision curve was drawn to evaluate the net benefit of the nomogram model.Results BPD occurred in 265 of 344 extremely preterm infants,with an incidence of 77.0%,including

关 键 词：支气管肺发育不良 超早产儿 影响因素 列线图模型 

分 类 号：R722.6[医药卫生—儿科]