Suppr超能文献

支气管肺发育不良:极低出生体重儿的风险预测模型。

Bronchopulmonary dysplasia: risk prediction models for very-low- birth-weight infants.

机构信息

Departamento de Neonatología, Facultad de Medicina, Pontificia Universidad Catolica de Chile, Santiago, Chile.

Unidad de Neonatologia, Instituto Universitario Hospital Italiano Buenos Aires, Potosí, Argentina.

出版信息

J Perinatol. 2019 Sep;39(9):1275-1281. doi: 10.1038/s41372-019-0430-x. Epub 2019 Jul 23.

Abstract

OBJECTIVE

Our objective is to develop risk prediction models for moderate/severe bronchopulmonary dysplasia (BPD) and BPD and/or death in very-low-birth-weight infants (VLBWI) at birth, 3, 7, and 14 postnatal days.

STUDY DESIGN

It is a multicenter study including 16,407 infants weighing 500-1500 g (2001-2015) from the Neocosur Network. BPD was defined as oxygen dependency at 36 weeks. Variables were selected using forward logistic regression models. Predictive values were evaluated using the ROC curve.

RESULTS

In total, 2580 (15.7%) presented BPD and 6121 (37.3%) BPD/death. The AUC values for the BPD models were 0.788, 0.818, 0.827, and 0.894 respectively. For BPD/death, the AUC values were 0.860, 0.869, 0.867, and 0.906. BW and gestational age had higher contribution at birth; at later ages, the length of oxygen therapy and ventilation had the highest contribution. All AUC values were statistically significant when compared with a neutral value of 0.5 (p-value < 0.001).

CONCLUSIONS

We developed high predictive power models for moderate/severe BPD and BPD/death at four postnatal ages.

摘要

目的

本研究旨在为出生时、生后第 3、7、14 天体重 500-1500g 的极低出生体重儿(VLBWI)中中重度支气管肺发育不良(BPD)和 BPD 及死亡建立风险预测模型。

研究设计

这是一项多中心研究,共纳入了来自 Neocosur 网络的 16407 例体重 500-1500g 的婴儿(2001-2015 年)。BPD 定义为生后 36 周氧依赖。采用逐步逻辑回归模型选择变量。通过 ROC 曲线评估预测值。

结果

共有 2580 例(15.7%)发生 BPD,6121 例(37.3%)发生 BPD/死亡。BPD 模型的 AUC 值分别为 0.788、0.818、0.827 和 0.894。对于 BPD/死亡,AUC 值分别为 0.860、0.869、0.867 和 0.906。BW 和胎龄在出生时的贡献较高,而在后期,氧疗和通气时间的长度具有最高的贡献。与中性值 0.5 相比,所有 AUC 值均具有统计学意义(p 值均<0.001)。

结论

我们建立了具有高预测能力的四个生后年龄的中重度 BPD 和 BPD/死亡预测模型。

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验