上海交通大学学报（医学版）

上海交通大学学报(医学版) ›› 2020, Vol. 40 ›› Issue (07): 908-914.doi: 10.3969/j.issn.1674-8115.2020.07.008

• 论著·临床研究 • 上一篇    下一篇

系统性红斑狼疮妊娠丢失预测模型的建立与验证

吴珈悦，蒋 萌，林思涵，狄 文   

  1. 上海交通大学医学院附属仁济医院妇产科，上海市妇科肿瘤重点实验室，上海 200127
  • 出版日期:2020-07-28 发布日期:2020-09-23
  • 通讯作者: 狄 文，电子信箱：diwen163@163.com。
  • 作者简介:吴珈悦（1989—），女，博士；电子信箱：janet_wu_jiayue@163.com。
  • 基金资助:
    国家自然科学基金青年科学基金（81901494）；上海市卫生和计划生育委员会课题（2017ZZ02016）。

Establishment and validation of predictive model of pregnancy loss in patients with systemic lupus erythematosus

WU Jia-yue, JIANG Meng, LIN Si-han, DI Wen   

  1. Shanghai Key Laboratory of Gynecologic Oncology, Department of Obstetrics and Gynecology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai 200127, China
  • Online:2020-07-28 Published:2020-09-23
  • Supported by:
    Youth Program of National Natural Science Foundation of China (81901494); Shanghai Municipal Commission of Health and Family Planning (2017ZZ02016).

PDF (PC)

225

摘要： 目的·建立系统性红斑狼疮（systemic lupus erythematous，SLE）妊娠丢失预测模型并加以验证。方法·选择2011年9月—2017年5月于上海交通大学医学院附属仁济医院收治的妊娠合并SLE患者（模型开发组）338例，以及2017年6月—2018年6月该院收治的妊娠合并SLE患者（模型验证组）131例为研究对象。采用多因素Logistic回归模型分析模型开发组，以确定妊娠丢失的预测变量及其系数。建立预测模型并进行风险评分分类，通过模型验证组对该预测模型进行外部验证。结果·多因素Logistic回归分析显示，非计划妊娠（P=0.032）、低补体C3（P=0.002）和24 h尿蛋白≥1.0 g（P=0.000）均是SLE妊娠丢失预测模型的危险因素。该模型风险评分为0～3分时表示SLE妊娠丢失低风险，>3分时则表示为高风险，其敏感度及特异度分别为60.5%和93.3%。经验证显示，该模型的SLE妊娠丢失预测准确度为90.1%。结论·SLE妊娠丢失预测模型可以帮助临床医师高效筛查SLE妊娠丢失高风险人群，以便尽早采取相关措施，获得较好的SLE妊娠结局。

关键词: 系统性红斑狼疮, 妊娠, 妊娠丢失, 预测模型

Abstract:

Objective · To establish and verity predictive model of pregnancy loss in systemic lupus erythematosus (SLE). Methods · A total of 338 SLE pregnant patients admitted to Renji Hospital, Shanghai Jiao Tong University School of Medicine from Sept. 2011 to May 2017 (model development group) and 131 SLE pregnant patients admitted from Jun. 2017 to Jun. 2018 (model validation group) were selected. Multivariable Logistic regression model was used to determine the predictive variables and their coefficients of pregnancy loss in model development group. The predictive model was established, the risk score classification was performed, and model validation group was used for external validation. Results · Multivariate Logistic regression analysis showed that unplanned pregnancy (P=0.032), low complement C3 (P=0.002) and 24 h urinary protein ≥ 1.0 g (P=0.000) were the risk factors of the predictive model of SLE pregnancy loss. When the risk score of the model was 0-3, the risk of SLE pregnancy loss was low, and when the risk score is more than 3, it is high risk, with a sensitivity and specificity of 60.5% and 93.3%, respectively. The model was used in the model validation group for external validation, and the prediction accuracy of SLE pregnancy loss was 90.1%. Conclusion · The predictive model of SLE pregnancy loss can help clinicians efficiently screen the high-risk population of SLE pregnancy loss in order to take relevant measures as soon as possible to obtain better pregnancy outcomes.

Key words: systemic lupus erythematosus (SLE), pregnancy, pregnancy loss, predictive model

中图分类号: