文献速递产科母胎医学

竞争风险模型预测妊娠36周小于胎龄儿

Competing-risks model for prediction of small-for-gestational-age neonate at 36 weeks' gestation.
PubMed 文献 · Ultrasound Obstet Gynecol · 2022-10-12

这项前瞻性观察研究纳入29 035例单胎妊娠妇女,在35+0至36+6周行常规超声检查,建立并验证纳入估计胎儿体重(EFW)及胎盘功能异常生物标志物的竞争风险模型预测小于胎龄儿(SGA),并与传统EFW低于第10百分位截断值比较。模型对接近评估时点分娩的SGA预测更好,加入EFW、子宫动脉搏动指数(UtA-PI)和胎盘生长因子(PlGF)可显著改善预测,且新模型优于固定EFW百分位截断值。

要点

  • 在10%假阳性率下,孕妇因素联合EFW对第10百分位以下SGA的预测率:38周前分娩77.6%,42周前分娩65.8%。
  • 对第3百分位以下SGA,相应预测率分别为85.5%和74.2%。
  • 加入UtA-PI和PlGF对需近期分娩的重度SGA预测仅有边际改善。

核心结论

  • 竞争风险模型可在35+0至36+6周对SGA新生儿提供有效风险分层,且优于EFW百分位截断值。
  • 在孕妇因素和胎儿生物测量基础上加入胎盘功能异常生物标志物,对重度SGA新生儿的预测性能仅有小幅改善。

研究设计与样本

  • 前瞻性观察研究,纳入29 035例单胎妊娠妇女,在35+0至36+6周妊娠接受常规超声检查。
  • 采用竞争风险模型预测SGA新生儿,先验历史模型参数取自既往研究,EFW采用交互连续模型,胎盘功能异常生物标志物采用折叠平面回归模型拟合,并制定分层方案。
  • 将新模型与EFW百分位截断值进行比较评估。

主要发现

  • 模型对接近评估时点分娩的SGA新生儿预测性能更好。
  • 仅孕妇因素的预测在加入EFW、UtA-PI和PlGF后显著改善,但加入平均动脉压或可溶性fms样酪氨酸激酶-1未改善预测。
  • 在10%假阳性率下,孕妇因素联合EFW对第10百分位以下SGA的预测率:38周前分娩为77.6%,42周前分娩为65.8%。
  • 对第3百分位以下SGA,孕妇因素联合EFW的预测率:38周前分娩为85.5%,42周前分娩为74.2%。
  • 加入UtA-PI和PlGF对需近期分娩的第3百分位以下SGA预测仅有边际改善;新模型校准良好。

关键数值与推荐

  • 在10%假阳性率下,孕妇因素联合EFW预测第10百分位以下SGA:38周前分娩77.6%,42周前分娩65.8%。
  • 在10%假阳性率下,孕妇因素联合EFW预测第3百分位以下SGA:38周前分娩85.5%,42周前分娩74.2%。
  • 加入EFW、UtA-PI和PlGF可显著改善仅用孕妇因素的预测,但平均动脉压或可溶性fms样酪氨酸激酶-1未改善预测。
  • 加入UtA-PI和PlGF对需近期分娩的第3百分位以下SGA预测仅有边际改善。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVES: To develop further a competing-risks model for the prediction of a small-for-gestational-age (SGA) neonate by including sonographically estimated fetal weight (EFW) and biomarkers of impaired placentation at 36 weeks' gestation, and to compare the performance of the new model with that of the traditional EFW < 10th percentile cut-off. METHODS: This was a prospective observational study in 29 035 women with a singleton pregnancy undergoing routine ultrasound examination at 35 + 0 to 36 + 6 weeks' gestation. A competing-risks model for the prediction of a SGA neonate was used. The parameters included in the prior-history model were provided in previous studies. An interaction continuous model was used for the EFW likelihood. A folded plane regression model was fitted to describe likelihoods of biomarkers of impaired placentation. Stratification plans were also developed. The new model was evaluated and compared with EFW percentile cut-offs. RESULTS: The performance of the model was better for predicting SGA neonates delivered closer to the point of assessment. The prediction provided by maternal factors alone was improved significantly by the addition of EFW, uterine artery pulsatility index (UtA-PI) and placental growth factor (PlGF) but not by mean arterial pressure or soluble fms-like tyrosine kinase-1. At a 10% false-positive rate, maternal factors and EFW predicted 77.6% and 65.8% of SGA neonates < 10th percentile delivered before 38 and 42 weeks, respectively. The respective figures for SGA < 3rd percentile were 85.5% and 74.2%. Addition of UtA-PI and PlGF resulted in marginal improvement in prediction of SGA < 3rd percentile requiring imminent delivery. A competing-risks approach that combines maternal factors and EFW performed better when compared with fixed EFW percentile cut-offs at predicting a SGA neonate, especially with increasing time interval between assessment and delivery. The new model was well-calibrated. CONCLUSIONS: A competing-risks model provides effective risk stratification for a SGA neonate at 35 + 0 to 36 + 6 weeks' gestation and is superior to EFW percentile cut-offs. The use of biomarkers of impaired placentation in addition to maternal factors and fetal biometry results in small improvement of the predictive performance for a neonate with severe SGA. © 2022 International Society of Ultrasound in Obstetrics and Gynecology.

来源:Ultrasound Obstet Gynecol,查看原文。版权归原作者所有。

引用

  1. Papastefanou I, Thanopoulou V, Dimopoulou S, et al. Competing-risks model for prediction of small-for-gestational-age neonate at 36 weeks' gestation. Ultrasound Obstet Gynecol. 2022. PMID: 36056735. doi:10.1002/uog.26057

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。