Moon Myoungjin, Odibo Anthony
a Department of Obstetrics and Gynecology , CHA Bundang Medical Center, CHA University , Seongnam-si , Korea and.
b Department of Obstetrics and Gynecology , Morsani College of Medicine, University of South Florida , Tampa , FL , USA.
J Matern Fetal Neonatal Med. 2015;28(17):2028-33. doi: 10.3109/14767058.2014.978758. Epub 2014 Nov 11.
The impact of maternal parity on screening efficiency for preeclampsia (PE) has been poorly studied. Our objective was to investigate the effect of maternal parity on models for screening for PE in the first-trimester and their effectiveness.
A secondary analysis of a prospective cohort study of women present between 11 and 14 weeks gestation. Maternal risk factors, uterine artery Doppler, mean arterial pressure (MAP) and serum markers including PAPP-A, ADAM12, PP13 in the first-trimester were used to create multi-parameter screening models for PE. The best models for screening in nulliparous versus parous women were developed using backward stepwise logistic regression approach. The area under the receiver operating characteristic curves (AUC) and the sensitivity for fixed false positive rates of 10% and 20%, respectively, were compared using non-parametric statistics.
Among 1177 women with complete outcome data (503 (42.7%) nulliparous and 674 (57.3%) multiparous), PE occurred in 102 (8.7%). There were significant differences in predicting variables in the final optimal models when stratified by parity; and screening performance also varied by parity. The AUC for the model for nulliparous women was 0.88(95% CI 0.80-0.95); and for multiparous was 0.84 (95% CI, 0.75-0.91). For fixed false positive rate of 10%, the sensitivity for predicting PE was 70% and 68% for nulliparous and multiparous, respectively. The screening performance for the models were however not statistically or clinically significantly different.
We found significant differences in prediction model parameters between nulliparous and multiparous women, but these did not significantly impact screening performance for PE in the first-trimester.
关于孕产妇胎次对先兆子痫（PE）筛查效率的影响，此前研究较少。我们的目的是研究孕产妇胎次对孕早期PE筛查模型及其有效性的影响。
对一项针对妊娠11至14周孕妇的前瞻性队列研究进行二次分析。使用孕产妇危险因素、子宫动脉多普勒、平均动脉压（MAP）以及孕早期血清标志物包括妊娠相关血浆蛋白A（PAPP - A）、解整合素金属蛋白酶12（ADAM12）、妊娠相关血浆蛋白13（PP13）来创建PE的多参数筛查模型。采用向后逐步逻辑回归方法分别建立初产妇和经产妇的最佳筛查模型。使用非参数统计方法比较接受者操作特征曲线（AUC）下的面积以及固定假阳性率分别为10%和20%时的敏感度。
在1177例有完整结局数据的女性中（503例（42.7%）为初产妇，674例（57.3%）为经产妇），102例（8.7%）发生了PE。按胎次分层时，最终最优模型中的预测变量存在显著差异；筛查性能也因胎次而异。初产妇模型的AUC为0.88（95%可信区间0.80 - 0.95）；经产妇模型的AUC为0.84（95%可信区间0.75 - 0.91）。对于固定假阳性率为10%的情况，初产妇和经产妇预测PE的敏感度分别为70%和68%。然而，这些模型的筛查性能在统计学或临床上并无显著差异。
我们发现初产妇和经产妇在预测模型参数上存在显著差异，但这些差异并未对孕早期PE的筛查性能产生显著影响。
