神经管缺陷产前筛查的AFP中位数倍数校正值的建立

Establishment of MOM Correction Models in NTD Prenatal Screening on IMMULITE 2000

目的探讨和构建北京人群神经管缺陷(neural tube defect,NTD)产前筛查的血清标志物甲胎蛋白(α-fetoprotein,AFP)浓度中位数倍数(multiple of the median,MOM)的校正模型。方法选取5668例孕中期孕妇,通过分析其AFP浓度与孕周和体重之间的相关性,分别进行回归分析,根据统计量和校正决定系数选取最优模型,校正AFP。结果 35岁以下(<35y)组,分别用指数和幂回归模型进行孕周和孕妇体重的校正;35岁及以上(≥35y)组,分别用S和二次回归模型进行孕周和孕妇体重的校正。校正前,孕周与AFP MOM值呈显著正相关(<35y:r=0.412,P=0.000;≥35y:r=0.435,P=0.000),体重与AFP MOM值呈显著负相关(<35y:r=-0.232,P=0.000;≥35y:r=-0.228,P=0.000);校正后,AFPMOM值与孕周(<35y:r=0.446,P=0.911;≥35y;r=0.001,P=0.985)和体重(<35y:r=0.010,P=0.483;≥35y:r=0.008,P=0.887)均无相关性。结论基于北京人群孕中期妇女血清标志物AFP水平建立的MOM校正模型可以有效地消除孕周和体重的影响。

Objective To establish appropriate correction models of multiple of the median of maternal serum marker ot-fetoprotein for prenatal screening on patients with neural tube defect at the second trimester in Beijing China. Methods Data were collected from 5668 second-trimester pregnancies in Beijing of China, the correla- tion between AFP concentration and gestational age, maternal weight were analyzed by means of regression method respectively, regression functions for adjustment of MOM by gestational age and maternal weight were established for our own population. Results The regression relationships of AFP concentration and gestational age, maternal weight were established by exponential and power model for 〈 35y group, S and quadratic model for ≥35y respectively. Before correction, a remarkable positive correlation （ 〈 35y ： r = 0. 412, P = 0. 000 ; ≥ 35y：r = O. 435, P = 0. 000） existed between AFP concentration and gestational age, a remarkable negative cor- relation （ 〈35y： r = -0. 232, P =0.000; ≥35y： r = -0. 228,P =0.000） existed between MOM and mater- nal weight. After correction, there were no correlations between MOM and gestational age （ 〈 35y ： r = 0. 446, P=0.911; ≥35y： r=0. 001, P=0.985）, maternal weight （ 〈35y： r=0.010, P=0.483; ≥35y： -r= 0. 008, P = 0. 887 ）. Conclusion The influence of gestational age and maternal weight on AFP concentration could be diminished by our correction models based on our own population.