孕中期产前筛查标志物中位数修正的探讨

A study on the correction of prenatal screening markers medians in the second trimester of pregnancy

目的以杭州萧山地区为例,建立产前筛查本地化中位数数据,并探讨应用国外软件内嵌中位数参数的适用性。方法收集了2013年5月-2014年4月共19 061例样本用于构建血清指标中位数浓度与孕龄间的回归关系,选取最优模型得到指标中位数浓度,并进行体质量的修正,完成中位数方程本地化。将其应用于孕二联19 133例筛查样本,并与应用Life Cycle内置公式得到的检出率、假阳性率等指标进行比较。结果在体质量未校正的条件下,萧山地区孕中期孕妇的甲胎蛋白平均水平比Lifecycle 4.0软件内嵌参数源人群(高加索人)平均水平高出6%,而游离β-人绒毛膜促性腺激素高出16%。修正本地化中位数后筛查实验的检出率虽未发生变化,但3种疾病的假阳性率却降低了2.5%,且筛查标记物的中位数倍数(MOM)值中位数更接近1,降低了实验的系统误差。结论通过本地化中位数公式的修正,不仅可以减少实验的系统误差,使结果更加准确可信,而且可以提高产前筛查工作的效率。

Objective With a case study of Xiaoshan in Hangzhou,to establish the median data of prenatal screening,and to explore the applicability of the foreign software with embedded medians. Methods 19 061 samples were collected from May,2013 to Apr,2014 to establish the regression between the median serum concentrations and gestational ages. The optimal model was used to get the median concentrations of the indexes,with which the weights were corrected and the localization of the medians was achieved. Then,the medians were applied to the 19 133 cases of pregnant women,and compared with the detection rate and false positive rate by using the Life Cycle formula. Results With the uncorrected weights,the average AFP level of the second trimester was 6% higher than the Lifecycle 4. 0 software embedded in the source population parameters（ Caucasian）,and the free beta human chorionic gonadotropin was up to 16%. After the median localization corrected,the detection rates of screening test didn＇t change,but the false positive rates of three kinds of diseases was reduced by 2. 5%,and also the median value of median of multiple（ MOM） was much closer to 1,which could help to reduce the system errors. Conclusion By modifying the median localization formula,not only the system errors can be reduced to make the results more accurate and credible,but also the efficiency of the prenatal screening can be improved.