ABSTRACT
<p><b>OBJECTIVE</b>To establish the median of serum markers for second trimester screening in Qingdao region and to assess the influence of median correction on the performance of screening.</p><p><b>METHODS</b>Maternal serum alpha-fetoproteins (AFP), human chorionic gonadotrophin, free beta subunit (β -HCG) and unconjugated oestriol (uE3) were assayed for prenatal screening of 18 188 singleton pregnancies at 15-20(+ 6) weeks gestation from January 2009 to July 2010. The median of serum markers was calculated based on above results and applied for risk estimation in screening for fetal aneuploidy from August 2010 to March 2011. The screening performance, specified in terms of detection rates (DRs), false positive rates (FPRs) and odds of being affected given a positive result (OAPR) were compared between the two groups. The risks of 45 affected pregnancies detected during the study were estimated with both Caucasian and corrected medians.</p><p><b>RESULTS</b>The average level of AFP in local pregnancies was similar to that of the Caucasian population, whilst β -HCG and uE3 were respectively 11% and 33% higher than those of Caucasians. The multiple of median (MoM) value was between 0.94 and 1.02 for the dataset based on the corrected median. At a cut-off of l in 270, FPR has decreased from 5.2% to 4.9%, and DR of Down syndrome has increased from 60% to 69.2%, and OAPR has increased from 1:79 to 1:59 when evaluating risk based on the corrected median. For the 45 affected pregnancies, three Down syndrome pregnancies could be missed because their risk estimates were lower than the cut-off level based on Caucasian median.</p><p><b>CONCLUSION</b>It is useful to establish and apply population and laboratory-specific medians in order to improve the performance of prenatal screening and diagnosis.</p>