A maternal serum screen for trisomy 18: an extension of maternal serum screening for Down syndrome

The feasibility of extending second-trimester maternal blood screening for Down syndrome so as to include screening for trisomy 18 was examined using stored maternal serum samples collected for neural tube-defect screening. There were 12 samples from trisomy 18 pregnancies and 390 controls. The median maternal serum concentration of alpha-fetoprotein, free alpha-subunit human chorionic gonadotrophin, free beta-subunit human chorionic gonadotrophin, intact human chorionic gonadotrophin, total estriol, unconjugated estriol, estradiol, human placental lactogen, and progesterone were lowered in those pregnancies affected by trisomy 18 when compared with unaffected pregnancies matched for racial origin, maternal age, gestational age, and sample-storage duration. At an estimated odds risk of 1:400, 83.3% of affected pregnancies were detected using an algorithm which combines the maternal age-related risk with the maternal serum concentrations of unconjugated estriol, free alpha-subunit human chorionic gonadotrophin, free beta-subunit human chorionic gonadotrophin, estradiol, and human placental lactogen. The associated false-positive rate was 2.6%. At high risk odds of 1:10, the detection rate was 58.3%, with an associated false-positive rate of 0.3%. beta-Subunit human chorionic gonadotrophin and unconjugated estriol were the most powerful discriminators. It is possible to incorporate into existing Down syndrome screening programs an algorithm for detecting trisomy 18 with high sensitivity and specificity.