Atypicality or specific screen: which is better at detecting non-Down's chromosomal anomalies?

I evaluated the value of adding a trisomy 18 screen to routine Down's screening and compared it with the benefits of atypicality screening. I studied 5080 unaffected pregnancies, 144 Down's syndrome and 190 non-Down's syndrome chromosome abnormalities (20 trisomy 13; 79 trisomy 18; 20 Turner's syndrome; 29 other sex chromosome abnormalities; 8 triploidy; and 34 miscellaneous). Using a one in 250 cut off, the Down's screen gave a screen positive rate of 4.07%; addition of atypicality without a trisomy 18 screen gave an extra 0.9% screen positives; trisomy 18 screening without atypicality gave an extra 0.51% screen positives; and atypicality screening after trisomy 18 screening gave 0.52% screen positives. Total screen positive rates were: Down's screening only, 4.07%; Down's screening + atypicality, 4.97%; Down's screening + trisomy 18 screen, 4.58%; Down's screening + trisomy 18 screen + atypicality, 5.09%. The detection rate for Down's syndrome using a one in 250 cut off was 58.9% and with addition of trisomy 18 and atypicality screening this increased to 59.3%, indicating that the extra screens add little to detection of Down's syndrome. For the other chromosomal abnormalities, Down's screening alone detected 22.6% of cases overall and addition of trisomy 18 and atypicality screening increased this to 49%. Examination of the marginal benefits of the extra screening tests revealed that the trisomy 18 screen was better at detecting chromosomal abnormalities than the Down's screen and that it would, therefore, be worthwhile adding this to all screening programs. Atypicality proved to be much less effective and it is suggested that this screen should only be applied in the early days of a screening program until sufficient data is available to design specific screens.