Effect on Down syndrome screening performance of adjusting for marker levels in a previous pregnancy.
Wald NJ., Barnes IM., Birger R., Huttly W.
OBJECTIVES: In prenatal screening for Down syndrome, serum marker values can be adjusted using values from a previous pregnancy to avoid the problem of women having a high chance of recurrent false-positive results. We investigate the effect of such adjustment on overall screening performance. METHODS: Monte Carlo simulation was used to investigate the effect of this adjustment on five widely used screening tests for Down syndrome (Triple, Quadruple, Combined, serum Integrated, Integrated tests). RESULTS: Adjustment for screening marker values (expressed in multiples of the median, (MoM)) in a previous pregnancy improved screening performance. The detection rate for a 1% false-positive rate (FPR) increased from 54 to 59% with the Triple test, from 63 to 68% with the Quadruple test, from 70 [corrected] to 75% for the Combined test, from 70 [corrected] to 76% for the serum Integrated test, and from 85 to 88% for the Integrated test. The FPR for an 85% detection rate decreased from 10 to 7.9%, 7.1 to 4.9%, 4.9 to 3.7%, 4.7 to 2.9% and 1.1 to 0.7% respectively for the five tests. Among women who have had a false-positive result in a previous pregnancy, adjustment substantially lowers the false-positive rate, for example, from 18 [corrected] to 7.3% with the Combined test using a 1 in 250 risk cut-off. CONCLUSION: MoM adjustment for values in a previous pregnancy improves overall screening performance and substantially reduces the high recurrent false-positive rate. This adjustment can be routinely applied in screening programmes through the screening software used to interpret a woman's screening results.