Inaccurate estimation of risk in second trimester serum screening for Down syndrome among women who have already had...

Problems can arise in prenatal screening for Down syndrome when tests are performed in the first and second trimester and some women who have a negative first trimester test have a second trimester serum test. The second test result does not usually take account of the previous one being negative. Even if it does, it is often inaccurate. Using published data the extent of the error was examined. The age-specific risk of an affected pregnancy in such women will be lower than if no first trimester test had been performed. The distributions of the screening markers in affected and unaffected pregnancies will be different from those in unscreened women. If the appropriate age-specific risk and marker distributions are not used, error will arise. For example, a 35-year-old woman with nuchal translucency (NT), pregnancy-associated plasma protein-A (PAPP-A) and free beta-human chorionic gonadotrophin (hCG) levels at the normal median would have a risk of 1 in 6500. If she then had the Triple Test with alpha-fetoprotein (AFP), unconjugated oestriol, and hCG levels of 0.7, 0.7 and 1.5 multiples of the median (MoM), respectively, her risk, ignoring the previous result, would be overestimated (1 in 95 compared with the correct estimate of 1 in 705). If the previous result was included, but the age-specific risk and second trimester marker distributions were not revised, her risk would be underestimated (1 in 820). If the correct age-specific risk and screening marker distributions were used, risk estimates would be accurate, but two tests would be less efficient than integrating all the screening information into a single test. The practice of offering second trimester serum screening to women who have already been screened is best avoided.