New Down syndrome screening algorithm: ultrasonographic biometry and multiple serum markers combined with maternal age.

OBJECTIVE

We compared the Down syndrome screening efficiency of a new algorithm that combines humerus length measurement and serum analytes versus that of the traditional triple-analyte serum screen.

STUDY

DESIGN

Humerus length measurements were obtained prospectively in 1743 midtrimester (14 to 24 weeks) singleton fetuses before genetic amniocentesis. All patients had triple-marker serum screening before amniocentesis. Data on humerus length were expressed as multiples of the median, and were normalized by log transformation. Backward multiple stepwise logistic regression analysis was performed to determine which combination of biometry and serum markers best predicted fetal Down syndrome. The screening efficiency of the traditional triple-analyte algorithm was compared with that of a new multivariate gaussian algorithm that combined biometry and serum markers.

RESULTS

There were 31 (1.8%) fetuses with Down syndrome in the study population. In the regression analysis humerus length, human chorionic gonadotropin, alpha-fetoprotein, and maternal age were significant predictors of Down syndrome, but unconjugated estriol was not. The combined algorithm (humerus length, human chorionic gonadotropin, and alpha-fetoprotein and age) was superior to the traditional triple screen for Down syndrome detection. The sensitivities at fixed false-positive rates were consistently higher in the combination than in the triple-screen protocol. For example, at a 10% false-positive rate the sensitivities were 65.0% and 52.3%, respectively. Similarly, at a 15% false-positive rate the sensitivities were 73.5% and 55.0%, respectively.

CONCLUSION

A new screening algorithm combining humerus length and serum analytes was superior to the traditional triple screen. Although we used a high-risk population in this study, it is expected that the observed superiority of the combination screen would persist in a population of younger women. The development of a combined biometric and serum analyte screening algorithm for estimating individual odds could represent an advance in prenatal Down syndrome screening.