Routine antenatal screening for Down’s syndrome (DS) is now capable of detecting nine out of 10 cases. In order to achieve this level of performance, multiple mater­nal serum and ultrasound markers are determined. Multi­variate Gaussian modeling techniques are used to convert information on the marker profile, maternal age and fam­ily history into a risk factor, and the result is considered positive if the risk exceeds a fixed cut-off.
The precision and accuracy of the risk estimate is dependent on the modeling parameters: means, standard deviations (SD) and correlation coefficients (R-values).

Risk is most imprecise when the parameters are de­rived from a single series as the number of DS cases will be relatively small. And it is inaccurate when the there are local factors such as ethnic distribution, assay precision and gestational accuracy which differ from those in the population used to derive the parameters. The best ap­proach is to use: 1) a meta-analysis of all the published results on DS pregnancies to estimate the DS means; 2) the local screened population to estimate the SDs and R-values in unaffected pregnancies, and 3) a meta-analysis to estimate the difference in variance and covariance be­tween DS and unaffected pregnancies, which are added to the unaffected estimates to derive DS SDs and R-values.
This has been done for first trimester maternal serum pregnancy associated plasma protein (PAPP)-A, free-β human chorionic gonadotrophin (hCG), β-fetoprotein (AFP) and unconjugated estriol (uE3) [1], and second trimester hCG, free-β hCG, AFP and uE3 [2]. In this paper the method is extended to second trimester inhibin, com­bining first trimester PAPP-A with second trimester mark­ers, and ultrasound first trimester nuchal translucency (NT) and second trimester nuchal skin-fold (NF). The existing and updated meta-analysis parameters are used to predict screening performance for policies using inhibin, NT, NF or PAPP-A together with second trimester serum.