ORIGINAL ARTICLE
Medians for second-trimester maternal serum markers: geographical differences and variation caused by median multiples-of-median equations
1 Analis NV, Leeuwerikstraat, Gent, Belgium
2 Queens Hospital Burton on Trent, Burton on Trent, UK
Correspondence to:
Correspondence to:
G Vranken
Analis NV, Leeuwerikstraat 28, B-9000 Gent, Belgium; gvn{at}analis.be
Objectives: To establish gestational age-specific mid-trimester normal medians for the prenatal serum markers
fetoprotein (AFP), human chorionic gonadotropin (HCG) and unconjugated oestriol (uE3) for a Belgian population by using the Beckman Coulter Access chemiluminiscent immunoassays; to compare these data with data obtained from other geographical regions; to propose regression coefficients for regressed medians and analyse variation induced by different regression equations; to evaluate the effect of formulas used for gestation correction on estimating risk in Downs syndrome.
Design: Data derived from 862 fresh serum samples from women being screened for Downs syndrome pregnancy, composed of selected pregnancies deemed to be normal, were examined in a retrospective study. Regressed medians were calculated by using a first-degree logarithmiclinear fit of the raw data. Multiples-of-median (MoM) values estimated by using a simple logarithmiclinear equation were compared with those calculated with higher-degree polynomials chosen with a goodness-of-fit analysis. Model-specific variation was estimated and the effect on risk for Downs syndrome was evaluated.
Results: Regressed medians (Y) for Access serum markers AFP (IU/ml), HCG (IU/ml) and uE3 (nmol/l) for a Belgian population can be estimated with the equation Y = 10(A+BX) with X = decimal weeks. The best fit was obtained with a third-degree and a second-degree polynomial for AFP and uE3, respectively. Differences between the medians and among the slopes of the geographical populations were found to be significant (analysis of covariance, p<0.001).
Conclusions: Belgian marker medians versus gestational time are found to show a pattern that is similar to that in the literature. The loglinear equation is observed to give a good fit and can be suggested as a tool for calculating median MoM values for Belgian laboratories that use Access biochemical prenatal markers.
Abbreviations: AFP,
fetoprotein; ANOVA, analysis of variance; HCG, human chorionic gonadotropin; uE3, unconjugated oestriol; MoM, multiples of median; VIF, variance inflation factor
![]()
CiteULike
Complore
Connotea
Del.icio.us
Digg
Reddit
Technorati What's this?
Register for free content
The full back archive is now available for all BMJ Journals. Institutional subscribers may access the entire archive as part of their subscription. Personal subscribers will also have access to all content when logged in. Non-subscribers who register have free access to all articles published before 2006 right back to volume 1 issue 1. Register here to access the free archive of all BMJ Journals.
Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.
