Ethnicity-specific normative models of quadruple test as a screening test for down syndrome

Abstract

Background and Objectives: To establish normative models for median levels of serum biomarkers of the second trimester quad test (alpha-fetoprotein: AFP; free beta-human gonadotropins: hCG; inhibin-A; and unconjugated estriol: uE3) specific to Thai women and to compare multiples of the median (MoMs) derived from ethnicity-specific models and those derived from Caucasian models with ethnic correction. Materials and Methods: A cross-sectional study was undertaken in a tertiary, medical teaching center among low-risk pregnant Thai women between 14 and 21 weeks of gestation to measure the levels of the four serum biomarkers. The measured values of each biomarker were analyzed using the multivariable factorial polynomial technique for quantile regression as a function of gestational age and maternal weight. Results: The Thai-specific normative models for the four biomarkers were generated and available for use. The MoMs of all individuals generated from our models were significantly different from conventional (Caucasian) models with ethnic correction (Wilcoxon signed-rank test; p < 0.0001 for all biomarkers). The MoMs of AFP and hCG from both methods were in agreement, but those from Thai-specific models were significantly higher. However, those of inhibin-A and uE3 were markedly different and ethnic correction was unlikely to be useful. Conclusions: The Thai-specific normative models of the quad test as a function of gestational age and maternal weight were constructed using multivariable factorial polynomial models, better than simple quantile regression or log-linear regression used in earlier decades. The analysis of MoMs supports the use of ethnicity-specific models instead of Caucasian models with ethnic correction.