• atopic dermatitis;
  • birth cohort;
  • cord blood IgE;
  • genetic polymorphisms;
  • prediction;
  • stress

To cite this article: Wen H-J, Wang Y-J, Lin Y-C, Chang C-C, Shieh C-C, Lung F-W, Guo YL. Prediction of atopic dermatitis in 2-yr-old children by cord blood IgE, genetic polymorphisms in cytokine genes, and maternal mentality during pregnancy. Pediatric Allergy Immunology 2011; 22: 695–703.


Atopic dermatitis (AD) is the most common skin disease in childhood and the first step of atopic march. This study aimed to investigate whether AD in children could be better predicted by biologic markers (cord blood IgE [cbIgE], LT-αNcoI alleles, and FcεRI-β E237G genotypes) and maternal mentality during pregnancy, taking into account gender, socio-demographic factors, and parental atopy. From 2001 to 2005, 1264 mother–infant pairs were recruited to participate in a birth cohort study. Prenatal questionnaire was used to collect family history, maternal gestational conditions and mentality, and environmental exposures. Cord blood was collected and assayed for genotypes and IgE levels. Phone interviews at 6 months and 2 yrs of age were conducted to inquire children’s health status, including AD occurrence. In addition to the known risk factors such as gender, maternal education, and parental atopy, biomarkers and maternal mentality during pregnancy were screened by logistic regression as candidate predictors of AD. Area-under-curve (AUC) statistic from receiver-operating characteristic (ROC) curve analysis was used to compare two predicting models with and without biomarkers and maternal mentality. A total of 730 pairs completed the prenatal questionnaire and phone interview and were included in final analysis. The prevalence of ever having physician-diagnosed AD by 2-yr-olds was 5.9%. Elevated cbIgE levels (≥0.5 kU/l), LT-αNcoI alleles, FcεRI-β E237G genotype, and maternal psychologic stress during pregnancy were significantly associated with AD. Comparison with AUCs of the classic model (including gender, maternal education, and parental atopy), the model adding cbIgE levels, genotypes in cytokine genes, and maternal stress (model 2) showed higher ability to discriminate between children with and without AD (AUC statistics: 0.63 [95% CI = 0.60–0.67] vs. 0.73 [95% CI = 0.70–0.76], respectively; model comparison, p = 0.027). We conclude that elevated cbIgE, LT-α and FcεRI-β genotypes, and maternal stress during pregnancy were associated with ever having physician-diagnosed AD in 2-yr-old children and increased the predictive ability for AD after taking into account gender, maternal education, and parental atopic history.