Cord serum IgE was assayed by particle counting immunoassay (PACIA) in an unselected series of European newborns (n = 190; geom mean = 0.37 IU/ml) and a cut-off limit established (greater than or equal to 1.20 IU/ml) for prediction of atopy. At control follow-up by questionnaire 18 months after birth, 38 infants (20.0%) had developed definite (9.5%) or probable (10.5%) atopy with a significant predominance of boys (P less than 0.03). Infants with a positive immediate family history (IFH) had a higher risk of developing atopy (P less than 0.0025) and also had a higher incidence of elevated cord IgE (P less than 0.02) than infants with a negative IFH. Maternal atopy influenced cord IgE levels significantly (P less than 0.00005), whereas paternal atopy did not (P = 0.23). No fetal IgE antibodies against five common allergens could be demonstrated in 36 cord sera tested. Breast-feeding for 3 months was not sufficient to prevent atopic symptoms. The predictive value of cord IgE was high since 26 of 36 newborns (positive predictive value = 72.2%) with elevated cord IgE had developed atopic symptoms before follow-up. Of the 38 infants who developed atopic symptoms, 26 had elevated cord IgE (sensitivity = 68.4%) compared to only 10 (6.6%) of the 152 atopy-free infants (P less than 0.00005). The data indicate that elevated cord IgE as determined by PACIA is a good predictor of early-onset atopy, better than family history (P less than 0.008), and that primarily maternal atopy seems to affect fetal IgE synthesis.