BACKGROUND: Alcohol use during pregnancy leads to detrimental effects on fetal development. As self-reports by mothers are known to be unreliable for assessing prenatal alcohol exposure, there is a need for sensitive and specific biomarkers for identifying those at risk for alcohol-affected offspring. METHODS: We measured serum gamma-glutamyl transferase (GGT), carbohydrate-deficient transferrin (CDT), a mathematically formulated combination of GGT and CDT (GGT-CDT), and ethylglucuronide (EtG) concentrations from 1,936 mothers with a positive (n = 480) or negative (n = 1,456) history of alcohol use at the time of pregnancy. The material included 385 alcohol-abusing mothers who subsequently gave birth to children with fetal alcohol syndrome (FAS) and 1,551 mothers without FAS children including 95 women who reported a median of 1.0 standard drinks of alcohol per day during pregnancy and 1,456 nondrinking controls. Among those without FAS outcome, there were 405 mothers with gestational diabetes mellitus (GDM) and 745 mothers representing lifelong abstainers. RESULTS: Mothers of FAS children had higher mean GGT, CDT, GGT-CDT, and EtG levels than abstainers (p < 0.001 for all comparisons) or mothers reporting some alcohol consumption but whose children were not diagnosed with FAS (p < 0.001 for all comparisons). In receiver operating characteristic analyses using cutoffs based on abstainers, the area under the curves (AUCs) for GGT-CDT (0.873) were higher than those of GGT (0.824), CDT (0.776), or EtG (0.584) for differentiating the mothers of FAS children and abstainers. Unlike CDT, this combination marker also differed significantly between drinking mothers without FAS outcome and abstainers (AUC = 0.730, p < 0.001). In comparisons adjusted for GDM and body mass index, the group of mothers who had reported a median of 1.0 standard drinks of alcohol per day during pregnancy also differed from the group reporting no current alcohol intake in GGT (p < 0.02) and GGT-CDT (p < 0.01) levels. CONCLUSIONS: Combination of GGT and CDT improves the identification of prenatal alcohol exposure and associated high-risk pregnancies. A more systematic use of biomarkers may help intervention efforts to prevent alcohol-induced adverse effects on fetal development.