The purpose of the Eurofetus Study was to evaluate the accuracy of the antenatal detection of malformations by routine ultrasonographic examination in unselected populations. All ultrasonographic diagnoses of malformations and the outcomes of the fetuses were prospectively recorded in 61 European obstetric units over a 3-year period (1990-1993). Also recorded were all cases of malformation diagnosed after abortion or birth for the mothers who underwent follow-up in these centers. Of 3685 malformed fetuses, 2262 had received diagnoses during pregnancy (sensitivity, 61.4%). Of a total number of 4615 malformations, 2593 were detected (sensitivity, 56.2%). The detection sensitivity was higher for the major than for the minor abnormalities (73.7% vs 45.7%), and the diagnosis was made earlier in the pregnancy (24.2 weeks vs 27.6, P <. 01). Overall, 55% of the major abnormalities were detected within 24 gestational weeks. Within each severity group the accuracy of detection depended on the system. For the major abnormalities it was better for the central nervous system (88.3%) and urinary tract (84. 8%) but lower for the heart and great vessels (38.8%). Detection of minor abnormalities was also effective for the urinary tract (89.1%) but not for the heart and great vessels (20.8%) or the musculoskeletal system (18%). Detection of abnormalities had an influence on the rate of termination of pregnancy. The rate of live births for the mothers bearing fetuses with major abnormalities was lower than that for the mothers in whom no abnormalities were detected, mainly because of the higher rate of elective terminations of pregnancy in the former group. Systematic ultrasonographic screening during pregnancy can now detect a large proportion of fetal malformations, although some still escape detection.