The aim of the present study was to assess the sensitivity of ultrasound diagnosis used as a screening test in detecting major congenital anomalies in the prenatal period in a large nation-based multicentre setting. Data from the IPIMC register were collected in the period 1986-1990. One hundred and thirty-five hospitals, located in 17 out of the 20 regions in Italy, participated in the register. Study cases were 3479 infants with major congenital anomalies diagnosed at birth or in the first week of life. Subjects with chromosomal anomalies or multiple defects were excluded. The sensitivity of ultrasound prenatal diagnosis was 49.5 per cent for central nervous system anomalies, 3.8 per cent for congenital heart diseases, 17.1 per cent for gastrointestinal tract defects, 46.6 per cent for abdominal wall defects, 74.8 per cent for urinary tract anomalies, and 22.9 per cent for skeletal abnormalities. The detection rate for diaphragmatic hernia was 24.2 per cent. Overall, only 18 per cent of the defects diagnosed in utero were detected before 24 weeks' gestation. The sensitivity of prenatal diagnosis was 30.1 and 19.0 per cent in the northern, central, and southern regions, respectively. In light of its low sensitivity, ultrasonography as a screening test in the general population should be abandoned, although some improvement in its performance should be expected following adequate training of the ultrasound staff and the use of good technical equipment.