INTRODUCTION: The aim of this study was to examine the performance of the selective screening programme for hepatitis B (HB) in pregnancy. MATERIAL AND METHODS: Notifications of women with acute HB in the period 1981-1997 were traced in the National Birth Registry. For children born after the HB event, details of the records on the results of HB testing and other information about the previous HB disease were obtained from the maternity wards. RESULTS: The study group comprised 129 mothers and their 185 deliveries. Of the 185 deliveries, 31% took place without any information about the previous HB disease and 51% without testing for HB. When these results were related to the information on the notification form regarding the mode of transmission, it transpired that the proportion with an unknown mode of transmission was higher in those with no information about previous HB in the maternity records than in those with information (43% vs 18%). Of 18 infants delivered of 13 chronically infected mothers, four did not receive the proper immunoprophylaxis. DISCUSSION: The study shows that a selective screening programme as used in Denmark is not good enough to identify pregnant women at risk of chronic HB infection. The main problems are related to failure to identify risk factors of hepatitis B transmission. Screening and treatment failure in siblings delivered of chronically infected mothers was also found. As in most other western countries a general programme should replace the selective screening programme.