AIMS: Maternal alcohol consumption is a major health hazard for the fetus. Sweden has an extensive system of public antenatal care clinics, whose mission is to detect and prevent this type of health hazards. However, very few cases of alcohol consumption during pregnancy are detected. The aim of this study was to examine the prevalence of hazardous or harmful alcohol consumption during pregnancy in a consecutive series of Swedish pregnant females. DESIGN, SETTING, PARTICIPANTS AND MEASUREMENTS: The Alcohol Use Disorders Identification Test (AUDIT) was used to collect anonymous data from consecutive pregnant subjects admitted during 1 year to an antenatal clinic in Stockholm, and signing up for parental education offered routinely (n = 1327). Data were obtained from 1101 subjects, typically in pregnancy week 30. A complete AUDIT form was filled out referring to alcohol use during the year prior to pregnancy. A separate form with the consumption items from AUDIT was filled out to report behaviour during pregnancy. FINDINGS: For the year preceding pregnancy, 17% of subjects reported AUDIT scores of 6 or higher, indicating hazardous or harmful alcohol use in women. Few individuals reported scores of 13 or higher (indicating abuse or dependence), but almost half the subjects (46%) reported binge drinking (six standard drinks on a single occasion) once/month or more often, and 6% reported binge drinking on every occasion of alcohol consumption. One-third of the subjects (30%) continued regular alcohol use during pregnancy, and 6% reported consumption two to four times/month. In a logistic regression model, AUDIT scores for the year prior to pregnancy and subject age, but not education level were significant predictors of continued alcohol use during pregnancy. CONCLUSIONS: Alcohol use during pregnancy is more extensive than has been presumed in Sweden. Simple, clinically useful screening methodology detects hazardous consumption during pregnancy in a manner which regular antenatal care does not. If this methodology can be shown to have similar sensitivity when administered under non-anonymous conditions, it should be made part of routine antenatal care.