OBJECTIVE: To examine the association between coffee consumption in pregnancy and foetal growth. DESIGN: Retrospective unmatched case-control study. SETTING: Maternidade de Campinas, Universidade Estadual de Campinas, Pontificia Universidade Católica de Campinas, Hospital Albert Sabin. SUBJECTS: 356 mother/baby pairs who had interauterine growth retardation (IUGR) and 356 mother/baby pairs who were appropriate for gestational age (AGA). INTERVENTIONS: Newborns were classified as being IUGR according to the Lubchenco classification. Gestational age of the newborns was evaluated by the Capurro method. Coffee consumption in pregnancy was assessed by a food frequency questionnaire. Coffee consumption and a range of risk factors for IUGR were stratified and entered into a logistic regression model. The final results were expressed by the attributable risk percent (AR%). RESULTS: More IUGR mothers (85.4%) than AGA mothers (70.5%) ingested coffee in pregnancy (OR = 2.45; P < 0.001). The proportion of mothers who delivered IUGR babies increased as the average consumption of coffee increased (test for trend = 31.76; P < 0.001). The tendency for heavy coffee drinkers to deliver IUGR babies remained after controlling for alcohol intake and cigarette smoking (P < 0.001). According to the logistic regression model and to the attributable risk percent (AR% = 28.0%), coffee consumption, (independent of average coffee consumption) was an important preventable cause of IUGR in this Brazilian population. CONCLUSIONS: We recommend moderation in the consumption of coffee in pregnancy, since intrauterine growth retardation increases the risk of perinatal and neonatal morbidity and mortality. Further large prospective studies evaluating the content of caffeine in the coffee consumed by this population is advised.