It has been suggested that a high caffeine intake in pregnancy may be a risk factor for fetal growth retardation. We have tested this hypothesis in a population-based case-control study. Caffeine intake among 111 mothers of small-for-gestational-age (SGA) infants (56 boys, 55 girls) was compared with the intake among 747 mothers of non-SGA infants (368 boys, 379 girls). Food records for 3 days were collected in the second (week 17–20) and in the third (week 33) trimester, and caffeine intake from coffee, tea, soft drinks and chocolate was calculated and dichotomised as low or high, based upon the median value. Mothers of SGA infants had higher mean intake of caffeine [281 ± 210 (SD) mg/day] in the third trimester than mothers of non-SGA infants (212 ± 150 mg/day; P < 0.001). The risk of SGA birth was nearly doubled if the mother had a high rather than a low caffeine intake in the third trimester [odds ratio (OR) 1.8; 95% confidence intervals (CI) 1.2, 2.5]. The increased risk was mainly found in boys (OR 2.8; 95% CI 1.5, 5.2), and not in girls (OR 1.2; 95% CI 0.7, 2.1). The increased risk for boys persisted after adjustment for cigarette smoking alone, or for smoking and various other SGA risk factors together. Our results suggest that a high caffeine intake in the third trimester may be a risk factor for fetal growth retardation, in particular if the fetus is a boy.