Whether caffeine consumption during pregnancy represents a fetal hazard remains uncertain. The authors report on a large prospective study designed to examine this question. In 1996-2000, 2,291 mothers with singleton livebirths in Connecticut and Massachusetts were evaluated after their first prenatal visit and were questioned about caffeine consumption and important confounding factors. Urine samples were provided to analyze urinary caffeine, cotinine, and creatinine levels. Mothers were followed throughout pregnancy to monitor changes in consumption. Pregnancy outcomes were obtained from medical records. Self-reports of caffeine consumption in the first and third trimesters were not associated with intrauterine growth retardation, low birth weight, or preterm delivery. For every 1 mg/g creatinine increase in urinary caffeine, risk of intrauterine growth retardation was essentially unchanged (odds ratio (OR) = 0.96, 95% confidence interval (CI): 0.85, 1.08). In contrast, a 0.005 mg/g creatinine increase in urinary cotinine significantly increased risk (OR = 1.003, 95% CI: 1.001, 1.005). Mean birth weight was reduced by reported caffeine consumption (-28 g per 100 mg of caffeine consumed daily, 95% CI: -0.10, -0.46, p = 0.001) but not mean gestational age. Decaffeinated coffee did not increase risk for any perinatal outcome. This small decrease in birth weight, observed for maternal caffeine consumption, is unlikely to be clinically important except for women consuming >/=600 mg of caffeine daily (approximately six 10-ounce (1 ounce = 28.3 g) cups of coffee).