This article reports on a systematic review of studies on caffeine intake during pregnancy and prevalence of low birth weight and prematurity, focusing on methods to quantify intake and control for confounding. The review consisted of an article search from 1996 to 2006 in MEDLINE, LILACS, and PubMed, using the key words: "caffeine", "coffee", "low birth weight", "birth weight", "preterm", "premature", and "prematurity". Ten articles were selected. Methods used to quantify caffeine consumption were: semi-quantitative food frequency questionnaires for diet or only caffeinated products, including one self-applied questionnaire; food recall; questions on type and method of preparation; analysis of samples; and urine and plasma caffeine levels. In three studies, high caffeine consumption was associated with low birth weight and/or prematurity. Contradictions between studies may be due to difficulties in measuring caffeine consumption; assessment of different caffeine sources; variations in the mode of preparation and amount consumed; and sample size. Association between moderate caffeine consumption and fetal growth was not demonstrated, so a more precise measurement of caffeine intake is necessary.