BackgroundEvidence about the effects of low levels of alcohol consumption during pregnancy has been inconclusive, and national guidelines vary from advising abstention to consuming no more than 32 g per week, the equivalent of 1–2 UK units, once or twice a week (in the UK). We aimed to determine the effects of low-to-moderate levels of maternal alcohol consumption on pregnancy and longer-term offspring outcomes. MethodsWe systematically reviewed and pooled estimates of the effects of low levels of alcohol consumption (up to 32 g per week) during pregnancy on pregnancy outcomes and offspring health. We searched Medline, Embase, Web of Science, and PsycINFO (Jan 1, 1950, to Feb 2, 2015), for studies published in English using a combination of MeSH terms and text words to reflect outcomes (eg, pregnancy, fetal development), exposure (alcohol), and relevant study designs. Prospective observational studies, including negative control and quasi-experimental studies of pregnant women estimating effects of light versus no alcohol consumption, were included. Outcomes included pregnancy outcomes such as birth weight, and features of fetal alcohol syndrome such as cognitive impairment. Data extraction was carried out by one of four reviewers and then checked by the lead reviewer. Random-effects meta-analyses were performed where applicable (ie, where there were enough studies for the meta-analysis). This study is registered with PROSPERO, number CRD4201501594. Findings23 cohort and two quasi-experimental studies were included. With the exception of birth size and gestational age, there were insufficient data for the meta-analysis or to make robust conclusions about effects. The odds for small for gestational age and preterm birth were higher for babies whose mothers consumed up to 32 g per week than for babies whose mothers consumed none but results were imprecisely estimated (pooled odds ratio 1·07, 95% CI 0·99–1·16, I2=4% [six studies] and 1·13, 0·96–1·33, 63% [eight studies], respectively). InterpretationWe found little evidence to support recommendations related to light drinking. Since there was some evidence that even light alcohol consumption in pregnancy is associated with preterm delivery and having a small for gestational age baby, guidance should explain the paucity of evidence and could advise abstention as a precautionary principle. FundingMRC Integrative Epidemiology Unit and National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care (NIHR CLAHRC) West.