Background: Women who smoke are less likely to breastfeed their children than nonsmokers. It is thought that nicotine has a negative effect on breastmilk supply by suppressing prolactin levels. The aim of this review was to assess the epidemiological evidence that maternal smoking has a negative physiological effect on breastfeeding. Methods: The following data sources were searched: The Cochrane Library, Medline, CINAHL, Current Contents, Psychinfo, Sociological Abstracts and the Lactation Resource Centre (Australian Breastfeeding Association) using the key words "smoking" and "breastfeeding" or "infant feeding." The Journal of Human Lactation and Birth were hand searched. Results: Women who smoke are less likely to intend to breastfeed, less likely to initiate breastfeeding, and likely to breastfeed for a shorter duration than nonsmokers. Several studies have found a dose-response relationship between the number of cigarettes smoked each day and breastfeeding intention, initiation, and duration that persists after adjusting for confounding factors. In some population groups a high proportion of smokers breastfeed successfully. Conclusions: The association between maternal smoking and lack of breastfeeding is consistent across different study designs in a range of countries. Given that women who smoke are less likely to intend to breastfeed, however, it cannot be assumed that the relationship between smoking and duration of breastfeeding is a physiological one. If smoking had a consistent negative physiological effect on lactation, one would not expect to see such wide variations in breastfeeding rates among women who smoke. Therefore, it is likely that psychosocial factors are largely responsible for the lower rates of breastfeeding found in women who smoke compared with those who do not.