Animal and in vitro studies provide evidence of an anticarcinogenic effect of active ingredients in garlic. This review of the epidemiologic literature on garlic consumption addresses cancers of the stomach, colon, head and neck, lung, breast and prostate. Nineteen studies reported relative risk estimates for garlic consumption and cancer incidence. Site-specific case-control studies of stomach and colorectal cancer, in which multiple reports were available, suggest a protective effect of high intake of raw and/or cooked garlic. Cohort studies confirm this inverse association for colorectal cancer. Few cohort and case-control studies for other sites of cancer exist. Garlic supplements, as analyzed in four cohort studies and one case-control report, from two distinct populations, do not appear to be related to risk. Low study power, lack of variability in garlic consumption categorization within studies and poor adjustment for potential cofounders may limit the reliability of any conclusions regarding garlic supplements. However, an indication of publication bias was also found by visual inspection of a funnel plot and in a log-rank test (P = 0.004). Evidence from available studies nevertheless suggests a preventive effect of garlic consumption in stomach and colorectal cancers. The study limitations indicate the need for more definitive research and improved nutritional epidemiologic analyses of dietary data.