Previous in vitro and animal experiments have shown that sulforaphane, which is abundant in broccoli, inhibits Helicobacter pylori (H. pylori) infection and blocks gastric tumor formation. This suggests that broccoli consumption prevents chronic atrophic gastritis (CAG) introduced by H. pylori infection and, therefore, gastric cancer. For an epidemiological investigation of the relationship between the broccoli consumption and CAG, a cross-sectional study of 438 male employees, aged 39 to 60 years, of a Japanese steel company was conducted. CAG was serologically determined with serum cut-off values set at pepsinogen I < or = 70 ng/ml and a ratio of serum pepsinogen I/pepsinogen II < or = 3.0. Broccoli consumption (weekly frequency) and diet were monitored by using a 31-item food frequency questionnaire. The prevalence of CAG among men who ate broccoli once or more weekly was twice as high as that among men who consumed a negligible amount (P < 0.05). Multiple logistic regression analysis indicated that broccoli consumption once or more weekly significantly increased the risk for CAG (odds ratio, 3.06; 95% confidence interval, 1.12-8.38; P < 0.05), after controlling for age, education, cigarette smoking, and alcohol consumption. The present study failed to show an expected association between frequent broccoli consumption and a low prevalence of CAG.