||BACKGROUND: Studies in laboratory animals have suggested inhibitory effects of green tea on the induction of some cancers, notably, esophageal cancer. However, only a few epidemiologic studies have evaluated green tea as a potential inhibitor of human esophageal cancer. PURPOSE: Our purpose was to evaluate the relationship between green tea consumption and the risk of esophageal cancer. METHODS: This esophageal cancer study was part of a larger multicenter, case-control study that included three other gastrointestinal sites (pancreas, colon, and rectum). Medical records of patients aged 30-74 years old who were diagnosed with esophageal cancer from October 1, 1990, through January 31, 1993, were identified from the Shanghai Cancer Registry, which covers 6.8 million people in the urban area of Shanghai, People's Republic of China. During the ascertainment period, records of 1016 eligible cases of esophageal cancer were identified. Control subject records were selected by frequency matching in accordance with the age-sex distribution of the four gastrointestinal cancers ascertained by the cancer registry during 1986-1987. Patient interviews were then conducted using a structured, standardized questionnaire to obtain information on demographic characteristics, residential history, height and weight, diet, smoking, alcohol and tea drinking, medical history, family history of cancer, occupation, physical activity, and reproductive history. RESULTS: Of the 902 patients interviewed, 734 (81.4%) had their disease pathologically confirmed. There were 1552 control subjects interviewed, including 240 alternates. All analyses of tea effects were conducted separately among men and women and all were adjusted for age. After further adjustment for other known confounders, a protective effect of green tea drinking on esophageal cancer was observed among women (odds ratio [OR] = 0.50; 95% confidence interval [CI] = 0.30-0.83), and this risk decreased (P for trend < or = .01) as tea consumption increased. Among men, the ORs were also below 1.00, although not statistically significant. ORs for green tea intake were estimated among those persons who neither smoked nor drank alcohol. In this subset, statistically significant decreases in risk among tea drinkers were observed for both men (OR = 0.43; 95% CI = 0.22-0.86; P for trend = .05) and women (OR = 0.40; 95% CI = 0.20-0.77; P for trend < .001). CONCLUSIONS: This population-based, case-control study of esophageal cancer in urban Shanghai suggests a protective effect of green tea consumption. Although these findings are consistent with studies in laboratory animals, indicating that green tea can inhibit esophageal carcinogenesis, further investigations are definitely needed.