BACKGROUND

The associations between dietary factors and gastric cancer risk have been analysed by many studies, but with inconclusive results. We conducted a meta-analysis of prospective studies to systematically investigate the associations.

METHODS

Relevant studies were identified through searching Medline, Embase, and Web of Science up to June 30, 2015. We included prospective cohort studies of intake of dietary factors with risk estimates and 95% confidence intervals for gastric cancer.

RESULTS

Seventy-six prospective cohort studies were eligible and included in the analysis. We ascertained 32,758 gastric cancer cases out of 6,316,385 participants in relations to intake of 67 dietary factors, covering a wide ranging of vegetables, fruit, meat, fish, salt, alcohol, tea, coffee, and nutrients, during 3.3 to 30 years of follow-up. Evidence from this study indicates that consumption of total fruit and white vegetables, but not total vegetables, was inversely associated with gastric cancer risk. Both fruit and white vegetables are rich sources of vitamin C, which showed significant protective effect against gastric cancer by our analysis too. Furthermore, we found concordant positive associations between high-salt foods and gastric cancer risk. In addition, a strong effect of alcohol consumption, particularly beer and liquor but not wine, on gastric cancer risk was observed compared with nondrinkers. Dose-response analysis indicated that risk of gastric cancer was increased by 12% per 5 g/day increment of dietary salt intake or 5% per 10 g/day increment of alcohol consumption, and that a 100 g/day increment of fruit consumption was inversely associated with 5% reduction of risk.

CONCLUSION

This study provides comprehensive and strong evidence that there are a number of protective and risk factors for gastric cancer in diet. Our findings may have significant public health implications with regard to prevention of gastric cancer and provide insights into future cohort studies and the design of related clinical trials.