Objective: The association between vegetable consumption and colorectal cancer risk remains unclear and may differ by region. We performed a systematic review and meta-analysis of epidemiologic studies on this issue among the Japanese population.
Methods: A systematic review and meta-analysis was performed by searching MEDLINE through PubMed and the Ichushi database for cohort and case-control studies that were published by the end of December 2014. Associations were evaluated based on their magnitude and the strength of the evidence. Meta-analysis was performed by using the random effects model to estimate the summary relative risk with 95% confidence interval according to the study design. The final judgment was made based on a consensus of the research group members with consideration for both epidemiological evidence and biological plausibility.
Results: We identified six cohort studies and 11 case-control studies on vegetable intake and colorectal cancer among the Japanese population. Of the cohort studies, one study showed a weak inverse association with colon cancer and another study showed a weak positive association with rectal cancer in men, but other studies found no associations between vegetable consumption and colon and rectal cancers. With regard to case-control studies, one study found a strong inverse association with colon cancer, and three studies showed a weak-to-strong inverse association with rectal cancer. In meta-analysis, the summary relative risk (95% confidence interval) for the highest vs. the lowest categories of vegetable consumption were 1.00 (0.92-1.10) and 0.75 (0.59-0.96) for cohort and case-control studies, respectively.
Conclusions: There was insufficient evidence to support an association between intake of vegetables and the risk of colorectal cancer among the Japanese population.
Keywords: Japanese; colorectal cancer; epidemiology; systematic review; vegetable.
© The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email: email@example.com.