The association between meat consumption and risk of colorectal cancer has been controversial. We examined this question in a large prospective cohort study in Japan. From June through August 1990, 47,605 residents, aged 40-64 years, of Miyagi Prefecture in northern Japan completed a self-administered questionnaire, including a food frequency questionnaire. In the study population, we observed 474 incident cases of colorectal cancer during 11 years of follow-up, to March 2001. We used the Cox proportional hazards model to estimate the relative risk of colorectal cancer (colorectum, colon, rectum and proximal colon and distal colon) according to each of the categories of meat intake (total meat, beef, pork, ham or sausage, chicken and liver), with adjustment for sex, age and other potentially confounding variables. The multivariate relative risk of colorectal cancer in the highest category of total meat consumption compared with the lowest was 1.14 [95% confidence interval (CI)=0.85-1.53; P-trend=0.22]. We also found no significant association between total meat consumption and the risk of sub-site of colorectal cancer. In conclusion, our data do not support the hypothesis that meat consumption is a risk factor for colorectal cancer.