Diet and Cancer in a U.S. Cohort Containing Many Vegetarians

Synnove KNUTSEN, Center for Nutrition, Healthy Lifestyle and Prevention, Loma Linda University, United States

1 Center for Nutrition, Healthy Lifestyle and Prevention, Loma Linda University, Loma Linda, CA, USA

PURPOSE: The purpose of this study is to evaluate associations between dietary patterns, foods, nutrients and minerals, and risk of common cancers.

METHODS: Adventist Health Study-2 is a cohort of 96,000 subjects from the U.S. and Canada, established between 2002-7. About half are vegetarians and the others eat meat at least weekly. This provides a wide range of dietary habits, and many cohort members represent a region of intake of certain foods/nutrients uncommonly found in other studies. Dietary habits were measured by a validated food frequency questionnaire at study baseline. A calibration study of 1100 representative subjects also completed 6 structured 24 hour recalls. Incident cancers were ascertained by matching with state and provincial cancer registries. Analyses used Cox proportional hazard analyses with energy-adjustment, and where appropriate regression calibration for measurement error correction.

RESULTS: Adventists have substantially lower incidence rates of several common cancers. Vegan vegetarians only, have 1/3 lower incidence of prostate cancer compared to non-vegetarians. Dairy consumption is significantly positively, but consumption of cooked tomatoes also intake of alpha linolenic acid are negatively, associated with risk of prostate cancer. Despite a very low average intake of processed red meat, a significant association is found with risk of colorectal cancer with a much stronger slope than reported from meta-analyses of other studies. Calcium consumption is significantly negatively associated with colon (but not rectal) cancer, and dairy kcals are significantly negatively associated with rectal (not colon) cancer. For breast cancer there is a negative association with intake of soy isoflavones (half of this population eat at Asian levels) and an apparent advantage of substituting these for dairy foods.

CONCLUSIONS: In this special population prostate, colorectal and breast cancers are associated with diet in ways that often have known plausible biological mechanisms.

FUNDING: National Cancer Institute, World Cancer Research Fund