Preprint version. Prepared for: U.S. Army Medical Research and Materiel Command Fort Detrick, Maryland 21702-5012, Volume Award Number: DAMD17-01-1-0057, June 1, 2002, pages 1-26.
NOTE: At the time of publication, the author Marilyn Tseng was not yet affiliated with Cal Poly.
We examined the association between dietary patterns and prostate cancer risk using data from the first National Health and Nutrition Examination Epidemiologic Follow-up Study. The study population included 3,779 men followed 1982-4 to 1992, with 136 prostate cancer cases identified during the follow-up. Principal components analysis on responses to a lOS-item diet questionnaire in 1982-84 was used to identify dietary patterns, and scores representing level of intake of each pattern were categorized into tertiles. Three distinct patterns were identified: a vegetable-fruit pattern, a red meat-starch pattern characterized by intake of beef, pork, potatoes, and sweets, and a southern pattern characterized by foods including cornbread, grits, bacon, beans, and okra. In proportional hazards models, prostate cancer risk was not associated with the red meat-starch pattern, but it was non-significantly higher with intermediate consumption of a fruit-vegetable pattern and non-significantly lower with high intake of a southern dietary pattern (RR=0.6, 95% CI 0.4-1.1 for highest vs. lowest tertile). Of the nutrients and foods that we examined, only calcium and dairy foods were associated with prostate cancer risk, but RR estimates for dietary patterns were unchanged after adjustment for either calcium or dairy. Features of the patterns that might contribute to the associations that we observed have yet to be elucidated.