Key words: carotenoids; vitamin A; adenomatous polyps; epidemiologyCarotenoids are fat-soluble pigments found primarily in fruits and vegetables. Besides the well-known provitamin A activity of hydrocarbon carotenoids, numerous mechanisms have been postulated for carotenoid effects on carcinogenesis including inhibition of growth and induction of differentiation of cancer cells by modulating the expression of cell cycle regulatory proteins, modulation of the IGF-1/IGFPB-3 system, enhancement of immune function, increased gap junction intracellular communication, modulation of redox signaling and prevention of oxidative DNA damage and modulation of carcinogen metabolizing enzymes. 1 Consumption of carotenoid-rich fruits and vegetables has been associated with a reduction in risk of colorectal, lung, ovarian and other cancers. 2-5 However, few epidemiologic studies have examined individual carotenoid or vitamin A intake or serum concentrations and risk of colorectal cancer or adenomatous polyps, which are precursors to colorectal cancer. In a large beta-carotene supplementation trial of adenoma recurrence, similar in design to the Polyp Prevention Trial (PPT), participants in the beta-carotene supplemented group (20 mg/day) and placebo group experienced a similar rate of adenoma recurrence by intent to treat analysis. 6 In a subsequent analysis, stratified by smoking and drinking status, a significant protective effect from beta-carotene supplementation was observed in nonsmoking and nondrinking individuals. 7 Colorectal cancer is the third leading cancer in men and women in the U.S., resulting in an estimated 56,600 deaths in the year 2002. 8 While not all adenomas develop into cancer, it is believed that at least 95% of all colorectal cancers are derived from adenomatous polyps. The prevalence of colorectal adenomas in the U.S. ranges from 35-60%, 9 and recurrence of adenomas has been estimated at 10% or more annually. 10 Factors associated with recurrence of these lesions, including adenoma and patient characteristics, are not fully understood; however, a number of studies have demonstrated an increased recurrence of adenomas in individuals with a history of multiple adenomas. 11,12 Thus, reducing the incidence or recurrence of any or multiple adenoma formation is a plausible means for preventing colorectal cancer. Since genetic and epidemiology data suggest differences in colorectal cancer 13Abbreviations: BMI, body mass index; CI, confidence interval; FFQ, food frequency questionnaire; GAM, generalized additive model; LRT, log likelihood ratio test; NSAIDS, nonsteroidal anti-inflammatory drugs; OR, odds ratio; PPT, Polyp Prevention Trial.