BACKGROUND: Epidemiologic evidence of associations between the high intake of fat and low intake of dietary fiber, beta carotene, and other dietary constituents and the risk of colorectal neoplasia has been inconsistent and has not provided a sufficient basis for recommendations concerning the dietary prevention of large-bowel cancer in humans. PURPOSE: We conducted a clinical trial to assess the effects on the incidence of adenomas of reducing dietary fat to 25% of total calories and supplementing the diet with 25 g of wheat bran daily and a capsule of beta carotene (20 mg daily). METHODS: We performed a randomized, partially double-blinded, placebo-controlled factorial trial in which half the patients were assigned to each intervention, resulting in seven intervention groups and one control group. Eligibility criteria included histologic confirmation of at least one colorectal adenoma and confidence expressed by the colonoscopist that all polyps had been removed. Dietary changes were individually initiated and monitored by dietitians and research nurses. At surveillance colonoscopy, the size and location of all polyps were recorded, and their histology was later centrally reviewed. Among 424 patients who were randomly assigned in the trial, 13 were found to be ineligible upon histologic review. Among the remaining 411, complete outcome data were collected from 390 at 24 months and from 306 at 48 months. All P values are from two-sided tests of statistical significance. RESULTS: There was no statistically significant prevention of total new adenomas with any of the interventions. We found a statistically non-significant reduced risk of large adenomas (> or = 10 mm) with the low-fat intervention: At 24 months, the odds ratio (OR) adjusted for potential confounders = 0.4 and 95% confidence interval (CI) = 0.1-1.1; at 48 months, OR = 0.3 and 95% CI = 0.1-1.0. Less and statistically nonsignificant reductions in the risk of large adenomas were found with wheat bran: At 24 months, OR = 0.8 and 95% CI = 0.3-2.2; at 48 months, OR = 0.8 and 95% CI = 0.3-2.5. Patients on the combined intervention of low fat and added wheat bran had zero large adenomas at both 24 and 48 months, a statistically significant finding (P = .03). CONCLUSIONS: Because only small numbers of patients were studied, our finding that the combination of fat reduction and a supplement of wheat bran reduced the incidence of large adenomas in this randomized, controlled trial must be treated with caution. The results do suggest, however, that these interventions may reduce the transition from smaller to larger adenomas, a step that may critically define those adenomas most likely to progress to malignancy.