A multivariate analysis of dietary and microbial interactions in the etiology of colorectal cancer
Cancer of the colon and rectum is a significant (15%) type of cancer in this country and other western nations. It appears from epidemiological data that the cause of this morbidity and mortality is directly attributable to diet. Many dietary influences have been studied for their correlation with the development of human colon cancer: fat levels and fat types, high cholesterol consumption, and reduced fiber and vegetable intake. All of these dietary factors have been studied individually in colon cancer animal models, but this approach does not accurately detect interactions occurring across and between dietary variables. The specific objective of this study was to design an experimental animal model for colon cancer which examined the interactions of dietary factors as they might coexist in the human diet, and to determine which variables (or combinations of variables) pose the greatest risk for tumor development.
The experimental model was a 2^ factorial design and the four dietary factors examined were wheat bran, cholesterol, beef tallow, and the vegetable compound, indole-3-carbinol (l-3-C). Colon tumors were induced in 160 male Fischer 344 rats with 1, 2-dimethylhydrazine. Hematocrit levels, white blood cell counts, fecal mutagen, and Bacteroides fragilis counts in feces were examined at seven week intervals. At necropsy additional variables examined included total weight gain, spleen and liver weights, incidence and severity of tumors, serum cholesterol levels and activity of the liver enzyme aryl hydrocarbon hydroxylase (AHH). A comprehensive multivariate analysis demonstrated that the combination of cholesterol with both beef tallow and corn oil, accompanied by an AHH inducer (1-3-C) posed the greatest risk for colon cancer. Wheat bran consumption decreased this risk only in diets with cholesterol. In other diets, bran was found to have an enhancing effect on colon tumorigenesis, so its role must be evaluated cautiously. These results are consistent with previous epidemiologic data.