A case-control study of diet and colo-rectal cancer.

A case-control study of cancer of the colon and rectum has been conducted in Calgary, Alberta and Toronto, Ontario, Canada. A total of 348 cases of cancer of the colon and 194 cases of cancer of the rectum were individually matched by age, sex and neighbourhood of residence to 542 population controls and frequency match to 535 hospital controls who had undergone an abdominal operation. Each subject received a personal medical history questionnaire and a quantitative diet history questionnaire. Data on a number of potential non-nutrient risk factors for bowel cancer and on the consumption of 9 nutrients in the 2-month period up to 6 months before interview were analysed. The dietary data thus refer to recent diet consumed in a period antedating the diagnosis of, and in most cases symptoms from, large-bowel cancer in the cases, and a corresponding time period in the controls. The major findings were an elevated risk for those with a history of bowel polyps, and for those with an elevated intake of calories, total fat, total protein, saturated fat, oleic acid and cholesterol. No association was seen with an elevated intake of crude fibre, Vitamin C and linoleic acid. The nutrients for which an increased risk was demonstrated were highly correlated, though multivariate analysis using logistic regression indicated highest risk for saturated fat, with evidence of a dose-response relationship. The findings in both cancer sites, both sexes and with both sets of controls were quantitatively very similar. The population-attributable risk for colon and rectal cancer combined was estimated from the neighbourhood controls to be 41% for males and 44% for females for saturated fat intake and 9.8% and 6.4% respectively for any history of polyps.