The relationship between nutrient intake and colorectal cancer in a Tecumseh cohort.

Abstract

In 1967-69 a subset of the third phase of the Tecumseh Community Health Study included dietary data collected by the methods of food frequency and 24 hour recall. The resultant dataset offered a unique opportunity to investigate the effects of specific nutrients on the occurrence of large bowel cancer 20 years later and constituted the problem addressed in this research project. The nutrients of interest were: (i) total fats; (ii) saturated fats; (iii) unsaturated fats; (iv) fiber; (v) calcium; (vi) carotenoids; (vii) retinoids. Data of cancer occurrence as well as age, sex and personal and family history of cancer were collected by questionnaire for all participants in 1986-88. Baseline values for education, marital history, smoking patterns and alcohol consumption were also used. The analytic method of choice was survival analysis: it provided a means of adjustment of age, sex and other important risk factors while testing for an association between colorectal cancer and nutrients of interest. Other procedures which gave additional information about the problem were stratified analysis, Pearson product moment correlation analysis and distributional analyses. There were 51 confirmed colorectal cancers in the food frequency cohort of 3453 participants and 31 in the recall cohort of 1921. The two datasets overlapped and 1837 persons had dietary data collected by both methods. Correlational analysis of this group for corresponding nutrient intake of both methods showed highly significant positive coefficients for alcohol intake and dietary calcium; the weakest coefficients were derived from the correlations of dietary fiber and dietary vitamin A. The major conclusions of study were that calcium, retinoids and fish were protective against the occurrence of colorectal cancer although only the retinoid-colorectal cancer association was significant. Fats and carotenoids showed no association with the occurrence of colorectal cancer in this study. Alcohol intake quartiles showed significant trends for risk; and fish and calcium, significant trends for protection.