Proportion of colon cancer risk that might be preventable in a cohort of middle-aged US men

Objective: Diet and lifestyle likely play major roles in colon cancer incidence; however, the proportion of colon cancer risk that might be preventable is unknown. Thus, we estimated the proportion of colon cancer risk among men in the prospective Health Professionals Follow-up Study that might be attributable to a constellation of modifiable risk factors, and thus might be preventable. Methods: We included 47,927 men aged 40-75 years in 1986, among whom we confirmed 411 colon cancer cases from 1986 to 1996. Risk factors considered were obesity, physical inactivity, alcohol consumption, early adulthood cigarette smoking, red meat consumption, and low intake of folic acid from supplements. We calculated a risk score that was the sum across the six risk factors of the values of 1 (better exposure) to 5 (worse exposure) corresponding to the exposure category. We entered the risk score into a logistic regression model and estimated the population attributable risk percent (PAR%) using the method of Bruzzi et al. Results: After adjusting for age and family history of colorectal cancer and comparing the risk score for the combined six modifiable colon cancer risk factors at or above the approximate 20th, 10th, or 5th percentiles vs. below, the PAR% increased from 39% (95% confidence interval (CI) = 23-58%), to 48% (95% CI = 25-71%), to 55% (95% CI = 27-80%), respectively. Using a second method in which we used cut-points consistent with general-good health behaviors for each risk factor, comparing men with at least one risk factor to men without any risk factors (3.1% of the men), the PAR% was 71% (95% CI = 33-92%). Conclusion: The findings from this analysis suggest that, if all the members of this cohort of middle-aged US men had a modifiable exposure distribution comparable to the men with low risk scores, a large proportion of colon cancer risk might be avoidable. Additional study is required to determine whether making changes in these six risk factors now would reduce the risk of colorectal neoplasia, or whether the proportion of colon neoplasia that might be avoidable would be similar in populations with different characteristics.