Re: Smoking and Colorectal Cancer: A 20-Year Follow-up Study of Swedish Construction Workers

Extract

Intriguing data on the absence of excess risk of colonic cancer (relative risk [RR] = 0.98; 95% confidence interval [CI] = 0.82–1.17) and moderately excess risk of rectal cancer (RR = 1.16; 95% CO = 0.94–1.44) in Swedish male construction workers who smoke have been presented by Nyrén et al. ( 1 ). The data are discrepant with U.S. studies ( 2 – 4 ), and the authors state that the association found in those studies is likely to be artifactual.

In the study by Nyrén et al. ( 1 ), a detailed smoking history and exposure data on approximately 200 items were collected at the establishment of the cohort. The report fails to mention exposure to oral snuff, a tobacco product that is particularly relevant in Sweden. In some age groups, more than 30% of Swedish males have used snuff during the time period relevant to this study ( 5 ). The control group may thus be heavily contaminated with users of a highly carcinogenic tobacco product ( 6 ) that is shown to be associated with colorectal cancer ( 2 ). Furthermore, in the present study ( 1 ), RRs in the analyses of dose-effect relationships of smoking are flawed. Nonsmokers are not used as the reference group, but nonsmokers, former smokers (shown in a preceding analysis in the same report to have the greatest risk for rectal cancer), and smokers of other tobacco products are lumped together in one group.