medical organizations recommend treatment with NSAIDs or postmenopausal estrogen and progestin hormone replacement to prevent cancer because of potential side effects of NSAIDs and hormones. Women who take hormone-replacement therapy may be more likely to have colorectal cancer diagnosed at a more advanced stage.

The 1- and 5-year survival rates for persons with colorectal cancer are 83% and 63%, respectively. Survival continues to decline beyond 5 years to 57% at 10 years after diagnosis. When colorectal cancers are detected at an early, localized stage, the 5-year survival rate is 90%; however, only 39% of colorectal cancers are diagnosed at this stage, mostly because of low rates of screening. After the cancer has spread regionally to involve adjacent organs or lymph nodes, the 5-year survival rate drops to 67%. The 5-year survival rate for persons with distant metastases is 10%.

EPIDEMIOLOGIC EVIDENCE CONSIDERED

Results varied from study to study (for both cohort and case-control designs) in whether they were presented as colorectal or for colon and for rectum separately. In reviewing the available data on cancers of the colon and rectum in association with exposure to asbestos, the committee conducted three preliminary meta-analyses on the information presented separately for colon, for rectum, and for colorectal cancer as already combined by the original researchers. The plots and summary tables for those runs are presented in Appendix F. For the 15 cohort populations with individual results for colon and for rectum there did not appear to be any systematic difference, and their aggregate results were similar to those for the studies that had precombined their observations into a colorectal category. The results when the case-control results were considered in this fashion were so sparse that strong contrasts could not be drawn, but no major difference was apparent.

Condensing these three datasets into a single analysis would provide a better chance of amassing adequate information to reach a conclusion, and did not seem contraindicated by the screening runs. Furthermore, the legislation driving the committee’s charge specified colorectal cancer as a single endpoint. Age and sex are the primary known risk factors for rectal cancer, while colon cancer also appears more clearly associated with family history, physical inactivity, and several other factors such as body mass index and dietary and alcohol intake (Wei et al. 2004). Less ability to detect risk factors may itself be a function of limited statistical power arising from the fact that rectal cancer comprises only about 30% of tumors of the large intestine. Because these subsites are not clearly or consistently distinguished from each other on death certificates, colon and rectum cancer are frequently combined in analyses based on mortality, as was the case of a majority of



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement