APPENDIX F
Initial Analyses of Available Data Concerning Cancers of Colon and/or Rectum and Asbestos Exposure
After culling all the data available concerning either colon or rectal cancers and exposure to asbestos from the cohort and case-control studies, the committee conducted preliminary analyses to determine whether conclusions should or could be derived separately for the colon and for the rectum. As presented by the original researchers, there were three sets of findings that grouped themselves:
-
findings on just the colon,
-
findings on just the rectum, and
-
findings on colon and rectum combined.
The summary tables and plots on the following pages are the result of the initial analyses of these three datasets; unfortunately the graphic quality of these preliminary runs was poor.
For almost all the cohort studies reporting on either of these two sites individually, if data on one were given, the corresponding information on the other was also present. Furthermore, because most of the cohort studies presented the expected number of cancers at a site as well as the number observed, the statistics for the combined sites could accurately be recalculated. Since adjustment for confounders had generally not been performed for the cohort studies, this approach posed no compromise on the refinement of the analysis.
For the (fewer) case-control studies (all of which had analyses involving some adjustment for confounders), when the results were not presented in combined form, only one of the two sites had been studied. The committee was, therefore, not in a position to derive statistics for the combined statistics where they were not already available.
Colon cancer is more prevalent than rectal cancer, representing more than 70% of the cases when the two sites are combined. Therefore, colon cancer can be considered to dominate the calculations. The risk factors for cancers at these two segments of the intestinal tract are not well enough understood to be distinguished.
Having scanned the plots for evidence of systematic differences between the results for colon and rectal cancers and discussed the options for deriving a valid and useful conclusions, the committee agreed to conduct its meta-analyses of the epidemiologic data on colon and rectal cancers on datasets with the sites combined and to draw a single conclusion about causality for the combined sites (as stated in the Fairness in Asbestos Injury Resolution, or FAIR, legislation and the committee’s charge).
A. Results for colon cancer and asbestos exposure