Endoscopy and Colorectal Cancer Mortality-Reply

Abstract

Many of the issues raised by de Saussure and Hadengue were addressed in a previous response to Weiss. 1 They are correct in pointing out that, by ignoring the colorectal procedure leading to the first diagnosis of cancer, we could have biased the results of our study toward a seemingly stronger protective influence of diagnostic procedures against death. As indicated previously, the theoretical possibility of a bias does not ascertain its existence, and its magnitude remains in question. In designing the study, it became obvious to us that this potential bias was unavoidable and that the study would not be feasible if procedures leading to the diagnosis of cancer were to be included in the comparison of cases and controls. Nevertheless, we find the arguments raised by de Saussure and Hadengue to be highly speculative. For instance, their hypothesis fails to explain why the protective influence of colorectal procedures pertained