Reply

Abstract

We thank Ausk and Dominitz for their insightful summary of and comment on our paper (Ann Intern Med 2011:154:22–30), and we agree with most of their conclusions. In particular, we agree that our study provides some evidence that colonoscopy offers significant benefit (in terms of protection from colorectal cancer [CRC]) in the proximal colon. To what extent the costs and risks of screening colonoscopy relative to sigmoidoscopy are justified cannot be judged based on our study alone, even though our study provides relevant data to inform careful, country-specific cost-effectiveness analyses required for such judgement (Gastrointest Endosc Clin N Am 2010;20:751–770), as well as evidence toward such justification. We also agree that randomized trials are the method of choice to ultimately establish efficacy of screening procedures. For screening colonoscopy, one such trial, the NORDICC study, aiming for inclusion of 66,000 participants, has recently been initiated in a number of European countries. However, results regarding primary endpoints, reduction of CRC incidence and mortality, will only be available 10–15 years after completion of recruitment. By then, colonoscopy technology and training will hopefully and most likely have made further major progress, and results will again pertain to technology that will most likely be outdated. Furthermore, in the population of the United States and some European countries, colonoscopy and sigmoidoscopy are already now so widespread for both screening and diagnostic purposes, and prevalence of previous lower gastrointestinal endoscopy has become so high in the target population of screening (Gastrointestinal Endoscopy 2011;73:435–443; Endoscopy 2010;42:546–556) that randomized trials in those populations would most likely be strongly affected by contamination of the control group. Also, although randomized trials are able to assess screening efficacy under highly standardized trial conditions, they are insufficient to ultimately evaluate effectiveness of a colonoscopy screening program in routine practice. Finally, randomized trials are typically able to evaluate a single, specific screening option (such as one first-time colonoscopy at the age of 55–65), and rarely permit comparative evaluation of different options, such as a different age at first-time colonoscopy, or different screening or surveillance intervals. Therefore, although randomized trials undoubtedly have their merits, they can and should not replace carefully conducted observational studies that evaluate the effectiveness of colonoscopy in routine practice, and that may help to inform recommendations and decisions on CRC screening offers. Such preliminary recommendations and decisions are clearly needed before the availability of results from recently commenced or future randomized trials on colonoscopy efficacy. Colonoscopy Prevents Colorectal Cancer in Both the Right and Left ColonGastroenterologyVol. 141Issue 1PreviewBrenner H, Chang-Claude J, Seiler CM, et al. (German Cancer Research Center, Heidelberg, Germany). Protection from colorectal cancer after colonoscopy. Ann Intern Med 2011;154:22–30. Full-Text PDF