Colonoscopic polypectomy and the incidence of colorectal cancer

It has long been believed that colorectal cancer evolves from a precursor lesion, the adenomatous polyp. This concept was based on the elegant pathology studies from the St Mark’s Hospital, London, published by LockhartMummery and Dukes in 1928, and culminating in the concept of the polyp-cancer sequence published in 1975 by Muto and colleagues. The introduction of colonoscopy in the early 1970s, followed by the demonstration of the feasibility of colonoscopic polypectomy, provided the technology for the application of this concept to clinical practice. The entire colon could be examined, polyps identified and removed and, it was believed, colorectal cancer prevented. Evidence for this belief was provided by observations from the prospective National Polyp Study (NPS) in the USA that demonstrated a reduction in the incidence of colorectal cancer by 76–90% following colonoscopic polypectomy. The paper in this issue of Gut by Citarda et al addressed a common question—that is, can eYcacy observed in a clinical trial be translated to eVectiveness in clinical practice (see page 812). The study was a retrospective analysis of the incidence of colorectal cancer following colonoscopic polypectomy performed at medical centres throughout Italy, with many endoscopists participating in routine clinical practice. The study provides further evidence of the robustness of the NPS observations and adds validity to the clinical practice of colonoscopic polypectomy for the prevention of colorectal cancer. The belief initially formulated in the St Mark’s studies has now been translated by evidence into clinical practice. Citarda et al add further fuel to the increasing sentiment that colonoscopy may be the best screening option for the prevention of colorectal cancer. Guidelines published recently have suggested faecal occult blood testing every year and flexible sigmoidoscopy every five years as one option; or screening colonoscopy every 10 years; or double contrast barium enema every 5–10 years along with flexible sigmoidoscopy. The radiological option has become less desirable considering the overall reduced interest in utilisation of the barium enema and the most recent observations of the NPS demonstrating that the double contrast barium enema, even when performed by experienced radiologists, has only a 50% sensitivity for the detection of adenomatous polyps, even those 1 cm in size or larger. Recent reports from the USA have demonstrated that screening colonoscopy is feasible, with a high completion rate to the caecum, a very low complication rate, and a manageable outcome of approximately 10% neoplasia with advanced pathology, including significant adenomas and cancer. 9 This rate of advanced pathology provides excellent stratification, separating those (90%) who may not need any further follow up from those that need further colonoscopy follow up. This gives us some indication of the resulting burden on the available medical resources. These observations, as well as the present study in Gut, strengthen the case for screening colonoscopy as a desirable option for asymptomatic general population screening. Do we need additional evidence to support screening colonoscopy which is rapidly becoming the preferred option based on available evidence? The evidence we now have is indirect and addresses feasibility. The studies of Lieberman and colleagues and Imperiale and colleagues of feasibility report neoplastic findings and performance of colonoscopy. The NPS and Citarda et al studies provide an estimate of the reduction in the incidence of colorectal cancer following colonoscopic polypectomy. However, these studies were in adenoma bearing cohorts as observational studies could overestimate the magnitude of the eVect in the general population. A randomised controlled trial of screening colonoscopy in the general population could answer the following questions: the magnitude of the reduction in incidence and mortality of colorectal cancer in the general population as a result of screening colonoscopy; complication rate of widespread screening colonoscopy; duration of the benefit; harms versus benefits; cost eVectiveness; acceptance of screening colonoscopy in the general population; and comparison of eYcacy of screening colonoscopy with that observed in randomised controlled trials of screening flexible sigmoidoscopy 12 and faecal occult blood testing. The age at which screening colonoscopy is to be oVered is an important consideration. If we adhere to present guidelines of screening colonoscopy every 10 years beginning at age 50, resources could be enormously overwhelmed. A properly timed “once in a lifetime” screening colonoscopy could provide an enormous impact while keeping the added burden on medical resources at a lower level. Why do we need a long term randomised trial of screening colonoscopy in the general population? Universal acceptance of screening colonoscopy by the public, medical community, and providers requires strong direct evidence of eVectiveness, data on benefit versus harms, data on cost eVectiveness and intervals between screening examinations, and follow up surveillance examinations in those with neoplastic findings. We need to have a full understanding of all aspects of this strategy to implement it optimally in the millions of people at risk. Issues on the eVectiveness of screening colonoscopy will arise when strong evidence of screening flexible sigmoidoscopy is reported over the next few years and added to the strong evidence of faecal occult blood testing. Questions will be asked regarding the “projected” higher eVectiveness of screening colonoscopy, and at what cost to the patient and to society in terms of harms, resources, and finances. We believe that screening colonoscopy is the preferred option for colorectal cancer screening but we must test out this belief. We reiterate the statement written in our guidelines paper published in 1997. “Because this strategy (screening colonoscopy) is potentially so eVective, there is an urgent need for better evidence of its eVects in prospective control trials.” This belief needs strong evidence to be translated into eVective universal implementation.