Are proximal colorectal cancers always associated with distal adenomas?

Of all the common cancers, colorectal cancer is the best suited to prevention through screening as it is derived from benign adenomas which can be easily detected and removed. The best screening investigation remains much debated. Many argue that colonoscopy is superior to other techniques because it has the highest sensitivity (>90%) and examines the whole colon.¹ However, it has a number of important disadvantages. Firstly, it is potentially dangerous. Perforation rates of 1 in 1000–1 in 20 000 have been found in large studies from the USA and Germany.^2,3 Colonoscopy is also expensive and requires highly skilled operators who are in short supply.⁴ For these reasons investigators have sought a screening strategy that reduces the number of colonoscopies undertaken. A study from St Mark’s Hospital of the long term risk of colorectal cancer in patients with rectosigmoid adenomas found that 88% of cancers developed in patients with high risk (villous, tubulovillus histology, or >10 mm in diameter) rectosigmoid adenomas.⁵ This study led Atkin et al to propose that a single examination with a flexible sigmoidoscopy leading to full colonoscopy in patients with high risk rectosigmoid adenomas would be a cost effective and safe screening protocol.⁶ This strategy is now being tested in a randomised controlled clinical trial. Baseline findings have already established the perforation rate of diagnostic flexible sigmoidoscopy to be considerably lower than that of colonoscopy at 1 in 40 000.⁷

The crucial assumption for the use of flexible sigmoidoscopy as the initial screening test is that all proximal cancers are associated with distal adenomas. If this is not true then the findings at flexible sigmoidoscopy will not trigger the colonoscopy required to make the diagnosis of proximal colorectal cancer. This assumption is brought into sharper focus by the increasing proportion of colorectal cancers arising in the right colon.⁸ In this issue of Gut, Gondal and colleagues⁹ investigated the association of distal adenomas with proximal colorectal neoplasia [see page 398]. The investigators took advantage of the Norwegian Colorectal Cancer Prevention study (NORCCAP) in which 20 780 individuals aged 54–64 years, selected randomly from the population registry of Oslo and Telemark County, were offered a once only examination by flexible sigmoidoscopy or a combination of flexible sigmoidoscopy and faecal occult blood testing. Individuals diagnosed as having an adenoma of any size were offered full colonoscopy. The current study examined the risk of proximal adenomas and carcinomas in the 2154 individuals (17% of the total screenees) who were found to have distal neoplasms. Of these, 1833 individuals were studied. Twenty one per cent of subjects had colonic neoplasms proximal to the level reached by flexible sigmoidoscopy and a further 5% of subjects had proximal advanced neoplasms (PAN) defined as high risk adenomas or carcinomas. The risk of PAN increased threefold in subjects with distal adenoma >10 mm in diameter or containing villous components. The investigators then calculated the number of PAN that would have been missed depending on the threshold criteria for offering colonoscopy. If the threshold criteria for colonoscopy had been more than one adenoma or a single high risk adenoma (as defined by a diameter >10 mm or villous components or showing severe dysplasia) then 38% of PAN would have been missed, including 17% of proximal carcinomas. Furthermore, the tendency to miss PAN was found to increase with the age of the subject. On the other hand, implementation of these strict threshold criteria would have resulted in 66% fewer colonoscopies being undertaken.

A particularly interesting feature of the study was that colonoscopes were used to perform many of the flexible sigmoidoscopic examinations with the extent of examination limited by the degree of bowel cleansing from a single sorbitol enema. In this way the investigators were able to examine a greater proportion of the colon than is usually possible with a conventional flexible sigmoidoscope. This additional examination above that of conventional flexible sigmoidoscopy resulted in a further 3% of patients being offered full colonoscopy and three proximal carcinomas being detected in subjects who would have otherwise been misclassified as having no neoplasia. The authors conclude that the finding of any adenoma at flexible sigmoidoscopy should trigger a full colonoscopy. They recommend that the initial examination should be an unsedated examination with a colonoscope after a simple enema.

This study is consistent with previous findings that 46–52% of PAN are not accompanied by distal polyps.^10,11 Addition of faecal occult blood testing to flexible sigmoidoscopy does not significantly increase the detection of advanced neoplasia.¹² One is therefore left with the conclusion that colonoscopy remains the most sensitive screening tool and, if performed by a skilled operator, is reasonably safe. No screening technique will entirely eliminate the risk of colorectal cancer. Risk reduction is all that can be achieved and this must be carefully explained to patients. Flexible sigmoidoscopy is safer, cheaper, and more convenient for patients than colonoscopy but at the cost of lower efficacy for preventing and detecting cancer.