Intended for healthcare professionals


New polyps, old tricks: controversy about removing benign bowel lesions

BMJ 2013; 347 doi: (Published 08 October 2013) Cite this as: BMJ 2013;347:f5843
  1. Geir Hoff, professor12,
  2. Michael Bretthauer, professor 2345,
  3. Kjetil Garborg, consultant4,
  4. Tor Jac Eide, professor2
  1. 1Telemark Hospital, Skien, Norway
  2. 2University of Oslo, Norway
  3. 3Oslo University Hospital, Oslo, Norway
  4. 4Sørlandet Hospital, Kristiansand, Norway
  5. 5Harvard School of Public Health, Boston, USA
  1. Correspondence to: G Hoff hofg{at}

Colorectal cancer screening programmes have increased the number of benign lesions being detected. Geir Hoff and colleagues argue that we need more evidence about their malignant potential to be sure that the risks of following current recommendations for removal do not outweigh the benefits of screening

Screening programmes for colorectal cancer aim to detect early colorectal cancer and benign lesions with malignant potential, such as adenomas, before symptoms develop. Screening has been shown to reduce mortality from colorectal cancer by 16% when faecal occult blood testing is used1 and by 28% with flexible sigmoidoscopy.2 Colonoscopy is the standard follow-up examination after positive screening using either flexible sigmoidoscopy or faecal occult blood tests. This has increased the demand for colonoscopy services internationally to provide definitive, detailed assessment of the colon and often to remove benign lesions.

Although removal of cancerous lesions is uncontroversial, there is debate and uncertainty about how to manage benign lesions. Expansion of screening programmes and the development of new technology that can detect inconspicuous polyps of unknown clinical importance increase the need for clear evidence on the natural course of polyps and their management. This is particularly important for serrated polyps, which are less understood and often more risky to remove than adenomas. Recent guidelines recommend removal of all serrated lesions except those in the sigmoid colon or rectum that are <5 mm in diameter,3 but evidence that the risks of removal are outweighed by the benefits is lacking.

Not all polyps are equal

Polyps can be classified to help predict their malignant potential. Although some features can be identified macroscopically—for example, pedunculated (stalked) polyps can be differentiated from sessile (flat) polyps—definitive classification requires microscopic examination of histology and cellular dysplasia. This means that the polyp must be removed.


The most common polyps found on screening are adenomas. Although adenomas may occur throughout the colorectum, a large proportion are found in the distal colon and they are often pedunculated, which facilitates complete removal in one piece with low risk to the patient. Adenomas measuring ≥10 mm are thought to be high risk. Other factors that increase the risk of malignant transformation are villous components (villous and tubulovillous adenomas) and the presence of high grade dysplasia. Small adenomas have a low malignant transformation rate, particularly those <5 mm in diameter. Nevertheless, because they are common they are still believed to contribute to the burden of colorectal cancer.

The adenoma detection rate is defined as the proportion of people screened in whom at least one adenoma is detected and varies widely depending on the endoscopist and equipment used. The prevalence of adenomas in the target population for colorectal cancer screening may be as high as 50% (fig 1). High risk adenomas are found at colonoscopy in 4-10% of the average risk population4 and 5% of 55-64 year old healthy individuals screened by flexible sigmoidoscopy.5 6 7


Fig 1 Prevalence of adenomas and proximal serrated polyps in the 55-64 year old healthy population far exceeds the average cumulative lifetime risk of colorectal cancer (CRC)

Guidelines suggest that endoscopists remove all adenomas. However, since epidemiological data show that less than 5% of adenomas develop into colorectal cancer,8 95% of polypectomies may be exposing patients to unnecessary risks.

Serrated polyps

Until recently, adenomas were considered to be the only type of polyp with potential to develop into colorectal cancer but attention has now fallen on serrated polyps. Serrated polyps are less well understood than adenomas. They were all previously classified as hyperplastic polyps and presumed to be harmless. In recent years, however, subtypes of serrated polyps have been described, and there has been a growing suggestion of a serrated neoplasia pathway. Hyperplastic polyps are thought to transform into sessile serrated polyps (a possibly premalignant subtype of serrated polyps) and colorectal cancer.4 The small hyperplastic polyps often found in the distal colon and rectum are still considered innocuous. Sessile serrated polyps are usually larger and located in the proximal colon. Case reports have described colorectal cancer within the borders of some of these lesions and molecular characteristics that are similar to those found in established cancer.9 10 3 Although these reports provide circumstantial evidence of the malignant potential of sessile serrated polyps, they tell us little about the rate of transformation. In one study, the association between colorectal cancer and serrated polyps ≥10 mm in diameter was not much stronger than the association between cancer and age ≥65 years (adjusted odds ratio 3.34 and 2.63, respectively).11

Understanding of the link between serrated polyps and colorectal cancer has been hampered by disagreement about how to define such polyps.12 In 2010, the World Health Organization launched a standard modified classification of serrated lesions of the large bowel, which may improve reproducibility among pathologists in diagnosing the spectrum of serrated lesions.13 This classification is also supported by evolving data on genetic changes, which further confirms different pathways in colorectal carcinogenesis.14

Serrated polyps are more difficult to detect than adenomas during colonoscopy. They are often pale pink and covered by colonic mucus, making them hard to distinguish from the surrounding normal bowel mucosa (fig 2). The broad based, non-pedunculated structure of sessile polyps also makes them difficult to remove. The bowel wall is thinner in the proximal colon than in the distal colon and the rectum, and their broad base often necessitates resection in multiple pieces. Consequently, the risk of perforation or major bleeding is greater for removal of proximal sessile lesions than distal stalked lesions. Endoscopic removal of proximal sessile polyps ≥20 mm is reported to be associated with an 11.7% risk of major complications; this compares with 5.3% for removing polyps with the same morphology in the distal colon or rectum.15 Risks for polyps of 10-19 mm were 3.5% and 0.9%, respectively.


Fig 2 Large (4 cm) sessile serrated polyp in caecal pole detected during screening. The polyp is barely visible with white light even with high definition image quality (left) but the margin becomes visible using narrow band imaging enhancement technology (right)

Operating in uncertainty

Better understanding of serrated polyps is needed. Screening for colorectal cancer is already increasing their diagnosis, and the introduction of new high definition endoscopy with modern enhancement techniques looks set to raise it further. One study detected proximal serrated polyps in up to 20% of the population.16

In the absence of evidence over how to act, consensus based recommendations have suggested that all serrated polyps should be removed at endoscopy except for small sigmoid or rectal lesions.3 This is done by first lifting the lesion by submucosal injection of fluid to separate it from the deeper layers of the bowel wall and then applying a diathermy snare around the elevated lesion for removal. In addition, patients with sessile serrated polyps will be recalled for repeat colonoscopy every 1-5 years.3

Missed opportunities

The risk is that by adopting consensus guidelines the medical profession misses out on the opportunity to learn about the natural course of serrated polyps and try different management options. With more polyps being detected as a result of screening, we need to be able to quantify the gains and harms and share this information with screening participants. It is not enough simply to share a belief that we are doing good, partly motivated by fear of not doing enough.

Experience with adenomas shows the dangers of this approach. When flexible endoscopy was first used (in the 1970s and 1980s), guidelines recommended short surveillance intervals after removal of adenomas by polypectomy (3-6 monthly at some centres). Later, surveillance intervals gradually stretched out to 3-5 yearly, 10 yearly, or never for low risk patients.17 18 19 20 21 The result of earlier frequent surveillance strategies is that, more than 120 years after the malignant transformation of adenomas was described,22 we still do not fully understand the natural course of adenomas or know what is the best surveillance strategy after polypectomy. Consensus based fear of missing cancer through inadequate surveillance intervals in the 1980s restricted research that could have given us definitive answers.23 24 25

Although we may need guidelines for management of serrated polyps for medicolegal reasons, we should not allow them to delay or prevent the generation of further evidence, particularly through randomised trials. If we are to avoid risking harm to patients and wasting resources, consensus guidelines should also provide guidance on future research—explicitly stating areas of uncertainty and highlighting the priorities. For serrated polyps in the proximal colon, we suggest randomised trials of polypectomy and different intervals of surveillance compared with watchful waiting, serial biopsies or “optical biopsy,” 26 and close surveillance. This could be done within the framework of screening programmes as comparative effectiveness research.27

The balance and quantification of harms and benefits are more important when screening a healthy population than when monitoring or testing patients with symptoms or at high risk of disease. The change in guidelines towards an aggressive intervention strategy for serrated polyps may tilt the balance against screening if the improvements in mortality and incidence of colorectal cancer cannot be shown to outweigh the harms of intervention. Because the risks of removing proximal serrated polyps are higher than for removal of adenomas, the evidence for intervention must be stronger than that we accepted for adenomas in the 1980s.

Key messages

  • Bowel cancer screening has increased the detection of benign polyps

  • Concern about malignancy, which had largely been limited to adenomas, has now been extended to sessile serrated polyps

  • Sessile serrated polyps are less understood than adenomas and more risky to remove

  • Current consensus guidelines recommend removal of serrated polyps >5 mm

  • These guidelines should not be allowed to impair essential research to determine the malignant potential and best management of serrated polyps

  • The higher risks associated with their removal may upset the balance of benefit and harm in screening programmes


Cite this as: BMJ 2013;347:f5843


  • Contributors and sources: GH and MB have studied and reported widely on colorectal cancer screening and quality improvement in gastrointestinal endoscopy. KKG has studied and reported on endoscopy technique and technology and is particularly engaged in challenges related to serrated polyps. TJE has studied and reported widely on colorectal polyps and cancer. GH had the idea and drafted the first version of the paper. MB, KKG, and TJE revised the manuscript.

  • Competing interests: All authors have read and understood the BMJ Group policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Not commissioned; externally peer reviewed.


View Abstract