Article Text


Population based endoscopic screening for colorectal cancer
  1. D A L Macafee,
  2. J H Scholefield
  1. Division of GI Surgery, University Hospital, Nottingham, UK
  1. Correspondence to:
    D A L Macafee, Queen’s Medical Centre, Derby Rd, Nottingham, UK;

Statistics from


Colorectal cancer is the second commonest cause of cancer death in the Western world and the median five year survival from this terrible disease remains at approximately 40%. In the UK approximately one third of cases still present as emergencies with intestinal obstruction or perforation

One method of reducing the mortality of this disease is through earlier diagnosis but the symptoms and signs of bowel cancer are often non-specific and therefore earlier diagnosis is unlikely to occur through increased awareness or patient education alone.

Colorectal cancer is an ideal disease for population screening: it is common, has a well recognised premalignant precursor lesion (the adenoma), and treatment of the premalignant condition reduces the risk of cancer.1 Several case control studies and four randomised trials have shown that population screening for colorectal cancer reduces disease specific mortality. One large trial has also shown that population screening can reduce the incidence of this disease.2 Overall, the cost effectiveness of screening for colorectal cancer compares favourably with other cancer screening strategies (that is, breast and cervical).3

The case for screening for colorectal cancer is increasingly compelling and with the launch of the National Pilots in Scotland and Warwickshire the politicians appeared convinced too! Over 12 months later, with the pilots nearing completion and preliminary data from the MRC Flexible Sigmoidoscopy study, another difficult decision looms.


The debate over which screening test to use encapsulates some of the major issues in the provision of healthcare in the 21st century (especially in the National Health Service). Major technological advances are occurring at an increasing rate, but manpower and finance for today’s technology are severely limited.

Although there are many potential screening tests for colorectal cancer, ones which are serious contenders for population screening are stool based tests or endoscopy. Of these, only faecal occult blood (FOB) testing has been thoroughly evaluated. FOB testing has been used in at least five case control studies4–8 and four large randomised trials.9–12 Endoscopic screening shows promise, but so far no endoscopic screening study has shown a reduction in mortality or incidence in a randomised population based setting. Can a national screening programme using endoscopy be launched without such data?

The pros and cons of FOB testing and endoscopic screening are summarised in table 1.

Table 1

Pros and cons of the three main screening modalities


Endoscopic screening is essentially a choice between colonoscopy (examining the whole colon) and flexible sigmoidoscopy (examining the left half of the colon). Colonoscopic screening is good for high risk groups, such as those patients with hereditary non-polyposis coli and adenomatous polyposis coli. Its ability to undertake full colonic examination and simultaneous polypectomy is a major benefit, making it both a diagnostic and therapeutic tool. However, these abilities and benefits do not necessarily make it the most appropriate population based screening tool. Population screening by colonoscopy would be very expensive both in terms of direct costs and the complications such a programme would undoubtedly yield.

In the UK we have neither the manpower nor the necessary facilities to undertake colonoscopic screening for the average risk population. Population screening by colonoscopy would require a doubling of our endoscopy facilities and manpower. In addition, the standard of colonoscopy in the UK appears to be very variable and although steps are being taken to address this, we are a long way from providing a uniformly high quality diagnostic colonoscopy service.13

The complications which would arise from screening using colonoscopy would be unacceptable. Our calculations, based on published complication rates, suggest that colonoscopic screening of the UK population at age 60 years would probably lead to over 500 severe haemorrhages, over 150 perforations, and 50 deaths each year (table 2). Complications on this scale would rapidly lead to failure of the screening programme.

Table 2

Complications of colonoscopy and flexible sigmoidoscopy

Taking all this into account we believe that population screening by colonoscopy is a non-starter in the UK for the foreseeable future.

By contrast, the US literature is increasingly supporting colonoscopic population screening.14–16 Screening on a private basis (albeit through insurance companies) may skew the data, by directing resources at the “worried well”. Nevertheless, several case control studies have shown a reduction in colorectal cancer incidence and mortality following polypectomy.17,18 Testing in a randomised controlled setting is still awaited.

While endoscopic screening by colonoscopy may be a non-starter in the UK, flexible sigmoidoscopy (FS) screening has some appeal although the evidence to confirm its ability to reduce the mortality of the disease is still some way off.

The large MRC multicentre trial of FS19 has recently completed recruitment. Early data have shown that population screening by this modality is safe and yields sufficient polyps and cancers to make it a worthwhile screening tool. However, data on the effect on mortality in this trial are not due until 2005.

A conceptual difficulty with FS screening is that it is analogous to mammographic screening of only one breast! A small proportion of screened patients will have lesions beyond the reach of the FS (resulting in “missed lesions”). While the advocates of FS screening minimise this risk, citing “marker polyps” as their salvation, it remains a medicolegal minefield if such a screening modality were introduced. Evidence of a shift in the site of colorectal cancer and an increase in the proportion of right sided and transverse colon lesions only fuels this anxiety.20–22

The facilities and manpower investment required for population screening by FS are substantial and they raise three other issues:

(i) In order to minimise the cost of introducing FS screening, “a one off” FS examination is proposed to be offered to the population at age 50 years. This is in contrast with an annual or biennial FOB test (likely to be offered to the population from ages 50 to 70 years). The “one off” strategy has several additional merits over and above making the cost of the two programmes comparable:

– (a) FS is more sensitive than FOB for cancers and polyps (table 1).

– (b) In an average risk population, a polyp probably takes 10–20 years to grow to a size where malignant transformation becomes a possibility.

However, introducing such a “one off” programme is likely to cause outrage among those who fall outside the age cut off at the time of commencement. If the “one off” age is 50 years, there may be a discontented cohort of several million who just missed out. A similar ageist policy caused problems in the breast screening programme.

(ii) The second major issue with FS is the proportion of patients who need referral for colonoscopy because of small polyps found on FS. The initial MRC study data found a 5% referral rate for colonoscopy. This is 2.5× the rate in the Nottingham FOB trial where a 2% colonoscopy referral rate required two extra colonoscopy lists per week. Extrapolating the published FS data,19 up to five extra colonoscopy lists per week in each hospital would be needed to meet the demand from screening patients. Even if one assumes funding will be available, there are too few endoscopists and endoscopy nurses to meet such demands. Furthermore, rolling out FS screening from trial to clinical practice will almost certainly necessitate nurses and technicians performing these examinations and it is widely recognised that nurses are more scrupulous in detecting small polyps than doctors, so potentially pushing the colonoscopy rates up even further. The cut off point at which colonoscopy becomes necessary is therefore a crucial issue, and one which remains to be resolved.23

(iii) Finally, a further worry about FS screening is the issue of compliance. While in the MRC study compliance was cited at approximately 60%, overall compliance was only 40%. Investigators enriched their study group by randomising only those who expressed interest in the trial (response to postal questionnaire). Only two thirds of those who expressed interest and were randomised to FS actually turned up and underwent the investigation. FOB screening trials achieved population compliance rates of 35–50% in the general population24 and >60% in clinical trials.25


All four randomised population based studies showing that an FOB testing programme saved lives used a guaiac based test (usually Haemoccult). Haemoccult, with its specificity of 98% and sensitivity of only 50% (table 1), could lead to a label of “old technology” but it is tried and tested. Indeed, the current National Pilots in Warwickshire and Scotland are using this type of test.FOB testing is painless, relatively inexpensive, and has no complications of its own. It also complies well with the WHO criteria of a screening tool: “a screening test should be inexpensive, rapid, and simple, and is not intended to be diagnostic; those with positive tests require further evaluation”.26 When using Haemoccult or a similar test, only 2% of the screened population would need further evaluation (normally by colonoscopy).11 This modality therefore places much less strain on under resourced endoscopy units than population screening by endoscopic means. Of those 2% who need endoscopy, approximately 40% will have pathology identified; mostly adenomas, some carcinomas, and a few incidental findings such as inflammatory bowel disease.More modern stool based tests are immunologically based (for example, Haemsensa) and have a better sensitivity than their guaiac based counterparts.27 These tests can also have an element of controlled sensitivity, using automated test reading devices, which is attractive. Even more promising, but some way off, are the stool based gene array tests, identifying mutations in the APC gene and others.28 Multi-target assay panels, with the capability to identify many point mutations, have been successful on freezer archived stools.29Further support for an FOB testing programme comes from the fact that we know we can do it and we know it works. The current National Pilots are using the tried and tested FOB test (Haemoccult), and the health economic data show that it will be as cost effective as either breast or cervical screening—costing approximately $13 500 per added year of life.30 If better tests come along in the future, the programme would provide an excellent test bed for their use and, if successful, they could easily be rolled out into the programme when appropriate.


As mentioned above, “once only” FS has its merits but may cause resentment among those who are above the age at inception. Combining the “one off” FS screen at age 50 years with FOB screening for 60–70 year olds may be politically expedient and from the data available scientifically sensible. The expense of combined modality screening would be more than either FS or FOB screening alone but combining the tests may yield more right sided lesions and may be worth investigating in a trial setting.


Unless a screening programme is accepted and uptake by the population is over 50%, the benefits of screening for the population are likely to be outweighed by the costs of implementing the programme.

Screening for bowel cancer has been handicapped by the taboo surrounding anything to do with bowels in general. Neither the FOB tests nor FS are pleasant tests, FOB requires the individual to do the test whereas FS allows the individual to be more passive but the test is more invasive (table 1).

Compliance with FOB may be slightly better than for FS. If a National Programme was introduced, compliance rates for either could be expected to increase as the public become more educated and aware of colorectal cancer.


Screening for bowel cancer works—it reduces the mortality from this awful disease and costs about the same as established screening programmes for breast and cervical cancer. If we are serious about improving outcome for colorectal cancer, screening for this disease ought to be high on our list of priorities.

At the present time population screening for colorectal cancer by FOB testing is the only modality which has been shown to be safe and effective. We should use the National Pilots as a model from which to roll out a nationwide programme, starting by introducing an FOB screening programme and reviewing the technology used as more evidence becomes available. Some centres could be used to pilot new screening technologies and different screening regimens such as “once only” FS and combined FS/FOB regimens.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles