Developing the Case for Population Screening for Bowel Cancer

Is Bowel Cancer an Important Health Problem?
In Australia, bowel cancer is indisputably an important health problem, being the most common internal cancer and the second most common cause of death from cancer, after lung cancer. In 1996, there were 10,988 new cases and 4,606 deaths from the disease, and a total of 31,000 person-years of life were lost before age 75.  ¹⁷   Direct health costs of treatment were AUS$ 205 million in the financial year 1993/94. The number of deaths from bowel cancer in 1996 was substantially higher than for either prostate cancer (n=2,644) or breast cancer (n=2,640).

Is there an Accepted Treatment Available?
Bowel cancer is treated by surgical resection of the cancer, adjacent large bowel and regional lymph nodes. On occasions, early stage polypoid cancers are treated by colonoscopic polypectomy alone. The pathological stage correlates closely with survival, early, stage A cancers having a 90% cancer-specific five-year survival.  ¹⁸   A feature of bowel cancer is that, in many cases, there are no symptoms until the cancer has already reached an advanced (incurable) pathological stage – hence the interest in and importance of pre-symptomatic diagnosis.

Most high incidence countries have sufficient facilities for surgical treatment of bowel cancer but many have insufficient resources – trained personnel and equipment – to provide colonoscopic back-up for a national FOBT-based screening programme. However, with over 900 accredited colonoscopists, Australia is considered to have sufficient resources in colonoscopy to support population-based screening.  ¹⁹  

Is there a Suitable Screening Test?
Faecal occult blood testing has been shown to reduce bowel cancer mortality in three randomised controlled trials (RCTs), providing Level 1 evidence for effectiveness. All three trials used the guaiac test, Hemoccult. In addition, there is supportive evidence coming from progress results of a fourth, smaller RCT and a non-randomised controlled trial and from several case-control studies. The first RCT was from Minnesota and was performed on volunteers. It revealed a 33% reduction in mortality with annual screening.  ¹   Initially, that trial failed to show a mortality benefit with biennial testing (see Table below) but mortality fell by 21% with longer follow-up.  ²   The other two RCTs were population-based and were performed in Nottingham, UK  ³   and Funen, Denmark.  ⁴   Mortality was reduced by 15% and 18% respectively, on an intention-to-treat basis. For those complying with testing, mortality was calculated to fall by 40%.  ⁴  

TABLE: Mortality Reduction in Randomised Controlled Trials Based on Faecal Occult Blood Testing

	Minnesota USA (1)	Nottingham UK (3)	Funen Denmark (4)
Participants	Volunteers	Population-based	Population-based
Numbers enrolled	46,551	150,251	61,933
Type of FOBT	Hemoccult	Hemoccult	Hemoccult
Development	Hydrated	Standard	Standard
Frequency	Group 1: 12mo Group 2: 24mo	24mo	24mo
Mortality reduction	Group 1: 0.67 Group 2: 0.94	0.85	0.82
Odds ratio (+/- 95%CI)	(0.50-0.87) (0.68-1.31)	(0.74-0.98)	(0.68-0.99)

When planning screening, choice of test will be an important issue. Factors affecting choice include test performance characteristics, need for special diet, cost of the test, and ability to automate test results. Test specificity is especially important. If a test has too low a specificity, ie it produces too many false-positive results, colonoscopy resources could be overwhelmed by the resulting demand for follow-up investigation and loss of confidence in the value of the programme could adversely affect participation. Low test specificity would also reduce cost-effectiveness.

The UK and Danish trials used the guaiac test, Hemoccult, developed without re-hydration. That test has a low sensitivity for cancer but has high (98–99%) specificity. Although testing missed 40–50% of early cancers in those trials, high specificity meant that only 4% of the screened population required colonoscopy over five rounds (ie, 10 years) of screening.  ^{3  ,   4}   While newer, more sensitive guaiac tests and immunochemical tests specific for human haemoglobin have obvious advantages, a test with an acceptably high specificity should be chosen.

Evaluation of screening by flexible sigmoidoscopy is in progress. Case control studies show evidence of benefit  ²⁰  ,   ²¹   but results of the RCTs needed to provide high level evidence are several years away.

Is the Test Acceptable to the Population?
Whatever method(s) are used for screening – FOBT or flexible sigmoidoscopy, alone or in combination, or a programme based on colonoscopy – the programme will only be successful if people are willing to undergo testing. In the European RCTs, test kits accompanied by a letter of invitation signed by the subject’s GP were distributed by mail. Although there could be no media publicity, encouragingly, 59.6% in the UK trial and 67% in both the Danish and the smaller Swedish trials completed tests in at least some rounds of testing. In the Danish trial, almost all eligible participants who completed the first round testing complied with testing in later rounds.  ⁴  

In Australia, initial recruitment is likely to be slow because of the little media attention paid to bowel cancer and the lack of public discussion about screening. Confusion and uncertainty among health care professionals add to potential problems. It will be essential to mount educational programmes about bowel cancer screening for medical practitioners and the target population as well as conducting studies on how to motivate them to promote or undergo screening. Even so, the high participation rates achieved in the European RCTs provide grounds for optimism.

Is the Natural History Adequately Understood and Who Should be Screened?
The natural history of bowel cancer has been well studied, colonoscopic biopsy and polypectomy specimens providing a remarkable opportunity to study pre-cancerous events. Most cancers arise from adenomatous polyps (adenomas), the likelihood of progression to cancer being greater in so-called advanced adenomas (10 mm or more in diameter or with villous features or severe dysplasia). It should be recognised that most adenomas never progress to cancer. This applies particularly to small tubular adenomas, which are a common colonoscopic finding in older people. When adenomas do progress to cancer, the average length of time from early adenoma to advanced cancer is about 10 years. The window of opportunity for detecting bleeding from an advanced adenoma or an early stage, curable cancer appears to be several years, at least in many cases. A wide window of opportunity means that a falsely negative FOBT on the first round of screening could be acceptable, provided that the test is positive on the next round and the lesion is still localised and curable. Discussion about programme sensitivity of particular FOBTs refer to this concept.

One as yet unresolved question relates to other possible pathways to cancer. In Japan, there have been many reports of small cancers arising either de novo or from flat or depressed adenomas.  ²²   Their time course appears to be short – of the order of one or two years. This type of cancer is often found in the familial form of cancer known as HNPCC (hereditary non-polyposis colorectal cancer). With that exception, fast track cancers are rarely recognised in Western countries. However, a recent report from UK suggests that colonoscopists may be missing these lesions.  ²³   The Japanese view is that flat and depressed lesions are responsible for about 15% of all bowel cancer.  ²²  

Who should be screened? In the USA, the early experience with screening showed disappointing results when people in their forties were included in the target population. Participation rates with both FOBT and sigmoidoscopy were low, as was the yield of significant lesions in those who did participate. For those reasons, the American Cancer Society changed its recommended starting age for those at average risk from 40 years to 50 years, retaining a starting age of 40 years for those who have one or two close relatives with bowel cancer.  ²⁴   International guidelines adopted the starting age of 50 years for those at average risk but, in Australia, the NH&MRC has recommended also starting at 50 years for most of those with one or two affected relatives, as their risk is only slightly above average.  ²⁵  

An unresolved question is when to discontinue screening.  ¹⁰   Clearly, the state of a person’s health will influence the decision. In general, it appears to be appropriate to restrict active promotion of screening to those aged up to age 75 years, but to include those over that age who request screening, provided that they are in good health.

Is there an Economically Balanced Case for Screening?
Many studies on cost-effectiveness of screening have been conducted. The studies have taken into account the current high costs of treating colorectal cancer, the cost of chemotherapy, radiotherapy and palliation making advanced bowel cancer more expensive to treat than early stage cancer.

One method of expressing results is to compare the cost-effectiveness of bowel cancer screening with that of screening for breast cancer and cervical cancer. Several studies have shown that FOBT-based screening compares well with that of mammographic screening for breast cancer. In the UK and Danish RCTs, cost-effectiveness was in the range of US$5,400–$9,650 per life year saved.  ²⁶   In the UK trial, it was concluded that cost-effectiveness of screening with Hemoccult was the same as that of mammographic screening in the short term and likely to be better in the long term.  ²⁶

Should Screening be a Continuing Process?
When people enter a cancer screening programme, they expect their life-time risk for that cancer to be reduced substantially. The approach with FOBT is to offer testing on a regular basis, either yearly or two-yearly, proceeding to full evaluation of the large bowel if testing should ever show occult bleeding. Participants should be made aware of the importance of continuing with screening as risk for cancer rises with advancing age.

While sigmoidoscopic and colonoscopic-based screening is offered on a regular (eg, 5-yearly) basis to groups at increased risk, a different endoscopic approach has been suggested for those at average risk. The suggestion is that a once-only examination around the age of 55 years could enable triage into high risk and low risk groups.  ²⁷   Those with normal findings would then be advised against having further screening. Trials examining this approach are in progress. However, as yet, there is no convincing evidence to support the concept.

Clearly, many of the conditions listed above should be met before embarking on a national screening programme. Information about some factors has to be obtained through local experience, especially those factors such as willingness to participate in screening, impact of screening on colonoscopic services and cost-effectiveness of the overall programme. These can be studied in pilot programmes, with modification of plans for a national programme in response to any problems identified.

How Were Decisions About Bowel Cancer Screening Made in Australia?

After release of the AGI/Australian Cancer Society guidelines on bowel cancer screening, a deputation from the Australian Cancer Network met with representatives of the Federal Health Department in 1994 to discuss views on screening at a national level. As a result, the Federal Health Minister established a working party under the direction of the Australian Health Technology Advisory Committee to review the subject. Members of the working party had expertise in gastroenterology, surgery, radiology, pathology, general practice, epidemiology and health economics; other members included a consumer representative and representatives of the NH&MRC.

Advertisements in the national press and invitations to medical colleges and societies called for submissions, a total of 40 submissions being received. Consultancy groups were responsible for producing a comprehensive literature review and a review of coverage of bowel cancer by the media. The working party met over the course of two years, during which time the results of the UK and Danish RCTs were published. The final report was completed in December 1997 and released in March 1998.  ⁵  

The AHTAC Report covered the needs of groups with above-average risk of bowel cancer, including members of families with inherited cancer syndromes. However, most of the recommendations applied to population screening. The main recommendation was for the introduction of population screening by faecal occult blood testing for the average risk population (the well population aged over 50 years) but commencing with pilot and feasibility studies to determine the most effective means of implementing a nation-wide programme.

In 1998, the Australian Cancer Network and the Clinical Oncological Society of Australia established a committee to develop best practice guidelines for prevention, early detection and management of bowel cancer. The process followed the NH&MRC guidelines for the development of guidelines. Over 90 experts contributed to various working parties, a conference was held to discuss the first draft report, later drafts were made available for public comment and the committee received a total of 250 submissions. The final document, which confirmed the AHTAC recommendations about screening, was then approved and given official endorsement by NH&MRC.  ²⁵  

The most recent development occurred in May 2000. In the Federal budget for 2000/2001, it was announced that funding will be provided for the pilot studies proposed in the AHTAC Report. The proposal, drawn up by the National Cancer Control Initiative, involves three studies, two of which will be conducted in urban areas and one in a rural area. A total of 50,000 people aged between 50 and 75 years will be invited to have FOBT-based screening and the studies will be closely monitored to address questions raised in the AHTAC Report. One question relates to the method of delivery of a national programme, including the role of general practitioners and their interaction with a central body responsible for quality assurance and measurement of outcome.