Faecal occult blood testing for colorectal cancer

C. S. McArdle,+

University Department of Surgery, Royal Infirmary, Glasgow, UK

Received 9 August 2001; accepted 10 October 2001.

Key words: colorectal cancer, faecal occult blood, screening

Rationale for screening

Colorectal cancer is the second most common cause of death from cancer in the western world. Each year there are almost 800000 new cases and over 400000 deaths attributable to the disease worldwide. In the UK alone, there are more than 32000 new cases and over 17000 deaths each year [1]. The incidence has increased by ~1% per year over the last 25 years in men, but rates in women have hardly changed. Recently, mortality has fallen (Table 1). For example, between 1971–75 and 1986–90, 5-year relative survival for colorectal cancer in England and Wales increased from approximately 27% to 39% [2]. Similar increases in 5 year relative survival have been observed in Scotland [3]. It is not clear from these registry-based data whether the improvement in survival is due to earlier presentation or improvements in treatment.


View this table:
[in this window]
[in a new window]
 
Table 1. 5 year relative survival: England and Wales
 
However, more detailed analyses of patients presenting to a single centre suggest that both factors have contributed to the improvement in survival. Between 1974 and 1979, approximately one-third of patients presenting with colorectal cancer did so as an emergency, and at the time of presentation, over one-quarter had evidence of distant metastases [4]. The overall resectability rate was ~70%, curative resection being obtained in just over half the patients. Overall mortality was ~13%. Less than half the patients undergoing apparently curative resection survived 5 years. These results were typical of those achieved in non-specialist centres at that time.

A subsequent audit of patients presenting between 1991 and 1994 showed that although approximately one-third still presented as an emergency and one in five had evidence of metastatic disease, fewer patients had Dukes’ C tumours. In contrast to the earlier period, 85% of patients underwent resection, of which 61% were considered to be curative. Overall post-operative mortality had fallen to ~9%. Almost 60% of those undergoing curative resection survived 5 years. It would therefore appear that the overall improvement in survival following surgery for colorectal cancer was due both to earlier presentation and to improvements in treatment, including improved surgical technique and better post-operative care.

Nevertheless, the overall outlook for patients with colorectal cancer remains poor. Although it is likely that increased specialisation, better quality surgery and the increased use of adjuvant therapy will continue to impact on survival, a substantial proportion of patients will continue to present as an emergency or with evidence of advanced disease.

However, current attempts to identify the disease at an earlier stage in symptomatic patients is unlikely to be productive, since even in those patients who undergo apparently curative resection, up to one-quarter die as a result of their cancer within 5 years. Previous studies have shown that many of these patients harbour ‘occult’ hepatic metastases at the time of surgery, and it is the presence or absence of these occult metastases that predominately determines death from disseminated disease [5]. Moreover, studies of the rate of growth of these metastases suggest that the metastases occur early in the development of colorectal cancer and have been present for many months, if not years, prior to the onset of symptoms [6]. Therefore, although earlier detection of symptomatic disease and further improvements in treatment may contribute to a further increase in survival, clearly substantial reductions in mortality will only be achieved if the disease is prevented or detected at an earlier, asymptomatic stage.

It is now generally accepted that most, if not all, adenocarcinomas of the large bowel arise from pre-existing adenomas. It is known from observational studies that the prevalence of adenomas increases markedly after the age of 50 to a plateau at around 60 years. The risk of colorectal cancer also increases with age, with an approximate doubling of the incidence in each decade from 40 to 80 years of age. It has been estimated that the lifetime risk of developing colorectal cancer is 1:50.

Patients at high risk of developing colorectal cancer include those with familial adenomatous polyposis and those with hereditary non-polyposis colorectal cancer, which account for <1% and 5–10% of colorectal cancers, respectively.

However, the vast majority (~90%) of colorectal cancers occur sporadically. Two methods are available for screening these low-risk individuals in the population: faecal occult blood testing (FOBT) and flexible sigmoidoscopy.

Faecal occult blood testing

The main aim of FOBT is to detect early asymptomatic cancers. Although most advanced colorectal cancers bleed, most polyps do not. As a result, sensitivity is limited and FOBT may fail to detect 20–50% of cancers and up to 80% of polyps. Furthermore, the ingestion of red meat, and some raw vegetables and fruit, may yield false-positive results. Sensitivity is increased by rehydrating the slides, but at the expense of specificity. The main advantage of FOBT is that the investigation is inexpensive, although the subsequent radiological and endoscopic investigation of those who test positive increases costs.

Four randomised studies involving almost 330000 people, based on the use of FOBT, have now reported long-term results.

The Minnesota study
In the Minnesota study, 46551 volunteers aged 50–80 years, recruited from the American Cancer Society and other highly motivated groups, were randomly assigned to screening once a year, screening every 2 years or a control group [7]. Participants in the two groups assigned to screening were asked to provide six guaiac-impregnanted slides (Hemoccult) whilst adhering to a period of dietary restriction. Ninety per cent of each of the screened groups completed at least one screening. All screenings were completed by 46.2% of the group screened annually and 59.7% of those screened biennially.

To increase the sensitivity of the test, most slides were rehydrated prior to analysis. This had the effect of increasing the sensitivity of the test for colorectal cancer from 80.8% to 92.2%, decreasing the specificity from 97.7% to 90.4% and reducing the positive predictive value from 5.6% to 2.2%. In the slides that were rehydrated, the positive test level increased from 2.4% to 9.8%; this increase was most marked in the elderly. Because of the high positive test level, 38% of those screened annually and 28% of those screened biennially had at least one colonoscopy during the course of the study.

During the first 13 years of follow up, there were 1002 cases of colorectal cancer, 10097 deaths and 320 deaths from colorectal cancer (Table 2). There were twice as many patients with Dukes’ stage D cancers in the control group compared with the annually screened group (65 versus 33). The favourable shift towards less advanced stage had a profound effect on the number of deaths from colorectal cancer. Compared with the control group, there was a 33% reduction in mortality [odds ratio 0.67; 95% confidence interval (CI) 0.50 to 0.87] in the group offered annual screening, equivalent to a reduction in the cumulative annual mortality rate from approximately nine deaths per 10000 in controls to six deaths per 10000 in the screening group, and a 6% reduction in mortality (odds ratio 0.94; 95% CI 0.68 to 1.31) in the group offered biennial screening. The authors acknowledged that one reason for the substantial reduction in mortality in the group offered annual screening might be the high colonoscopy rate. They also commented that the cost of screening was substantial and would have to be weighed against the potential reduction in mortality from colorectal cancer.


View this table:
[in this window]
[in a new window]
 
Table 2. Mortality (Mandel et al. 1993)
 
A recent update has confirmed the 33% reduction in mortality (odds ratio 0.67; 95% CI 0.51 to 0.83) in the group offered annual screening and a 21% reduction in mortality (odds ratio 0.79; 95% CI 0.62 to 0.97) in those offered biennial screening [8].

The Nottingham study
In the Nottingham study, to date the largest population-based randomised trial of screening, 152000 individuals, aged 45–74, were randomised to either screening using FOBT or a control group [9]. The study design differed from the Minnesota study in that FOBT was performed biennially, slides were not rehydrated before testing, patients with weakly positive tests were re-tested to reduce the number of false-positive results and dietary restriction was imposed only when testing borderline results. A total of 38.2% of participants completed all the FOB tests, 21.4% completed at least one test and 40.4% did not complete any tests. Only 2.1% of patients tested positive on initial screening and 1.2% on re-screening. As a result, only 4% of subjects underwent colonoscopic or radiological investigation.

After a median follow-up of almost 8 years, 893 and 856 cases of colorectal cancer were detected in those allocated to the screening and control groups, respectively (Table 3). In the screening group, 236 (26.4%) cases of colorectal cancer were detected by screening, 249 (27.9%) presented as interval cancers and 400 (44.8%) were detected in non-responders. A further eight cases were diagnosed during endoscopic follow-up of a screen-detected lesion. Seven hundred and ten participants had screen-detected adenomas; of these, 207 (27%) were 20 mm or more in diameter and 97 (13.7%) were severely dysplastic. The positive predictive value of FOBT was 12% for colorectal cancer and 46% for all neoplasia. Detection rates were higher in those >65 years old and in men.


View this table:
[in this window]
[in a new window]
 
Table 3. Stage of CRC (Hardcastle et al. 1996)
 
Overall, the proportion of Dukes’ A tumours was significantly higher in the screening group compared with the control group (20% versus 11%; P <0.001). The proportion of patients with Dukes’ A tumours varied from 51% in those who responded to the initial invitation to 11% in those who declined to participate. In contrast to the Minnesota study, there was no difference in the proportion of patients with Dukes’ D tumours between the two groups, although the proportion of Dukes’ C tumours was lower in the screening group.

Three hundred and sixty patients in the screening group died of colorectal cancer compared with 420 in the control group (Table 4), a 15% reduction in cumulative colorectal cancer mortality in the screened group (odds ratio 0.85; 95% CI 0.74 to 0.98, P = 0.026). The reduction in colorectal cancer mortality in individuals who accepted the first FOB test compared with the control group was 39% (odds ratio 0.61; 95% CI 0.50 to 0.74).


View this table:
[in this window]
[in a new window]
 
Table 4. Mortality (Hardcastle et al. 1996)
 
The Scandanavian studies
(i) The Funen study. The results of the Funen study were remarkably similar to those of the Nottingham study [10]. Over 60000 individuals, aged 45–75, were randomised to bienniel FOBT with dietary restriction but without rehydration of slides. Only those participants who completed the first screening round were invited for further screening; 67% completed the first round and >90% accepted further screenings.

After 10 years, 481 and 483 cases of colorectal cancer were detected in the screening and control groups, respectively (Table 5). In the screening group, 138 (29%) cases were detected by screening, 148 (31%) presented as interval cancers and 195 (41%) were detected in non-responders. The proportion of Dukes’ A tumours was significantly higher in the screening group (22% versus 11%; P <0.01), whereas the proportion of patients with advanced tumours (Dukes’ C or D) was higher in the controls.


View this table:
[in this window]
[in a new window]
 
Table 5. Stage of CRC (Kronborg et al. 1996)
 
During the course of the 10-year study, there were 205 deaths attributable to colorectal cancer in the screening group compared with 249 deaths in controls (Table 6). Colorectal cancer mortality in the screening group was significantly lower than in controls (odds ratio 0.82; 95% CI 0.68 to 0.99; P = 0.03). Survival rates were higher in patients with screen-detected colorectal cancers (P <0.01) or interval cancers (P <0.05) compared with controls, whereas survival in the non-responders was poorer than in controls.


View this table:
[in this window]
[in a new window]
 
Table 6. Mortality (Kronborg et al. 1996)
 

(ii) The Gothenburg study. In the Gothenburg study, ~68000 inhabitants were randomised to either two screens 16–22 months apart or a control group [11]. The results have not been formally reported, but unpublished results suggest that after 8 years of follow up there was a 12% reduction in mortality in the screened group (odds ratio 0.88; 95% CI 0.69 to 1.12).

The results of the above population-based studies show remarkable consistency. Approximately two-thirds of those offered FOBT attended at least one screening. In each of the studies, more Dukes’ A tumours and fewer advanced tumours were detected in the screening group. Furthermore, although there were differences in the background risk of death from colorectal cancer in the different study populations, varying from 40 deaths per 100000 population in the Gothenburg study to 80 deaths per 100000 population in the Funen study, the relative reduction in cancer-related mortality was very similar, being 15% overall.

Indeed, a recent meta-analysis based on the three European studies and original data from the Minnesota study calculated that, overall, the effect of biennial FOBT was to reduce mortality from colorectal cancer by 16% (odds ratio 0.84; 95% CI 0.77 to 0.93) [12]. In the UK, this would translate into ~2700 fewer deaths per year. For those who completed at least one screen, the overall reduction in mortality was 23% (odds ratio 0.77; 95% CI 0.57 to 0.89). A greater reduction in cancer-specific mortality might be achieved by increasing the frequency of the examination, as in the Minnesota study, or by initiating a public health eduction programme to improve compliance.

Flexible sigmoidoscopy

An alternative approach to population-based screening would be the use of flexible sigmoidoscopy. Whereas FOBT aims to detect asymptomatic cancers, flexible sigmoidoscopy offers the possibility of detecting not only asymptomatic cancers, but also polyps at a pre-malignant stage. Since it is thought that progression from an adenoma to invasive carcinoma is a slow and orderly process taking 10–15 years, it has been postulated that the detection and removal of these pre-malignant adenomas may be a more effective method of reducing deaths from colorectal cancer than the detection of early localised tumours [13].

Since <10% of colorectal cancers occur in people <55 years old, a single flexible sigmoidoscopy towards the end of the sixth decade, with colonoscopic surveillance for those patients (approximately 3% to 5%) found to have high risk adenomas, could prove to be effective. However, it is likely to take 10–15 years before the results of the multicentre ‘once only’ flexible sigmoidoscopy study currently being undertaken in the UK become available.

Conclusion

In conclusion, there is good evidence to show that FOBT can reduce mortality from colorectal cancer. However, the cost of the initial testing and the subsequent colonic assessment has not been subjected to a proper cost–benefit analysis. Launching a national screening programme should therefore not be undertaken lightly. Apart from the enormous logistical problems associated with such a programme, there are multiple other factors to consider. Undoubtedly, false-positive tests will cause excess emotional distress in some individuals; some will be inappropriately reassured by a false-negative result. Furthermore, some asymptomatic, apparently healthy individuals may die prematurely as a result of their investigations or treatment.

Bearing in mind the continuing debate about the cost-effectiveness of breast and cervical cancer screening, it would seem prudent to delay launching such a programme until there are reliable data on the acceptability and compliance, especially in deprived areas, the financial costs, the potential complications of investigation and treatment, and the emotional side-effects of screening. Two pilot studies of FOBT currently taking place in the UK may help to define the above factors.

Footnotes

+ Correspondence to: University Department of Surgery, Alexandra Parade, Glasgow G31 2ER, UK. Tel: +44-(0)141-211-5435; Fax: +44-(0)141-552-3229. Back

References

1. Cancer Research Campaign. CRC CancerStats: large bowel—UK. London: CRC November 1999.

2. Coleman MP, Babb P, Damiecki P et al. Cancer survival trends in England and Wales, 1971–1995: deprivation and NHS region. London: The Stationery Office, 1999; pp. 199–225.

3. Scottish Cancer Intelligence Unit. Trends in cancer survival in Scotland 1971–1995. Edinburgh: Scottish Cancer Intelligence Unit 2000; pp. 54–71.

4. McArdle CS, Hole D, Hansell D et al. Prospective study of colorectal cancer in the West of Scotland: 10 year follow-up. Br J Surg 1990; 77: 280–282.[ISI][Medline]

5. Finlay IG, McArdle CS. Occult hepatic metastases in colorectal carcinoma. Br J Surg 1986; 73: 732–735.[ISI][Medline]

6. Finlay IG, Meek D, Brunton F, McArdle CS. Growth rate of hepatic metastases in colorectal carcinoma. Br J Surg 1988; 75: 641–644.[ISI][Medline]

7. Mandel JS, Bond JH, Church TR et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 1993; 328: 1365–1371.[Abstract/Free Full Text]

8. Mandel JS, Church TR, Ederer F, Bond JH. Colorectal cancer mortality: effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 1999; 91: 434–437.[Abstract/Free Full Text]

9. Hardcastle JD, Chamberlain JO, Robinson MH et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996; 348: 1472–1477.[ISI][Medline]

10. Kronborg O, Fenger C, Olsen J et al. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996; 348: 1467–1471.[ISI][Medline]

11. Kewenter J, Brevinge H, Engaras B et al. Results of screening, rescreening, and follow-up in a prospective randomized study for detection of colorectal cancer by fecal occult blood testing. Results for 68,308 subjects. Scand J Gastroenterol 1994; 29: 468–473.[ISI][Medline]

12. Towler B, Irwig L, Glasziou P et al. A systematic review of the effects of screening for colorectal cancer using the faecal occult blood test, hemoccult. BMJ 1998; 317: 559–565.[Abstract/Free Full Text]

13. Atkin WS, Cuzick J, Northover JM, Whynes DK. Prevention of colorectal cancer by once-only sigmoidoscopy. Lancet 1993; 341: 736–740.[ISI][Medline]