Evaluation of Flexible Sigmoidoscopy Screening as an Adjunct to the National FOBT Screening Programme in Scotland

Screening for colorectal cancer is now being introduced in many countries worldwide, but there is still considerable uncertainty as to the ideal modality. Population based trials of guaiac faecal occult blood testing (gFOBT) have consistently demonstrated significant reductions in disease specific mortalities and three randomised population based trials of biennial gFOBT have demonstrated reductions in deaths from colorectal cancer of around 16%. As a result of these trials, a demonstration pilot was performed in the United Kingdom which has led to the introduction of gFOBT screening in all four countries of the United Kingdom.

However, gFOBT has significant limitations. It is clear that this form of screening is associated with a substantial interval cancer rate in the region of 50 %, indicating a severe deficiency in sensitivity for cancer. Furthermore, as it is now well established that the majority of colorectal cancers arise from pre-existing adenomas, it is important for any colorectal screening programme to detect adenomas in order to reduce the incidence of the disease as well as the associated mortality. Although gFOBT does detect some adenomas, the randomised studies have not demonstrated a reduction in colorectal cancer incidence with the exception of the Minnesota Study that used rehydrated gFOBT resulting in a high positivity rate and a large number of colonoscopies. It should be borne in mind however, that the newer faecal immunochemical tests (FIT), which, unlike gFOBT, are specific for human haemoglobin, perform better in terms of both cancer and adenoma detection.

It is also of interest that recent scrutiny of the interval cancer data from the Scottish demonstration pilot has clearly demonstrated that gFOBT screening tends to under-detect cancers in women when compared with men. In addition, it is relatively insensitive for rectal cancer when compared with colon cancer. Analysis of quantitative FIT data in our laboratory as part of an evaluation of FIT as a first line test in Scotland, has shown that the mean faecal haemoglobin concentration in women is lower than that in men, and that the cutoff value required for women to give a 2% positivity rate (similar to that achieved by the gFOBT currently in use in the UK) is less than half that for men. Thus, for gender at least, FIT will have the same limitations as gFOBT.

In a recent randomised trial carried out in 14 UK centres, a single flexible sigmoidoscopy (FS) between the ages of 55 and 65 years has been shown to bring about a significant reduction in colorectal cancer mortality. In addition, and most importantly, after a period of four years a significant reduction in colorectal cancer incidence was observed, presumably as a result of the routine removal of adenomas at FS. Interestingly, the reduction in incidence was restricted to left-sided cancers despite the fact that total colonoscopy was carried out in all those with a significant index lesion found at FS (5% of the screened population). FS does not suffer from low specificity since false positives do not occur, and there is independent evidence that it is more sensitive than a single gFOBT or FIT. In addition, FS is ideally suited to detecting rectal cancers and adenomas, and it is unlikely that there would be a gender difference in the sensitivity.

This landmark study (henceforth referred to as the "UK FS trial") was, however, an efficacy study since it was carried out in a population who had already indicated an interest in participating and, as a result, the uptake in those randomised to FS was an impressive 71%. This, however, leaves significant questions surrounding the introduction of FS screening, as it is not clear how it would perform as a population screening tool. Extrapolation of the results of the FS trial to the general population would suggest an uptake in the region of around 30% and although it is not clear what the uptake of FS would be in the Scottish population, data from the Glasgow centre that participated in the FS indicate a likely uptake of 24%. This compares with an overall uptake of around 60% in the current Scottish Bowel Screening Programme based on gFOBT.

A population based randomised trial of FS from Norway achieved a participation rate of 67% but a randomised study from The Netherlands achieved an uptake of 32.4% for FS compared with 49.5% and 61.5% for gFOBT and FIT respectively. In addition, there is evidence that participants perceive the personal burden of FS to be greater than that of either type of faecal testing. On the other hand, a study from Italy found a similar participation rate or FIT and FS, although both were low at 32% of those invited. Two small studies conducted in the London area observed an uptake of screening FS of around 50% but a similar study carried out in Tayside, Scotland achieved an uptake of only 24%. It is not clear why there should be such discrepancies in uptake of FS, but both cultural issues and differences in levels of deprivation are likely to be important.

The randomised study from The Netherlands demonstrated that the diagnostic yield of advanced neoplasia (cancers and significant adenomas) per 100 invitees was greater for FS than for either of the faecal tests suggesting that the overall performance of FS may be better than faecal testing despite a lower participation rate. This introduces an important ethical dimension; namely, whether or not it is acceptable to use a population screening tool that reaches a relatively small proportion of the population rather than a test that is associated with a higher participation rate but has an overall poorer performance in terms of disease detection. This is further complicated by the adverse effect of deprivation on uptake of screening. It is known that, in Scotland, the difference in uptake of gFOBT population screening between the most deprived and the least deprived quintile is around 20%. The effect of deprivation on uptake of FS population screening is not known, although in the UK FS trial there was a 16% difference in intention to participate and a 20% difference in actual uptake in those invited between the most and least deprived quartiles in Glasgow.

In a recent re-appraisal of the options for colorectal cancer screening commissioned by the UK National Screening Committee, and based on the UK FS trial, data from the first two rounds of the English Bowel Screening Programme and data on the sensitivity and specificity of FIT, modelling has suggested that a single FS would perform better than biennial gFOBT and that FS at age 62 results in the greatest reduction in CRC incidence, CRC mortality and CRC treatment costs, whereas FS at age 54 results in the greatest gain in life years and QALYs. In addition it was suggested that biennial FIT may outperform both biennial gFOBT and one off FS.

As result, FS screening has been introduced into England for all at the age of 55. Thus it is offered 5 years before gFOBT screening starts, as this is offered between the ages of 60 and 74. In Scotland, however, the age range for gFOBT screening is 50 to 74, and the main issue surrounding the introduction of FS screening in Scotland is that there is no information regarding the utility of FS in a population that has already been exposed to FOBT screening.

Nevertheless, given the high degree of efficacy of FS screening, particularly in terms of disease prevention via adenoma detection, and the relatively higher participation associated with gFOBT screening and its potential to detect proximal cancers, it is hypothesised that offering a combination of both approaches would provide an enhanced screening algorithm that would be associated with better outcomes than either modality alone. In order to test this hypothesis and answer some key unresolved questions around FS, it is necessary to carry out a pilot of FS screening integrated into the current faecal test-based Scottish Screening Programme, and to maximise the information from this pilot it is proposed to carry it out as a random evaluation. It is also proposed that FS is offered at around the age of 60, as this appears to be the age at which adenoma prevalence peaks, and therefore the age at which adenoma detection and removal is likely to confer the maximum benefit. There is also evidence from the gFOBT pilot that 80% of interval cancers are diagnosed over the age of 60.

As the Scottish Bowel Screening Programme offers gFOBT from the age of 50, and as screening started in Grampian, Tayside and Fife in 2000 as part of the UK demonstration screening pilot, carrying out a study in these areas will demonstrate whether or not FS adds value to a mature biennial FOBT screening programme. Further evaluation in Greater Glasgow will test FS in a challenging Health Board with areas of high urban deprivation.

We will seek to establish the value and feasibility of flexible sigmoidoscopy in populations which have been exposed to gFOBT screening for colorectal cancer.

In this way it will be possible to estimate the added value of adding FS to the FOBT programme. This will inform the structure of the Scottish Bowel Screening Programme and provide information that will have international implications.

It will also be possible to estimate the practicalities of introducing flexible sigmoidoscopy into the Scottish population against a background of ongoing gFOBT screening.