NICE referral guidelines for suspected cancer: colorectal cancer and faecal occult blood testing

In June 2015, the National Institute for Health and Care Excellence (NICE) published revised guidelines on referral for suspected cancer. ¹ This included a section on lower gastrointestinal cancers including colorectal cancer (CRC). An extensive review of the literature was carried out by NICE. Positive predictive values (PPVs) for a number of symptoms were considered based on the available evidence and symptoms with a PPV greater than 3% led to a recommendation for referral. A combination of evidence and personal experience was used to consider what age threshold should be used for these referrals.

Controversially, the revised guidelines recommended that faecal occult blood tests (FOBT) be performed in the following symptomatic groups:

Those aged 50 years and over with unexplained abdominal pain or weight loss;

Those aged under 60 years with change in bowel habit or iron-deficiency anaemia;

Those aged 60 years and over who have anaemia, even in the absence of iron deficiency.

The guidelines acknowledge that there is a possibility of false-negatives and include a separate ‘safety netting’ section which includes a statement advising general practitioners (GPs) to be aware of the possibility of false-negative FOBT results and to review patients within an agreed time period or should symptoms recur, persist or worsen.

Laboratories offering an FOBT have used the traditional guaiac method (gFOBT). This is a crude chemical method that has very poor diagnostic sensitivity and specificity,^2,3 and for this reason a large number of laboratories across the United Kingdom have withdrawn the assay from their diagnostic repertoire.

During the consultation period, there were many stakeholder responses to the guidance expressing concern about inclusion of the gFOBT. ⁴ These objections focused predominantly around the risk of false-negatives, although there were also concerns about the number of false-positives and subsequent impact on colonoscopy resource. Despite the considerable number of objections received, NICE proceeded to keep gFOBT in the guidance with the reassurance that negative results will be subject to the ‘safety netting’ discussed above to minimise the risk of cancers being missed. The decision basis of the guideline is focused around the PPV of the gFOBT, and the evidence is based on the guaiac method. Despite recognition in the evidence statement from the summary data that the negative predictive value (NPV) of gFOBT can be as low as 16% in the primary care symptomatic population, we do not feel that adequate emphasis has been placed on the NPV and the impact of these potentially missed diagnoses. A number of the quoted papers used as evidence in the guidance for PPV do actually highlight the risks of NPV. Two of the papers conclude that the test should not be used in symptomatic patients because of its limited sensitivity for CRC.^5,6 Another states that due to the small sample numbers and high number of variables the results from their paper cannot be used as evidence of diagnostic value. ⁷ A further paper (published in 1984 and only as an abstract) while concluding that gFOBT can be useful if positive because of a decreased delay in diagnosis, also reports a 44% false-negative rate for both advanced cancer and polyps. ⁸

Despite gFOBT’s poor PPV and NPV, NICE proceeded to recommend the test for patients presenting to general practice with symptoms. A response from the NICE Guideline Development Group to a letter recently published in the British Medical Journal highlighting concerns associated with the inclusion of FOBT ⁹ explained that the new guidelines bring in some symptom groups not in the previous suspected cancer guidelines published in 2005 and highlight the economic evaluation carried out that demonstrated that a gFOBT strategy was markedly more cost-effective than colonoscopy.

NICE use the incremental cost-effectiveness ratio (ICER), a statistic used to summarise the cost-effectiveness of a health-care intervention. It is defined by the difference in cost between two possible interventions, divided by the difference in their effect (quality-adjusted life years gained (QALY)). There is a considerable difference in cost between gFOBT and a colonoscopy (NICE quote £4.86 for gFOBT and £368 plus additional adverse incident risk costs for colonoscopy). This huge cost difference is very difficult to balance out even taking in to account the varying diagnostic accuracy data of the two tests which NICE have done. For the QALY for gFOBT, NICE have assumed that any patients with a false-negative will re-present and be referred within one year. The assumptions around one year re-presentation have no evidence-base and are disputable.

Evidence actually suggests that over-reassurance and under-support of patients given the all-clear after presenting with symptoms can undermine them seeking medical advice in the case of new or recurrent potential cancer symptoms in the future. ¹⁰ This is supported by both anecdotal evidence and more systematic research ¹¹ from the English Bowel Cancer Screening Programme of people presenting to their GPs with symptoms suggestive of CRC. Because they have received a negative gFOBT result, the GP reassures them. A number of these patients, or relatives of the patients, have reported a subsequent diagnosis of CRC with often devastating outcomes due to the delayed referral.

A suitable biomarker to support GPs in their decision-making for patients presenting with symptoms that would include CRC in the differential diagnosis would be invaluable. Colonoscopy resources are limited and expensive and not all patients with these non-specific symptoms can be referred. However, we know that the gFOBT is flawed. It has a very high risk of providing false reassurance if a false-negative is obtained and the high false positive rate may lead to over investigation and unnecessary colonoscopy in patients with low risk symptoms.

Another test with potential in the primary care symptomatic population is the quantitative faecal immunochemical test (FIT) for haemoglobin for which there is an accruing evidence of its utility in this patient group.^12–15 This is discussed in detail by Fraser and Strachan in the accompanying Editorial. ¹⁶

The dilemma for many laboratories is whether to re-implement gFOBT testing in order to satisfy the requests from Clinical Commissioning Groups that will be arising as a result of the NICE guideline. One of the major considerations should be whether a diagnostic laboratory should be offering a test with such poor diagnostic capability as gFOBT. Additional consideration should be given to the strength of the cost-effectiveness analysis, bearing in mind the clinical assumptions that have been made (patients will re-present within one year) along with the considerable cost difference between colonoscopy and FOBT that will undoubtedly impact the overall outcome of the analysis.

In short, the new NICE guidance is ill-judged. Our advice to those responsible for routine clinical biochemistry laboratories is to resist calls to introduce or re-instate gFOBT. In the future, laboratories may wish to consider offering a FIT service for which there is already reasonable evidence that absence of detectable blood has an extremely high NPV for bowel cancer notwithstanding uncertainty as to precisely what cut-off levels should be used. If new evidence does continue to support the use of FIT in symptomatic patients, then the algorithm proposed by NICE is more likely to be robust.