Background

The National Institute for Health and Care Excellence (NICE) published guidance on Suspected Cancer: Recognition and Referral (NG12) in 2015.1 This guideline was aimed at identifying people with symptoms that could be caused by cancer, with recommendations on the investigations to aid referral. A controversial statement in this document, relating to investigation of lower bowel symptoms, was the recommendation that testing for occult blood in faeces should be undertaken in cases where there was a low risk of colorectal (bowel) cancer (CRC). This was universally interpreted as advocating the use of traditional guaiac-based faecal occult blood tests (gFOBT), the evidence based on a small number of studies; faecal immunochemical tests for haemoglobin (FIT) were not widely available in routine laboratory use at the time. The qualitative gFOBT has many documented disadvantages,2 was not previously recommended for use in patients with symptoms and was considered obsolete; in consequence, gFOBT had been withdrawn from the repertoire of many laboratories and their use in other settings such as wards and clinics was much discouraged. Laboratories were then faced with the dilemma of how to comply with the recommendations put forward within the NG12 guideline, either reintroduce gFOBT or investigate the possibility of introducing the reportedly superior FIT. In reality, a large number of laboratories offered neither and, in consequence, general practitioners (GPs) were unable to comply with the NG12 guidance. This review aims to address some of the issues likely to be encountered by laboratories and the factors that are required to be taken into account when introducing a service providing FIT to aid in the assessment of patients presenting with lower bowel symptoms, it is concerned only with patients presenting with symptoms and not with the asymptomatic population-based bowel cancer screening, already carried out in national programmes in operation within the United Kingdom.

FIT in a diagnostic referral pathway

Since it is proposed that the use of FIT can aid a decision to refer made using a pathway for the investigation of suspected CRC and other SBD, consideration should be made when developing such a pathway as to who can request the test and by whom the test results are interpreted. In those centres in the UK who have implemented the test, current practice varies considerably, with results either going back to the GP or returned to secondary care gastroenterology or gastric surgery, depending on the referral pathway being used. The reasons for this variation in practice are considered below and are dependent on the clinical question being addressed and the key drivers for the introduction of FIT. At present, there is no single best approach that can be recommended for ubiquitous adoption.

There have been increasing referrals for colonoscopy from primary care, putting the service under considerable pressure, particularly following the introduction of bowel screening programmes in all four UK nations and with the introduction of local, regional and national publicity regarding the importance of seeking healthcare when lower bowel symptoms are experienced. As pointed out in the NICE: DG30 guideline, FIT has the potential to correctly rule out CRC and other SBD and avoid colonoscopy. Thus, measurement of f-Hb is potentially a useful tool to avoid unnecessary invasive procedures that themselves carry an inherent risk of bleeding, perforation and possible death. Although the risk is small for diagnostic colonoscopy, it increases markedly when combined with therapeutic procedures carried out at the same time including polypectomy and, in four recent studies, overall perforation rates ranged from 0.03% to 0.09%.18

One example of a referral pathway is illustrated in figure 1. Research studies give evidence that FIT, as an adjunct to clinical information and full blood count, can provide a very useful tool to assist with the triage of patients presenting in primary care with lower abdominal symptoms. An analysis of the three studies carried out to date using FIT in symptomatic patients in Scotland suggested that, within this cohort, there was potential to have avoided a referral and subsequent invasive investigation in 82.2% patients without SBD19 if a pathway similar to this is followed. However, it is vital to note that FIT are not 100% sensitive and rely on the pathological lesion to be bleeding at the time of faecal sampling and require the faeces to transit the colon before any haemoglobin present is degraded; in the 1514 cases reviewed by Quyn et al 19 three cases of CRC would have been missed using f-Hb with a 10 µg Hb/g faeces cut-off, all three cases had f-Hb above the limit of detection of the FIT system used but f-Hb below the recommended 10 µg Hb/g faeces cut-off. This has confirmed that a robust safety-netting procedure should be built into any clinical pathway for patients with symptoms. Watching and waiting could be supplemented by referral to gastroenterology or other appropriate clinic in secondary care, particularly if there are ongoing symptoms such as recurrent diarrhoea and/or iron-deficient anaemia, although patients with such serious symptoms should probably be referred for further investigation directly, irrespective of the f-Hb found, since these symptoms have higher positive predictive value for SBD than other low risk symptoms documented in the NICE NG12 guidelines. Some have suggested that a repeat FIT could be done, particularly if there are ongoing problems, but there are no data on this approach to date to our knowledge; studies on outcomes using this strategy would be of much interest.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Example of a referral pathway for colonoscopy incorporating FIT. f-Hb, faecal haemoglobin concentration; FIT, faecal immunochemical tests for haemoglobin; GP, general practitioner.

In order to have an impact on the number of referrals to secondary care, the FIT result should be available to the referring clinician to enable an informed decision to be made. If this is the GP, then those implementing the clinical pathway need to consider the education and training of their colleagues in primary care. Supplementary information could also be available in the form of interpretive comments on reports that accompany the results. Moreover, regular newsletters giving, for example, number of requests, use by different GP, clinical outcomes, audit information and synopses of interesting cases encountered, publications or new relevant guidelines might be of value, especially if web based. This approach has the advantage that all patients with symptoms receive a FIT and the f-Hb can, along with the impression of the GP and the results of other tests, particularly the full blood count, result in an objective decision to refer or not refer, saving secondary care from deciding on the need for colonoscopy or a secondary care clinic appointment. The disadvantage is the FIT workload would be high since patients presenting with lower bowel symptoms are very commonly seen in primary care.

An alternative implementation strategy might be to make FIT only available as part of the referral mechanism to secondary care. Here, the GP would make a referral to secondary care for a colonoscopy or outpatient review with clinical information as required by local practice and IT systems but would also request a FIT. The result of the FIT would be returned to the secondary care clinician who could use it to decide on the relevance of the referral. Options could include fast tracking the patient straight to colonoscopy, review in an outpatient clinic or discharge back to the care of the GP, perhaps with guidance on further follow-up. The advantage of this approach is the smaller pool of clinical staff required to be trained to make an informed decision based on the f-Hb result and therefore have the opportunity to gain more experience in interpretation and comparing this with outcome data. The disadvantage in this approach is that this will not reduce the number of patients referred to secondary care; however, it is still likely to impact the number of colonoscopies carried out.

When planning an implementation strategy for FIT, the other key factors to consider are the logistics of supply of FIT kits and specimen transport. The most widely used methods for FIT in use in the UK use their own individual proprietary specimen collection devices. These are certainly preferred to the traditional faecal collection pots since there is evidence that f-Hb decreases rapidly with time when not protected by collection into the preservative buffer in the devices.20 21 In order for the requestor to be able to give these devices to the patient, adequate supplies must be distributed to each requesting location, along with information (preferably mainly pictorial) on how to correctly collect the faecal sample into the device and a return envelope or container. This can be a labour-intensive process and one that requires continuous management and monitoring including keeping a record of the expiry dates of collection devices. The devices are approved for sending through the mail in the UK, and posting the sample to the laboratory may provide a patient-friendly alternative to returning the sample to the GP for onward transmission to the laboratory.

Analytical considerations

As well as the procedural considerations that must be appraised and the challenges solved when setting up a service for assessment of symptomatic patients using FIT, there are many analytical considerations which must also be taken into account.

FIT use polyclonal antibodies raised against the globin moiety of the Hb molecule. These antibodies are specific to human globin so, unlike the traditional gFOBT, they are not prone to interference from any haemoglobin derived from the diet or from any substance that interferes in the redox reaction of guaiac, such as peroxidases. Moreover, a recent systematic review has shown that cessation of oral anticoagulants, aspirin and other non-steroidal anti-inflammatory drugs is not required.22 There are a large number of FIT on the market; many of these are qualitative tests with very little evidence to support their use in any clinical setting. There are an increasing number of quantitative FIT analytical approaches on the market, mainly designed for CRC screening,23 and four of these were subject to an independent analytical evaluation that ratified their potential use in the diagnosis of SBD.24

While there are a large number of FIT methods available in the UK, both qualitative and quantitative, the latter are generally used, as recommended in DG30. Faecal samples for quantitative FIT are collected at home by the patient. A small mass of faeces (2 mg or 10 mg) is loaded on to a grooved or dimpled probe attached to the cap of the specimen collection device and then this is reinserted into the body of the device that contains a preservative buffer that stabilises any Hb present. Excess faeces on the probe is removed by an internal collar leaving a consistent mass of faeces in the buffer.25

There are a number of preanalytical and analytical challenges with FIT. There is no harmonisation of the sample collection devices. All have different designs of grooves or dimples26 in which to collect the sample and, due to the heterogeneity of faecal samples, potentially the amount of f-Hb collected will vary dependant from where in the faeces the sample is collected. Manufacturers’ recommendations on how to collect faeces using their devices varies, with some recommending sticking the sample probe into the faeces a number of times and others advocating swiping across the surface of the faeces. Faeces has variable consistency, and there are no studies to date that investigate the impact of this on f-Hb results. Analytically, there is no standardisation of FIT. There is currently no primary reference material or method to which FIT methods are standardised. Work is being done to address these problems internationally through an International Federation of Clinical Chemistry Scientific Division Working Group in which the four major FIT manufacturers participate, and a reference material is in the process of being investigated. A small study carried out recently has demonstrated that the use of polyclonal antibodies appears to limit problems in antibody detection of most haemoglobin variants; those variants with a missing alpha or beta globin chain are under-recovered with FIT and hence patients with such Hb variants are at risk of falsely low or negative f-Hb results.27

There are no established third-party internal quality control (IQC) materials currently available in the UK; however, materials have recently been made available elsewhere in Europe. If introducing an IQC protocol, factors considered in a previous review in this journal should be taken into account.28 It has also been challenging for external quality assessment scheme (EQAS) organisers to develop meaningful EQAS to assess FIT analytical performance; EQAS that send out faecal like matrices spiked with Hb exhibit large variation in results from the participant laboratories, and it is difficult to ascertain whether these differences are preanalytical or analytical. Conversely, those schemes that send out lyophilised aqueous samples of Hb, while providing a means of assessing comparative analytical bias, are not conforming to the traditional requirements of a commutable EQAS sample being in a matrix that reflects from sample patients.

Harmonisation of reporting units is also an issue, and while the international guidelines29 recommended that FIT be reported in µg Hb/g faeces rather than ng Hb/mL buffer, this still requires manufacturers to have a conversion in their analysers which is dependent on the both the mass of faeces collected and the volume of buffer in the specimen collection device in order to report comparable results.

DG30 guidance recommends a cut-off of 10 µg Hb/g faeces for referral of patients who present with symptoms in primary care. All three FIT systems recommended have performance characteristics that allow reporting of a quantitative f-Hb result, since this cut-off is equal to or higher than the documented limits of quantitation (LoQ), defined as the lowest concentration at which the performance meets the defined analytical performance specifications of the assay.30 Some studies have reported at lower f-Hb cut-offs than the LoQ. It is important that a local analytical verification of a FIT system is carried out, including confirmation of the LoQ, before reporting f-Hb lower than the 10 µg Hb/g faeces cut-off and any publications including data below 10 µg Hb/g faeces need to be evidenced to confirm the validity of the cut-off applied. It may be that f-Hb results lower than the LoQ but above the limit of detection, the highest measured result likely to be observed (typically at 95% certainty) for a sample containing no f-Hb could be reported as f-Hb detected but without a numerical result. Detailed recommendations for reporting f-Hb data at low concentrations are made in a recent opinion paper along with the interim proposal that the analytical performance specification for f-Hb should be that imprecision (and measurement uncertainty) be based on the current state of the art, that CV be <10%.31