Abstract
Guaiac-based faecal occult blood tests (gFOBT) are still used in asymptomatic population bowel screening programmes but are being replaced by faecal immunochemical tests (FIT) for haemoglobin. gFOBT have many well-documented disadvantages and there is little evidence for their use in assessment of the symptomatic. Many laboratories have eliminated gFOBT from their approved repertoires by invoking the authoritative published guidelines. Data continue to accumulate that gFOBT are obsolete. FIT are available in two formats, qualitative and quantitative, the latter having advantages that the faecal haemoglobin concentrations are measured and cut-off concentrations that stimulate further investigation can be user-defined. There is growing evidence that FIT would be useful in a spectrum of clinical settings in addition to screening. All laboratories should have FIT in their existing repertoire. For some uses, qualitative FIT would be adequate. However, much evidence has accumulated that measurements of faecal haemoglobin concentrations are beneficial for the assessment of both disease severity and the future risk of colorectal neoplasia. Interpretation requires appreciation that faecal haemoglobin concentrations are higher in men than women and rise with age. It might well be that risk scoring systems that take gender and age into account, possibly with other factors including symptoms, will benefit individuals. Laboratories should consider how quantitative faecal haemoglobin measurements could be brought into routine practice and included in their forward planning. External quality assessment is needed. Specialists in laboratory medicine are urged to play a significant role in the research and development still required to make this a truly mature investigation.
Introduction
There is no doubt that use of guaiac-based faecal occult blood tests (gFOBT) does reduce mortality from significant colorectal neoplasia in asymptomatic population screening. This was demonstrated in randomized controlled trials (RCTs): combined results from the four eligible RCTs showed that participants had a 16% reduction in the relative risk of colorectal cancer mortality. 1 Pilot studies in Scotland 2 and England 3 clearly demonstrated that the results of these trials could be replicated in routine practice and, in consequence, what is now generally termed ‘bowel screening’ was rolled out across the four countries of the UK, although with some differences in the ages of those invited and the screening algorithm used. The results attained continue to show the significant benefits of this approach. 4,5 Other countries, for example, France, also use gFOBT in a successful nationwide screening programme. 6
However, the interest stimulated in laboratory medicine by the advent of the UK screening programmes, along with significant publicity in the media and the intensive efforts of charities involved in this area, led to assessment of whether gFOBT were appropriate for investigation of those individuals with symptoms of colorectal disease. Only a few years ago, the consensus was undoubtedly that there was little evidence for the use of gFOBT in such investigations and, indeed, national guidelines were unanimous that gFOBT had no role in assessment of the symptomatic. 7
As a result of these considerations and the many well-documented problems with gFOBT in faecal sample collection, storage and handling prior to analysis, analysis and clinical interpretation of results, 8 many laboratories eliminated gFOBT from their approved repertoires, usually invoking the authoritative published guidelines as objective justification. 9 Although many gFOBT are still purchased for use in wards, clinics and general practices, if specialists in laboratory medicine had, as they should have, considered these tests as point-of-care test procedures and, as for glucose, cholesterol, international normalized ratio and other quantities, applied the many local and national regulations that cover the performance of these, then this usually most inappropriate use (and probably less than ideal performance) of gFOBT would also have been minimized.
Laboratories still wishing to eliminate use of gFOBT will find assistance from others’ documented efforts to attain this laudable outcome. 9 The Yorkshire External Quality Assessment Scheme Annual Report 2010 for Occult Blood Detection shows that the number of participating laboratories has decreased year on year and also gives advice as to how to eliminate gFOBT from repertoires. This has been achieved mainly through circulation of laboratory bulletins giving information on the lack of evidence for use of gFOBT in assessment of the symptomatic, the existing guidelines, the many disadvantages of this test and the appropriate referral pathways. 10 Moreover, an easy additional strategy to cut down workload and provide educational opportunities at the same time is to reject all samples that are submitted in the traditional faecal sample collection pots as ‘unsuitable for analysis’ since haem, as measured by the surrogate peroxidase activity of gFOBT, is unstable in native faeces and false-negative results will occur. 11
Thus, the view held widely over recent years is that gFOBT should be eliminated from medical laboratories and point-of-care test settings. However, in contrast, data in a recent comprehensive systematic review and meta-analysis on the value of symptoms and additional diagnostic tests for colorectal cancer in primary care do suggest that relatively good results for the diagnostic performance of FOBT have been found. 12 The risk for colorectal cancer was significantly higher among patients with a positive test result than among those with a negative test result: for gFOBT, the median risk was 0.28 among those with a positive test result and only 0.01 for those with a negative test result. 12 This seems to contradict current dogma and, in consequence, this Personal View will pose (and answer) the questions (a) whether advice to eliminate a faecal test for occult blood was a mistake and (b) whether the opportunity to update rather than eliminate FOBT was missed and is still worthy of pursuit.
Further data supporting the view that gFOBT are obsolete
In spite of their ongoing use in very successful asymptomatic population screening programmes and the small quantum of work in the symptomatic as discussed by Jellema et al., 12 guidelines promulgated in 2011 do affirm the view that gFOBT have no role in investigation of the symptomatic individual. Scottish Intercollegiate Guideline Network, SIGN 126: Diagnosis and Management of Colorectal Cancer, states: ‘Faecal occult blood testing (FOBT) is not indicated and should not influence decision making in symptomatic patients’. 13 Moreover, the British Society of Gastroenterology have published new guidelines on the management of iron deficiency anaemia (IDA) which state: ‘Faecal occult blood testing is of no benefit in the investigation of IDA’. 14 Furthermore, the National Institute for Health and Clinical Excellence (NICE) Clinical Guideline 131, namely, Colorectal Cancer; Diagnosis and Management of Colorectal Cancer, does not even mention faecal tests. 15
In addition to the well-documented problems in sample collection and analysis, 8 recent work has shown that gFOBT may not be used correctly in clinical practice. In the USA, a recent survey showed that clinicians in primary care did not adhere to published guidelines: 16 traditional gFOBT were still widely used despite the fact that newer tests with lower analytical detection limits were recommended. Importantly, many still collected one sample during a digital rectal examination, although it has been unequivocally proven that this practice misses many cases of significant colorectal neoplasia. Furthermore, some simply repeated gFOBT on finding a positive test result rather than referring such patients for lower gastrointestinal tract endoscopy. A study performed in a hospital setting in Australia also detailed gross misuse of gFOBT. 17 gFOBT were requested in clinically inappropriate situations without consideration of possible confounding factors, often leading to inappropriate clinical decisions and additional costs to the hospital and patient: it was concluded that gFOBT had no utility in an acute hospital setting. An accompanying editorial commented: ‘The study shows how poor use of technology generates ant-like activity of limited benefit that drives the misallocation of resources, and leads to substandard non-evidence-based patient care’. 18 A recent study 19 investigated the frequency and findings of gastrointestinal endoscopy in patients with IDA attending a tertiary hospital and associations of endoscopy with patient and clinician-related factors and results of FOBT. Endoscopy was documented in 50% of patients and, perhaps unsurprisingly, rates ranged from 96% of patients seeing gastroenterologists to 31% of those seeing nephrologists. In the 275 patients who underwent colonoscopy, the prevalence of colon cancer and high-risk adenomas was not significantly different regardless of whether gFOBT and faecal immunochemical tests (FIT) were both positive, either one was positive, both were negative or neither had been performed.
Ongoing information from screening programmes continues to show other negative aspects of gFOBT that are undoubtedly relevant to their potential use in the symptomatic. A recent paper by Scholefield et al. 20 details the results of a 20-year follow-up of the RCT done in Nottingham that used traditional gFOBT. The observed impact on colorectal cancer mortality was small with a reduction of just 13% and, perhaps more importantly, there was no significant impact on the incidence of colorectal cancer despite removal of many advanced adenomas during colonoscopy of those with positive gFOBT results. This is likely due to the fact that gFOBT has an analytical detection limit (ca. 0.5 mg haemoglobin/g faeces 10 ) such that many lesions due to significant neoplasia are not detected: this is likely to hold in the symptomatic as well and many significant lesions that do bleed, but in small amounts less than the analytical detection limit, will not be detected using gFOBT.
This is confirmed by the clinical characteristics found for gFOBT use in the symptomatic in which sensitivity ranged from 0.33 to 1.00, while specificity ranged from 0.72 to 0.94. 12 This has led to a plea that more high-quality studies on the diagnostic performance of FOBT and symptom combinations in primary care settings should be supported. 21 However, the ongoing evidence is that gFOBT should indeed be considered as obsolete 22 and use should continue to be eliminated, although the body of published work does give many valuable insights on the role of detection of blood in faeces. Efforts should be directed to the wider and appropriate use of better tests for the presence of blood in faeces. And, better tests clearly do exist! Indeed, the authors of many recent publications on results obtained using gFOBT all state that the newer FIT for haemoglobin have significant advantages and are now the way ahead for screening programmes and other uses in routine practice. 5,7,8,12,20,21
FIT for haemoglobin
FIT, sometimes termed immunochemical FOBT or iFOBT, have been available for many years. The terminology is important and the FIT for haemoglobin descriptor has been supported by members of the Expert Working Group on FIT of the Colorectal Cancer Screening Committee of the World Endoscopy Organization.
23
There are two types of FIT, qualitative and quantitative, and both have many very well-documented advantages over gFOBT,
7,8
the major being that:
FIT are specific for intact haemoglobin and its early degradation products; Dietary restriction prior to testing is definitely not required; Use of medicines such as aspirin, non-steroidal anti-inflammatory drugs and anti-clotting agents such as warfarin do not need to be ceased and, in contrast to most views on this aspect of faecal testing, their intake possibly aids in the detection of significant neoplasia:
24
this may be because clotting is inhibited rather than gastrointestinal bleeding stimulated; The format of the sample collection devices encourages people to complete the test since generally little handling of faeces is involved; Instead of the two samples from each of three faeces that are required for gFOBT, only one or two are needed for FIT.
From a clinical point of view, there have been many studies on the use of FIT in screening. These have generated such overwhelmingly positive evidence that now FIT are the recommended faecal test in the recent comprehensive European Guidelines for Quality Assurance in Colorectal Cancer Screening and Diagnosis
25
and of other professional groups in Europe
26
and the USA.
23
Qualitative FIT
Qualitative FIT have a wide variety of sample collection devices using cards or, more commonly, tubes containing buffer with a stick or probe attached to the lid for faecal sampling. Analysis is usually based upon immunochromatographic test cassettes. These are simple to use: the results appear as coloured bands on the test strips that are fairly straightforward to interpret, although experience and factors such as good lighting do help, and each cassette has an integral control so as to monitor individual result quality. However, they do have some disadvantages including that they take longer to analyse and are more expensive than gFOBT, analysis cannot be automated and the stability of haemoglobin in the buffer tubes generally used in the collection devices may be limited. 8
Previously, it has been advocated that FIT should be used in a number of clinical situations other than screening in which investigation for the presence of blood in faeces would probably be helpful: these include the detection of human faecal blood in the paediatric setting in situations where the decision to undertake lower gastrointestinal tract endoscopy is often difficult and, similarly, investigation of those symptomatic individuals when this is difficult or impossible (or perhaps even when the individual patient is reluctant to undertake colonoscopy). 27 Since small numbers of FIT are likely to be required in these situations, qualitative FIT are generally satisfactory for adoption by laboratories. Such procedures are best performed in the laboratory as an integral component of the repertoire and subject to the ISO 15189-based standards of Clinical Pathology Accreditation (UK) Ltd 28 or similar. Note, however, that it is vital that the samples be collected into the sample collection devices of the manufacturer of the selected FIT: this is because haemoglobin in faeces collected in traditional pots is very unstable and false-negative results will occur in unpreserved samples. 29 Moreover, it has been stated 10 that use of gFOBT is probably inappropriate for clinical applications other than screening and that more specific tests with lower analytical detection limits would be advantageous, in other words, FIT.
A major disadvantage of qualitative FIT is that, exactly as for gFOBT, the single cut-off faecal haemoglobin concentration of the device is set by the manufacturer. Moreover, selection of a qualitative FIT for the clinical uses advocated above is difficult and it is very important to recognize that all FIT are not the same. Just as for calprotectin as comprehensively documented recently by Ayling, 30 different FIT have different cut-off faecal haemoglobin concentrations to determine positive and negative results. This has been nicely shown by Brenner et al. 31 who examined six qualitative FIT and demonstrated that the positivity rates using the same panel of faecal samples varied. Clinical sensitivity and specificity for advanced neoplasia also varied directly and inversely, respectively, with the positivity rate, as expected. We have examined the relationship between positivity rate and cut-off haemoglobin concentration as documented by Hundt et al. 32 and found that when the data were expressed as nanogram haemoglobin per millilitre buffer, there was no direct relationship. 23 For this and other reasons, we have firmly advocated that all data on FIT should be expressed in units of microgram haemoglobin per gram faeces and have urged manufacturers and all users to adopt these units, as well as improving many other aspects so as to make results more comparable over time, geography and methodology. 33
Quantitative FIT
Quantitative FIT involve techniques that allow the estimation of the faecal haemoglobin concentration. Most of the methods available use automated immunoturbidimetry on either dedicated faecal analysers or on commonly available routine clinical biochemistry analysers. Apart from the obvious advantages of high throughput of samples with good analytical quality and elimination of the inter-observer variability seen with gFOBT and to some extent with qualitative FIT, these methods allow the selection of the faecal haemoglobin concentration that triggers referral for colonoscopy.
There are very many publications on the use of FIT in asymptomatic population screening 25,26 and it would be impossible and inappropriate to review all of these here. Moreover, new informative publications appear regularly. Suffice it to state that, as expected, these studies have shown unequivocally that uptake by potential participants in screening programmes is improved using FIT as compared with gFOBT due to the more user-friendly nature of most of the faecal sample collection devices and the need to collect only one or two samples. In addition, as cut-off haemoglobin concentration is decreased and/or the number of samples is increased, the sensitivity increases and the specificity decreases, exactly as would be predicted.
Other possible future routine, but rather specialist, uses of FIT are likely but outside the scope of this article, for example, in the follow-up of those with proven disease such as in adenoma surveillance programmes, 34 and, perhaps with considerable advantage in the follow-up of those with familial colorectal cancer, which is currently generally done using colonoscopy. 35,36 Some further research is under way at present on these potential uses of quantitative FIT and more is required.
Quantitative faecal haemoglobin concentration estimation
If there is a future for quantitative faecal haemoglobin concentration measurements in the routine laboratory, then the main application would undoubtedly be in the assessment of the symptomatic. Jellema et al. 12 have documented that the available evidence is that combinations of symptoms and results of FIT showed good diagnostic performance for colorectal cancer. For the qualitative FIT used to date in published studies, sensitivity ranged from 0.70 to 1.00 and specificity ranged from 0.71 to 0.93. As for gFOBT, the risk for colorectal cancer was significantly higher among patients with a positive test result than among those with a negative test result. For FIT, the median risk was 0.21 among those with a positive test result and 0.00 for those with a negative test result. They concluded that diagnostic tests as first line investigation in primary care need to be valid, easy to perform, well tolerated by patients and sensitive, especially in the case of serious diseases: their systematic review showed that FIT might prove to be such tests. Now is the time to translate the theory into practice.
Interpretation of faecal haemoglobin concentration results
Automated systems and available reagents allow high-quality estimation of faecal haemoglobin concentration. 37 However, as well as the fact that most information available to date on FIT has been generated on asymptomatic individuals in screening programmes, studies generally use a single cut-off haemoglobin concentration so as to divide the participant population into those who are referred for colonoscopy and those who are not. Essentially, quantitative analyses are being used as qualitative techniques: thus, most estimates of faecal haemoglobin with low uncertainty of measurement are being wasted. It has been stated that current data support the view that better use of faecal haemoglobin concentrations in individuals is warranted. 38
There is considerable evidence that the faecal haemoglobin concentration is related to severity of colorectal disease. Indeed, even results obtained with gFOBT support this since those with five or six windows positive (strong positive, abnormal) have more disease and the cancers detected are at more advanced stages than those who have one to four windows positive (weak positive, equivocal) and are then positive again on either gFOBT 39 or qualitative FIT repeat investigation. 4 Similarly, with FIT, many studies have shown that the sensitivity for cancer is higher than that for higher risk (advanced, significant) adenomatous polyps through to low risk polyps, implying that more advanced neoplastic disease has higher faecal haemoglobin concentrations: a recent study demonstrates this very nicely. 40 This is supported by the fact that the sensitivity falls as the cut-off haemoglobin concentration is increased. More focused studies 41,42 have demonstrated that faecal haemoglobin concentration increases as disease stage advances from normal through low-risk polyps, through higher risk polyps to cancer. As for many quantities examined in laboratory medicine, there is much overlap between these groups and receiver operating characteristic curves demonstrate this very well. Interestingly, faecal haemoglobin concentrations in individuals with the less important clinically hyper- and metaplastic polyps, haemorrhoids and diverticular disease appeared to be similar to those in people with normal colonoscopy. 42 In addition, Ciatto et al. 43 have reported that faecal haemoglobin is significantly related to the presence of the very lesions, cancer and advanced adenoma, that screening is aimed at detecting. An important recent study by Chen et al. 44 showed that, in participants with faecal haemoglobin concentrations below the selected cut-off, the concentration found at first screening did predict subsequent risk of incident colorectal neoplasia. An accompanying commentary stated that, taking these data together, there is likely a continuum of risk of colorectal neoplasia as faecal haemoglobin increases from zero. 38
Thus, faecal haemoglobin is related to colorectal disease severity. However, the interpretation of a single faecal haemoglobin concentration measurement requires further knowledge concerning the individual. Again, some underpinning knowledge can be gained from consideration of the large quantum of data on gFOBT. The overall positivity rates found with gFOBT in Scotland, 2,45 England 3 and France 6 show that men have higher positivity rates than women and positivity rates increase with age. Ethnicity and deprivation, as defined by complex indices which take a variety of factors into account, also have effects on positivity rates 3,45 but, in contrast to age and gender, data on such variables are not usually available to the routine laboratory.
Although men have more significant colorectal neoplasia than women and older people have more than younger, 46 the levels of disease do not fully explain the gender and age findings with gFOBT. Interval cancers, that is, cancers that arise in the population who have had a negative screening test in the interval before the subsequent invitation for screening is issued, are substantial in gFOBT screening. There are few publications on interval cancers but a recent study showed that, of the cancers diagnosed in the screened population, interval cancers comprised 31.2% in the first round, 47.7% in the second and 58.9% in the third: the number of interval cancers increased slightly over screening rounds and, as expected, there was a concomitant decline in the numbers of screen-detected cancers. 47 Thus, gFOBT screening was associated with substantial interval cancer rates that increased with screening round and appears to preferentially detect cancers in men and the left side of the colon at the expense of cancers in women and in the right colon and rectum. Interval cancers are thus commoner in women and it may be that women are considerably disadvantaged by the use of currently available gFOBT. As discussed above, the cut-off haemoglobin concentration for gFOBT is set by the manufacturer and therefore the same gFOBT is used for men and women of all ages: this is probably quite inappropriate and is another reason for considering gFOBT as obsolete and adopting quantitative FIT measurements, since gender and age could be taken into account in the interpretation of results.
Using FIT, it has been shown clearly that there are major gender differences. In one example, as shown in a randomized trial of gFOBT versus FIT, the positivity rate and the detection rate for all colorectal neoplasia was higher for men than women and for those at or above 60 years of age. 48 In a study on the effects of gender on FIT, it was found that, at any cut-off for haemoglobin concentration, sensitivity and positive predictive value were substantially higher, and the specificity and negative predictive value were substantially lower among men than women: the authors considered that careful attention in the interpretation of test results is required and discussed the difficult question of having different cut-off faecal haemoglobin concentrations for men and women. 49 It has been stated that similar findings on gender and age mean that there is a need for more tailored screening strategies. 50 A number of possible explanations for the gender difference have been postulated. 46,47 Men have higher blood haemoglobin concentrations than women in their reproductive years of life but most women who are screened are probably postmenopausal when such differences are less. It might be that men take more medicines that inhibit clotting or cause gastrointestinal tract bleeding than women. Men may have a lower amount of residue in the diet than women. However, perhaps the most plausible explanation is that women have a longer colonic transit time than men 51 and blood present in the faeces of women is degraded before defaecation.
Percentiles of faecal haemoglobin concentration (nanogram haemoglobin per millilitre buffer) and potential upper reference limits (URL, with 90% CI)
CI, confidence interval
Adapted from McDonald et al. 52
Note: 100 ng haemoglobin/mL buffer = 20 μg haemoglobin/g faeces
Ramifications for populations and individuals
At any single potential cut-off faecal haemoglobin concentration decision limit, more men are declared positive than women and more older people are declared positive than younger people. Using the approaches of Chen et al., 44 the future risk of neoplasia is higher in men than in women and in older as compared with younger people. In consequence, the ramifications for screening programmes are that different cut-off faecal haemoglobin concentrations might well be used for the different genders and different age groups and that screening intervals could be different for different risk groups as postulated by Chen et al. 44 Low, intermediate, high and extremely high-risk groups were defined using faecal haemoglobin concentrations of 1–19, 20–39, 40–79 and 80–99 ng haemoglobin/mL buffer, respectively, using the OC Sensor (Eiken Chemical Co., Tokyo, Japan).
It is considered that these findings also have consequences for individuals. There is a growing interest in what are often termed risk scores and there have been a number of efforts describing risk stratification approaches which include variables such as gender, age, obesity, family history, smoking habits and alcohol consumption that impact on the development of colorectal neoplasia. A recent example, for asymptomatic individuals, the very population at whom screening is targeted, is a proposed score that enables risk stratification using elementary clinical information on age, gender, family history and smoking. This is mathematically simple in that the score can range from 0 to 6 and is calculated from age (<50 y, 0 points; 50–69 y, 2 points; ≥70 y, 3 points), gender (women, 0 points; men, 1 point), family history (absent, 0 points; present, 1 point) and smoking (never, 0 points; current or past, 1 point). It was shown that this particular scoring system successfully predicted the risk of advanced colorectal neoplasia and, impressively, that high-risk groups had four-fold higher risk compared with the average risk group. 54 Recently, a more sophisticated risk scoring nomogram has been published that takes, for the first time, faecal haemoglobin concentration as well as gender, age and body mass index (BMI) into account: points are allocated to each of these variables and then summed and the probability of significant neoplasia, or any adenoma plus significant neoplasia, is then interpolated from the nomogram. 40 It has been stated that the evidence supports the concept that faecal haemoglobin concentration could, with advantage, be included in such scoring systems and that this should prove a fruitful field of research for the future. 52
There are also scoring systems for use in assessment of the symptomatic. Here, the main need is assistance with the difficult decision as to which of the many patients with lower gastrointestinal tract symptoms would benefit from colonoscopy. It has been stated that, in spite of the availability of screening, the majority of colorectal cancers will continue to be diagnosed after presentation with symptoms.
55
In consequence, the diagnostic performance of scoring systems to identify symptomatic colorectal cancer compared with current referral guidance was recently investigated by Marshall et al.
55
They suggested that automated identification of patients with symptoms of suspected colorectal cancer from electronic primary care records using either the Bristol–Birmingham equation or the CAPER (
An aspect worthy of consideration and study is the use of faecal haemoglobin concentration measurements as rule-out rather than rule-in tests. In contrast to fecal calprotectin, which has been demonstrated to be useful in the differential diagnosis of lower abdominal symptoms and has potential for use in a number of ways in clinical management, particularly the concept that a negative calprotectin concentration in a low-risk patient might allow discharge without further invasive investigation, 30 there have been few studies examining the use of FIT for this purpose. Recently, Kok et al. 57 performed calprotectin and FIT on faecal samples collected from patients in primary care: firm diagnoses were determined from endoscopic and histological data and estimates of clinical characteristics calculated. The negative predictive values for each test and for the two combined were above 90% and, in consequence, it was stated that the ability to identify a large population for which the presence of colorectal disease could be ruled-out to a reasonable extent was evident. Thus, a strategy that might be adopted in primary care would be to perform calprotectin and FIT analyses on presentation and then refer for colonoscopy only if either or both of the results were positive. An accompanying editorial 58 reminded that calprotectin and FIT assays were not all the same and that, using the tests applied, 57 some important colorectal disease would be missed by the above strategy. It was considered that quantitative calprotectin and FIT assays perhaps could be used in this setting and cut-off concentrations calculated to improve the clinical outcomes: again, this is a topic very worthy of future research.
For the future
With these more complex approaches and even the risk scoring approach based on faecal haemoglobin concentration alone, 40 there is clearly a thrust towards consideration of the individual as such rather than simply as one of a group. Although the dogma is that everyone has some blood in their faeces, 59 the data in Table 1 show that at least 90% of both men and women have faecal haemoglobin concentration <20 μg haemoglobin/g faeces, to date the most often used cut-off concentration applied to refer individuals for colonoscopy. Indeed, more than half of the overall population screened have no detectable haemoglobin in their faeces. In consequence, it has been proposed that the truly healthy individual has no detectable faecal haemoglobin by conventional quantitative methodology and that any haemoglobin present is unusual. 52
Thus, as well as using faecal haemoglobin concentrations in better designed and executed population screening programmes and in evaluation of the symptomatic, it is highly probable, in my view, that there could be great advantages for individuals, just as for weight, waist circumference, BMI, blood pressure, cholesterol and glucose or glycated haemoglobin, to be knowledgeable about their own faecal haemoglobin concentration. Perhaps absolute change or rate of change in faecal haemoglobin concentration could be a marker of the presence of disease. A problem with this suggestion is that the abovementioned variables are modifiable by individuals, but there are no data on how, and even if, the faecal haemoglobin concentration of an individual could be reduced and, if it could be, what effect this would have on future risk of colorectal neoplasia. Perhaps smoking cessation, lowering of alcohol intake, increasing exercise and improvement in diet might affect faecal haemoglobin concentration, but this is simply speculation at present. Such ideas may be worthy of future research. Of course, if faecal haemoglobin became to be used in this way, there would be a real need for results to be comparable over time, geography and, ideally, methodology. At present, as recently documented by Members of the Expert Working Group on FIT, Colorectal Cancer Screening Committee, World Endoscopy Organization, 23,33 the numerical results obtained by FIT cannot be currently directly compared across different products for a number of reasons, including that current methods do not all use a haemoglobin calibrator that is traceable to an international reference preparation with concentration assigned by a higher order metrological method. This is another area in which progress is urgently needed to make faecal haemoglobin estimations better tests.
Recommendations for routine medical laboratories
Laboratories should continue to eliminate the use of gFOBT and should participate in, or initiate, efforts to eliminate the inappropriate and wasteful use of gFOBT in point-of-care test situations such as wards, clinics and primary care through education on both the existing professional and governmental guidelines and the many well-documented disadvantages. Some might argue that gFOBT are useful because bleeding from anywhere in the gastrointestinal tract is detected since haem, detected by its pseudo-peroxidase activity in gFOBT, is stable until defecation takes place. This has been proven many times to be fallacious and there is little evidence that gFOBT-positive individuals with normal colonoscopy have significant upper tract disease. 60 However, I now firmly believe that over the last few years, many in the community of laboratory medicine simply took the opportunity to eliminate a less than well-done and poorly understood faecal test and, in large part, the ‘baby was thrown out with the bathwater’.
FIT should be considered as quite a different type of test from gFOBT and, like the recent wider adoption of faecal calprotectin concentration estimations in the assessment of inflammatory bowel disease as against irritable bowel syndrome, 30 should be considered a new and valuable addition to the repertoire of the routine laboratory. This approach is commended so as to facilitate the wide introduction of FIT. Use of the FIT for haemoglobin nomenclature rather than the unfortunately often used iFOBT would be advantageous in changing perceptions and opinions.
As discussed earlier, some clinical applications could be adequately done using qualitative FIT. All who use laboratory services should have access to a FIT. For use in paediatrics and in adult medicine where the decision to undertake lower gastrointestinal tract endoscopy is difficult, use of a qualitative FIT could be satisfactory. A FIT should be selected with low analytical detection limit. Adherence by manufacturers to the recommendations made in the CLSI User Protocol for Evaluation of Qualitative Test Performance; Approved Guideline EP12-A would help in this regard since data expressed in the recommended format would facilitate objective comparison of the analytical performance characteristics. 61 Use of the strategies in this excellent document would also aid laboratories in validation of performance characteristics and comparison of qualitative FIT. Samples should be taken as soon as the faeces are passed into the FIT sample collection device and should be analysed without significant delay since there is some evidence that faecal haemoglobin is unstable in certain FIT sample collection devices. 36,62
Quantitative FIT have advantages over qualitative FIT and estimation of faecal haemoglobin concentration would be more useful in the clinical settings described earlier and for other purposes. The evidence is that such measurements would be useful in assessment of the symptomatic either alone or along with simple or more complex scoring systems that include variables shown to be important in the assessment of the possibility of significant colorectal disease. Further research is warranted and it is suspected that good-quality publications in the peer-reviewed literature will appear over the next few years. Laboratories should be aware of these laudable developments and consider now how quantitative faecal haemoglobin concentration measurements could be brought into routine practice: such a development should be included in forward planning initiatives.
There is a growing need for an External Quality Assessment Scheme that provides challenges for FIT, both qualitative and quantitative.
Finally, it is considered that this is an exciting, seminal point in the evolution of FIT as a routine laboratory test for clinical applications other than screening and it is hoped that specialists in laboratory medicine play a significant role in the research and development still required to make this a really mature evidence-based investigation. Moreover, as stated recently, the availability of methods for estimation of faecal haemoglobin and the wide interest in the use of the results do present a plethora of opportunities for such professionals to collaborate with clinical colleagues. 58
DECLARATIONS