Faecal calprotectin diagnostic tests for inflammatory diseases of the bowel

The Diagnostics Advisory Committee discussed the focus of the evaluation and the evidence available for faecal calprotectin testing. It noted that evidence existed on faecal calprotectin testing in differing populations with differing conditions. For example, some study populations included large numbers of adults with irritable bowel syndrome (IBS) and inflammatory bowel disease (IBD) (for example, Li et al. 2006), while others included children with a much wider range of organic and non‑organic conditions (for example, Tomas et al. 2007). It also noted that, while the evaluation was concerned with the role of faecal calprotectin testing for distinguishing between inflammatory and non‑inflammatory conditions of the bowel, the External Assessment Group suggested that the role of faecal calprotectin in 2 specific scenarios is likely to be the most important in clinical practice. These are IBD and IBS in the adult population presenting in primary care and IBD and non‑IBD in children who are referred for specialist investigation. Furthermore, the Committee noted that, although the use of faecal calprotectin testing is most relevant for helping to distinguish between inflammatory and non‑inflammatory conditions of the bowel, the number of conditions involved placed a prohibitively large burden on the data requirements for a cost‑effectiveness analysis. Therefore, the scenarios above represent a reasonable proxy for the likely clinical use of faecal calprotectin testing, balanced against the demands of the economic analysis. The Committee agreed with the External Assessment Group that it was appropriate for the evaluation to focus on the clinical and cost effectiveness of faecal calprotectin testing in these 2 scenarios. Faecal calprotectin testing is used in symptomatic patients to distinguish between 2 different types of disease. Diagnostic sensitivity refers to the proportion of patients whose test is positive in the presence of an inflammatory disease of the bowel (such as IBD); diagnostic specificity refers to the proportion of patients whose test is negative in the absence of inflammatory disease of the bowel. Patients whose test is negative may be found to have IBS. The Committee also noted that, although there is a growing focus on faecal calprotectin testing in primary care, there were limited data on faecal calprotectin testing in this environment.

The Committee understood that several different types of faecal calprotectin tests are available to the NHS in England and that such tests are continually improving. The Committee noted that there were limited data on the comparative effectiveness of different faecal calprotectin tests and agreed with the External Assessment Group that sufficiently robust data suggesting considerable differences in clinical reliability and performance between the tests were not available. The Committee recommended research on the comparative performance of different faecal calprotectin tests.

The Committee discussed pre‑analytical factors that may affect the results of faecal calprotectin testing. The Committee heard from specialists that several factors can affect the result of faecal calprotectin testing, including: sample storage, stool consistency, stool sampling, extraction and extract dilution. The Committee also heard that a UK National External Quality Assessment Service scheme (UK NEQAS scheme) has been set up for faecal calprotectin testing. The Committee encouraged participation in the UK NEQAS scheme and, when possible, standardisation of sample preparation methodology.

The Committee discussed the evidence on the clinical effectiveness of faecal calprotectin testing in IBD and IBS in adults. It noted that multiple studies of diagnostic accuracy were identified, which assessed faecal calprotectin testing when interpreted using different thresholds. The Committee noted that the evidence mainly concerned the use of quantitative ELISA tests in a secondary care adult population. The most commonly used threshold value in these studies was 50 micrograms/g, which allowed the results of 5 studies to be meta‑analysed. The Committee noted that the results of the meta‑analysis showed that faecal calprotectin testing performed well, with a sensitivity of 93% and a specificity of 94%. The Committee was aware that a study was also published on the performance of a point‑of‑care test (POCT), CalDetect, when used in secondary care (Otten et al. 2008), which showed that the test performed well, with a sensitivity of 100% and a specificity of 95%. However, the Committee noted that the Otten et al. study included relatively few patients. The Committee concluded that, on the whole, faecal calprotectin was a reliable marker for distinguishing between IBD and IBS in a secondary care adult population, but that further data are needed on point‑of‑care testing, to verify the results seen in the Otten et al. study. The Committee recommended further research on the use and clinical utility of faecal calprotectin testing for the diagnosis and long‑term management of these conditions in the community. The Committee also recommended that support pathways be developed for faecal calprotectin testing to support consistent and appropriate use.

The Committee discussed the de novo model constructed by the External Assessment Group. The Committee understood that limitations in the available data and/or variability in clinical practice meant that the model did not account for: (1) the way in which people with indeterminate results would be followed up before receiving a colonoscopy (all are assumed to be determinate and, therefore, receive a colonoscopy) and (2) the costs associated with the more common but relatively minor adverse events associated with colonoscopy (the costs of relatively rare but serious adverse effects are accounted for in the model). The Committee noted that, despite these limitations, the outcomes of the External Assessment Group's model were similar to those observed in previously conducted economic analyses. The Committee thought that, although the analysis may have benefitted from further sensitivity analysis, the results of the model are likely to be reasonably robust. The Committee concluded that the model was acceptable for decision‑making.

The Committee went on to discuss the economic analysis that assessed the cost effectiveness of faecal calprotectin for distinguishing between IBD and IBS in a primary care adult population. It noted that data from the Durham Dales implementation project were used to inform the sensitivity and specificity estimates of GP current practice (see section 5.47). The Committee noted from the data that GPs were currently very good at identifying those patients with IBD who needed to be referred for specialist investigation (near‑perfect sensitivity); however, a lower specificity of GP assessment meant that a significant proportion of people with IBS are being referred for specialist investigation, which may be avoided. The cost‑effectiveness analysis compared GP current practice plus quantitative ELISA testing, GP current practice plus point‑of‑care testing with CalDetect and GP current practice without faecal calprotectin testing as separate diagnostic strategies in adults presenting in primary care with lower gastrointestinal symptoms of abdominal pain or discomfort, bloating or change in bowel habit for at least 6 weeks. The Committee noted that the main goal of faecal calprotectin testing in primary care is to help reduce the number of unnecessary referrals of people with IBS (given the relatively high prevalence compared with IBD) and so reduce the number of unnecessary colonoscopies. The Committee noted that the model demonstrated that the differing diagnostic accuracies of the different strategies resulted in 19.8% (GP current practice), 5.6% (ELISA strategy) and 5.1% (CalDetect strategy) of the total modelled population being classified as having a false positive result and referred for colonoscopy. Furthermore, the Committee noted that the lower number of referrals after faecal calprotectin testing meant that both ELISA and CalDetect strategies dominated current practice (produced greater health benefits at reduced cost); however, the Committee agreed that the greatest benefit of faecal calprotectin testing is in reduced costs. Both ELISA and CalDetect strategies led to cost savings of £82 and £83 per patient respectively. The Committee concluded that faecal calprotectin testing is a cost‑effective use of NHS resources for distinguishing between IBD and IBS in adults in primary care and that sensitivity analysis showed these results to be robust.

The Committee discussed the use of faecal calprotectin testing of adults in secondary care. The Committee heard from clinical specialists that most of the faecal calprotectin testing in adults is expected to take place in primary care rather than secondary care. The Committee also recognised that there is a trend towards reducing the number of referrals to secondary care. It noted, however, that testing may also be appropriate for adults who have been referred for specialist assessment if testing has not already been done in primary care, in order to inform the decision on whether to do further investigations such as colonoscopy. Furthermore, the Committee concluded that cost savings from reduced numbers of colonoscopies are likely in this situation. Therefore, the Committee recommended faecal calprotectin testing as an option to aid differential diagnosis in adults with recent onset of lower gastrointestinal symptoms for distinguishing between IBD and IBS.

The Committee discussed the use of faecal calprotectin testing for clinical decision‑making. It agreed with the clinical specialists that faecal calprotectin should be used with other clinical information to support a physician's assessment and that physicians should be aware that inflammatory and non‑inflammatory diseases other than IBD and IBS respectively may affect levels of faecal calprotectin. The Committee emphasised that faecal calprotectin testing has the potential to falsely reassure GPs when used in people suspected of having bowel cancer, and so these people should be excluded from the recommendations. The Committee strongly emphasised that, when uncertainty remains in primary care around whether to refer a patient for specialist assessment based on faecal calprotectin testing, the clinician will benefit from further specialist input (clinical or laboratory) before making a decision. The Committee also considered that a repeat testing strategy may be considered as part of patient follow‑up (see section 6.14). The Committee was aware that there are limited data on the use of faecal calprotectin testing in primary care. However, the Committee concluded that the assessment had demonstrated the benefit of using faecal calprotectin testing in adults who meet the specific criteria set out in section 1.1 and the benefits were, on balance, generalisable to testing in primary care.

The Committee discussed the evidence on the clinical effectiveness of faecal calprotectin testing for distinguishing between IBD and non‑IBD in children. It noted that multiple studies of diagnostic accuracy were identified that assessed faecal calprotectin testing using different thresholds. The Committee noted that the evidence mainly concerned the use of quantitative ELISA tests in a secondary care paediatric population. The most commonly used threshold values in these studies were 50 micrograms/g and 100 micrograms/g, which allowed the results of 6 studies for each threshold to be meta‑analysed. The Committee noted that the results of the meta‑analysis showed that faecal calprotectin testing performed reasonably well at both thresholds (50 micrograms/g with a sensitivity of 99% and a specificity of 74%, and 100 micrograms/g with a sensitivity of 94% and a specificity of 82%). The Committee concluded that, on the whole, faecal calprotectin was a reliable marker for distinguishing between IBD and non‑IBD in a secondary care paediatric population.

The Committee went on to discuss the economic analysis that assessed the cost effectiveness of faecal calprotectin for distinguishing between IBD and non‑IBD in a secondary care paediatric population. It noted that this model was an adaptation of the primary care model for IBD and IBS in a primary care adult population. The Committee agreed with the External Assessment Group that the secondary care paediatric model was limited because it did not fully account for the non‑IBD case mix in the paediatric population (the prevalence of IBS in the non‑IBD group is lower than that seen in adults). The Committee concluded that, despite this and other limitations in the model (see section 6.5), this analysis would provide a reasonable proxy for the expected costs and benefits of faecal calprotectin testing in a secondary care paediatric population. The cost‑effectiveness analysis compared quantitative ELISA testing interpreted using a threshold of 50 micrograms/g followed by colonoscopy; quantitative ELISA testing interpreted using a threshold of 100 micrograms/g followed by colonoscopy; and direct referral for colonoscopy as separate diagnostic strategies in children with lower gastrointestinal symptoms of abdominal pain or discomfort, bloating or change in bowel habit, for at least 6 weeks, who had been referred for specialist investigation. The Committee noted that the main goal of faecal calprotectin testing in people who have been referred for specialist investigation is to help identify those who are likely to have IBD and will need further diagnostic tests (because the prevalence of IBD in this population is much greater than that seen in primary care), for example, colonoscopy. The Committee noted that the model demonstrated the different strategies resulted in 100% (direct referral for colonoscopy), 61.5% (ELISA with a 50 micrograms/g threshold) and 54.4% (ELISA with a 100 micrograms/g threshold) of the total modelled population receiving a colonoscopy. These estimates include 13.5% of people with false positive results being referred to colonoscopy with the 50 micrograms/g cut‑off strategy, and 9.4% of people with false positive results being referred to colonoscopy with the 100 micrograms/g cut‑off strategy. Furthermore, the Committee noted that the lower number of people expected to receive colonoscopies with the faecal calprotectin strategies meant that ELISA testing at both thresholds dominated current practice (produced greater health benefits at reduced cost); however, the Committee agreed that the greatest benefit of faecal calprotectin testing is in reduced per‑patient costs. Both ELISA interpreted at a threshold of 50 micrograms/g and ELISA interpreted at a threshold of 100 micrograms/g led to cost savings, of £205 and £240 per patient respectively. The Committee concluded that faecal calprotectin testing is a cost‑effective use of NHS resources for distinguishing between IBD and non‑IBD in a secondary care paediatric population and that sensitivity analysis showed these results to be robust. Therefore, the Committee recommended faecal calprotectin testing as an option in children with suspected IBD who have been referred for specialist assessment.

The Committee discussed the clinical interpretation of test results in children. The Committee heard from, and agreed with, the clinical specialists that faecal calprotectin should be used with other clinical information to support a specialist's assessment and that the specialist should be aware that inflammatory and non‑inflammatory diseases, other than IBD and IBS respectively, may affect levels of faecal calprotectin.

The Committee was aware that most of the data on faecal calprotectin identified for this assessment came from studies of ELISA testing in a secondary care population. Therefore, in the absence of robust primary care data (in particular, robust primary care data for POCTs), the Committee recommended that faecal calprotectin testing is performed in accordance with appropriate quality assurance processes and locally agreed care pathways to ensure results are reliable and replicable, and to increase the likelihood that the benefits and cost savings estimated by the model are delivered in the NHS. In addition, the Committee recommended that additional expertise should be sought when the faecal calprotectin tests are used in general practice, as outlined in section 6.8.

Given the lack of robust evidence comparing different tests, the Committee thought it appropriate that preferred faecal calprotectin tests might be selected locally in the NHS but that people should be aware that differences between tests may exist. The Committee noted that POCTs currently have a smaller evidence base and are not yet widely used in routine practice. The Committee therefore concluded that robust evidence is needed on the comparative performance of different faecal calprotectin tests, including the performance of POCTs compared with laboratory-based tests.

The Committee discussed the different thresholds for interpreting faecal calprotectin results. The Committee heard from clinical specialists that, while faecal calprotectin has been studied when interpreted using different thresholds (and investigated in the economic analysis), further research is needed on the impact of testing on clinical decision‑making when added to current practice before a recommendation on a particular cut‑off can be made. The Committee was aware of emerging evidence from a study of faecal calprotectin in primary care by Pavlidis et al. (publication in press, provided to the Committee as academic in confidence). However, the Committee concluded that it was too early to make judgements on these data. The Committee was aware that the assessment did not account for people with minimally elevated (intermediate) levels of faecal calprotectin who, as suggested by clinical specialists, may have low‑grade IBD and might be better off following a repeat testing strategy with faecal calprotectin to monitor levels of bowel inflammation through time as opposed to being subjected to invasive colonoscopies. The Committee heard from clinical specialists that faecal calprotectin levels can vary markedly between the time a person is tested in primary care and then subsequently retested (likely to be after several weeks) either by their GP or a specialist. The Committee noted that differences in tests may exist in the intermediate range of faecal calprotectin levels but may not have been measured in studies to date because of selective sampling of study populations. Therefore, the Committee recommended further research on the impact of faecal calprotectin testing on clinical decision‑making when added to current practice in both primary and secondary care. The Committee also recommended research into optimal cut‑off values for tests and the investigation of repeat testing strategies in people with intermediate levels of faecal calprotectin. Development of a consistent definition for the 'intermediate range' is encouraged by the Committee. The Committee further recommended that test result cut‑offs should be discussed and agreed locally as part of the implementation process for this testing pathway.

The Committee noted some general points: (1) the clinical-effectiveness estimates for faecal calprotectin testing summarised in this evaluation have been corroborated by faecal calprotectin databases around the country (for example, the Edinburgh Faecal Calprotectin Registry and the database maintained by King's Health Partners); (2) the Durham Dales project data may represent a best‑case scenario for GP current practice and, if this is the case, faecal calprotectin may have an even greater benefit in primary care (this additional benefit may be offset by losses in benefit if more than 50% of people with lower gastrointestinal symptoms are tested); (3) a significant proportion of people with IBD (particularly children with Crohn's disease), largely because of the similarity in symptoms to those in people with non‑IBD conditions, face delays in their diagnosis of up to several years, and the introduction of faecal calprotectin may help to reduce such delays.

The Committee was encouraged by the results of the assessment because it is likely that the use of faecal calprotectin testing will result in significant capacity being generated in colonoscopy departments to allow them to focus on people with greater need for a colonoscopy (for example, those suspected of having bowel cancer). Furthermore, the Committee noted that the good diagnostic performance of faecal calprotectin has the ability to provide reassurance to both physicians and patients alike given the heterogeneous and overlapping symptoms in lower gastrointestinal disease.

The Committee considered the impact of this guidance on groups of people with characteristics protected by UK equality legislation. During scoping, it was noted that IBS is most common in people in the 20 to 40 years age range, and is twice as common in women as men. Additionally, IBD is more common in white people than in African‑Caribbean people or those of Asian origin. The condition is most prevalent among Jewish people of European origin. The Committee considered that the guidance did not present any restrictions in access to diagnosis or treatment in the above groups.