NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Coeliac Disease: Recognition, Assessment and Management. London: National Institute for Health and Care Excellence (NICE); 2015 Sep. (NICE Guideline, No. 20.)
5.1. Accuracy of serological testing
5.1.1. Review questions
What is the sensitivity and specificity of the serological tests for coeliac disease?
Are the sensitivity and specificity of results different in any specified subgroups?
A number of different serological tests exist to test for coeliac disease. Each of these serological tests works by testing for the presence of a positive antigenic response of a given antibody against either IgA or IgG. Each of these serological tests has different sensitivity and specificity to detect coeliac disease. Determining the optimal serological test will ensure a balance of high sensitivity, whereby all those with the disease are accurately diagnosed, and specificity, whereby all those who do not have the disease are excluded.
5.1.2. Methods
The aim of this review question was to determine the sensitivity and specificity of the different serological tests available in the diagnosis of coeliac disease. This is an update of the chapter on ‘serological tests in the diagnostic process for coeliac disease’ in the 2009 guideline for coeliac disease (CG86). This updated review incorporates studies that were included in the previous guideline together with newly-published evidence.
The second component of this chapter focuses on whether specificity and sensitivity of serological test results is altered in any specified subgroups. Selective IgA deficiency is 10 to 15 times more common in patients with coeliac disease than in healthy subjects (Chow et al., 2012). Therefore, there is a risk of false negative serological results if IgA-dependent assays are used to assess the presence of CD. It has been reported that there has been inadequate evaluation of IgA deficiency while testing for coeliac disease, which has resulted in the underdiagnoses of both coeliac disease and IgA deficiency (McGowan et al., 2008). Therefore, this guideline considered the use of IgA-deficiency testing and IgG-based serological testing in the diagnostic process for coeliac disease.
Studies were only included if they met the following criteria: the population examined was children or adults suspected of having coeliac disease; all participants received both serological testing and an intestinal biopsy. The serological tests considered were:
- IgA tTG
- IgG tTG
- IgA EMA
- IgG EMA
- IgA DGP
- IgG DGP
- HLA DQ2/DQ8 genotyping
The comparator test was an endoscopic intestinal biopsy (reference standard in practice) and outcomes of interest were sensitivity and specificity of the different serological tests to detect coeliac disease.
The GDG expressed the need for a uniform histological reference standard to be used when examining sensitivity and specificity of the serological tests. Marsh grade 3 was identified as the optimal histological reference standard, and thus only studies which used this histological criterion to diagnose coeliac disease were considered. Furthermore, the GDG expressed that only studies within Europe should be considered. This is due to the high prevalence of non-coeliac enteropathies in countries outside of Europe, particularly India, Africa, and Israel, which is often difficult to discriminate from true coeliac disease.
Within the studies, different diagnostic kits and different cut-off values were used for the analysisd. Further differences between studies were differing or incompletely reported biopsy strategies, possible variability between laboratories or operators, and studies taking place in several different countries.
The included studies were cohort studies, which provided the best quality evidence within a modified GRADE framework.
For full details of the review protocol please see Appendix C.
Included studies
A systematic search was conducted (see Appendix C) which identified 2527 references. This search was restricted to studies published from 2008 onwards to avoid duplicates of studies considered in the previous coeliac disease guideline (CG86). The references were screened on their titles and abstracts and full papers of 76 references were obtained and reviewed against the inclusion and exclusion criteria in the review protocol (see Appendix C).
Overall, 70 studies were excluded as they did not meet the eligibility criteria such as inappropriate study design (case-series), not a primary study (descriptive narrative, opinion, etc.), examined the prevalence of coeliac disease in certain populations, studies in which the study population was not suspected of coeliac disease (but may have had an increased risk for developing coeliac disease, such as a commonly comorbid condition, or a family history of coeliac disease), and studies which did not use Marsh grade 3 for the histological diagnosis of coeliac disease. A detailed list of excluded studies and reasons for their exclusion is provided in Appendix C.
The 43 studies included studies in the previous coeliac disease guideline (CG86) were reviewed against the current protocol. Of these, 42 were excluded based on data extracted in the CG86 evidence tables as they did not meet eligibility criteria. Primary reasons for exclusion included inappropriate index tests (IgA AGA and IgG AGAe), populations that were not suspected of coeliac disease, and studies in which 100% of the participants did not receive both serological testing and an intestinal biopsy.
The search for this question was also designed to identify studies in which there was evidence that the serological tests for coeliac disease performed in any way differently from the general population. No studies of interest were found in this area.
5.1.3. Evidence review
The 6 remaining published papers did meet the stated eligibility criteria and were included. Data was extracted into detailed evidence tables (see Appendix D) and are summarised below). A single study (Hopper et al., 2008) included in the previous coeliac disease guideline did meet eligibility criteria for the current guideline and was included. Data is available in detailed evidence tables derived from the previous coeliac disease guideline (see Appendix D) and are summarised below.
The overall quality of the evidence from these 7 published papers ranged from very low to high, with the majority of evidence to be of moderate quality.
5.1.3.1. Evidence for the sensitivity and specificity of IgA tTG in the detection of coeliac disease in adults, children, and mixed age populations
Two studies (Mubarak et al., 2011; Panetta et al., 2011) of 376 children (mean age 3.9 years) with suspected coeliac disease conducted IgA tTG serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. Five children with IgA deficiency were excluded from one study (Panetta et al., 2011), and IgA deficiency was not commented on in by Mubarak and colleagues (2011). Three studies (Hopper et al., 2010; Volta et al., 2010; Swallow et al., 2012) of 2900 adults (mean age 40.4 years) with suspected coeliac disease conducted IgA tTG serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. A total of 16 adults with IgA deficiency were excluded from the analyses. One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA tTG serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.1.3.2. Evidence for the sensitivity and specificity of IgA EMA in the detection of coeliac disease
Two studies (Mubarak et al., 2011; Panetta et al., 2011) of 376 children (mean age 3.9 years) with suspected coeliac disease conducted IgA EMA serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. Five children with IgA deficiency were excluded from one study (Panetta et al., 2011), and IgA deficiency was not commented on in by Mubarak and colleagues (2011). Three studies (Hopper et al., 2010; Volta et al., 2010; Swallow et al., 2012) of 2900 adults (mean age 40.4 years) with suspected coeliac disease conducted IgA EMA serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. A total of 16 adults with IgA deficiency were excluded from the analyses. One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA EMA serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.1.3.3. Evidence for the sensitivity and specificity of IgA DGP in the detection of coeliac disease
One study (Mubarak et al., 2011) of 212 children with suspected coeliac disease conducted IgA DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. One study (Volta et al., 2010) of 144 adults with suspected coeliac disease (mean age 25 years) conducted IgA DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. Two patients with IgA deficiency were excluded from this analysis. One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study
5.1.3.4. Evidence for the sensitivity and specificity of IgG DGP in the detection of coeliac disease
One study (Mubarak et al., 2011) of 212 children with suspected coeliac disease conducted IgG DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. One study (Volta et al., 2010) of 144 adults with suspected coeliac disease (mean age 25 years) conducted IgG DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgG DGP serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population.
5.1.3.5. Evidence for the sensitivity and specificity of HLA DQ2/DQ8 genotyping in the detection of coeliac disease
One study (Clouzeau-Girard et al., 2011) of 170 children suspected of coeliac disease (median age 18 months) conducted HLA DQ2/DQ8 genotyping and endoscopic intestinal biopsy to determine the association between these coeliac-associated haplotypes and the presence of coeliac disease in this population. A total of 8 children were excluded from the analyses: 2 children were already consuming a GFD; one child had previously been on a gluten-free diet (GFD) and reintroduced gluten only eight weeks earlier; two children had selective IgA deficiency; three children had intestinal biopsies which could not be classified because of bad orientation of the sample.
5.1.3.6. Evidence for the sensitivity and specificity of serological testing in any specified subgroups
One study (Mubarak et al., 2011) was identified which examined serological test accuracy in children under the age of two. This study directly compared the sensitivity and specificity of serological tests in children under two years old to children over the age of two years old.
5.1.4. Health economic evidence
Any economic evaluations regarding the diagnosis of coeliac disease would be more appropriately categorised under the question of which test (or sequence of tests) to use (see 5.2.4), rather than a question concerned with the accuracy of the testing strategies alone. However, the evidence generated in the clinical review for this question forms the basis of the estimates of diagnostic outcome used in the original economic model described in 5.2.4
5.1.5. Evidence statements
5.1.5.1. Evidence for the sensitivity and specificity of IgA tTG in the detection of coeliac disease
Moderate to low quality evidence from 2 studies (Mubarak et al., 2011; Panetta et al., 2011) of 275 children reported that IgA tTG has high levels of sensitivity [96 %(95% CI: 93 to 99)] and moderate specificity [86 %(95% CI: 78 to 91)] in the diagnostic process for children suspected of coeliac disease.
High to moderate quality evidence from 3 studies (Hopper et al., 2008; Volta et al., 2010; Swallow et al., 2012) of 2900 adults reported that IgA tTG has moderate levels of sensitivity [91% (95% CI: 85 to 95)] and specificity [91% (95% CI: 90 to 92)] in the diagnostic process for adults suspected of coeliac disease.
Moderate to low quality evidence from a single study (Burgin-Wolff et al., 2010) of 268 children and adults reported that IgA tTG has high levels of sensitivity [97% (95% CI: 94 to 99) ] and moderate specificity [87% (95% CI: 80 to 92)] in the diagnostic process for children and adults suspected of coeliac disease.
5.1.5.2. Evidence for the sensitivity and specificity of IgA EMA
Moderate to low quality evidence from 2 studies (Mubarak et al., 2011; Panetta et al., 2011) of 275 children reported that IgA EMA has high levels of sensitivity [97% (95% CI: 94 to 99)] and moderate specificity [76% (95% CI: 67 to 83)] in the diagnostic process for children suspected of coeliac disease.
High quality evidence from 3 studies (Hopper et al., 2008; Volta et al., 2010; Swallow et al., 2012) of 2900 adults reported that IgA EMA has moderate levels of sensitivity [85% (95% CI: 78 to 90)] and high specificity [ 98% (95% CI: 98 to 99)] in the diagnostic process for adults suspected of coeliac disease.
Moderate to low quality evidence from a single study (Burgin-Wolff et al., 2010) of 268 children and adults reported that IgA EMA has high levels of sensitivity [98% (95% CI: 96 to 100)] and moderate specificity [85% (95% CI: 78 to 91)] in the diagnostic process for adults and children suspected of coeliac disease.
5.1.5.3. Evidence for the sensitivity and specificity of IgA DGP
High quality evidence from a single study (Mubarak et al., 2011) of 212 children reported that IgA DGP has moderate sensitivity [82% (95% CI: 72 to 89)] and specificity [80% (95% CI: 71 to 88)] in the diagnostic process for children suspected of coeliac disease.
High quality evidence from a single study (Volta et al., 2010) of 144 adults reported that IgA DGP has moderate sensitivity [83% (95% CI: 73 to 93)] and specificity [80% (95% CI: 71 to 88)] in the diagnostic process for adults suspected of coeliac disease.
Moderate quality evidence from a single study (Burgin-Wolff et al., 2010) of 268 children and adults suggests that IgA DGP has moderate sensitivity [78% (95% CI: 71 to 85)] and high specificity [97% (95% CI: 93 to 99)] in the diagnostic process for children and adults suspected of coeliac disease.
5.1.5.4. Evidence for the sensitivity and specificity of IgG DGP
High quality evidence from a single study (Mubarak et al., 2011) of 212 children reported that IgG DGP has moderate sensitivity [89% (95% CI: 80 – 95)] and specificity [81% (95% CI: 71 to 88)] in the diagnostic process for children suspected of coeliac disease.
High quality evidence from a single study (Volta et al., 2010) of 144 adults reported that IgG DGP has moderate sensitivity [83% (95% CI: 73 to 94)] and high specificity [97% (95% CI: 95 to 100)] in the diagnostic process for adults suspected of coeliac disease.
Moderate quality evidence from a single study (Burgin-Wolff et al., 2010) of 268 children and adults reported that IgG DGP has moderate sensitivity [85% (95% CI: 80 to 90)] and specificity [92% (95% CI: 86 to 97)] in the diagnostic process for children and adults suspected of coeliac disease
5.1.5.5. Evidence for the sensitivity and specificity of IgG tTG
There were no published studies that examined the sensitivity and specificity of IgG tTG in populations suspected of coeliac disease.
5.1.5.6. Evidence for the sensitivity and specificity of IgG EMA
There were no published studies that examined the sensitivity and specificity of IgG EMA in populations suspected of coeliac disease.
5.1.5.7. Evidence for the sensitivity and specificity of human leucocyte antigen (HLA DQ2/DQ8) genotyping
High quality evidence from a single study (Clouzeau-Girard et al., 2011) of 170 children considered the sensitivity and specificity of HLA DQ2/DQ8 genotyping in a population of children suspected of coeliac disease. This paper reported a very high sensitivity of 99% (95% CI: 96 to 100) and low specificity of 69% (95% CI: 59 to 79) of HLA DQ2/DQ8 genotyping in children.
5.1.5.8. Evidence for the sensitivity and specificity of serological testing in any specified subgroups
Moderate quality evidence from a single study showed increased specificity in the younger children for IgA EmA (< 2, 93% (95% CI: 66 to 100); >2, 82% (95% CI: 72 to 89)), and increased sensitivity and specificity for IgA DGP (<2, 100% (95% CI: 84 to 100), >2, 87% (95% CI: 77 to 93); <2, 100% (95% CI: 75 to 100), >2, 81% (95% CI: 71 to 88)) and IgG DGP (<2, 100% (95% CI: 84 to 100), >2, 87% 95% CI: (77 to 93; <2, 100% (95% CI: 75 to 100), >2, 81% (95% CI: 71 to 88)), suggesting that the DGP antibodies in particular may have maximum diagnostic accuracy in this population
5.1.6. Evidence to recommendations
Both questions relating to the evidence for serological testing in coeliac disease were presented in tandem and discussed together. Therefore, the linking evidence to recommendation information will be presented for the two components of this question at the end of this chapter.
5.1.7. Recommendations & research recommendations
Both questions relating to the evidence for serological testing in coeliac disease were presented in tandem and discussed together. Therefore, the associated recommendations will be presented for the two components of this question at the end of this chapter.
5.2. Order and sequencing of serological tests
5.2.1. Review questions
Which serological test is the most appropriate to diagnose coeliac disease?
Depending on test results, should more than one test be used, and if so, what should be the sequence of testing?
Following which sequence of tests and test results is it appropriate to refer onwards for endoscopic intestinal biopsy for confirmatory diagnosis?
In section 5.1 the sensitivity and specificity of different serological tests to detect coeliac disease was explored. The purpose of this section is to examine whether a combination of those serological tests investigated in section 5.1 can achieve a greater sensitivity and specificity to detect coeliac disease than when those tests are used in isolation.
5.2.2. Methods
The aim of this review question was to determine when serological test results would indicate a diagnosis of coeliac disease without the need for intestinal biopsy. The second part of this question was designed to determine when serological tests results would indicate a referral for endoscopic intestinal biopsy for confirmatory diagnosis is appropriate. This is an update of the chapter on ‘serological tests in the diagnostic process for coeliac disease’ in the 2009 guideline for coeliac disease (CG86). This updated review incorporates studies that were included in the previous guideline together with newly-published evidence.
Combinations (including parallel or sequential combinations) of serological and IgA deficiency testing were compared to intestinal biopsy, or other combinations of tests, or test algorithms.
Studies were only included if they met the following criteria: the population examined was children or adults suspected of having coeliac disease; all participants received both serological testing and an intestinal biopsy.
- The serological tests considered (in any combination) were:
- IgA tTG
- IgG tTG
- IgA EMA
- IgA DGP
- IgG DGP
- HLA DQ2/DQ8
- Total IgA (for IgA deficiency)
The outcomes of interest were sensitivity and specificity of the different combinations of serological tests to detect coeliac disease.
The GDG also expressed the need for a uniform histological reference standard to be used when examining sensitivity and specificity of the serological tests. Marsh grade 3 was identified as the optimal histological reference standard, and thus only studies which used this criterion to diagnose coeliac disease were considered. Furthermore, the GDG expressed that only studies within Europe should be considered. This is due to the high prevalence of non-coeliac enteropathies in countries outside of Europe, particularly India, African and Middle-Eastern countries, which is often difficult to discriminate from true coeliac disease and may skew sensitivity and specificity estimates of serological tests in these populations.
Within the studies, different kits and different cut-off values were used for the analysisf. Further differences between studies were different or incompletely reported biopsy strategies, possible variability between laboratories or operators, and studies taking place in several different countries.
The included studies were cohort studies and case-control studies, which provided the best quality evidence.
For full details of the review protocol please see Appendix C.
Included studies
A single systematic search was conducted (see Appendix C) for both review question three and review question four together, which identified 2527 references. This search was restricted to studies published from 2008 onwards to avoid duplicates of studies considered in the previous coeliac disease guideline (CG86). The references were screened on their titles and abstracts and full papers of 17 references were obtained and reviewed against the inclusion and exclusion criteria in the review protocol (see Appendix C).
Twelve studies were excluded as they did not meet the eligibility criteria such as inappropriate study design (case-series), not a primary study (descriptive narrative, opinion, etc.), examined the prevalence of coeliac disease in certain populations, or studies in which the study population was not suspected of coeliac disease (but may have had an increased risk for developing coeliac disease, such as a commonly comorbid condition, or a family history of coeliac disease). A detailed list of excluded studies and reasons for their exclusion is provided in Appendix F.
The 5 remaining published papers did meet eligibility criteria and were included. Data was extracted into detailed evidence tables (see Appendix D) and are summarised below).
The 6 studies included in the previous coeliac disease guideline (CG86) were reviewed against the current protocol. Of these, 5 were excluded as they did not meet eligibility criteria. Primary reasons for exclusion included inappropriate index tests (IgA AGA and IgG AGA),g populations that were not suspected of coeliac disease, and studies in which 100% of the participants did not receive both serological testing and an intestinal biopsy.
The 1 remaining published paper included in the previous coeliac disease guideline did meet eligibility criteria for the current guideline and was included. Data is available in detailed evidence tables derived from the previous coeliac disease guideline (CG86) and are summarised below.
The overall quality of the evidence from these 6 published papers ranged from low to high, with the majority of evidence to be of moderate quality. Evidence was downgraded due to methodological issues such as unclear recruitment strategy, inconsistency between studies, or imprecision.
Sensitivity and specificity values presented here for one of the included studies (Burgin-Wolff et al., 2013) were calculated from raw data values. These differ from the sensitivity and specificity results presented in the paper. The paper presents ‘non-classified’ data, which relates to the number of participants per test combination that were unable to be classified due to inconsistency between two or more tests (i.e. positive result on one test and negative result in another test(s)). This ‘non-classifiable’ data was incorporated into the analyses presented here as false negative data, as it is assumed that the ‘non-classified’ data was classed as negative.
5.2.3. Evidence review
5.2.3.1. Sensitivity and specificity of combination IgA tTG + IgG DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA tTG + IgG DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.2. Sensitivity and specificity of combination IgA EMA + IgG DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA EMA + IgG DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.3. Evidence for the sensitivity and specificity of combination IgA tTG + IgG DGP + IgA DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA tTG + IgG DGP + IgA DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.4. Evidence for the sensitivity and specificity of combination IgA EMA + IgG DGP + IgA DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA EMA + IgG DGP + IgA DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.5. Evidence for the sensitivity and specificity of combination IgA EMA + IgA tTG + IgG DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA tTG + IgA EMA + IgG DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.6. Evidence for the sensitivity and specificity of combination IgA tTG + IgA EMA + IgA DGP + IgG DGP testing
One study (Burgin Wolff et al., 2013) of 268 children and adults (median age 29 years) with suspected coeliac disease conducted IgA tTG + IgA EMA + IgG DGP + IgA DGP combination serological testing and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. IgA deficient patients were excluded from the outset in this study.
5.2.3.7. Evidence for the sensitivity and specificity of combination IgA tTG + IgA EMA + HLA DQ2/DQ8
One study (Clouzeau-Girard et al., 2011) of 170 children suspected of coeliac disease (median age 18 months) conducted HLA DQ2/DQ8 genotyping with IgA tTG + IgA EMA serological testing and endoscopic intestinal biopsy to determine the association between these coeliac-associated haplotypes, serology, and the presence of coeliac disease in this population. A total of 8 children were excluded from the analyses: 2 children were already consuming a gluten-free diet (GFD); 1 child had previously been on a GFD and reintroduced gluten only 8 weeks earlier; 2 children had selective IgA deficiency; 3 children had intestinal biopsies which could not be classified because of bad orientation of the sample.
5.2.3.8. Evidence for the sensitivity and specificity of combination IgA + IgG h-tTG/DGP
One study (Mubarak et al., 2011) of 212 children with suspected coeliac disease conducted serological testing using a combination test with a human recombinant tissue substrate of IgA + IgG tTG/DGP and endoscopic intestinal biopsy to determine the presence of coeliac disease in this population. One case-control study (Porcelli et al., 2011) of 201 adults serologically tested for coeliac disease was also examined. This study population was comprised 41 recently diagnosed people with coeliac disease; 145 ‘disease-controls’ with various other conditions, including autoimmune hepatopathies; viral hepatitis, and other gastrointestinal diseases; and 24 healthy blood donors. All participants underwent serological testing using a combination test with a human recombinant tissue substrate of IgA + IgG tTG/DGP and endoscopic intestinal biopsy to diagnose or exclude the presence of coeliac disease.
5.2.3.9. Evidence for the sensitivity and specificity of the test algorithm If IgA tTG is positive, and then IgA EMA is positive
One study (Hopper et al., 2008) of 2000 adult participants (mean age 55.8 years) with suspected coeliac disease examined a 2-step serological screening strategy in which participants were screened first with IgA tTG and then with IgA EMA if the IgA tTG test was positive. Participants were considered serologically positive for coeliac disease if both serological tests were positive. All patients also underwent an endoscopic intestinal biopsy to confirm the diagnosis of coeliac disease. Fourteen participants were excluded from the analyses due to IgA deficiency.
5.2.3.10. Evidence for the sensitivity and specificity of the test algorithm If IgA tTG is positive or equivocal, and then IgA EMA is positive
One study (Swallow et al., 2012) of 756 adult participants (mean age unknown) with suspected coeliac disease examined a 2-step serological screening strategy in which participants were screened first with IgA tTG and then with IgA EMA if the IgA tTG test was positive or equivocal, according to the strategy recommended by NICE in the coeliac disease guideline CG86. Participants were considered serologically positive for coeliac disease if both serological tests were positive. All participants also underwent an endoscopic intestinal biopsy to confirm the diagnosis of coeliac disease. Fourteen participants were excluded from the analyses due to IgA deficiency.
5.2.3.11. Evidence for the sensitivity and specificity of the test algorithm If both IgA tTG is positive and IgA EMA is positive
Two studies (Hopper et al., 2008; Swallow et al., 2012) of 2756 adult participants (mean age 55.8 years) with suspected coeliac disease examined a 2-step serological screening strategy in which participants were screened with both IgA tTG and IgA EMA. Participants were considered serologically positive for coeliac disease if both serological tests were positive. All participants also underwent an endoscopic intestinal biopsy to confirm the diagnosis of coeliac disease. Fourteen participants were excluded from the analysis due to IgA deficiency.
5.2.3.12. Evidence for the sensitivity and specificity of the 2-step test algorithm If either IgA tTG is positive, or IgA EMA is positive
One study (Hopper et al., 2008) of 2000 adult participants (mean age 55.8 years) with suspected coeliac disease examined a 2-step serological screening strategy in which participants were screened first with IgA tTG and then with IgA EMA. Participants were considered serologically positive for coeliac disease if either or both serological tests were positive. All participants also underwent an endoscopic intestinal biopsy to confirm the diagnosis of coeliac disease. Fourteen participants were excluded from the analyses due to IgA deficiency.
5.2.4. Health economic evidence
5.2.4.1. Systematic review of published cost–utility analyses
An economic evaluations filter was applied to the search protocol for this research question (an update of a review question considered within the 2009 NICE coeliac disease guideline CG86) with the aim of finding economic evaluations that explored the cost effectiveness of diagnostic strategies for people with signs and symptoms suggestive of coeliac disease.
The search identified 135 references. The references were screened on their titles and abstracts and 10 full-texts were ordered. None of the studies met the inclusion criteria.
No cost–utility analyses were found to address selection criteria
5.2.4.2. Original health economic analysis
An original cost–utility model was developed to explore the benefits, harms and costs associated with different strategies for serological investigation of people with symptoms suggestive of coeliac disease. The model used a cohort (Markov) structure to estimate lifetime costs and effects, incorporating the tests themselves, endoscopic investigation of serologically positive cases, treatment for coeliac disease and the long-term complications of treated and untreated disease (including impact on mortality). Long-term complications modelled were osteoporosis, subfertility and cancer (divided into non-Hodgkin’s lymphoma and other cancer). Separate analyses were conducted for adults and children. Full details of the methods and results of the model are provided in Appendix G.
5.2.5. Evidence statements
5.2.5.1. Sensitivity and specificity of combination IgA tTG + IgG DGP testing
Moderate quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA tTG + IgG DGP has low sensitivity [72%; 95% CI: 65 to 80] and high specificity [96%; 95% CI: 92 to 99] in the diagnostic process for children and adults suspected of coeliac disease.
5.2.5.2. Sensitivity and specificity of combination IgA EMA + IgG DGP testing
Moderate quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA EMA+ IgG DGP has low sensitivity [73%; 95% CI: 66 to 80] and high specificity [95%; 95% CI: 91 to 98] in the diagnostic process for children and adults suspected of coeliac disease.
5.2.5.3. Evidence for the sensitivity and specificity of combination IgA tTG + IgG DGP + IgA DGP testing
High quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA tTG+ IgG DGP + IgA DGP has low sensitivity [73%; 95% CI: 66 to 80] and high specificity [99%; 95% CI: 98 to 100] in the diagnostic process for children and adults suspected of coeliac disease.
5.2.5.4. Evidence for the sensitivity and specificity of combination IgA EMA + IgG DGP + IgA DGP testing
High quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA EMA+ IgG DGP + IgA DGP has low sensitivity [58%; 95% CI: 50 to 66] and high specificity [99%; 95% CI: 98 to 100] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.5. Evidence for the sensitivity and specificity of combination IgA EMA + IgA tTG + IgG DGP testing
High quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA EMA+ IgA tTG + IgG DGP has low sensitivity [70%; 95% CI: 48 to 64] and high specificity [96%; 95% CI: 98 to 100] in the diagnostic process for children and adults suspected of coeliac disease.
5.2.5.6. Evidence for the sensitivity and specificity of combination IgA tTG + IgA EMA + IgA DGP + IgG DGP testing
High quality evidence from a single study (Burgin-Wolff et al., 2013) reported that the combination test for IgA tTG + IgA EMA+ IgG DGP + IgA DGP has low sensitivity [56%; 95% CI: 48 to 64] and high specificity [99%; 95% CI: 98 to 100] in the diagnostic process for children and adults suspected of coeliac disease.
5.2.5.7. Evidence for the sensitivity and specificity of combination IgA tTG + IgA EMA + HLA DQ2/DQ8
High quality evidence from a single study (Clouzeau Girard et al, 2011) reported that the combination test for IgA tTG + IgA EMA + HLA DQ2/DQ8 has high sensitivity [99%; 95% CI: 96 to 100] and high specificity [96%; 95% CI: 92 to 100] in the diagnostic process for children suspected of coeliac disease.
5.2.5.8. Evidence for the sensitivity and specificity of combination IgA + IgG h-tTG/DGP
High quality evidence from a single study (Mubarak et al., 2012) reported that the combination test for IgA + IgG h-tTG/DGP has high sensitivity [99%; 95% CI: 95% CI: 93 to 100] and specificity [99%; 95% CI: 96 to 100] in the diagnostic process for children suspected of coeliac disease.
Moderate quality evidence from a single study (Porcelli et al., 2011) reported that the combination test for IgA + IgG h-tTG/DGP has 100% sensitivity and moderate specificity [90%; 95% CI: 86 to 95] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.9. Evidence for the sensitivity and specificity of the test algorithm If IgA tTG is positive, and then IgA EMA is positive
High quality evidence from a single study (Hopper et al., 2008) reported that the 2-step algorithm of positive IgA EMA following positive IgA tTG has moderate sensitivity [87%; 95% CI: 65 to 97] and high specificity [97%; 95% CI: 95 to 98] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.10. Evidence for the sensitivity and specificity of the test algorithm If IgA tTG is positive or equivocal, and then IgA EMA is positive
High quality evidence from a single study (Swallow et al., 2012) reported that the 2-step algorithm of positive IgA EMA following positive or equivocal IgA tTG has moderate sensitivity [86%; 95% CI: 76 to 92] and high specificity [99%; 95% CI: 98 to 99] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.11. Evidence for the sensitivity and specificity of the test algorithm if both IgA tTG is positive and IgA EMA is positive
High quality evidence from 2 studies (Hopper et al., 2008; Swallow et al., 2012) reported that the 2-step algorithm of both positive IgA EMA and positive IgA tTG has moderate sensitivity [86%; 95% CI: 76 to 92] and high specificity [99%; 95% CI: 98 to 100] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.12. Evidence for the sensitivity and specificity of the 2-step test algorithm If either IgA tTG is positive, or IgA EMA is positive
High quality evidence from a single study Swallow et al., 2012) reported that the 2-step algorithm of either positive IgA EMA, or positive IgA tTG has moderate sensitivity [92%; 95% CI: 84 to 96] and specificity [90%; 95% CI: 89 to 92] in the diagnostic process for adults suspected of coeliac disease.
5.2.5.13. Health economic evidence statements
An original, directly applicable cost–utility analysis with minor limitations suggested that, in adults, the most effective testing strategy is to consider people serologically positive if they are positive on either IgA tTG or IgA EMA. However, the incremental benefit of this strategy comes at a very high cost (base-case ICER in excess of £170,000 per QALY), and much better value for money is achieved by a strategy that tests IgA tTG in all people and reserves IgA EMA to classify cases in which IgA tTG results are weakly positive.
An original, directly applicable cost–utility analysis with minor limitations suggested that, in children, the most effective testing strategy is a combination of IgA tTG, IgA EMA and HLA DQ2/DQ8. However, the incremental benefit of this strategy comes at additional cost, with an ICER of approximately £34,000 per QALY. Of the modelled options, the most cost effective – when QALYs are assumed to be worth £20,000 – was a combination of IgG DGP and IgA tTG. No evidence was available to analyse the combination of IgA tTG and IgA EMA without additional tests in children.
5.2.6. Evidence to recommendations
| Relative value of different outcomes | The GDG discussed and agreed that a single histological standard for the diagnosis of coeliac disease was needed in order to ensure consistency between studies. The ESPGHAN (European Society for Paediatric Gastroenterology, Hepatology and Nutrition) criterion relies in part on serological results to inform diagnosis; however, the GDG decided that these criteria would be inappropriate as a reference standard to investigate the accuracy of serological tests in the diagnosis of coeliac disease. The GDG agreed that Marsh criteria grade 3 was the most appropriate reference standard for the histological diagnosis of coeliac disease and that only studies using these criteria should be considered as part of the evidence review. The GDG further raised the necessity of providing guidance on the operationalisation of ‘equivocal’ when examining the output of tTG results in order to optimize diagnostic accuracy. There is currently no guidance in interpreting weakly positive (or negative) results, and what defines these weak or equivocal ranges. The group was concerned that it was common for laboratories to assess IgA deficiency when testing for IgA tTG, essentially using IgA tTG as a surrogate marker for total IgA. The benefits of conducting total IgA testing were highlighted as important, because critical immunodeficiency’s such as CVID and myeloma can be picked up, as well as those with IgA deficiency, which in itself is relatively common (1/500), and highly important to a person’s immunological health and wellbeing. |
|---|---|
| Trade-off between benefits and harms | Inter-test variability and internal validity of serological testing The GDG discussed the number of different testing kits that are available for each antibody, in particular the transglutaminase ELISA kits and the wide variability in the sensitivity and specificity to diagnose coeliac disease between each of these different kits. The GDG agreed that there is a strong need for each laboratory to internally validate their serological testing assays in order to ensure optimal diagnostic utility. The GDG expressed concern that many laboratories may be using poor testing kits, and that internal validity of testing kits in each laboratory is not being examined, and further, that there is no evidence of quality assurance in labs to make sure that optimal internal validity is being achieved. The GDG further discussed the continual improvement in these ELISA testing kits for the detection of tTG, and expressed that the emergence of new immunofluorescence techniques for the detection of tTG look particularly promising. The GDG also noted that quality assurance procedures are now being replaced by ISO15189 (http://www Equivocal range in tTG serological results The group recognised that each laboratory will have their own definitions of how to define equivocal tTG results, depending on the specific anti-tTG assay used. GDG members believed that, in most laboratories, equivocal can be interpreted as weakly positive. In this circumstance, the majority of laboratories will then conduct an EMA as a confirmatory test (as indicated in the current NEQAS report, 2014). The GDG expressed the purpose of the secondary EMA to be to make sure that the tTG was performed correctly in the first place, as it is known that one can conduct testing using the same ELISA on the same sample 3 separate times, and get slightly different result every time (see Egner et al., 2011). Laboratories tend to do anti-tTG because it is a test which is performed in large numbers and it is both comparatively easy to automate and can be quantified. It is often done in biochemistry labs, who have much experience in automating quantifiable assays, but little experience in more subjective immunological assays such as EMA immunofluorescence. It is also cheaper and more amenable for most laboratories to do IgA tTG. Actually anti-EMA testing is not significantly more expensive than tTG testing; however the procedure requires specialist immunological training. The problem was raised that theoretically ‘equivocal’ could be interpreted as slightly below the positive titre threshold, and therefore ‘weakly negative’. The group agreed that those who were weakly negative were highly unlikely to have CD whilst those with unambiguously positive tTG were very likely to have CD, and it is those who are ‘weakly positive’ that they are unsure about and would be most important to have a secondary serological screen of EMA conducted. The group further discussed the difficulty in determining how to define weakly positive, as each different ELISA uses different manufacturer-recommended cut-off points. It was noted that the ESPGHAN criteria in children uses 10 × the upper limit of normal as strongly positive, and that a similar algorithm could be used to define weakly positive. However ‘normal’ is still an ambiguous term which will differ between laboratories. Testing for total IgA The group discussed that it was very common for laboratories to examine IgA levels when testing for IgA tTG, essentially using IgA tTG as a surrogate marker for IgA deficiency. This was raised as poor practice, and highly problematic, because total IgA levels are measured using an assay that is not designed or validated for measuring these levels (the ELISA is designed to measure tTG, not total IgA). The benefits of conducting total IgA testing were highlighted as highly important, primarily to maximise diagnostic accuracy, but also because critical immunodeficiency’s such as common variable immunodeficiency can be picked up, as well as IgA deficiency, which in itself is relatively common in people with coeliac disease The costs of conducting a separate test for total IgA were discussed and agreed to be negligible (when the test is undertaken as part of multiple tests on the same sample, as would be the case here), further supporting the notion that this test should be carried out separately to IgA tTG in all circumstances. The group also raised the accreditation process that will be introduced under ISO15189, whereby each individual assay must undergo an accreditation process to prove optimal practice and replicable results. Under this accreditation process, it is highly likely that IgA tTG assays will only be accredited to test for IgA tTG. If laboratories that do not have total IgA assays undergo accreditation, they will not have an accredited means of testing for total IgA, which would be a potential violation of good clinical practice standards Need for biopsy (in adults) The group recognised that one component of this review was to determine under which circumstances serology would be accurate enough to diagnose CD, and therefore not require a biopsy to make a diagnosis. The group felt strongly that a biopsy should always be used to confirm a diagnosis for the following reasons:
Deamidated gliadin peptides (DGP) The GDG discussed and agreed that the evidence for deamidated gliadin peptides (DGP) looks promising; however, evidence for these antibodies is still emerging and needs to be strengthened by more studies before any recommendations can be made as to the diagnostic utilities for these. Combination testing The GDG discussed combination testing for achieving optimal sensitivity and specificity of diagnosis. It was agreed that there is a lack of evidence for the utility of second-line testing after initial serology is negative in people strongly suspected of coeliac disease. HLA DQ2.5 DQ2.2 and DQ8 genotyping The GDG discussed HLA DQ2/DQ8 genotyping and the variability between different centres in terms of how results are reported and fed back to clinicians. In particular, the GDG expressed a need for the standardisation of which HLA DQ2/DQ8 haplotypes were examined. The GDG further expressed that many less common HLA haplotypes exist, however the most common, accounting for around 99% of HLA genotypes relevant to coeliac disease, are HLA DQ2.5, DQ2.2, and DQ8. The GDG wished to specifically refer to these HLA variants when recommending HLA genotyping. |
| Economic considerations | In the original health economic analysis, the lifetime effectiveness of each strategy – in terms of QALYs accrued – was found to be strongly correlated with the strategy’s sensitivity. This is because false-negative diagnoses are associated with reduced QALYs (as a function of both persistent coeliac symptoms and increased likelihood of long-term complications, some of which may impact on life expectancy). Therefore, strategies with fewest false- negative diagnoses are those that accrue most QALYs. Conversely, the total costs of each strategy are strongly correlated with their specificity. This is predominantly because false-positive serological diagnoses incur additional costs due to unnecessary endoscopic biopsies that would be avoided with a more specific approach. Weighing these factors against each other leads to somewhat different conclusions in adults and children. In adults, greatest value for money (maximal net monetary benefit at £20,000 per QALY) tends to be achieved by strategies that are most sensitive (that is, those that minimise false-negative diagnoses and, therefore, maximise QALYs). In children, the approaches that demonstrate greatest value are those that have higher specificity (that is, those with fewest false-positive diagnoses that, therefore, minimise costs). The reason for this difference is that endoscopic biopsies are much more expensive in the paediatric population, as they are invariably performed under general anaesthesia. However, the GDG also emphasised the importance of correctly identifying children with coeliac disease, and did not believe that minimising false-positive diagnoses could, alone, be the overriding objective of best practice. In adults, the most effective strategy was the most sensitive – that is, considering people serologically positive if they are positive on either IgA tTG or IgA EMA. However, the incremental benefit of this approach came at a very high cost: the base-case ICER exceeded £170,000 per QALY. However, the model suggested that almost all the benefit of this approach could be achieved at lower cost by a strategy that tests IgA tTG in all people and reserves IgA EMA to classify cases in which IgA tTG results are weakly positive. Indeed, accounting for the costs of the tests themselves and the downstream consequences of true and false diagnoses over the lifetime of the cohort, the model estimated that this approach is associated with lowest net costs of all options. The GDG concluded that it should be recommended as the preferred approach. Although sensitivity was the main determinant of value in the adult population, small differences in sensitivity between strategies could be outweighed by larger differences in specificity. This was the case with the recommended approach: although there were 2 strategies in the model that had higher sensitivity than IgA tTG with IgA EMA to determine weakly positive cases, the benefits associated with those strategies’ superior true-positive rates were smaller than the harms and costs associated with their inferior false-positive rates (lower specificities). The GDG noted that current provision of serological testing is variable, with different laboratories relying on different assays, either singly or in combination. This means that, in order to recommend the routine use of any particular strategy (especially one involving more than 1 test); it would be necessary to take account of the implications for standardising practice. In particular, the additional costs associated with the new equipment required by some laboratories should be accounted for. Therefore, in addition to the unit cost of each test, the original model included an approximate estimate of additional capital costs that would be incurred, by some laboratories, in expanding their provision to enable them to undertake those tests. Data from a national audit of current provision (NEQAS) were used to estimate the proportion of laboratories for which such additional investment would be necessary. These additional costs had no impact on base-case findings: the strategy using IgA tTG as a first-line test with IgA EMA to discriminate in cases of weak tTG positivity remained optimal. Further exploration of this parameter suggested that the total costs of increasing capacity would have to increase the unit cost of every tTG test undertaken in England and Wales by over £4 per test before it would be preferable to rely on a single-test strategy. This figure was very substantially higher than the base-case estimate of 9p per test; therefore, the GDG considered that any capital costs required would be clearly justified. In children, the most effective strategy was one that combined serological assays for IgA tTG and IgA EMA and HLA DQ2/DQ8 genotyping, an approach that had been shown to benefit from very high sensitivity and specificity in the clinical evidence review. However, because HLA DQ2/DQ8 genotyping is a relatively expensive test (over £70 each, some 5–8 times more expensive than any of the serological assays), its routine use is associated with significant costs, with the consequence that the 3-test strategy was associated with a relatively high ICER, around £34,000 per QALY gained compared with the next-cheapest non-dominated option. The GDG advised that, in addition to its relative expense, HLA DQ2/DQ8 genotyping is subject to practical difficulties in non-specialist settings, both in gaining access to the test and in interpreting its results. The GDG noted that many of the strategies that appear attractive in children, from a cost effectiveness point of view, are combination approaches that include one or more DGP assay. The group was aware that all evidence for these tests in children came from a single study (Burgin Wolff et al., 2013). While these results appeared promising, especially as regards the high specificity of the strategies, the group felt that further research would be necessary before such approaches could be universally recommended. Therefore, the GDG chose to make a research recommendation for further research into the accuracy of DGP assays, particularly in younger children. If DGP-containing strategies are excluded from the paediatric decision-space, IgA tTG is the least expensive strategy in the model. It is also estimated to be more effective than IgA EMA alone (that is, it dominates it). Under this circumstance, the model predicts that the 3-test combination of IgA tTG, IgA EMA and HLA DQ2/DQ8 would become the optimal approach, generating more QALYs than any of the individual tests alone, with an ICER of approximately £6000 per QALY compared with IgA tTG as a single test. However, the GDG reiterated that there are significant practical barriers to the adoption of routine HLA DQ2/DQ8 genotyping in the primary care settings in which most initial testing would be undertaken. In theory, one first-line testing strategy that might be considered in children would be to combine IgA tTG and IgA EMA assays (in a serial algorithm – as in the recommended approach for adults – or in parallel). However, there was no evidence on the accuracy of any strategies combining these tests in children. Moreover, the GDG felt that positive results on any single assay warranted consideration by a specialist in paediatric gastroenterology. Therefore, the GDG chose to split their recommendation into 2 parts, with first- line testing comprising IgA tTG assay (which should be available and familiar in primary care), followed by referral to a specialist for further investigation (which might include IgA EMA testing and/or HLA DQ2/DQ8 genotyping as well as consideration for endoscopic biopsy). By doing this, the GDG believed that costs associated with genetic testing would be minimised – as specialists would be able to use their experience in ordering the test, rather than taking the more indiscriminate approach that a routine testing strategy implies. In this way, the group inferred that the effectiveness of the recommended strategy would be similar to that achieved under routine 3-step testing, and the costs would be reduced. A further exploratory analysis using the original health economic model attempted to simulate the benefits, harms and costs of a diagnostic algorithm for children that enables a diagnosis of coeliac disease to be made without the need for confirmatory biopsy (as proposed by ESPGHAN). This analysis was more speculative than other simulated strategies, as it was not based on direct evidence of the diagnostic accuracy of the algorithm; instead, it combined evidence on various tests used in isolation and assumed independence between them. The model suggested that this approach is extendedly dominated by some sequences with routine biopsy that had been simulated. However, results were broadly comparable, in terms of costs and effects, to some of the better-value approaches. The GDG was aware that primary evidence on the accuracy of the ESPGHAN algorithm is due to be published in 2015, and concluded that any explicit recommendation endorsing or rejecting the approach should await the availability of this evidence. It should be noted that the original health economic analysis assumed that everyone who is serologically positive undergoes confirmatory biopsy and that biopsy is 100% accurate. Therefore, there is no such thing as a long-term false-positive diagnosis in the model: biopsy immediately corrects the false- positive serology, and the only disbenefits incurred are the costs and disutility associated with an unnecessary endoscopy. |
| Quality of evidence | The GDG discussed and agreed that the evidence ranged from high to low quality due to methodological issues such as small sample size and potential biases in retrospective studies where the intention to biopsy may be been driven by serological results, or strong clinical suspicion. However, the GDG agreed that, in general, the measures taken to ensure optimal quality results, such as the single histological reference criteria, homogeneous European sample, a single suspected coeliac disease population, and the 100% serology and an endoscopic intestinal biopsy in all participants, ensured that evidence obtained was of the best quality available. |
| Other considerations | As detailed above, the GDG noted that IgA tTG alone has suboptimal specificity. Therefore, the group concluded that, in adults, a strategy using IgA EMA to classify cases that are weakly positive on IgA tTG should be recommended. Evidence suggested that this approach has excellent specificity and only slightly reduced sensitivity, and the health economic modelling showed that this represented an optimal trade-off between the risks of false- positive and false-negative diagnoses (see below). However, the GDG was aware that evidence on diagnostic accuracy of serological tests can only reflect standards that have been achieved historically. With developments in IgA tTG assays in mind, the GDG noted that it might be possible for a laboratory to demonstrate through its internal validation processes that it is able to achieve levels of specificity from a single IgA tTG test that are comparable with the expected specificity of the 2-step strategy (that is, 99% or better). If this were the case, the GDG thought it would be reasonable for such a laboratory to rely on IgA tTG assays alone, as that approach would benefit from better sensitivity, and specificity would not be compromised. However, the group chose not to make an explicit recommendation detailing this eventuality, as it believed that the circumstances would be rare. The GDG noted that its recommendations for adults and children were closely comparable, with the only difference coming in the handling of weakly positive IgA tTG results – adults with this finding would be subject to an additional serological test before referral to a gastrointestinal specialist, whereas children would be referred straight away. The group suggested that, in practice, the paediatricians to whom children with weakly positive IgA tTG results are referred may choose to perform additional tests such as IgA EMA in the same way that the group had recommended for adults. However, in the absence of evidence of this approach in children, the group felt that a prescriptive recommendation for either primary care professionals or secondary care specialists would not be helpful. Nevertheless, it was felt that the emergence of these related recommendations from discrete evidence-bases and modelling with somewhat different assumptions provided a degree of convergent validity. |
5.2.7. Recommendations & research recommendations
5.2.7.1. Recommendations
- 4.
All serological tests should be undertaken in laboratories with clinical pathology accreditation (CPA) or ISO15189 accreditation.
- 5.
When healthcare professionals request serological tests to investigate suspected coeliac disease in young people and adults, laboratories should
- test for total immunoglobulin A (IgA) and IgA tissue transglutaminase (tTG) as the first choice
- use IgA endomysial antibodies (EMA) if IgA tTG is weakly positive
- consider using IgG EMA, IgG deamidated gliadin peptide (DGP) or IgG tTG if IgA is deficienta.
- 6.
When healthcare professionals request serological tests to investigate suspected coeliac disease in children, laboratories should:
- test for total IgA and IgA tTG as the first choice
- consider lgG EMA, lgG DGP or lgG tTG if IgA is deficienta.
- 7.
When laboratories test for total IgA, a specific assay designed to measure total IgA levels should be used.
- 8.
Refer young people and adults with positive serological test resultsh to a gastrointestinal specialist for endoscopic intestinal biopsy to confirm or exclude coeliac disease.
- 9.
Refer children with positive serological test results to a paediatric gastroenterologist or a paediatrician with a specialist interest in gastroenterology for further investigationi for coeliac disease
- 10.
Refer people with negative serological test results to a gastrointestinal specialist for further assessment if coeliac disease is still clinically suspected.
- 11.
Healthcare professionals should have a low threshold for re-testing people identified in recommendations 1 or 2 if they develop any symptoms consistent with coeliac disease.
- 12.
Laboratories should clearly communicate the interpretation of serological test results and recommended action to healthcare professionals.
- 13.
Do not use human leukocyte antigen (HLA) DQ2 (DQ2.2 and DQ2.5)/DQ8 testing in the initial diagnosis of coeliac disease in non-specialist settings.
- 14.
Only consider using HLA DQ2 (DQ2.2 and DQ2.5)/DQ8 testing in the diagnosis of coeliac disease in specialist settings (for example, in children who are not having a biopsy, or in people who already have limited gluten ingestion and choose not to have a gluten challenge).
5.2.7.2. Research Recommendations
1. What is the sensitivity and specificity of IgA DGP and IgG DGP in the detection of coeliac disease in children aged under 2 years?
Why this is important
The deamidated gliadin peptide (DGP) antibodies are emerging as promising antibodies for the detection of coeliac disease. There is evidence which suggests that these antibodies may be particularly useful in children under the age of two years old (Mubarak et al., 2011). Further research into the sensitivity and specificity of the DGP antibodies in children under two years of age will strengthen this preliminary evidence and may lead to these antibodies being the first point of call in serological testing for coeliac disease in children under two years old.
2. What is the sensitivity and specificity of IgG tTG, IgG EMA and IgG DGP tests in detecting coeliac disease in people with IgA deficiency?
Why this is important
IgA deficiency is significantly more common in people with coeliac disease than in the general population. People with IgA deficiency will have a false negative result when tested for IgA antibody, which may lead to a missed diagnosis of coeliac disease. A missed diagnosis may result in increased use of NHS resources and the person experiencing the risks associated with undiagnosed coeliac disease. IgG antibodies are recommended for use in place of IgA antibodies in people who have IgA deficiency, but there is limited evidence to demonstrate the sensitivity and specificity of tests for IgG antibodies – that is, IgG tTG, IgG EMA and IgG DGP – in people suspected of having coeliac disease with IgA deficiency.
3. What is the sensitivity and specificity of IgA EMA and IgA DGP tests in detecting coeliac disease in people who test negative for IgA tTG?
Why this is important
In people with suspected coeliac disease, IgA tTG is most commonly used as the first-choice test to detect the presence of coeliac disease antibodies but some people with coeliac disease will get a false negative result this happens, and if there is a strong and ongoing clinical suspicion of coeliac disease, serological testing for IgA EMA or IgA or IgG DGP antibodies should also be requested. However, there is little evidence for the sensitivity and specificity of these antibodies in people who have tested negative for IgA tTG antibodies. A clearer understanding of the sensitivity and specificity of EMA and DGP antibodies in people who have tested negative for IgA tTG will allow clinicians to better interpret test results and make a more informed diagnosis.
5.3. Criteria for referral for endoscopic intestinal biopsy
5.3.1. Review question
What are the referral indications for endoscopic intestinal biopsy for further investigation in people with coeliac disease?
Once diagnosed and treated with a gluten-free diet, most people with coeliac disease experience significant improvement in their clinical symptoms, which typically resolve after 3 to 6 months. In some circumstances, people with CD may experience little or no improvement of their symptoms, or a resurgence of clinical symptoms after a period of resolution. In these situations an endoscopic biopsy may be required for further investigation.
5.3.2. Methods
The aim of this review question was to determine what factors may indicate appropriate referral for endoscopic intestinal biopsy for people with coeliac disease.
Studies were only included if the population examined were adults or children with a diagnosis of coeliac disease who were being monitored while on a gluten-free diet and in whom an endoscopic intestinal biopsy may be useful for further investigation.
Outcomes of interest were as follows: complications of coeliac disease; mortality; health-related quality of life; and resource use and cost.
For full details of the review protocol please see Appendix C.
A systematic search was conducted (see Appendix C) which identified 925 references. The references were screened on their titles and abstracts and full text papers of 20 references were obtained and reviewed against the inclusion and exclusion criteria in the review protocol (see Appendix C).
All 20 studies were excluded as they did not meet the eligibility criteria such as inappropriate study design (case-studies), not a primary study (descriptive narrative, opinion, etc.), or studies in which an endoscopic intestinal biopsy was being conducted for the purposes of initial diagnoses rather than for follow-up investigation. A detailed list of excluded studies and reasons for their exclusion is provided in appendix F.
No data were available in relation to the a-priori specified outcomes of interest. A post-hoc decision was made by the GDG to analyse the following data, which was available from other related review questions (see sections 5.4 and 6.1);
- Change in serological markers on routine monitoring on a gluten-free diet as an indication of histological recovery
- Presenting symptoms of nonresponsive coeliac disease (NRCD) as an indication for the need for further assessment
As these outcomes of interest directly relate to review questions in sections 5.4 (monitoring of people with coeliac disease) and 6.1 (presenting features of non-responsive coeliac disease), relevant evidence was extrapolated from each of these review questions to inform the present review. Overall, 5 papers from the present literature review met the inclusion criteria and were included for analyses. A further 5 studies from review question in section 5.4, and 4 studies from review question 6.1 met the inclusion criteria and were therefore included within the present review.
Cohort studies were considered the best quality evidence for this question and were therefore considered high quality according to a modified GRADE framework. The quality for each outcome could be downgraded due to risk of bias in terms of methods, inconsistency between studies, indirectness in terms of population, tests and outcomes used, or imprecision in terms of outcomes.
Data were extracted into detailed evidence tables (see Appendix D)
5.3.3. Evidence review
5.3.3.1. Resolution of gastrointestinal and non-gastrointestinal symptoms
Two studies (Dickey et al, 2000; Midhagen et al., 2004) contributed data to the analysis. Dickey (2000) included 53 young people and adults (16 to 81 years of age). None of the study population had IgA deficiency. Midhagen (2004) included 21 adults but was unclear on the age range of participants. None of the study population had IgA deficiency.
5.3.3.2. Change in IgA EMA while on GFD
Four studies (Dickey et al., 2000, Fotoulaki et al., 1999, Midhagen et al., 2001, Trigoni et al., 2014) contributed data to this analysis. Samples sizes ranged from 17 to 70, people with coeliac disease with positive EMA at diagnosis, the percentage with IgA deficiency ranged from 0% to 10%. The age of study participants ranged from 1 to 86 years. The study participants were on a gluten-free diet for between 3 months and 3 years.
5.3.3.3. Change in IgA anti-reticulin antibodies (IgA ARA) while on GFD
A single study (Fotoulaki et al., 1999) contributed data to this analysis. This study included 30 children, young people and adults (age range 1 to 24 years) with coeliac disease diagnosed using ESPGHAN criteria, and 3 (10%) had IgA deficiency. The study participants were on a gluten-free diet for up to 12 months.
5.3.3.4. Change in IgA tTG while on GFD
Four studies (Martin-Pagola et al., 2007, Midhagen et al., 2001, Samasca et al., 2011, Trigoni et al., 2014) contributed data to this analysis. Samples sizes ranged from 14 to 93, people with coeliac disease. None of the study populations had IgA deficiency. Two studies included children and young people (0.95 to 17.5 years of age but the range was not reported in the second study) and 2 included adults (19 to 86 years). The study participants were on a gluten-free diet for between 3 months and 3 years.
5.3.3.5. Change in IgG tTG while on GFD
One study (Martin-Pagola et al., 2007) contributed data to this analysis. This study included 93 children and young people (aged between 0.95 to 17.5 years diagnosis) of whom none of the study populations had IgA deficiency. The study participants had been on a gluten-free diet for an average of 24 months at study follow-up.
5.3.3.6. Change in IgA AGA while on GFD
A single study (Midhagen et al., 2004) contributed data to this analysis. This study included adults (age range 29 to 86 years) with coeliac disease who tested positive for IgA AGA at diagnosis. It was unclear how many participants had IgA deficiency. The study participants were on a gluten-free diet for up to 12 months.
5.3.3.7. Proportion of patients suffering persistent symptoms whilst on gluten-free diet
Four studies (Dewar et al., 2010; Leffler et al., 2007; Abdulkarim et al., 2002; Van Weyenberg et al., 2013) contributed data to this analysis. These studies included adults (age range 29 to 79 years) with coeliac disease who presented with persistent symptoms despite being on a gluten-free diet for a minimum of 12 months.
5.3.3.8. Presenting symptoms of non-responsive coeliac disease (NRCD)
Four studies (Dewar, 2012; Leffler, 2007; Abdulkarim, 2002; Van Weyenberg, 2013) contributed data to this analysis. Samples sizes ranged from 48 to 113 adults with non-responsive coeliac disease. The mean age of study participants ranged from 42 to 63 years. The study participants were on a gluten-free diet for between 6 months and 6 years.
5.3.4. Health economic evidence
An economic evaluations filter was applied to the search protocol for this research question with the aim of finding economic evaluations that explored the cost effectiveness of referral indications for endoscopic intestinal biopsy for further investigation of people with coeliac disease.
The search identified 97 references. The references were screened on their titles and abstracts however none of the studies met the inclusion criteria.
No cost–utility analyses were found to address selection criteria
5.3.5. Evidence statements
5.3.5.1. Evidence for proportion in clinical remission while on GFD
Very low quality evidence from 2 studies (N = 71) found that between 89% and 91% of people diagnosed with coeliac disease were in clinical remission after 12 months on a gluten-free diet
5.3.5.2. Evidence for proportion with negative IgA EMA while on GFD
Very low quality evidence from 2 studies (N = 60) found that between 41% and 58% of adults with coeliac disease were IgA EMA negative after 3 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 57% of mixed age-groups were IgA EMA negative after 3 months on a gluten-free diet
Very low quality evidence from 3 studies (N = 130) found that between 39% and 75% of adults with coeliac disease were IgA EMA negative after 6 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 93%% of mixed age-groups were IgA EMA negative after 6 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 90% of mixed age-groups were IgA EMA negative after 9 months on a gluten-free diet
Very low quality evidence from 3 studies (N = 130) found that between 73% and 87% of adults with coeliac disease were IgA EMA negative after 12 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 100% of mixed age-groups were IgA EMA negative after 6 months on a gluten-free diet
Very low quality evidence from a single study (N = 70) found that 94% of adults were IgA EMA negative after 6 months on a gluten-free diet
5.3.5.3. Evidence for proportion with negative IgA ARA while on a GFD
Very low quality evidence from a single study (N = 30) found that 77% of adults with coeliac disease were IgA ARA negative after 3 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 87% of adults with coeliac disease were IgA ARA negative after 6 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 100% of adults with coeliac disease were IgA ARA negative after 9 months on a gluten-free diet
Very low quality evidence from a single study (N = 30) found that 100% of adults with coeliac disease were IgA ARA negative after 12 months on a gluten-free diet
5.3.5.4. Evidence for proportion with negative IgA tTG while on a GFD
Very low quality evidence from a single study (N = 50) found that 68% of children with coeliac disease were IgA tTG negative after 3 months on a gluten-free diet.
Low quality evidence from a single study (N = 14) found that 57% of adults with coeliac disease were IgA tTG negative after 3 months on a gluten-free diet.
Very low quality evidence from 2 studies (N = 143) found that between 49% and 68% of children with coeliac disease were IgA tTG negative after 6 months on a gluten-free diet.
Low quality evidence from 2 studies (N = 84) found that between 20% and 71% of adults with coeliac disease were IgA tTG negative after 6 months on a gluten-free diet.
Very low quality evidence from a single study (N = 50) found that 82% of children with coeliac disease were IgA tTG negative after 12 months on a gluten-free diet.
Very low quality evidence from 2 studies (N = 84) found that between 49% and 100% of adults with coeliac disease were IgA tTG negative after 12 months on a gluten-free diet.
Very low quality evidence from a single study found that 88%% of children with coeliac disease were IgA tTG negative after 24 months on a gluten-free diet.
Very low quality evidence from a single study (N = 70) found that 80% of adults with coeliac disease were IgA tTG negative after 3 years on a gluten-free diet.
5.3.5.5. Evidence for proportion with negative IgG tTG while on a GFD
Very low quality evidence from a single study (N = 93) found that 63% of children with coeliac disease were IgG tTG negative after 6 months on a gluten-free diet.
Low quality evidence from a single study (N = 93) found that 96% of children with coeliac disease were IgG tTG negative after 24 months on a gluten-free diet
5.3.5.6. Evidence for proportion with negative IgA AGA while on a GFD
Low quality evidence from a single study (N = 15) found that 74% of adults with coeliac disease were IgA AGA negative after 3 months on a gluten-free diet
Low quality evidence from a single study (N = 15) found that 93% of adults with coeliac disease were IgA AGA negative after 6 months on a gluten-free diet
Low quality evidence from a single study (N = 15) found that 100% of adults with coeliac disease were IgA AGA negative after 12 months on a gluten-free diet
5.3.5.7. Evidence for symptom presentation of nonresponsive coeliac disease (NRCD)
Five high quality studies found that between 50 – 84% of people experiencing NRCD experience persistent diarrhoea while following a gluten-free diet for a period of at least 6 months.
Five high quality studies found that between 14 – 55% of people experiencing NRCD experience persistent abdominal pain while following a gluten-free diet for a period of at least 6 months.
Five high quality studies found that between 5 – 47% of people experiencing NRCD experience persistent weight loss while following a gluten-free diet for a period of at least 6 months.
Three high quality studies found that between 5 – 43% of people experiencing NRCD experience persistent lethargy while following a gluten-free diet for a period of at least 6 months.
Two high quality studies found that between 10 – 17% of people experiencing NRCD experience persistent nausea with or without vomiting while following a gluten-free diet for a period of at least 6 months.
Three high quality studies found that between 4 – 37% of people experiencing NRCD experience persistent anaemia while following a gluten-free diet for a period of at least 6 months.
5.3.6. Evidence to recommendations
| Relative value of different outcomes | The GDG discussed and agreed that currently there was very limited evidence on the a priori selected outcomes of complications of CD, mortality, health-related quality of life, and resource use and cost, in relation to indications for endoscopic intestinal biopsy for further investigation in those with CD. The value of using serological testing to drive indication for biopsy was discussed extensively in terms of the low sensitivity and specificity of these tests as markers of diet adherence and histological recovery (see section 5.4). For example, symptoms may disappear while on a gluten- free diet however an endoscopic intestinal biopsy shows no histological recovery. The value of using symptomology to drive indication for an endoscopic intestinal biopsy was also discussed in detail. It was recognised that not all people with coeliac disease have symptoms at diagnosis, and therefore, symptoms are not a useful marker of clinical response to gluten-free diet. These people may require an endoscopic intestinal biopsy to ensure that there is histological response to the gluten-free diet. The group also cited published research and anecdotal experience of cases that had shown symptomatic improvement despite patients having persistent severely damaged villi, or vice versa, suggesting that clinical improvement should also not be relied upon as an indication for biopsy. |
|---|---|
| Trade-off between benefits and harms | The benefit of routine monitoring was discussed and the group agreed this was useful in order to identify those in whom a gluten-free diet is not having an optimal outcome. (See section 5.4). It was discussed by the GDG that often clinicians will chose to re-biopsy patients after 18 – 24 months as part of a monitoring strategy. Persistently high titres in serology can indicate when further investigations are needed; however persistently high titres may also be misleading and be unrelated to histological outcome. The group agreed however, that in a person with persistently high antibodies, an endoscopic intestinal biopsy would be useful to inform the full clinical picture. People who present with symptoms of non-responsive coeliac disease should be referred to an expert dietitian, as the most common cause of persisting clinical symptoms is gluten ingestion (see section 6.1). Due to the invasive nature of biopsy, the group agreed that a person with non- responsive CD should only be considered for re-biopsy for further investigation after continued gluten ingestion has been ruled out. Most people with coeliac disease report clinical improvement within 2–6 months; anything beyond this was agreed to represent an outlier. |
| Economic considerations | No economic evidence on referral indications for endoscopic intestinal biopsy for further investigation of coeliac disease was found. |
| Quality of evidence | Overall the evidence identified for serological monitoring was of very low quality. This is because although it is possible to design a randomised controlled trial comparing two different monitoring strategies. No such study was identified and only lower quality evidence with design limitations was used in this review. The quality of evidence for presenting symptoms for non-responsive coeliac disease was high. |
| Other considerations | Current clinical practice in the UK is to monitor with serology on an annual basis. Biopsies are often repeated 12 to 24 months after diagnosis. The GDG did not consider the evidence sufficient to change current practice regarding serology and biopsy, despite the fact that they did not believe that serology was an accurate biomarker for response to gluten-free diet. |
5.3.7. Recommendations & research recommendations
- 15.
Consider referring people with coeliac disease for endoscopic intestinal biopsy if continued exposure to gluten has been excluded and:
- Serological titres are persistently high and show little or no change after 12 months or
- they have persistent symptoms, including diarrhoea, abdominal pain, weight loss, fatigue or unexplained anaemia.
5.4. Routine monitoring
5.4.1. Review questions
How frequently should people with coeliac disease be routinely monitored?
Should the frequency of routine monitoring differ for patients at risk of developing certain complications?
What should routine monitoring consist of?
There exists a great variation in current clinical practice in terms of what constitutes the routine monitoring of those with coeliac disease. Some centres may monitor patients more frequently and choose to biopsy patients within a year of diagnosis to assess histological recovery, while other centres may offer initial dietetic advice and follow-up patients only in the event of continued complications. The relationship between frequency and type of monitoring and the clinical symptoms of celiac disease is yet to be explored. Optimal follow-up and monitoring should aim to aid those with coeliac disease to achieve histological and symptomatic recovery. The aim of this review is therefore to identify what optimal monitoring practice should consist of.
5.4.2. Methods
The aim of these review questions was to determine how often people with coeliac disease should be followed up. The second component of this question was designed to investigate if any subgroups at risk of developing any particular complications of coeliac disease should be followed up more frequently. The final component of this question examined what assessments and checks should be carried out to monitor coeliac disease, particularly in those at risk of developing complications.
Studies were only included if the population examined included people of any age who were diagnosed with coeliac disease and who were being monitored while on a gluten-free diet. GDG-selected outcomes of interest were as follows; resolution of gastrointestinal and non-gastrointestinal symptoms, growth in children and young people, complications of coeliac disease, dietary adherence, impact on carers and health-related quality of life. All forms of monitoring were considered appropriate for inclusion.
For full details of the review protocol please see Appendix C.
A systematic search was conducted (see Appendix C) which identified 4851 references. The references were screened on their titles and abstracts and full text papers of 63 references were obtained and reviewed against the inclusion and exclusion criteria in the review protocol (see Appendix C).
Fifty three studies were excluded as they did not meet the eligibility criteria such as inappropriate study design (case-series with variable length of follow-up), not a primary study (descriptive narrative, opinion, etc.). A detailed list of excluded studies and reasons for their exclusion is provided in Appendix F.
Randomised controlled trials were considered to be the highest quality evidence available to answer this question and are graded as high in a GRADE framework. No randomised controlled trials were identified so studies with the next best design (cohort studies) for this question were used. The quality for each outcome could be downgraded due to risk of bias in terms of methods, inconsistency between studies, indirectness in terms of population, tests and outcomes used, or imprecision in terms of outcomes.
Ten papers met this revised criteria and were included. Another 2 studies were identified through searches for other questions or reference checking and so a total of 12 studies were included. No randomised controlled trials were identified, and cohort studies were rated as low quality. Data were extracted into detailed evidence tables (see Appendix D)
Data were available for 2 of the a priori GDG specified outcomes resolution of gastrointestinal and non-gastrointestinal symptoms and dietary adherence but no data were identified for the other four GDG requested outcomes.
A post-hoc decision was taken to analyse the following as data were available;
- Change in serological markers on routine monitoring on a gluten-free diet
- Change in histology on routine monitoring on a gluten-free diet
- Nutritional status while on gluten-free diet
- Healthcare professionals involved in routine monitoring
5.4.3. Evidence review
5.4.3.1. Resolution of gastrointestinal and non-gastrointestinal symptoms
Two studies (Dickey et al., 2000; Midhagen et al., 2004) contributed data to the analysis. Dickey et al., 2000 included 53 young people and adults (16 to 81 years of age), of whom 39 (74%) were female. None of the study population had IgA deficiency. Midhagen and colleagues (2004) included 21 adults but was unclear on the proportion of female participants and the age range. None of the study population had IgA deficiency.
5.4.3.2. Adherence to gluten-free diet
Five studies (Dickey et al., 2000, Monzani et al., 2011, Trigoni et al., 2014, Zanchi et al., 2013; Galli et al., 2014) contributed data to this analysis. The sample sizes ranged from 28 to 315 and the proportion of female participants ranged from 61% to 74%. None of the study populations had IgA deficiency. The studies had mixed age groups with Monzani et al.,(2011) including children, (1 to 16.8 years of age), Zanchi et al., (2013) including children and adults (6 to 45 years of age) while the remaining 3 studies included only adults (16 to 81 years of age).
5.4.3.3. Growth in children and young people
- No studies reported this outcome
5.4.3.4. Complications of coeliac disease
- No studies reported this outcome
5.4.3.5. Impact on carers
- No studies reported this outcome
5.4.3.6. Health-related quality of life
- No studies reported this outcome
5.4.3.7. Diagnostic accuracy to detect non-adherence
Two studies (Monzani et al., 2011, Zanchi et al., 2013) reported on the accuracy of serological tests to detect non-adherence to a gluten-free diet. Monzani et al., 2011 included 28 children and young people, of whom 17 (61%) were female between the ages of 1 and 16.8 years. None of the children had IgA deficiency. Zanchi et al., (2013) included 315 children and adults with an age range of 6 to 45 years of whom 227 (65%) female. IgA deficiency status was not reported. The findings of both studies are summarised in table 1 below.
5.4.3.8. Change in IgA EMA while on GFD
Four studies (Dickey et al., 2000, Fotoulaki et al., 1999, Midhagen et al., 2001, Trigoni et al., 2014) contributed data to this analysis. Samples sizes ranged from 17 to 70, including people with coeliac disease with positive EMA at diagnosis, the percentage of female participants ranged from 55% to 74%; the percentage with IgA deficiency ranged from 0% to 10%. The age of study participants ranged from 1 to 86 years. The study participants were on a gluten-free diet for between 3 months and 3 years.
5.4.3.9. Change in IgA anti-reticulin antibodies (IgA ARA) while on GFD
A single study (Fotoulaki et al., 1999) contributed data to this analysis. This study included 30 children, young people and adults (age range 1 to 24 years) with coeliac disease diagnosed using ESPGHAN criteria. 17(57%) were female and 3 (10%) had IgA deficiency. The study participants were on a gluten-free diet for up to 12 months
5.4.3.10. Change in IgA tTG while on GFD
Four studies (Martin-Pagola et al., 2007, Midhagen et al., 2001, Samasca et al., 2011, Trigoni et al., 2014) contributed data to this analysis. Samples sizes ranged from 14 to 93, people with coeliac disease. The percentage of females ranged from 54.5% to 71% and none of the study participants had IgA deficiency. Two studies included children and young people (0.95 to 17.5 years of age but the range was not reported in the second study) and 2 studies included adults (19 to 86 years). The study participants were on a gluten-free diet for between 3 months and 3 years
5.4.3.11. Change in IgG tTG while on GFD
One study (Martin-Pagola et al., 2007) contributed data to this analysis. This study included 93 children and young people (0.95 to 17.5 years of age at time of diagnosis) of whom 62% were female and none of the study participants had IgA deficiency. The study participants were on a gluten-free diet for 24 months at the time of follow-up.
5.4.3.12. Change in IgA AGA while on GFD
A single study (Midhagen et al., 2004) contributed data to this analysis. This study included adults (age ranged from 29 to 86 years) with coeliac disease who tested positive for IgA AGA at diagnosis. It was unclear how many were female or had IgA deficiency. The study participants were on a gluten-free diet for up to 12 months
5.4.3.13. Change in histology on routine monitoring on a gluten-free diet
Four studies (Dickey et al., 2000, Midhagen et al., 2004, Martini et al., 2002, Galli et al., 2014) contributed data to this analysis. Samples sizes ranged from 18 to 101 people with coeliac disease. The percentage of females ranged from 55% to 78% and none of the study participants had IgA deficiency. The age of study participants ranged from 18 to 86 years. The study participants were on a gluten-free diet for at least 12 months.
5.4.3.14. Nutritional status while on GFD
A single study (Shepherd et al., 2012) contributed data to this analysis. This study included 50 adults (18 to 71 years) diagnosed with coeliac disease of whom 38 (76%) were female. All were adherent to a gluten-free diet.
5.4.3.15. Health care professional involvement
A single study (Wylie et al., 2005) contributed data to this analysis. This study included 99 adults (23 to 86 years of ages) diagnosed with coeliac disease of whom 69 (70%) were female. This study examined the change after the introduction of a dietitian to coeliac disease team.
5.4.4. Health economic evidence
5.4.4.1. Systematic review of published cost–utility analyses
An economic evaluations filter was applied to the search protocol for this research question with the aim of finding economic evaluations that examined the cost effectiveness of monitoring people with coeliac disease.
The search identified 632 references. The references were screened on their titles and abstracts and three full texts were ordered. None of the studies met the inclusion criteria.
No cost–utility analyses were found to address selection criteria.
5.4.4.2. Original health economic analysis
An original cost–utility model was used to explore the benefits, harms and costs associated with serological investigation of people at increased risk of coeliac disease. A modified version of the model developed to analyse the serological investigation of people with symptoms suggestive of coeliac disease was used (see section 5.2.4.2). The long-term consequences of disease were modelled, with the single parameter of adherence to GFD varied to reflect the effectiveness of dietitian-led follow-up (as reported by Wylie 2005). Full details of the methods and results of the model are provided in Appendix G.
5.4.5. Evidence statements
5.4.5.1. Evidence for resolution of gastrointestinal and non-gastrointestinal symptoms
Very low quality evidence from 2 studies (N = 71) found that 90.1% (95%CI 80.7% to 95.2%) of people diagnosed with coeliac disease were in symptomatic remission after 12 months on a gluten-free diet
5.4.5.2. Evidence for dietary non-adherence on GFD
Very low quality evidence from four studies (N = 486) found that 23.6% (95%CI 9.2 to 48.5%) of people diagnosed with coeliac disease were not adhering to a gluten-free diet.
5.4.5.3. Evidence for diagnostic accuracy to detect partial adherence to GFD
Very low to low evidence from a single study (N = 28) reported that DGP IgA/G had high sensitivity to discriminate between strictly adherent and partially adherent at 6 – 8 months and 9 – 12 months. The same study reported that anti TTG IgA had high sensitivity at 2 – 4 months and that AGA IgA was highly sensitive at 2 – 4 months. Where calculable, all other levels of sensitivity and specificity were low.
Very low to low evidence from a single study (N = 315) reported that both the anti TTG ELISA test and a ‘rapid’ version had low sensitivity and high specificity to discriminate between strictly adherent and partially adherent at 24 months.
5.4.5.4. Evidence for proportion with negative IgA EMA while on GFD
Very low to low quality evidence from 4 studies (N = 150) including people with newly-diagnosed coeliac disease found that the proportion who tested negative for IgA EMA antibodies increased over time on a gluten-free diet. The proportion ranged from between 30.3% (95%CI 17.3%, 47.4%) at 3 months to between 89.4% (95%CI 82.2% to 93.9%) at 12 months.
5.4.5.5. Evidence for proportion with negative IgA ARA while on a GFD
Low quality evidence from a single study (N = 30) including people with newly-diagnosed coeliac disease found that the proportion who tested negative of IgA ARA antibodies increased over time on a gluten-free diet. The proportion ranged from 76.7% (95% CI 57.3% to 89.4%) at 3 months to 100% (No 95% CI) at 12 months.
5.4.5.6. Evidence for proportion with negative IgA tTG while on a GFD
Very low quality evidence from 3 studies (N = 115) including people with newly-diagnosed coeliac disease found that the proportion who tested negative of IgA tTG antibodies increased over time on a gluten-free diet from 85%% (95%CI 53.1 to 76.1%) at 3 months to 76.5% (95% CI 42.2% to 93.6%) at 12 months.
5.4.5.7. Evidence for proportion with negative IgG tTG while on a GFD
Low quality evidence from a single study (N = 93) including people with newly-diagnosed coeliac disease found that the proportion who tested negative of IgG tTG antibodies increased over time on a gluten-free diet. The proportion ranged from 63.4% (95%CI 52.8% to 73.0%) at 6 months to 96.7% (95%CI 90.2% to 99.2%) at 24 months.
5.4.5.8. Evidence for proportion with negative IgA AGA while on a GFD
Low quality evidence from a single study (N = 15) including people with newly-diagnosed coeliac disease found that the proportion who tested negative of IgG AGA antibodies increased over time on a gluten-free diet. The proportion ranged from 60.0% (95% CI 32.9% to 82.5%) at 3 months to 100% (No CI) at 12 months.
5.4.5.9. Evidence for histological recover while on a GFD
Low quality evidence from four studies (N = 237) including people with newly-diagnosed coeliac disease found between 56% and 89% demonstrated either mucosal recover or improvement in Marsh grading.
5.4.5.10. Evidence for nutritional status while on a GFD
Very low quality evidence from a single study (N = 50) found that 10% of adults with coeliac disease had nutritionally deficient diets after 12 months on a gluten-free diet
5.4.5.11. Evidence for healthcare professional involvement in monitoring people with coeliac disease
Very low quality evidence from a single study (N = 99) found an increase of 12% in those adhering to a gluten-free diet after 12 months with a dietitian-led coeliac disease clinic. The same study (N = 80) found a 52% increase in those satisfied with their care after 12 months with a dietitian-led coeliac disease clinic.
5.4.5.12. Health economic evidence statements
An original, partially applicable cost–utility analysis with potentially serious limitations suggested that the introduction of a dietitian-led coeliac disease clinic results in increased health benefits at increased cost, with an ICER of £15,200 per QALY gained. The model was reliant on a single, very low-quality study for its effectiveness evidence.
5.4.6. Evidence to recommendations
| Relative value of different outcomes | The GDG discussed and agreed that currently there was very limited evidence on the a priori selected outcomes such as resolution of gastrointestinal and non-gastrointestinal symptoms and adherence to a gluten-free diet. Both outcomes were deemed to be critical but also difficult to interpret on their own. For example, symptoms may disappear while on a gluten-free diet but biopsy results still show no change. Adherence to a gluten free-diet is also difficult to monitor accurately as commonly used tools such as 3- or 7-day diaries and self-report questionnaires are not very accurate as people may believe that they are adherent but inadvertently are ingesting some gluten. Sensitivity to gluten is an important factor in terms of serology or biopsy monitoring as some people with coeliac disease are highly sensitive to gluten while others are less so. |
|---|---|
| Trade-off between benefits and harms | One of the major benefits of routine monitoring is the increased level of contact between the person with coeliac disease and healthcare professionals. This facilitates the provision of information and ongoing support essential for a condition such as coeliac disease. The value of knowing that the gluten-free diet appears to be working, through reduction in antibody titres over time, can be reassuring to people with coeliac disease and/or their carers. This is a powerful motivation factor in helping people with coeliac disease adhere to a gluten-free diet. Routine monitoring can also identify those in whom the gluten-free diet is not having an optimal outcome (see sections 5.3 and 6.1). Persistently high serological titres can indicate when further investigations are indicated and, in many cases, when further dietary education and counselling is needed (see section 7). However, the group discussed that the evidence does not indicate a strong nor conclusive relationship between serological titres and dietary adherence, in that a number or patients who were shown to be strictly adhering to a GFD still had persistently high serological titres. The GDG also discussed shared clinical experience in which patients who were very strict adherers to the GFD still had high serological titres. In these patients, serological testing does not accurately reflect histological recovery. It can therefore be potentially harmful to the patient’s adherence to the GFD when they feel that they are doing everything in their power to exclude gluten from their diet and this is not being reflected in their serological testing. The group therefore recognised that serology may be used to inform a clinical picture of a patient, but it should not be used alone to determine GFD adherence. The group raised the notion that it is also important to consider what the routine monitoring consists of. Satisfaction with routine monitoring increased in one study when a dietitian was involved. A qualitative study cited by the GDG (Rajani et al., 2013) also supports this evidence and found that regular clinics with the full coeliac disease multi- disciplinary lead to increase satisfaction with the service. The chapter on nutritional status and diet (section 7.3) also highlighted a need for additional nutritional supplements particularly in newly diagnosed people with coeliac disease. The GDG agreed that nutritional deficiencies in diet should be identified through appropriate ongoing monitoring based on an individual’s needs and supplementations should not be initiated without a full assessment from the healthcare professional team including the dietitian. |
| Economic considerations | The GDG understood that the original health economic modelling undertaken for this question was exploratory in nature, and totally reliant on a single parameter from a very low-quality study (Wylie et al., 2005) to estimate the effectiveness of dietitian-led follow-up (in terms of improved adherence to GFD). The GDG discussed that, if the improvement in adherence to GFD reported by Wylie et al. (2005) can be believed, then dietitian-led follow- up is very likely to be a cost-effective strategy. However, the shortcomings of this evidence make it difficult to be confident of the size of effect that would be seen in practice. Therefore, the GDG concluded it was critical to recommend additional research on this topic. The GDG also noted that the package of follow-up care reported by Wylie et al. (2005) comprised multiple elements, including dietetic review, DEXA scanning, blood tests and gastroenterological referral for a proportion of patients. It was not possible to identify what contribution each of these components made to the reported effect. However, when it came to the outcome that was critical to the health economic model – adherence to GFD – the GDG was content to assume that the involvement of a dietitian was an important factor. |
| Quality of evidence | Overall, the evidence identified for routine monitoring was of very low quality. This is because although it is possible to design a randomised controlled trial comparing two different monitoring strategies, no such study was identified and only lower quality evidence with design limitations was identified. The GDG highlighted, based on their own expertise and clinical experience, that people with coeliac disease benefit from regular dietetic assessment to monitor adherence to the gluten free diet, review symptoms, and provide nutritional advice, and that this would represent a gold-standard for annual review. However, in the current context of a lack of good quality evidence to support this, and the significant implementation ramifications of offering all people with coeliac disease access to a specialist dietitian, the GDG agreed that the recommendation should reflect that dietetic input should be considered as part of an annual review. The GDG highlighted this as a priority for further research to stimulate investigation into the utility of dietetic support as an integral component of an annual review for people with coeliac disease. |
| Other considerations | Current clinical practice in the UK is to monitor with serology on an annual basis. Intestinal biopsies are often repeated 12 to 24 months after diagnosis. The GDG did not consider the evidence sufficient to change current practice regarding serology and biopsy. However, the GDG acknowledged that nutritional deficiencies at baseline may dictate follow up with a dietitian. The GDG also acknowledged that, while dietitians are routinely employed in secondary care, there are fewer in primary care, leading to a resource gap. |
5.4.7. Recommendations & research recommendations
5.4.7.1. Recommendations
- 16.
Do not use serological testing alone to determine whether gluten has been excluded from the person’s diet.
- 17.
Offer an annual review to people with coeliac disease. During the review:
- measure weight and height
- review symptoms
- consider the need for assessment of diet and adherence to the gluten-free diet
- consider the need for specialist dietetic and nutritional advice.
- 18.
Refer the person to a GP or consultant if concerns are raised in the annual review. The GP or consultant should asses all of the following:
- the need for a dual-energy X-ray absorptiometry (DEXA) scan (in line with the NICE guideline on osteoporosis: assessing the risk of fragility fracture) or active treatment of bone disease
- the need for specific blood tests
- the risk of long-term complications and comorbidities
- the need for specialist referral.
5.4.7.2. Research recommendations
4. How can the role of the dietitian contribute most effectively within a coeliac disease team?
Why this is important
As a gluten-free diet is the primary treatment option for people with coeliac disease, it is important that a dietitian with a specialist interest in coeliac disease should play a significant role in their care and follow up. Many of the common problems associated with the long-term management of coeliac disease happen because of non-adherence to a gluten-free diet. It is important to explore how to maximise the effectiveness of the dietitian role in helping people with coeliac disease to adhere to a gluten-free diet.
5. What is the effectiveness of more frequent monitoring compared with monitoring at 12 months after diagnosis in people with newly diagnosed coeliac disease?
Why this is important
It is currently not known how often people with coeliac disease should have their condition monitored. No research adequately investigated the effectiveness of different monitoring frequencies. There is variation across the UK in how often people with coeliac disease have their condition monitored. Further research within this area is important to ensure that people with coeliac disease are having their condition adequately monitored.
Footnotes
- d
If studies used different cut-off levels, those used were that of the manufacturer’s recommended cut-off levels
- e
IgA AGA and IgG AGA were assessed in the previous coeliac disease CG86 guideline. Due to low sensitivity and specificity outcome for these tests, they were not recommended for the diagnosis of coeliac disease. On this basis, IgA AGA and IgG AGA were excluded from the present guideline.
- f
If studies used different cut-off levels, the used were that of the manufacturer’s recommended cut-off levels.
- g
IgA AGA and IgG AGA were assessed in the previous coeliac disease CG86 guideline. Due to low sensitivity and specificity outcome for these tests, they were not recommended for the diagnosis of coeliac disease. On this basis, IgA AGA and IgG AGA were excluded from the present guideline.
- h
In young people and adults, a positive serological result is defined as: unambiguously positive IgA tTG alone, or weakly positive IgA tTG and a positive IgA EMA test result. Note: In people who have IgA deficiency, a serologically positive result can be derived from any one of the IgG antibodies.
- i
Further investigation may include, but is not limited to, one or more of the following: an IgA EMA test to confirm serological positivity, HLA genetic testing, an endoscopic biopsy
- Evidence for testing for coeliac disease - Coeliac DiseaseEvidence for testing for coeliac disease - Coeliac Disease
Your browsing activity is empty.
Activity recording is turned off.
See more...