This expert review article highlights the progress that has occurred in the development and application of serological tests used in clinical practice in the diagnosis and management of coeliac disease (CD). It also examines the possibility of serological tests obviating the need for a duodenal biopsy in the diagnosis of adult CD in the future.
Diagnosis of CD has been based on clinical symptoms and duodenal biopsies since the 1950’s. In the late 1950’s, the first serological test for CD, which measured circulating IgA and IgG anti-gliadin antibodies (AGA), became available. This became a first-line step in the diagnosis of CD to help determine which patients should then undergo a biopsy until the early 2000s. In the last two decades the AGA test has been superseded with more accurate antibody tests- endomysial antibody (EMA), tissue transglutaminase (tTG) and deamidated gliadin peptide (DGP) antibodies.
The EMA is regarded by some as the superior serological test for CD because of its high sensitivity (>86%) and specificity (~100%) for the disease in adults. However, EMAs are detected by indirect immunofluorescence, which has limitations including being more costly, labour intensive and open to inter-observer variability compared with the enzyme-linked immunosorbent assay (ELISA) method used to detect AGA, tTG and DGP antibodies. This makes the EMA test a useful confirmatory test alongside tTG rather than the most appropriate first-line test.
In 1997, a study demonstrated that tTG was the target substrate of the EMAs and the autoantigen in CD. There are eight isoforms of the tTG enzyme, which are found at different sites throughout the body. The tTG2 isoform is located in the intestines and is the primary autoantigen in CD; tTG3 is found in the skin and is associated with dermatitis herpetiformis and tTG6 is located in the central nervous system (CNS) and is associated with gluten ataxia. The tTG test is performed using the ELISA method and is cheaper, more objective and can be used in a higher throughput manner than the EMA test. In adult studies using human recombinant tTG, the average sensitivity and specificity are 93.5% and 97.4% respectively. For these reasons, IgA tTG is widely used as the first-line serological test in CD.
A recent meta-analysis showed that the IgA tTG test outperforms the IgA DGP test with pooled sensitivities of 87.8% (95% confidence interval (CI), 85.6-89.9) and 93% (95% CI, 91.2-94.5) and pooled specificities of 94.1% (95% CI, 92.5-95.5) and 96.5% (95% CI, 95.2-97.5) respectively. Whilst marginal, the lower specificity of the IgA DGP test increases the chances of false positives and adds no additional value to the tTG/EMA testing combination in the diagnosis of CD. IgG DGPs are highly specific for CD (upwards of 98%) and this test is reported as a reliable serological marker of disease in IgA deficient patients. Therefore, IgG DGP should be performed in IgA deficient individuals who are suspected of having CD.
Point of Care Tests
There has been a rise in the development of point of care tests (POCTs) that detect IgA tTG or IgA/IgG DP antibodies. Results are rapid and can provide ‘real-time’ feedback to help direct patient investigation and management in settings such as primary care and/or endoscopy and may confer a cost-benefit by reducing the requirement for unnecessary investigations. A recent comparative study showed a combined IgA/IgG DGP-based POCT outperformed two other commercial IgA tTG-based POCTs and had similar performance characteristics to the serology-based tTG tests. Despite this, larger studies in low prevalence populations are required to further assess the place for POCTs in clinical practice.
Follow-up of patients post diagnosis
Serology is often performed during follow-up of patients with CD in those with ongoing symptoms to assist in determining adherence to the diet. However, a recent meta-analysis showed that IgA tTG and EMA levels had a low sensitivity (at, or below, 50%) for detecting persisting villous atrophy on a gluten-free diet (GFD) and therefore do not accurately reflect mucosal inflammation or healing after initiation of a GFD. In the absence of an appropriate serologic test, the current recommendation is to assess adherence and/or persisting mucosal inflammation with a duodenal biopsy.
Serology in subtypes of coeliac disease
Seronegative coeliac disease
Patients with classical symptoms, genetic risk and villous atrophy but negative serological (EMA and tTG) test results are classed as having seronegative coeliac disease (SNCD). The prevalence of SNCD is around 2-5% of coeliac cases. Individuals are IgA competent but fail to show a systemic antibody response to coeliac antigens and diagnosis is based on duodenal biopsy and clinical and histological response to a GFD. The presence of HLA risk also helps to support a diagnosis. Whilst patients with SNCD do not have detectable circulating antibodies to coeliac antigens, anti-tTG antibodies are produced locally in the small intestine and therefore the detection of these in biopsy tissue, supernatants of cultured biopsies or within the faeces, could aid in diagnosis in this group. However, these approaches lack robust supportive data or require a high level of expertise meaning they are not routinely used in practice.
Whilst SNCD is the most common cause of seronegative villous atrophy, it is important patients are not commenced on treatment until other causes have been ruled out such as immune-mediated (e.g. autoimmune enteropathy), inflammatory (e.g. eosinophilic gastroenteritis), infectious (e.g. giardiasis) and drug-related (e.g. Olmesartan) causes.
Potential Coeliac Disease
Individuals with potential CD have circulating antibodies to tTG, EMA and/or DGPs without demonstrable villous atrophy on biopsy. It accounts for approximately 10% of cases of CD. Treatment of this group should be guided by symptoms as symptomatic individuals appear to derive benefit from a GFD, whilst it is less clear if a GFD should be imposed on asymptomatic individuals. A pragmatic approach has been suggested that asymptomatic individuals with potential CD should continue on a gluten-containing diet but with monitoring for the development of symptoms and/or of changes to their serological status, which may prompt further investigation or advice. It is important to note that evidence supports a GFD for asymptomatic patients when villous atrophy is detected on duodenal biopsy.
Could serology replace the need for duodenal biopsies in the diagnosis of adult CD?
Despite being highly accurate, these serological tests are not 100% specific for CD, therefore positive coeliac serology is confirmed by typical histological changes on duodenal biopsy. However, biopsies are expensive, invasive and poorly tolerated by many and the procedure carries some risks.
In 2012, the European Society for the Study of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN)
guidelines for the diagnosis of CD challenged the necessity for duodenal biopsies to diagnose CD in paediatric patients. The guidelines proposed that an IgA tTG antibody titre greater than 10 times the upper limit of normal (ULN), along with a positive IgA EMA test in a second blood sample and compatible HLA genotype was sufficient to
support a diagnosis in symptomatic individuals. The updated guidance, published last year, negate the need for the presence of symptoms and HLA testing in this diagnostic pathway.
The studies looking at whether this approach can be applied to diagnosis in adult CD have shown some promise, however, many of these are subject to methodological flaws including subjects with a high pretest probability for CD being used, which can falsely elevate the predictive value of the diagnostic pathway. Large prospective studies are required to address this and other issues to inform as to whether this is an appropriate approach in adult CD.
The applicability of the uniform 10 x ULN threshold across different sites is one of the major challenges of implementation of this pathway. Different commercial assays, or the same assay used at different sites, can show a wide dispersion of results meaning that a universal threshold may have different implications for individuals tested at different locations. Solutions could include assay cut-offs being validated at a local level, although this is complex, or, alternatively, standardisation of tTG assays across different sites enabling the setting of a more refined cut-off value. The feasibility of either has yet to be established.
Reliance on EMA testing in a second blood sample also needs to be considered as it has been suggested that the future availability of this test might be limited due to the issues relating to the test. Incorporation of DGP testing as an alternative could be considered as simultaneous double positivity for IgA tTG and IgG DGP has a high specificity for CD in both adult and paediatric populations.
It is important to note that there remain reasons to perform endoscopies in adult patients with very high tTG levels and this proposed approach may not be suitable for all, however, there is growing evidence that implementation of this approach into adult gastroenterology services is feasible.
Serological markers in CD have some of the best-reported performance characteristics when compared with antibody-based tests in other conditions. There are clear reasons for a non-biopsy approach including reduced time to diagnosis, improved patient care as endoscopy can be poorly tolerated and is not without risk, cost-effectiveness with studies suggesting upwards of 50% of individuals referred to specialist clinics fulfilled the biopsy avoidance criteria.
However, there are also reasons that advocate a diagnostic duodenal biopsy in adult CD including that adults >50 years presenting with non-specific symptoms associated with CD, such as weight loss and/or iron-deficiency anaemia, should undergo GI endoscopy regardless of coeliac serology results. The most important consequence of implementing this strategy will be to ensure that suspected cases are still referred to gastroenterology services in secondary care rather than a presumptive diagnosis in the primary care setting. It is envisaged that the next five years will see an increasing reliance on serology in the diagnosis of adult CD.
Expert Rev Gastroenterol Hepatol. 2020 Mar;14(3):147-154. doi: 10.1080/17474124.2020.1725472. Epub 2020 Feb 13.