Based on the most recent definition of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN),1 coeliac disease (CD) is an immune-mediated systemic disorder elicited by the consumption of gluten and related prolamins (secalins, hordeins and possibly avenins) in genetically susceptible individuals (human leukocyte antigen [HLA] haplotype). It manifests with a variable combination of clinical symptoms, specific serological markers, a HLA-DQ2/DQ8 haplotype and enteropathy.

Treatment consists in the life-long strict elimination of gluten from the diet, which achieves resolution of symptoms, negative titres of autoantibodies and histological recovery of the intestinal mucosa.

In 1969, the ESPGHAN established the first stringent criteria for diagnosis of CD in the paediatric population, the so-called Interlaken criteria2 (Table 1). They established as mandatory criteria evidence of villous atrophy in the initial intestinal biopsy (IB) while the patient consumes a gluten-containing diet, histological recovery after elimination of gluten and recurrence of the lesion after gluten reintroduction (challenge test) to differentiate CD from other causes of enteropathy.

The application of these criteria along with the discovery in the early 1980s of a marker of active CD, antigliadin antibodies (AGAs),3 prompted their revision, leading to the establishment of the new 1990 ESPGHAN criteria for diagnosis of CD.4 These criteria allowed making the diagnosis in older children with symptoms suggestive of CD after a single IB with characteristic (albeit nonspecific) abnormal mucosal histology in the context of a gluten-containing diet and improvement of symptoms with exclusion of gluten from the diet. The gluten challenge was to be reserved for children aged less than 2 years at the time of the initial IB, cases diagnosed without an IB and cases of uncertain diagnosis.

Endomysial antibodies (EMAs)5 were discovered in the mid-1980s, and tissue transglutaminase (tTG) was identified in 1997 as the autoantigen of CD, followed by the discovery of the presence of circulating anti-tTG type 2 antibodies. Both types of antibodies had a similar sensitivity and specificity and performed better as markers of active CD than AGA.6

Clinical signs and symptoms associated with CD (Table 2) are the features that, on appearing, lead to suspicion of CD and the need for differential diagnosis.

The information available on the clinical presentation of CD has experienced drastic changes with the development of highly sensitive and specific serological tests. New categories of patients have been identified that have mild gastrointestinal symptoms, symptoms considered atypical (extraintestinal manifestations) or no symptoms, as well as factors that increase the risk of CD, which has led to changes in the clinical forms of disease currently diagnosed, with an increase in atypical and silent forms. Coeliac disease has gone from being considered a paediatric disease to being diagnosed throughout the lifespan.14

When it comes to the specificity of its features, a recent study in a cohort of patients with positive serological tests for CD found that the combination of a clinical picture of malabsorption (chronic diarrhoea, weight loss, failure to thrive and anaemia) and high titres of specific antibodies (anti-tTG antibody titre > 10 times the upper limit of normal [ULN] and a positive EMA result) had a 100 % positive predictive value for intestinal involvement compared to a slightly lower value if the antibody titres were combined with symptoms other than malabsorption.12

Screening for CD should also be performed in specific situations that carry an increased risk for it, such as autoimmune disease, certain chromosomal disorders, IgA deficiency and history of CD in a first-degree relative (Table 2).13

On the other hand, the notion of universal screening in the general population remains controversial, as CD does not fulfil the criteria established by the World Health Organization for screening tests. Different groups of experts support the recommendation of the ESPGHAN of active case finding, given the lack of evidence on the outcomes of asymptomatic patients that do not receive a diagnosis.15,16

Given the high accuracy of the antibody tests used for diagnosis of CD, serological testing is recommended for initial screening of patients with suspected CD.

Coeliac disease is associated with a specific high-risk variants of the HLA class II region, whose alleles encode molecules involved in the presentation of gluten peptide antigens to CD4+ T cells, thus playing a role in the immune cascade that results in gluten intolerance. More than 90 % of patients with CD are found to be homozygous or heterozygous for the HLA DQ2.5 variant encoded by the DQA1*05 (α-chain) and DQB1*02 (β-chain) genes. The isolated HLA-DQ2.2 haplotype (DQA1*02-DQB1*02) carries a much lower risk, for while the resulting heterodimer is homologous to the DQ2.5 heterodimer it forms less stable complexes with the peptides. Patients that do not have the DQ2 type have the DQ8 type, encoded by the DQA1*0301-DQB1*0302 alleles.1 There are several combinations of the 2 alleles that carry a specific risk of developing CD.30 Class II HLA genes account for 40 % of the genetic risk of CD, while other parts of the genome have been shown to have an influence of less than 10 %.31 In recent years, the complexity of the interpretation of HLA typing has led to efforts to standardise the nomenclature.32

High-risk HLA haplotypes are relatively frequent in some populations, which is associated with a higher incidence of CD if consumption of wheat and other grasses is customary in the population, as is the case in Spain.30 In this sense, HLA genotyping is of little use for diagnosis, as it has a very low positive predictive value. However, its high negative predictive value makes it very useful in ruling out CD in case of uncertain diagnosis and to establish the risk of developing CD in population groups with a high prevalence of the disease.12,19,33,34

Recent studies have reported that nearly 100 % of individuals with anti-tTG antibody titres above 10 times the ULN, a positive EMA result and an abnormal mucosa have a DQ2/DQ8 haplotype, which means that HLA typing would not be necessary to confirm the diagnosis without an IB,12,19 as only individuals with DQ2/DQ8 haplotypes can produce these antibodies.35

Due to all of the above, it seems reasonable to consider that the interpretation of the HLA haplotype and the associated risk is just one among many tools in the management of CD and needs to be performed by paediatricians specialised in paediatric gastroenterology and with experience in the management of CD,1,13 and therefore we do not recommend HLA genotyping for initial screening of CD at the primary care level.

An intestinal biopsy is no longer required for diagnosis of CD in every patient,1 but it unarguably continues to be an essential tool.

The interpretation of histological findings poses some challenges and therefore may vary between pathologists. To optimise the analysis of the intestinal mucosa, we recommend obtaining several biopsy samples through endoscopy, with at least 4 specimens of duodenal mucosa and 1 of the duodenal bulb, as in some cases this is the only affected area.1,13

A standardised protocol must be used to obtain correctly oriented cuttings of duodenal tissue. A villus height and crypt depth ratio of less than 2 in at least 1 of the biopsy samples is considered diagnostic of CD.36 The scheme used most widely to classify the degree of mucosal injury is the Marsh classification modified by Oberhüber37 (Table 3). The diagnosis of CD requires the presence of Marsh 2 or 3 histological changes, although no degree of enteropathy is pathognomonic.

Table 3.

Marsh–Oberhüber classification.37

Stage 0 Preinfiltrative		Histologically normal mucosa
Stage 1 Infiltrative		Increased IELs with preserved villous architecture
Stage 2 Hyperplastic		Hyperplastic crypts and increased IELs
Stage 3 Destructive		Villous atrophy with hyperplastic crypts and increased IELs	Subdivision stablished by Oberhüber based on the degree of VA:
			3a partial VA
			3b subtotal VA
			3c total VA
Stage 4 Hypoplastic		Irreversible villous atrophy	Does not respond to gluten-free diet, screen for establishment of malignant T-cell clone in the intestinal tract.

IEL, intraepithelial lymphocyte; VA, villous atrophy.

It is important to include the following information in the pathology report:

• •

  Whether the amount of tissue in the sample was sufficient

• •

  Whether the specimens were well oriented and allowed assessment of villous height

• •

  Percentage of intraepithelial lymphocytes (IELs) in relation to the total epithelial cell count

• •

  Presence of absence of crypt hyperplasia (mitosis)

• •

  Degree of villous atrophy.

The presence of Marsh 1 changes (a density of more than 25 IELs per 100 epithelial cells) is not sufficient to diagnose CD, as this amount of injury can be attributed to other causes, such as food allergy, infection (viral aetiology, Giardia, Cryptosporidium, Helicobacter pylori), drugs, autoimmune disease, inflammatory bowel disease, etc.38 Mild changes can also be found in patients with possible CD (patients with mild Marsh o or 1 changes and positive anti-tTG and EMA results combined with a high-risk HLA haplotype), in whom prescription of a gluten-free diet is controversial.39 In these cases, it may be useful to analyse the cytometric pattern of IEL subpopulations (elevation of CD3+TCRαβ T cells associated with a decrease in NK-like cells) or the presence of anti-tTG antibody deposits in the mucosa.13,39

The above notwithstanding, in some cases there is still disagreement between the results of anti-tTG antibody testing and histological findings. In such cases, we recommend a histological re-evaluation by examination of additional biopsy specimens or getting a second opinion from an expert pathologist.13

In light of the excellent correlation found between anti-tTG and EMA IgA antibody test results and the presence of mucosal changes, the 2012 ESPGHAN guidelines1 allowed the possibility of diagnosing CD without an IB exclusively in patients meeting the following criteria:

• •

  Suggestive symptoms

• •

  Anti-tTG IgA titres greater than 10 times the ULN

• •

  Positive EMA IgA test, which has the highest specificity, in a second serum sample to confirm the diagnosis and minimise the impact of potential methodological or labelling errors in the initial measurement of anti-tTG IgA titres.

• •

  Compatible HLA-DQ haplotype (DQ2 and/or DQ8).

The accuracy of this approach was confirmed in 2 large-scale studies published in 2017.12,19 In everyday clinical practice, this approach prevents at least 35 % of IBs, mainly in children aged less than 5 years and in patients with significant symptoms.40

In recent years, the role of serological testing and the 2012 ESPGHAN criteria for diagnosis of CD has been analysed in scenarios other than the symptomatic patient. Evidence has emerged that anti-tTG IgA titres greater than 10 times the ULN predict the presence of mucosal injury in asymptomatic patients, although the positive predictive value of this finding may be lower compared to symptomatic patients.12,13

Due to all of the above, the 2020 guidelines13:

• •

  Confirm that the initial evaluation of CD should involve measurement of anti-tTG IgA and total serum IgA levels. In case of IgA deficiency, the second step should be testing for IgG antibodies.

• •

  Allow the option of diagnosing CD in asymptomatic patients without an IB following the same steps used in symptomatic patients. However, since the positive predictive value of high anti-tTG titres is lower in asymptomatic patients, the decision to use the no-biopsy approach should be made on a case-by-case basis and with the agreement of the parents and also the patient if age allows.

• •

  Do not allow the no-biopsy approach for diagnosis of CD in asymptomatic patients with T1D, as the scientific evidence in this patient subset is currently insufficient.

• •

  Maintain the need for an IB in patients with IgA deficiency due to the lack of data on the predictive value of IgG antibodies for the presence of mucosal injury.

• •

  Establish that HLA genetic testing is unnecessary in patients that require an IB or with anti-tTG IgA titres greater than 10 times the ULN. This test would only be indicated for screening of at-risk individuals and in case of uncertain diagnosis.

The evolution of the ESPGHAN criteria for diagnosis of CD a in the past 50 years has led to the current recommendations, with 7 core recommendations (Table 4). The process to follow in case of suspected CD or presence of risk factors is summarised in a diagnostic algorithm (Fig. 1). Primary care paediatricians play an essential role in suspecting the diagnosis and ordering the initial serological tests. The ESPGHAN recommends that the diagnosis always be made by a paediatric1,13 given that a diagnosis of CD involves the prescription of a gluten-free diet for life, and, in most cases, the diagnosis is confirmed within a short time frame, all with the ultimate purpose of avoiding both overdiagnosis and underdiagnosis and their consequences (Table 5).41 At the primary care level, this calls for maintenance of a gluten-containing diet until the diagnosis is confirmed.

Figure 1.

Diagnostic algorithm (adapted from ESPGHAN 202013).

1. IgA deficiency: values below the thresholds established by the laboratory for age or < 0.2g/L in children aged more than 3 years.

2. Ask the family not to initiate a low-gluten or gluten-free diet before the patient is evaluated in the Paediatric Gastroenterology department.

3. Convey the message that regardless of how the diagnosis is reached, treatment with a gluten-free diet is life-long.

4. In case of a positive anti-tTG IgA test with a low titre, confirm sufficient dietary gluten intake. Consider repeating serological tests adding the EMA test.

5. Consider: a. revision of biopsy results; b. false positive anti-tTG result and testing for EMA (if positive EMA=potential CD); c. additional tests (HLA, anti-tTG deposits, cytometry, etc); d. consider followup and retesting ensuring normal gluten intake; e. assess the relevance of symptoms.

tTG=tissue transglutaminase antibodies, EMA=endomysial antibodies; DGP=deamidated gliadin peptide antibodies, ULN=upper limit of normal.

(0.35MB).

The authors have declared no conflicts of interest.