New entity of adult ultra-short coeliac disease: the first international cohort and case–control study

Background Ultra-short coeliac disease (USCD) is defined as villous atrophy only present in the duodenal bulb (D1) with concurrent positive coeliac serology. We present the first, multicentre, international study of patients with USCD. Methods Patients with USCD were identified from 10 tertiary hospitals (6 from Europe, 2 from Asia, 1 from North America and 1 from Australasia) and compared with age-matched and sex-matched patients with conventional coeliac disease. Findings Patients with USCD (n=137, median age 27 years, IQR 21–43 years; 73% female) were younger than those with conventional coeliac disease (27 vs 38 years, respectively, p<0.001). Immunoglobulin A-tissue transglutaminase (IgA-tTG) titres at index gastroscopy were lower in patients with USCD versus conventional coeliac disease (1.8×upper limit of normal (ULN) (IQR 1.1–5.9) vs 12.6×ULN (IQR 3.3–18.3), p<0.001). Patients with USCD had the same number of symptoms overall (median 3 (IQR 2–4) vs 3 (IQR 1–4), p=0.875). Patients with USCD experienced less iron deficiency (41.8% vs 22.4%, p=0.006). Both USCD and conventional coeliac disease had the same intraepithelial lymphocytes immunophenotype staining pattern; positive for CD3 and CD8, but not CD4. At follow-up having commenced a gluten-free diet (GFD) (median of 1181 days IQR: 440–2160 days) both USCD and the age-matched and sex-matched controls experienced a similar reduction in IgA-tTG titres (0.5 ULN (IQR 0.2–1.4) vs 0.7 ULN (IQR 0.2–2.6), p=0.312). 95.7% of patients with USCD reported a clinical improvement in their symptoms. Interpretation Patients with USCD are younger, have a similar symptomatic burden and benefit from a GFD. This study endorses the recommendation of D1 sampling as part of the endoscopic coeliac disease diagnostic workup.


INTRODUCTION
Coeliac disease is a common autoimmune disorder that affects individuals worldwide, with a global prevalence between 0.7% and 1.4%. 1 Despite being an increasingly significant global health problem, a significant proportion of individuals with coeliac disease remain undiagnosed (5%-76%). 2 3][6] Furthermore, 5%-12.4% of patients have had a

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ The first study of ultra-short coeliac disease in adults in 2016 identified patients with villous atrophy confined to the duodenal bulb with positive coeliac serology.⇒ Systematic review and meta-analyses suggest that taking a duodenal bulb biopsy can increase the diagnostic yield of adult coeliac disease by 8%.⇒ There are limited further studies and no data on how these patients respond to treatment.

WHAT THIS STUDY ADDS
⇒ Our study provides the first international data of patients with ultra-short coeliac disease.⇒ At presentation, adult patients with ultrashort coeliac disease are significantly younger, have a similar symptomatic burden but lower serological titres.⇒ Adult patients with ultra-short coeliac disease improve both clinically and serologically when on a gluten-free diet.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
⇒ Our data support adherence to undertaking a bulb biopsy.⇒ Once identified these patients can be treated effectively with a gluten-free diet.

Coeliac disease
previous gastroscopy (where no biopsies were taken) prior to their diagnosis representing a missed opportunity to diagnose coeliac disease. 7 8A gluten-free diet (GFD) remains the only treatment, and adherence improves quality of life for the individual and potentially reduces the burden to the healthcare system. 9 10he conventional form of coeliac disease is characterised by villous atrophy (VA) and crypt hyperplasia in the second part of the duodenal mucosa (D2) with concurrent positive coeliac serology (figure 1). 11Historical early reports of the value of acquiring biopsies from the duodenal bulb (D1) were disregarded in favour of biopsies from the distal duodenum as it was suggested that histological interpretation was potentially impaired by the presence of Brunner's glands, gastric heterotropia and duodenitis. 12Ultra-short coeliac disease (USCD) is defined as patients with VA only present in the duodenal bulb (D1) and concurrent positive coeliac serology (figure 1).This term was coined in 2016 and systematic review and meta-analyses suggest that taking a duodenal bulb biopsy can increase the diagnostic yield of adult coeliac disease by 8%.For paediatric populations, this was shown to be 4% (95% CI 1% to 9%; p<0.001). 12ubsequently, the American College of Gastroenterology and British Society of Gastroenterology (BSG) now recommend bulb biopsy as standard practice when an upper endoscopy is undertaken to assess for 'suspected coeliac disease/malabsorption'.However, adherence to biopsy protocols is low (37.0%-39.5%). 11 13-15One reason for this may be a perceived increase in cost or alternatively a perception that a patient with histological changes confined to the bulb may not benefit from a GFD. 7he value of taking D1 biopsies remains controversial and there is only a single centre study of 26 adult patients from the UK describing the clinical presentation of VA confined to D1. 11 14-17 Understanding the phenotype and therapeutic outcomes of patients with USCD is crucial for optimising diagnostic approaches and treatment strategies, as well as improving overall patient care and long-term health outcomes.There is a paucity of data regarding the phenotype of patients with USCD by comparison to conventional coeliac disease.Furthermore, there are limited data on the benefit of a GFD in patients with USCD.
To address this, we present the first, multicentre, international study of patients with USCD.

METHODS
The study was proposed after the 19th International Society for the Study of Celiac Disease Conference, Sorrento, October 2022, as an international, multicentre, observational cohort study enrolling all patients with USCD between 2009 and 2022.Patient data were collected from hospital databases that prospectively record information about patients with coeliac disease.One centre collected information retrospectively from hospital records based on positive histological findings.Ten tertiary coeliac disease centres participated in the study: Sheffield, UK; Caceres, Spain; New York City, USA; Palmerston North, New Zealand; Ankara, Turkey; Bologna, Pavia, Milan and Salerno, Italy; Tehran, Iran.

Patients
Group 1: For this cohort study, adult (≥16 years) patients were identified from 10 tertiary hospitals between January 2009 and December 2022.Patients were defined as having USCD if they had a combination of positive serological markers (immunoglobulin A-tissue transglutaminase (IgA-tTG) or immunoglobulin A-endomysial antibody (IgA-EMA)) and histologically confirmed VA confined to D1 while on a gluten containing diet.D2 biopsies were architecturally non-diagnostic of coeliac disease (Marsh grades 0-2) and diagnoses were made locally by gastroenterologists with expertise in coeliac disease.
Group 2: For the age-matched and sex-matched case-control study, controls were identified from databases of adult patients with coeliac disease diagnosed in each centre.Each age-matched and sex-matched control was from the same centre as the patient

Coeliac disease
with USCD.Age-matched and sex-matched adult coeliac disease controls were then randomly selected using IBM SPSS V.27.0 (IBM) case control matching function.
Both USCD and age-matched and sex-matched controls had D1 and D2 biopsies.
Data were collected following assessment by a clinician with a special interest in coeliac disease at each centre.Data were reviewed in case notes, endoscopy records and the referral.Data were collected on presenting symptoms, serology at time of presentation (including haemoglobin, vitamin B 12 , folate, iron, vitamin D, IgA-tTG and IgA-EMA), human leucocyte antigen (HLA) typing and histology of duodenal biopsies.Patients with USCD and the case-control patients were then followed up to determine the effects of a GFD on their serological markers and symptoms.All patients were assessed for commonly occurring symptoms in coeliac disease both at presentation and follow-up.
In order to assess for any potential differences between USCD and conventional coeliac disease when specifically considering age and sex at presentation, a further analysis was undertaken comparing all patients with USCD (n=137) to those with conventional coeliac disease from the Sheffield, UK coeliac database (n=434).

Serology
IgA-tTG antibody levels were measured by different ELISA kits (Aeskulisa Diagnostics (Wendelsheim, Germany), ELiA Celikey (Thermo Fisher, Freiburg, Germany), ARUP Laboratories (Utah, USA), QuantaLite (Inova Diagnostics, San Diego, California), Eu-tTG (Eurospital, Italy) and Euroimmune (Luebeck, Germany)).Therefore, levels were standardised using the upper limit of normal (ULN) based on the manufacturer's supplied reference ranges.IgA-EMA was detected by immunofluorescence on primate oesophagus sections (Binding Site, Birmingham, UK).The normal ranges of blood tests differed by centre, and therefore, to allow for direct comparison, the lower limit of normal (LLN) was used for ferritin, vitamin B 12 , folic acid and vitamin D based on the manufacturer's supplied reference ranges of each test.All blood tests were complete prior to endoscopy as part of the referral process.
HLA typing was performed for HLA-DQ2 and DQ8 at six centres and full genomic HLA typing at three.One centre did not perform HLA typing (table 1).

Biopsies and histology
Multiple biopsies were taken in D1 and quadrantic biopsies in D2 and the most severe histological findings used for diagnosis and analysis.The biopsy specimens were first preserved in formalin and then embedded in paraffin wax.Afterwards, they were thinly sliced into sections measuring 3 µm in thickness.These sections were subsequently stained using H&E.Duodenal biopsies were assessed by experienced histopathologists with an interest in gastroenterology.The biopsies were all orientated by experienced biomedical scientists in the histopathology laboratory and three levels were cut from each specimen.This ensured

Coeliac disease
that in at least some of the levels there were full length villi present and the interpreting histopathologists looked for the longest villi that were present in all three levels.Grading was completed using the modified Marsh criteria: Marsh 1 lesions demonstrated increased intraepithelial lymphocytes (IEL), Marsh 2 lesions demonstrated crypt hyperplasia and Marsh 3 lesions demonstrated VA. 18 CD3 antibody was measured using streptavidin biotin peroxidase method by automated Ventana Benchmark XT system (Roche, Ventana Medical Systems, Tucson), Clone:LN10: Leica Concentrate and immunohistochemistry anti-human Cd3 Dako or GenScript.CD8 antibody was measured with clone: C8/44B; Dako RTU Link and CD4 with clone:4B12; Dako RTU Link.An average of two biopsies was tested for both D1 and D2.

Follow-up
Follow-up data were collected in each centre based on clinical improvement (Likert scale) divided into four categories: 'symptoms worse', 'symptoms the same', 'symptoms improved' and 'symptoms completely resolved' after clinical assessment as part of routine care.Serological follow-up was completed using the blood tests as described above.The length of follow-up varied based on the time the patient was known to the centre.

Statistical analysis
Data handling was completed using Microsoft Excel (2016); statistical analysis was conducted in IBM SPSS V.27.0 (IBM).
The prevalence of each presenting symptom was compared between cohorts using χ 2 test of two proportions where there was adequate sample size and if not, Fisher's exact test was used. 19Shapiro-Wilk test was used to assess for Gaussian distribution of continuous data such as age and IgA-tTG titre.Where normally distributed and no outliers, a t-test was used, otherwise the Mann-Whitney U test used.A p<0.05 was considered statistically significant.

Role of the funding source
No funding was acquired to complete this study.
For other comparisons, group 1 patients with USCD (n=137) were compared with group two patients with conventional coeliac disease (n=137) (table 1).Group 1 patients were referred from primary care, were self-referring, referred from other departments within the same hospital (with symptoms or positive coeliac disease serology) or referred from other hospitals (68%, 19%, 10%, 3%, respectively).
The immunophenotype of the intraepithelial lymphocytes was the same in both D1 and D2 with all the intraepithelial lymphocytes staining with CD3 and CD8, but not with CD4 (figure 3).

Follow-up
Serological and clinical assessment occurred after a median of 1181 days (IQR 440-2160 days).Following recommendation of a GFD patients with both USCD and the age-matched and sex-matched controls experienced a similar reduction in IgA-tTG titres (0.5 ULN (IQR 0.2-1.4)vs 0.7 ULN (IQR 0.2-2.6),p=0.312) and similar levels of IgA-EMA positivity (26.9% vs 23.1%, p=0.598) (figure 2).Levels of vitamin B 12 , iron, folate and vitamin D all improved after undertaking a GFD (figure 2).Symptomatic improvement occurred in both patients with USCD and in the age-matched and sex-matched controls (95.7% vs 89.1%, p=0.115).In total, 16.1% of patients with USCD had complete resolution of their symptoms, 79.6% reported a partial improvement, 3.2% reported no change in their symptoms and 1.1% reported symptoms to be worse after following a GFD.

DISCUSSION
This is the first multicentre international study of USCD.We have demonstrated that patients with USCD are younger than those with conventional coeliac disease and have lower IgA-tTG titres.Despite only having VA in the duodenal bulb patients with USCD are both symptomatic and derive benefit from a GFD.This study endorses the recommendation of taking samples from D1 as a mandatory component of coeliac disease diagnostic workup.
A single-centre study (n=26) from our centre previously identified patients with USCD as younger and having lower IgA-tTG titres.Furthermore, this work demonstrated that an additional D1 biopsy can increase the diagnostic yield by 9.7%. 16Despite endorsement from the American College of Gastroenterology and BSG, the adherence to biopsy protocols in general remains low (37.0%-39.5%). 11 13-155][6] Of the patients with USCD, 41.6% and 16.8% had Marsh 1 and Marsh 2 lesions in D2, respectively, therefore, if only D2 biopsies were taken, these patients may have been incorrectly diagnosed as having potential coeliac disease. 11The implications for both the patient and the clinical recommendation to follow a GFD are different in 'real-world' practice when faced with a patient with potential coeliac disease by comparison to VA (Marsh 3) confirmed coeliac disease.The BSG guidelines have made no recommendation for the role of a GFD in patients with 'potential coeliac disease'.
Conversely in a prospective randomised controlled study of 23 patients with potential coeliac disease (Marsh grades 1-2, raised IEL only or raised IEL and crypt hyperplasia but no VA)

Coeliac disease
individuals who were randomised to commence a GFD showed both symptomatic benefit and a reduction in their tTG titres. 20n this historical study from 2003 to 2008, none of the patients had a duodenal bulb biopsy.It could be suggested that these patients may have had USCD.
It is perceived that a bulb biopsy strategy may increase healthcare utilisation costs.This may be due to the use of a second histopathology pot (for the bulb biopsy), processing costs and increased pathology reporting time.This may explain some of the reluctance to take biopsies, however, adequate duodenal biopsy strategies potentially avoid diagnostic delays for patients with undiagnosed adult coeliac disease and are a cost-effective approach in improving the quality-adjusted life-years of patients with coeliac disease. 21 22t may be possible to place D1 and D2 biopsy samples in the same histopathology pot.A historical paediatric study (n=198) found that 'intraepithelial lymphocytosis was easily recognised in bulb biopsies, and that although the normal villous-to-crypt ratio is lower in the bulb than in the more distal duodenum, significant VA was usually apparent'.When samples were reviewed by experienced histopathologists, the changes of coeliac disease were still identifiable and the risk of interobserver variability was low. 23e found that patients with conventional coeliac disease were more likely to have iron deficiency than those with USCD, which may correlate with more extensive mucosal inflammation and impaired absorptive capacity of the duodenum in the former.Interestingly, there was no difference in iron deficiency anaemia.
When considering the paediatric population, in a study of 834 paediatric patients diagnosed with coeliac disease, 11% were diagnosed with USCD, these USCD paediatric patients were also found to have lower tissue transglutaminase antibody titres and less iron deficiency than patients with conventional coeliac disease. 24This suggests that the paediatric and adult USCD cases are similar.This is corroborated by capsule endoscopy studies that demonstrate an association between iron deficiency anaemia, increased age and extent of disease in conventional coeliac disease. 25ur study demonstrates that HLA DQ2 homozygosity is more common in conventional CD by comparison to USCD (40.4% vs 25.8% p=0.038).This could suggest that the HLA genotype may have a quantitative relationship between the DQ heterodimer and phenotype.Supporting this, a study of seven patients with USCD that found the HLA-DQ2 haplotype to be less common in patients with USCD and no difference in HLA-DQ8 haplotype. 26However, all these findings are based on small sample sizes and further investigation is required to determine the significance of a possible different HLA haplotype in USCD.
A limitation of this study is that histology samples could be affected by interindividual variability between histopathologists, however, as all histopathologists have a specialist interest in coeliac disease the risk of this is reduced.It is uncertain how this would translate to 'real-world' clinical practice beyond centres with an interest in coeliac disease.
Another limitation is the lack of a central reference lab so there was no standardisation between IgA-tTG assays; to address this, the results were evaluated in relation to the ULN as stated by the manufacturer for each assay.Centres involved have a special interest in coeliac disease, and therefore, there may be a referral bias.
In conclusion, this is the first multicentre international study to evaluate the new entity of USCD.Patients with USCD are younger than those with conventional coeliac disease and have lower serological markers of coeliac disease.Despite only having VA in the duodenal bulb, patients with USCD are both symptomatic and derive benefit from a GFD.This study endorses the recommendation of taking samples from D1 as a mandatory component of coeliac disease diagnostic workup.

Author affiliations
1 Academic Unit of Gastroenterology, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK Figure 3 Biopsies of D2 and D1 from a patient with ultra-short coeliac disease.In the H&E-stained sections of D2 (A) there is a normal villous height, no significant crypt hyperplasia but there is an increased number of intraepithelial lymphocytes.In the H&E-stained sections of D1 (E) there is complete villous atrophy, gross crypt hyperplasia and an increased number of intraepithelial lymphocytes.The immunophenotype of the intraepithelial lymphocytes is the same in both sites with all the intraepithelial lymphocytes staining with CD3 (B, F) and CD8 (D, H) but none of them stain with CD4 (C, G).

Figure 1
Figure 1 Subtypes of coeliac disease divided by extent of villous atrophy.

Figure 2
Figure 2 Comparison of serological markers in coeliac disease at baseline and follow-up.Serological comparisons made between patients with ultra-short coeliac disease and age-matched and sex-matched controls at baseline and follow-up for: (A) immunoglobulin A-tissue transglutaminase titre, (B) ferritin, (C) folate, (D) vitamin B 12 , (E) vitamin D, (F) immunoglobulin A-endomysial antibody.LLN, lower limit of normal; ULN, upper limit of normal.

Table 1
Number of cases of ultra-short coeliac disease from each centre *Denotes staining complete if complicated case.†Denotes approximate value.HLA, human leucocyte antigen.

Table 2
Presentation of patients with ultra-short coeliac disease (USCD) and age-matched and sex-matched patients with conventional coeliac disease