In contrast to recommendations for the diagnosis of adult coeliac disease (CD), European paediatric guidance has long advocated a no-biopsy protocol (NBP) for symptomatic children with a high titre serology. In 2014, the British Society of Gastroenterology (BSG) recommended a small bowel biopsy for all adults in the diagnosis of CD. However in 2020, the BSG issued new interim guidance recommending a NBP in those aged <55 years, with IgA-based anti-tissue transglutaminase (tTG-IgA) titres >10× times the upper limit of normal (ULN), a positive anti-endomysial antibody (EMA) and no alarm symptoms. The aim and objectives of this retrospective study were to firstly evaluate the predictive rate of tTG-IgA >10× ULN on histological features of CD, and secondly to understand the extent of the population with suspected CD based on a positive tTG-IgA but without referral for histological assessment.
As part of a service review to see if a NBP could be implemented locally, a case notes review was performed of all patients aged ≥18 years with a positive tTG-IgA over the period from May 2012 to May 2019. CD cases were identified from the laboratory database with positive tTG-IgA titres. Data was collected on symptoms at initial presentation, age, sex, EMA-IgA results, tTG-IgA titres, duration between tTG-IgA testing and endoscopic assessment for those with tTG-IgA >10× ULN, and Marsh histological grading.
A total of 433 new adult patients were identified with a positive tTG-IgA. Non-referral by primary care clinicians and/ or gastroscopy refusal resulted in 265 patients having a gastroscopy and small bowel biopsy. Diagnosis of CD was histologically confirmed in 213 of the 265 (80%). Of the histologically confirmed CD patients, 66% were women. The mean age of the study group was 47.8 years. The authors of the study analysed the outcomes of their observations if they had followed the interim BSG protocol, performing both a tTG-IgA and EMA-IgA prior to considering a biopsy. Ninety eight of the initial 433 patients had a high titre (tTG-IgA >10× ULN). Of these, 76 underwent gastroscopy and 22 were not referred. Seventy two of the 76 with a high titre had positive histology (95%). Of the four with high tTG-IgA titres and non-diagnostic biopsies there were comments of poorly oriented biopsies in three out of the four samples, suggesting a potential false negative sampling error. Out of the 72 patients with tTG-IgA >10× ULN and histological confirmation, 69 had EMA-IgA positive status, and this assessment did not impact on the likelihood of a diagnostic biopsy. Whereas, a high tTG-IgA titre had significantly greater likelihood of a positive biopsy than a low tTG-IgA titre (p=0.00019). The majority of cases had either gastrointestinal symptoms, extraintestinal manifestations or a mixture of these (85%). In the asymptomatic group, 74% were opportunistic diagnoses, and the majority of these had low titres. Of 433 patients, 33% (n=145) were not referred for a gastroscopy. It was not possible to assess the reasons for non-referral in this retrospective study, although tTG-IgA titre did not appear to be a factor.
This study has validated the interim BSG guidance of not performing biopsies in patients with a high titre tTG-IgA, EMA-IgA positivity, aged less than 55 years and no alarm features. In this study, tTG-IgA titres >10× ULN had a 95% probability of translating into a positive biopsy diagnosis of CD. Endoscopic biopsy sampling limitations are likely to have explained some of the negative findings in this study. The addition of EMA-IgA analyses did not greatly alter the findings. Implementing the NBP for tTG-IgA >10× ULN in the centre studied would lead to around a quarter of all patients (98/443) not requiring a gastroscopy. This would improve the cost effectiveness of the service and the patient experience of avoiding unnecessary and invasive tests.
These findings are similar to those published by gastroenterology departments in other parts of the UK, India and Italy. However, variability within these studies suggest that the reliability of local tTG-IgA assay needs correlating to histopathology findings as part of a local audit before considering implementation of NBP in clinical practice. Those with a low-titre tTG-IgA benefitted from a biopsy assessment as 25% did not have histological features of CD. Endoscopy therefore clearly has a role in assessing those with low titres. Alarmingly, this study found that 33% of all tTG-IgA positive patients were not referred by their GPs for histological assessment. An American patient-reported study described 21% of patients being diagnosed based on serology alone (against national guidance). These patients were more likely to be diagnosed by non-gastroenterologists, use dietary supplements, and were less likely to seek nutritional counselling. As non-referral for histological confirmation appears to be common current practice despite historical guidance advocating a biopsy, then it is of concern that the new NBP guidelines may be misinterpreted.
There is the potential that this high rate of non-referral may yet further increase following such misinterpretation of NBP. It is important that following NBP introduction secondary care coeliac services generate appropriate pathways to manage and monitor referral practice from their local primary care providers.
Johnston RD, Chan YJ, Mubashar T, et al. Frontline Gastroenterology Epub ahead of print: 6/10/20. doi:10.1136/flgastro-2020-101624