1 in 100 people in the UK is estimated to have coeliac disease but only 30% of those with the condition have been diagnosed, meaning there is around half a million people in the UK living with undiagnosed coeliac disease. Obtaining a diagnosis of coeliac disease can be challenging and there are limitations to the current diagnostic tests. Recent research published in The Journal of Pathology by Coeliac UK funded researcher, Dr Elizabeth Soilleux, brings us closer to new and improved diagnosis methods for coeliac disease.
Diagnosis of coeliac disease is currently a two-step process, first a blood test to look for antibodies, followed by an endoscopy with biopsy to look for damage to the intestine. In up to around 40% of adults and children, an endoscopy might not be needed and a second blood test can confirm the diagnosis.
These current methods for diagnosis have two key limitations. First, to obtain accurate results from both the blood test and endoscopy, gluten a protein found in wheat, barley and rye must be eaten in more than one meal a day for at least six weeks before testing. This can be very difficult for people who are having severe symptoms, particularly as they may be facing long waiting times for their endoscopy, a situation which has worsened during the Covid-19 pandemic. Secondly, sometimes the results of these tests can be unclear. Biopsy samples taken during endoscopy are inspected by trained pathologists to see if the cells have the damage which is characteristic of coeliac disease. This process is subjective, with different opinions in up to 25% of cases.
Dr Soilleux has developed a machine learning algorithm to identify people with coeliac disease, even if gluten has already been removed from their diet. In people with coeliac disease, gluten is recognised by immune cells called T cells, which trigger an immune response leading to inflammation and damage to the gut. Although most of the inflammation of the small intestine disappears on a gluten free diet, we now know that there is a permanent change to the gluten specific T cells.
The researchers looked at the DNA of the T cells taken from biopsy samples and, using the machine learning algorithm, successfully classified people with and without coeliac disease. Most importantly, the algorithm correctly identified people with coeliac disease who have been on a gluten free diet for at least six months.
This is an exciting piece of research and it is hoped that with further research it could provide an improved diagnostic method which is less subjective, accurate even for people already on a gluten free diet and importantly it may allow people to adopt a gluten free diet earlier and potentially feel better sooner. You can hear more about the research from Dr Soilleux on our YouTube channel at https://www.youtube.com/channel/UCxy5w7waAdseR1p7Iuq065w/search?query=liz
“At the moment, getting an accurate diagnosis of coeliac disease means keeping gluten in the diet throughout the testing process – which can prove difficult when waiting lists are long and people feel unwell. This new research is an exciting step forward to help improve accuracy and provide an answer for those who are already on a gluten free diet. Pioneering research is essential to aid in developing new test methods and we are thrilled to have combined forces with Innovate UK to jointly fund this research to improve the lives of people with coeliac disease and we hope that one day this will become part of the diagnosis pathway for coeliac disease,” said Dr Heidi Urwin, Director of Evidence and Policy from Coeliac UK