Improving the diagnosis of coeliac disease

Project aims

This project was aimed at investigating the best testing strategy for identifying adults and children with coeliac disease. This strategy would have to be both cost-effective and acceptable to patients.

We wanted to establish who should be tested for coeliac disease, what tests should be offered, and whether a biopsy is necessary for all patients.

What we did

We searched for studies on coeliac disease testing to see if we could assess how accurate these tests were. Our team also reviewed and analysed the results of 191 studies looking at diagnostic indicators such as symptoms or risk conditions for coeliac disease.

In addition, we reviewed data from general practices. This was because we wanted to understand whether certain combinations of risk factors and symptoms could help identify people with an increased risk of having coeliac disease.

We interviewed 20 adults diagnosed with coeliac disease. We did this to improve our understanding of the patient experience. We also created an online survey to explore how confident people felt about a diagnosis given in different situations.

Finally, we assessed the cost-effectiveness of different diagnostic strategies.

What we found and what this means

Blood test accuracy

To assess the accuracy of blood tests for coeliac disease in adults and children, we looked at 113 studies. We reviewed studies comparing the results of blood tests with duodenal biopsies. Blood tests aimed at diagnosing coeliac disease look for antibodies in a person’s blood that might be associated with this disease. A duodenal biopsy involves a small section of a patient’s duodenum (the initial C-shaped segment of the small intestine) being taken so it can be examined under a microscope.

We found that the accuracy of blood tests for detecting coeliac disease was high. They are a useful first step towards diagnosis in patients suspected of having coeliac disease. Our findings suggest that blood tests alone could be used for diagnosis in some people, without the need for a biopsy.

Identifying people at risk of developing coeliac disease

During this part of the study, we identified and focused on 26 signs, symptoms and risk factors associated with coeliac disease.

We found strong evidence that people with a family history of coeliac disease, dermatitis herpetiformis (a skin condition caused by a reaction to gluten ingestion), anaemia, type 1 diabetes, migraines, certain genes, osteoporosis, or chronic liver disease are more than twice as likely to have coeliac disease than the general population. Additionally, close relatives of people with coeliac disease are three times as likely to have it themselves. These signs and symptoms could therefore help identify patients who would benefit from testing.

Migraine and chronic liver disease are not yet included as a risk factor in all guidelines. We think it may be appropriate for these to be added to guidelines.

Other signs including gastrointestinal symptoms (such as diarrhoea, constipation and abdominal pain), psoriasis, epilepsy, inflammatory bowel disease, systemic lupus erythematosus, fractures, type 2 diabetes and multiple sclerosis were not shown to be reliable indicators of the disease.

Patient experience

We found that patients experience uncertainty when following the traditional diagnostic pathway for coeliac disease. This uncertainty particularly affected patients when they wanted an explanation for their symptoms and during investigations for a potential diagnosis of coeliac disease.

Coeliac disease is traditionally diagnosed through a combination of blood tests and an endoscopic biopsy. An endoscopic biopsy involves a camera being passed into the body with a small sample of body tissue being collected so it can be examined under a microscope. One of the challenges for patients is the need to maintain a diet containing gluten in the lead up to the endoscopy.

Interviewees reported waiting a long time for an endoscopy and found it challenging to manage their diet before the procedure. However, having the procedure reassured most interviewees that they should be following a gluten-free diet for the rest of their lives. For some it increased their likelihood of adhering to the diet and reduced post-diagnostic uncertainty if their symptoms returned.

If we move toward a no biopsy approach for some patients, it is important to ensure patients receive adequate follow up and specialist support to reduce the risk of a misdiagnosis. Any uncertainty patients experience during the diagnostic process must also be addressed.

Developing prediction models

We reviewed data from general practices to see whether we could develop diagnostic prediction models to help decide who should be offered testing for coeliac disease in GP practices. Diagnostic prediction modelling uses statistics to predict whether someone has a certain condition. In this case we wanted to see whether we could predict how likely someone would be to have coeliac disease based on other conditions they may have or things we might know about them.

Our prediction model included conditions and factors that might mean either an adult or a child might have coeliac disease. Type 1 diabetes, Turner syndrome, IgA deficiency or having a close relative with coeliac disease were the strongest predictors that a child would be likely to have the disease. Anaemia and having a close relative with the disease were predictors for adults.

Although these types of models could be used by GPs to help identify people with an increased risk of coeliac disease, it would involve offering blood tests to more people than is currently done. Overall, we found that our model tended to overestimate coeliac disease risk in both adults and children.

Confidence in diagnosis and following a gluten-free diet

We organised focus groups and an online survey to find out how confident people needed to feel about their diagnosis before they started a gluten-free diet or underwent a biopsy and understanding what factors influenced their attitude. We also included questions about quality of life among people with coeliac disease and the difficulties they may face when following a gluten-free diet.

Four hundred and seventy-two people completed our survey. We found that the level of certainty people wanted varied a lot. In scenarios where we asked respondents to imagine that they had no symptoms, they wanted a higher level of certainty before committing to a lifelong gluten-free diet then when they imagined having debilitating symptoms.

Instead of starting a gluten-free diet immediately, a higher number of our respondents opted to wait for a confirmation biopsy, when given the option. This was the case even if a hypothetical blood test gave them 75–90 per cent certainty they had the disease.

Assessing how cost-effective diagnostic strategies are

We found that the immunoglobulin A tissue transglutaminase test, currently the most common blood test for coeliac disease, was the most cost-effective at diagnosing the disease in adults, when using blood tests alone. However, using specific blood test combinations was similarly cost-effective and could be the most practical strategy for the NHS to implement (a combination of human leucocyte antigen and immunoglobulin A tissue transglutaminase testing).

What next?

We expect the results of this study to change the way people with coeliac disease are identified in the UK.