Is celiac disease becoming more common?

A review of 50 population studies found that new celiac disease diagnoses have been rising by about 7.5% a year for decades, most of all among women and children. How much of that is more disease, and how much is better testing, is still an open question.

A diagnosis of celiac disease used to be thought of as uncommon and mostly something that happened to children. Anyone who has watched the gluten-free shelf grow over the past twenty years might suspect that has changed. A team at the University of Calgary set out to measure it properly, and published the result in the American Journal of Gastroenterology in 2020.

Their question was not how many people have celiac disease — that is prevalence, and earlier work puts it at about 0.7% of people with biopsy-confirmed disease and 1.4% when blood tests are used. They asked how many people are newly diagnosed each year, which is incidence, and whether that number has been moving.

How the review was built

The authors searched two large medical databases, MEDLINE and EMBASE, up to September 2019 with no restriction on language, looking for population-based studies — ones that count every diagnosis in a defined region, rather than patients at a single clinic or in a high-risk group such as people with type 1 diabetes. From 11,189 citations, 86 studies qualified. Fifty reported their results in a form that could be pooled: 42 from Europe, seven from North America and one from Oceania.

Where a study covered enough years, the team calculated the average annual percent change in the rate of new diagnoses, then combined those figures across studies. The review was registered in advance on PROSPERO, the international register of systematic reviews.

What it found

Two patterns stand out. First, diagnoses are about twice as common in women as in men: the pooled female-to-male ratio was 2.1. Second, children are diagnosed more often than adults, at a ratio of 1.7 — although adults still make up a large share of new cases.

The trend is the headline. Of the 33 studies with enough years to test, 24 — 72.7% — showed a statistically significant rise. Pooled together, new diagnoses grew by an average of 7.5% a year over the past several decades. Individual regions show what that looks like on the ground: in Olmsted County, Minnesota, incidence rose by about 8.1% a year between 1950 and 2010; among children in southeast Scotland, by 12.8% a year from 1990 to 2016; among adults in Cardiff and the Vale of Glamorgan in Wales, by 11.3% a year from 1981 to 2005.

The rise is not universal. Child incidence in Sweden climbed steeply from 1998 to 2003 and then levelled off; in Tampere, Finland, it was stable after 2007. The only falling trend was among adults in Finland, down 3.4% a year between 2005 and 2014. The authors suggest some of the highest-incidence countries may have reached a plateau.

More disease, or more diagnosis?

The study counted diagnoses, so the first explanation to rule out is better detection. The authors are explicit that it accounts for part of the increase. Accurate blood tests arrived near the end of the 20th century and made diagnosis cheaper and easier — in Calgary, the rate of diagnosis in children tripled after antiendomysial antibody testing was introduced. Guidelines have shifted too: European paediatric guidance now allows some children to be diagnosed without a biopsy when they meet specific criteria.

The authors also discuss possible reasons for a genuine increase, while stressing that the causes are not settled. Hypotheses they mention include infant feeding patterns, antibiotic use in the first year of life, which was linked to higher risk in cohorts from Norway and Denmark, and the hygiene hypothesis. Each is a line of research, not a conclusion.

There is a pull in the other direction as well. More people now follow a gluten-free diet without a diagnosis. The authors note that people who are tested after cutting out gluten may not be picked up, because removing gluten reduces the sensitivity of the tests — so some diagnoses may simply never be made. Celiac disease versus gluten sensitivity explains how the tests work.

Reading the numbers carefully

The pooled figures carry very high statistical heterogeneity, meaning the underlying studies differed a great deal — in diagnostic practice, data sources, years covered and, probably, the real epidemiology of each region. A pooled rate is best read as an average to compare against, not a figure that applies to any one country. Almost all the data came from Europe and North America; the authors found no population-based incidence studies from Africa, Asia or Latin America suitable for analysis, and call for them.

The research was funded by a young investigator award from the Canadian Celiac Association and by Alberta Health Services’ Digestive Health Strategic Clinical Network. The authors reported no competing interests.

Why it matters for everyday shopping

A growing number of diagnosed people means a growing number of households reading labels for gluten, many for the first time. For them, the practical questions are the familiar ones: which ingredient names signal gluten, what a “gluten-free” label actually guarantees, and how much is too much. The names gluten hides under and the trial behind the 20 ppm limit cover those.

Source: King JA, Jeong J, Underwood FE, et al. Incidence of Celiac Disease Is Increasing Over Time: A Systematic Review and Meta-analysis. Am J Gastroenterol, 2020;115(4):507–525.

Common questions

Keep reading