Type 1 diabetes and gluten

This week’s interesting paper was genuinely attention grabbing for the simple reason that it reported an association of double. We are so used to seeing claims such as “red meat increases risk by 10%” or “whole grains reduce risk by 15%” that an association of double is of real interest. This is the starting point for the Bradford Hill criteria – if the strength of association is double, then the other eight criteria are worth looking at for potential causation and not just association (Ref 1). Bradford Hill would not consider any of the epidemiological reports of 10-20% here or there to be worth any further investigation.

On 20th September 2018, print and television media reported a study that showed that pregnant women who had the highest consumption of gluten had double the risk of having a child with type 1 diabetes compared to those who ate the least gluten (Ref 2).

The study and background

The study was published in The BMJ and it was called “Association between maternal gluten intake and type 1 diabetes in offspring: national prospective cohort study in Denmark” (Ref 3). Most of the researchers listed on the paper were from Denmark and Iceland. The last named author (usually the senior supervisory researcher) was connected to Copenhagen University and the Harvard T.H. Chan School of Public Health, which amused me, as the latter is the factory for epidemiological papers supporting whole grains and attacking red meat. I wonder if someone will be having words with Dr Olsen!

The rationale for the study was interesting. In the section “what is already known on this topic”, the authors reported that “in an animal model of type 1 diabetes, a gluten free maternal diet during pregnancy almost completely prevented type 1 diabetes in offspring.” I did not know that. The researchers therefore aimed to see if there were an association between gluten consumption and T1D in offspring in humans.

The introduction to the paper enhanced this context by explaining that T1D is highest in countries with a western lifestyle and it has been growing at a rate of 3-4% per year, especially in children below the age of 5 in Europe. That rate is faster than can be accounted for genetically, which suggests that other factors are related. A paper was referenced to support the rationale for concern about gluten. Gluten contains proteins that have been found to be more immunogenic (able to produce an immune response) and these have been previously connected to the development of T1D (Ref 4).