Background: T1D concordance in MZ twins being ~40%, non-heritable factors play a major causal role in this autoimmune disease. T1D has recently increased in young European children. Collecting prospective environmental data in a cohort of millions children-years starting soon after birth seems unpracticable. Retrospective casecontrol studies are an alternative, provided biased controls and recall bias can both be avoided.
Objectives: To develop a virtual control (VC) geographical approach to unravel environmental factors significantly associated with T1D.
Methods: Four dimensions of environmental exposures were tested by mapping socioeconomic, infectious, climatic and land cover databases at the geolocalized address of the child before T1D diagnosis. Levels of exposures were compared between T1D patients and age-matched geographic VCs. A test was considered significant (**) when the median p value computed over 100 comparisons of cases with 100 sets of VCs was below the Bonferroni limit, and indicative (*) of a possible difference when it was <0.05.
Patients: 3548 children (age-at-onset 7.2±3.7 years) with diagnosis after 1984.
Results: The socioeconomic and land cover environment of T1D children was comparable to controls. The T1D children showed a greater past exposure to influenza (**) and acute diarrheas (*) and a lower past exposure to varicella (*). T1D children were more frequently exposed to heatwaves (**).
Conclusion: Our exploratory approach with four databases provides a proof-of-concept to space-time environment associations studies. Environmental markers (not causes) of T1D can be found. By using more databases, a larger part of a childs environment can be covered.
10 - 12 Sep 2016
European Society for Paediatric Endocrinology