Abstract

Background: Exposure to infections, particularly in early life, may modify the risk of developing childhood diabetes. Population mixing, based on the number and diversity of incoming migrants to an area can be used as a proxy measure for exposure to infections. We tested the hypothesis that incidence of childhood Type I diabetes is higher in areas of low population mixing. Methods: Children (<15 years) diagnosed with diabetes between 1986-1994 in Yorkshire, UK (n = 994) were analysed with demographic data and denominator populations from the 1991 UK Census. Population mixing was estimated separately for 'any age' (>1 year) and children (1-15 years) for each area, using the proportion of migrants and an index of diversity based on numbers and origins of migrants. Regression models calculated the effect of 'any age' and childhood population mixing on the incidence of diabetes, controlling for population density, ethnicity and proportion of migrants. Results: Areas with low levels of population mixing of children (bottom decile), were significantly associated with higher incidence of childhood diabetes for 0-14 years (incidence rate ratio IRR = 1.46, 95% CI : 1.01-2.11). When stratified by age different effects were observed for childhood population mixing with raised IRR for ages 5-9 (2.23, 95% CI: 1.20-4.11) and 10-14 (1.47, 95% CI: 0.89-2.42), and decreased IRR for 0-4-year-olds (0.56, 95% CI: 0.17-1.82). Conclusion: The incidence of childhood diabetes is highest in areas where limited childhood population mixing occurs and the diversity of origins of incoming children is low; those over 4 years are at greatest risk. This is consistent with an infectious hypothesis where absence of stimulation to the developing immune system increases vulnerability to late infectious exposure, which may precipitate diabetes.