Objective: To test the hypothesis that the risk of childhood diabetes type 1 increases with migration from a low to a high incidence region.
Methods: Register study of a national cohort of 783 547 children born between 1987 and 1993 who remained in Sweden in 2002, including 3225 children with childhood type 1 diabetes identified in hospital discharge data. Logistic regression analysis was used to test the hypotheses.
Results: Offspring of two parents born in very low (Asia excluding Middle East and Latin America) and low (southern and eastern Europe and the Middle East) incidence regions had the lowest adjusted odds ratios (ORs) of childhood type 1 diabetes; 0.21 (0.11-0.41) and 0.37 (0.29-0.48), respectively, compared with the Swedish majority population. When one parent was born in a low incidence country and one parent was Swedish born, the adjusted ORs increased but remained lower than the Swedish majority population.
Conclusions: Parental country of birth is an important determinant of childhood type 1 in Sweden. Heritable factors seem most likely to explain this pattern.