Variations in Incidence of Type 1 Diabetes in Different Municipalities of Stockholm

Abstract

This article reports a test of the hypothesis that municipalities within the County of Stockholm have varying incidence rates of type 1 diabetes (T1D), suggesting a strong etiologic environmental component to the disease. The study group included T1D patients in the age group from birth to 18 years who were diagnosed each year from 20 municipalities in Stockholm County during the 1990-2003. Specific incidence rates by age, sex, and socioeconomic characteristics (income level, proportion of taxpayers, proportion of foreigners, population density and green cover) were estimated annually together with age standardization. chi(2) analyses were used for the statistical assessment of variability in incidence. During the study period, 733 newly diagnosed T1D patients aged 0-18 years were recorded from the 20 municipalities under study. The overall age-standardized incidence in these 20 municipalities was 24.38 (22.65-26.21) per 100,000, with 45.35 (32.08-62.29) as highest and 13.41 (9.53-18.35) as lowest estimated incidence. For all socioeconomic variables statistically significant heterogeneity was demonstrated in the standardized incidence rate. High green index was positively associated with the incidence of T1D, as was low population density. For the three remaining socioeconomic variables no clear patterns of associations with incidence of T1D were seen. This study demonstrates a considerable and statistically significant variation between the lowest and highest values in the incidence and prevalence rates for T1D in municipalities of Stockholm County. Such variation seems unlikely to be explained by genetic differences since the population is homogeneous. Our study provides support for the hypothesis that environmental factors have a major influence on the pathogenesis of T1D.