Perinatal determinants of type 1 diabetes - A social epidemiological perspective

Sammanfattning: Background and Aims: While the development of type 1 diabetes (T1D) in childhood is in part genetically conditioned, non-genetic/environmental factors seem to play a decisive role in its aetiology. The incidence of T1D has been increasing rapidly during the last years, and its prevalence is higher in countries of high economical welfare like Sweden. Therefore, both from an aetiological and a public health perspective, there is a growing need to understand the role of non-genetic factors on the risk of T1D. A social epidemiological approach would appear very relevant in this scenario as many of the non-genetic/environmental factors, known to be associated with T1D risk, are conditioned by socioeconomic circumstances at different levels. This approach is especially pertinent for pre- and perinatal circumstances influencing such a critical period of the individual’s life course. Adopting a social epidemiological perspective, this thesis aims to increase our understanding of the role of contextual factors in the genesis of T1D. It also revisits previously described associations between, on the one hand, socioeconomic and perinatal factors and, on the other, T1D risk. When doing so, it uses an interpretative framework that can be useful for future social epidemiological studies. Concretely, this thesis investigates geographical variation in cumulative incidence of T1D, the association between maternal smoking habits and T1D risk, and seasonal trends in cord blood islet autoantibodies (Abs). Underlying all the empirical analyses, this thesis aims to develop register-based epidemiology and apply modern statistical methods for the analysis of geographical variation, contextual effects and causality in observational studies. Material and Methods: Linking the Swedish Medical Birth Registry with the Hospital Discharge Registry and other population databases with socioeconomic information, all newborns born between 1987 and 1993 (n = 811,599) were followed for development of T1D until the age of 14. Geographical variation in cumulative incidence of T1D was estimated by different statistical methods including multilevel logistic regression, and the results of this analysis interpreted within a critical essay to obtain a deeper theoretical and conceptual understanding of the findings. To investigate the association between maternal smoking habits and T1D risk, we applied a propensity score matching analysis and investigated T1D risk in siblings from the same mother. Additionally, we made use of a population based genetic study (Diabetes Prediction in Skåne (DiPiS)) to examine for seasonal trends in cord blood islet Abs. Results: In study I, we performed multilevel analyses and found T1D risk ranged from 4.7 to 5.7 and from 4.4 to 6.0 per 1000 newborns in counties and municipalities respectively and the area variation was small and without practical relevance (counties, ?2=0.006; municipalities, ?2=0.017). Nevertheless, there was sufficient variation across municipalities to observe a lower T1D risk in metropolitan municipalities. In Study II, we discussed in more detail this issue of association in the presence of low variance. In Study III, confirming a previous study in Sweden, we observed children born to mothers who smoked had a 25% reduced risk of developing T1D (OR = 0.75 (95%CI: 0.68 to 0.82)). The reduced risk remained when matching offspring by propensity score for maternal smoking (OR = 0.77 (95% CI: 0.69 to 0.85)), by area of birth (OR = 0.83 (95% CI: 0.71 to 0.96)) and by mother (OR = 0.53 (95% CI: 0.36 to 0.76)). Finally, in Study IV, the prevalence of newborn islet Abs was found to consistently vary with season over four years (p<0.0001); lowest in 1st quarter (1.2%) and highest in 3rd (2.4%). For children born in the summer, HLA-DQ genetic risk and gestational infections were independently associated with the presence of islet Abs in the cord blood. Conclusions: From a public health perspective, our results suggest that possible strategies of prevention of T1D in Sweden should not be focused on specific areas but cover the whole country. From a causal perspective, taken together, our observations add support to the hygiene hypothesis that speculates an increase in the levels of hygiene may be conditioning newborns to be more at risk of developing T1D when later exposed to an infection(s) or other environmental agents. However, it is also likely the increasing trends in incidence are due to other similar mechanisms of a genetic-environmental mismatch (i.e. the human genotype not fully adapted to the actual environment). We focused only on factors in the gestational period, but our proposed framework could be useful in future social epidemiological investigations of T1D, particularly in prospective studies of genetic risk children, such as DiPiS, that are currently underway.