Perinatal determinants of type 1 diabetes - A social epidemiological perspective

摘要: Background and Aims: While the development of type 1 diabetes (T1D) in childhood is part genetically conditioned, non-genetic/environmental factors seem to play a decisive role its aetiology. The incidence T1D has been increasing rapidly during last years, prevalence higher countries high economical welfare like Sweden. Therefore, both from an aetiological public health perspective, there growing need understand non-genetic on risk T1D. A social epidemiological approach would appear very relevant this scenario as many factors, known be associated with risk, are conditioned by socioeconomic circumstances at different levels. This especially pertinent for pre- perinatal influencing such critical period individual’s life course. Adopting thesis aims increase our understanding contextual genesis It also revisits previously described associations between, one hand, and, other, risk. When doing so, it uses interpretative framework that can useful future studies. Concretely, investigates geographical variation cumulative T1D, association between maternal smoking habits seasonal trends cord blood islet autoantibodies (Abs). Underlying all empirical analyses, develop register-based epidemiology apply modern statistical methods analysis variation, effects causality observational studies. Material Methods: Linking Swedish Medical Birth Registry Hospital Discharge other population databases information, newborns born 1987 1993 (n = 811,599) were followed until age 14. Geographical was estimated including multilevel logistic regression, results interpreted within essay obtain deeper theoretical conceptual findings. To investigate we applied propensity score matching investigated siblings same mother. Additionally, made use based genetic study (Diabetes Prediction Skane (DiPiS)) examine Abs. Results: In I, performed analyses found ranged 4.7 5.7 4.4 6.0 per 1000 counties municipalities respectively area small without practical relevance (counties, σ2=0.006; municipalities, σ2=0.017). Nevertheless, sufficient across observe lower metropolitan municipalities. Study II, discussed more detail issue presence low variance. III, confirming previous Sweden, observed children mothers who smoked had 25% reduced developing (OR 0.75 (95%CI: 0.68 0.82)). remained when offspring 0.77 (95% CI: 0.69 0.85)), birth 0.83 0.71 0.96)) mother 0.53 0.36 0.76)). Finally, IV, newborn Abs consistently vary season over four years (p<0.0001); lowest 1st quarter (1.2%) highest 3rd (2.4%). For summer, HLA-DQ gestational infections independently blood. Conclusions: From suggest possible strategies prevention Sweden should not focused specific areas but cover whole country. causal taken together, observations add support hygiene hypothesis speculates levels may conditioning later exposed infection(s) or environmental agents. However, likely due similar mechanisms genetic-environmental mismatch (i.e. human genotype fully adapted actual environment). We only period, proposed could investigations particularly prospective studies children, DiPiS, currently underway.