Prenatal and adult exposure to smoking and incidence of type 1 diabetes in children and adults-a nationwide cohort study with a family-based design.

Background: Prenatal exposure to smoking is linked to a reduced risk of type 1 diabetes in children. We wanted to find out if the risk of adult-onset type 1 diabetes is reduced in individuals who are exposed to smoking prenatally or during adulthood. Methods: We linked Swedish, nationwide registers and prospectively analyzed incidence of type 1 diabetes in relation to maternal smoking during pregnancy and adult smoking. Everyone was followed until age 30 or year 2019. We employed cohort and sibling design and used adjusted Cox regression and conditional logistic regression. Findings: For analyses of maternal smoking there were 3,170,386 individuals (18,745 cases of type 1 diabetes) and for adult smoking 1,608,291 individuals (1274 cases). Prenatal exposure to smoking was associated with lower incidence of type 1 diabetes during childhood and young adulthood (age 20-24, Hazard ratio (HR) 0.76, 95% Confidence interval 0.67-0.87), but not at higher ages. The HR associated with adult smoking was estimated at 1.14 (CI 1.00-1.31) overall and 1.34 (CI 1.03-1.75) in those with family history of diabetes. In sibling analyses, the odds ratio (OR) of type 1 diabetes in relation to prenatal exposure was 0.71 (CI 0.62-0.81) in children and 1.06 (CI 0.75-1.51) in adults (age 19-30), while adult smoking conferred an OR of 1.59 (CI 1.08-2.35). Interpretation: These findings indicate that a reduced risk conferred by tobacco exposure is limited to the prenatal period and type 1 diabetes developing during childhood. Adult smoking may be a risk factor for adult-onset type 1 diabetes, especially in people with family history of diabetes. Funding: Swedish Research Councils, Swedish Diabetes and Novo Nordisk Foundations, China Scholarship Council.