Risk factors and prognosis of diabetes with and without an autoimmune component

Abstract

Diabetes is heterogenous and encompasses different types such as type 1 diabetes (T1D), type 2 diabetes (T2D), and latent autoimmune diabetes in adults (LADA), which shares clinical characteristics with T2D and genetic similarities with T1D. Smoking is associated with an increased risk of T2D while the underlying mechanism remains unclear. Evidence on modifiable risk factors for T1D and LADA is scarce and the prognosis in LADA is also understudied.

Study I assessed the association between maternal smoking during pregnancy and childhood-onset T1D. We followed 2,995,321 children born in Sweden in 1983-2014 from birth until age 18 years or 2020. Apart from a traditional cohort analysis, we used a sibling and cousin comparison design which compares T1D risks between siblings (or cousins) with and without exposure to maternal smoking during pregnancy, to account for confounding from genetic and environmental factors shared within families. We identified 18,617 children with T1D during follow-up of 18.2 years. Maternal smoking during pregnancy was associated with a decreased T1D incidence in the offspring in the traditional cohort analysis (hazard ratio: 0.78, 95% CI: 0.75. 0.82]), sibling comparison (odds ratio [OR]: 0.78, 95% CI: 0.69, 0.88), and cousin comparison (OR: 0.72, 95% CI: 0.66, 0.79) analyses.

Study II assessed the potential causal relationship between birth weight, adult body mass index (BMI) and LADA using a two-sample Mendelian randomization (MR) design. We used 49 independent SNPs associated with birth weight through fetal genome and not maternal genome as the instrumental variables for the exposure of birth weight and used 820 independent SNPs associated with adult BMI as instrumental variables for the exposure of adult BMI. Summary statistics from the only genome-wide association study (GWAS) study of LADA were used to provide outcome information. We found that one standard deviation (SD) decrease in birth weight was associated with a 68% increased LADA risk (OR: 1.68, 95% CI: 1.01, 2.82) while every SD increase in genetically determined BMI in adulthood was associated with a 40% increased risk of LADA (OR: 1.40, 95% CI: 1.14, 1.71).

Study III assessed the risk of mortality and vascular complications, and clinical trajectories in individuals with LADA. We included 550 people with LADA (further classified LADAhigh and LADAlow by median autoimmune level), 2,001 people with T2D, and 1,573 people with adult-onset T1D diagnosed in 2007-2019, as well as 2,355 diabetes-free population controls. As compared to population controls, people with LADA (HR: 1.44, 95% CI: 1.03, 2.02), T2D and T1D all had increased all-cause mortality; Both LADAhigh (HR 1.67; 95% CI 1.04, 2.69) and T2D, but not LADAlow or T1D, were associated with elevated incidence of cardiovascular diseases. People with LADA had a higher incidence of retinopathy and worse glycemic control than people with T2D.

Study IV investigated the metabolomic pathways linking smoking to T2D and assessed potential interaction between smoking-related metabolic changes and genetic susceptibility. Smoking was associated with changing levels of 131 metabolites in both the cross-sectional analysis of 93,772 UK Biobank participants and MR analysis. We created a smoking-related metabolic signature based on the smoking-related metabolites. In prospective analysis, 38.3% of the smoking-T2D association was mediated by the metabolic signature. The metabolic signature and its mediation role were externally validated in TwinGene. We observed additive interactions between the metabolic signature and genetic risk scores for T2D/insulin resistance.

In conclusion, this doctoral project provides evidence in support of a causal link between maternal smoking during pregnancy and childhood-onset T1D, and between low birth weight and LADA. This indicates that early-life factors play a role in the development of autoimmune diabetes during childhood as well as adulthood. We also provide support on the causal link between adult BMI and both LADA and T2D, highlighting the importance of keeping a normal weight during adulthood to prevent or delay the development of adult-onset diabetes with and without an autoimmune component. In addition, smoking affects T2D risks through changes in a wide range of metabolites, and the adverse effects of such changes are more detrimental to people with high genetic susceptibility to T2D and insulin resistance, indicating the importance of avoiding or quitting smoking for the prevention of diabetes. In people with LADA, the equally high risks of mortality and CVD and higher risk of retinopathy coupled with worse glycemic control than people with T2D highlight the importance of differentiating people with LADA from those with T2D for more effective diabetes management.