Childhood adiposity and novel subtypes of diabetes in adults: a Mendelian randomisation and genome-wide genetic correlation study

Abstract

BackgroundFive novel subtypes of adult-onset diabetes were identified in 2018. We aimed to investigate whether childhood adiposity increases the risks of these subtypes using a Mendelian randomisation design, and to explore genetic overlaps between body size (self-reported perceived body size [ie, thinner, about average, or plumper] in childhood, and BMI measured in adulthood) and these subtypes. MethodsThe Mendelian randomisation and genetic correlation analyses were based on summary statistics from European genome-wide association studies of childhood body size (n=453 169), adult BMI (n=359 983), latent autoimmune diabetes in adults (n=8581), severe insulin-deficient diabetes (n=3937), severe insulin-resistant diabetes (n=3874), mild obesity-related diabetes (n=4118), and mild age-related diabetes (n=5605). We identified 267 independent genetic variants as instrumental variables for childhood body size in the Mendelian randomisation analysis of latent autoimmune diabetes in adults and 258 independent genetic variants as instrumental variables for other diabetes subtypes. The inverse variance-weighted method was used as the primary estimator in the Mendelian randomisation analysis, supplemented by other Mendelian randomisation estimators. We calculated overall genetic correlations (rg) between childhood or adult adiposity and different subtypes using linkage disequilibrium score regression. FindingsA large childhood body size was associated with increased risk of latent autoimmune diabetes in adults (odds ratio [OR] 1·62, 95% CI 1·95–2·52), severe insulin-deficient diabetes (OR 2·45, 1·35–4·46), severe insulin-resistant diabetes (OR 3·08, 1·73–5·50), and mild obesity-related diabetes (OR 7·70, 4·32–13·7), but not mild age-related diabetes in the main Mendelian randomisation analysis. Other Mendelian randomisation estimators gave similar results and did not support the existence of horizontal pleiotropy. There was genetic overlap between childhood body size and mild obesity-related diabetes (rg 0·282; p=0·0003), and between adult BMI and all diabetes subtypes. InterpretationThis study provides genetic evidence that higher childhood adiposity is a risk factor for all subtypes of adult-onset diabetes, except mild age-related diabetes. It is therefore important to prevent and intervene in childhood overweight or obesity. There is shared genetic contribution to childhood obesity and mild obesity-related diabetes. FundingThe study was supported by the China Scholarship Council, the Swedish Research Council (grant number 2018-03035), Research Council for Health, Working Life and Welfare (grant number 2018-00337), and Novo Nordisk Foundation (grant number NNF19OC0057274).