Evaluation of the causal associations between brain imaging-derived phenotypes and type 2 diabetes: a bidirectional Mendelian randomization study

OBJECTIVETo investigate whether the association between brain imaging-derived phenotypes (IDPs) and Type 2 diabetes (T2D) related traits is causal. DESIGNTwo sample, bidirectional Mendelian randomization study. SETTINGGenome wide association study (GWAS) summary data taken from various cohorts comprised of the general population (mainly composed of Europeans). PARTICIPANTSSummary data were used from previous GWAS. For IDPs, the data included up to 33,224 European individuals from the UK Biobank. For T2D-related traits, the number of participants ranged from 63,396 to 455,313. MAIN OUTCOME MEASURESA total of 587 reliable IDPs and five T2D-related traits (T2D, fasting glucose, 2h-glucose post-challenge, glycated hemoglobin, and fasting insulin). RESULTSWe identified 3 IDPs with potential causal effects on T2D or fasting insulin. For example, we observed that the area of the right rostral middle frontal cortex was negatively associated with the T2D risk (OR = 0.74, 95% CI 0.65 to 0.85, P = 1.31 x 10-5). In addition, we identified potential causal effects of T2D-related traits on 6 IDPs. For example, T2D was negatively associated with the volumes of the right superior frontal gyrus ({beta} = -0.05, 95% CI -0.08 to -0.03, P = 2.17 x 10-5) and the right paracentral lobule ({beta} = -0.05, 95% CI -0.07 to -0.02, P = 1.74 x 10-4). CONCLUSIONSOur results revealed strong genetic evidence for the bidirectional causal associations between brain neuroimaging phenotypes and T2D-related traits. This will contribute to better prediction and intervention for the risk of T2D.