Type 2 diabetes and cortical thickness in older Mexican Americans and non-Hispanic Whites

Background: Type 2 diabetes mellitus (T2DM) at midlife increases dementia risk. Metabolic factors contribute more to Alzheimer’s disease (AD) risk in Mexican Americans (MA) than in non-Hispanic Whites (NHW). However, most work relating T2DM to brain structure in older adults has been performed in NHWs. We evaluated how T2DM related to regional cortical thickness in cognitively intact older MA or NHW adults. We hypothesized that T2DM would relate more to regional cortical thickness in MA. Method: We assessed 819 cognitively intact controls (344 MA, mean age=62.2; 475 NHW, mean age =68.8) in the Health and Aging Brain-Health Disparities (HABS-HD) cohort who had 1) a T1-weighted structural brain MRI (3T Siemens Magnatom Skyra) 2) type 2 diabetes diagnostic assessment and 3) successful FreeSurfer (v5.3.0) cortical segmentations (34 cortical regions on the right and left). Regions that failed visual quality checks were imputed using MICE (R studio). We related mean bilateral regional cortical thickness to T2DM diagnosis, covarying for age and sex and controlling for multiple comparisons (Bonferroni correction.) For regions significantly associated with T2DM diagnosis, we evaluated the relationship of HbA1c and fasting blood insulin to cortical thickness, controlling for age and sex. Result: T2DM was significantly associated with cortical thickness in bilateral gyri: supramarginal, precuneus, superior and middle temporal, banks of the superior temporal sulcus, transverse temporal, temporal pole, lingual, pericalcarine, and left lateral occipital (Table 1). HbA1c levels, but not fasting insulin, were significantly correlated with cortical thickness in all identified regions after Bonferroni correction (p<0.0012). There was a significant effect of ethnicity in most of these regions, after controlling for education and household income (Table 2). Supramarginal gyrus, precuneus, transverse temporal gyrus, and left lateral occipital thickness showed a nominally significant interaction (p<0.05) between ethnicity and T2DM, with stronger T2DM effects in MA. Ethnicity interaction effects were diminished in all regions when HbA1c was added to the model. Conclusion: T2DM is associated with cortical thickness in temporal, parietal, and occipital regions. Those effects are stronger in MA than NHW, independent of age, sex, and economic status, but may be partly explained by differences in blood glucose level control.