The association of prediabetes and type 2 diabetes with hippocampal subfields volume: The Maastricht study

Aims/hypothesis: We investigated whether prediabetes, type 2 diabetes, and continuous measures of hyperglycemia are associated with tissue volume differences in specific subfields of the hippocampus. Methods: We used cross-sectional data from 4,724 participants (58.7 +/- 8.5 years, 51.5% women) of The Maastricht Study, a population-based prospective cohort. Glucose metabolism status was assessed with an oral glucose tolerance test, and defined as type 2 diabetes (n = 869), prediabetes (n = 671), or normal glucose metabolism (n = 3184). We extracted 12 hippocampal subfield volumes per hemisphere with FreeSurfer v6.0 using T1w and FLAIR 3T MRI images. We used multiple linear regression and linear trend analysis, and adjusted for total intracranial volume, demographic, lifestyle, and cardiovascular risk factors. Results: Type 2 diabetes was significantly associated with smaller volumes in the hippocampal subfield fimbria (standardized beta coefficient +/- standard error (beta +/- SE) = -0.195 +/- 0.04, p-value < 0.001), the hippocampus proper, i.e. Cornu Ammonis (CA) 1, CA2/3, CA4, dentate gyrus, subiculum and presubiculum (beta +/- SE < -0.105 +/- 0.04, p-value < 0.006); as well as the hippocampal tail (beta +/- SE = -0.162 +/- 0.04, p-value < 0.001). Prediabetes showed no significant associations. However, linear trend analysis indicated a dose-response relation from normal glucose metabolism, to prediabetes, to type 2 diabetes. Multiple continuous measures of hyperglycemia were associated with smaller volumes of the subfields fimbria (beta +/- SE < -0.010 +/- 0.011, p-value < 0.001), dentate gyrus (beta +/- SE < -0.013 +/- 0.010, p-value < 0.002), CA3 (beta +/- SE < -0.014 +/- 0.011, p-value < 0.001), and tail (beta +/- SE < -0.006 +/- 0.012, p-value < 0.003). Conclusions/interpretation: Type 2 diabetes and measures of hyperglycemia are associated with hippocampal subfield atrophy, independently of lifestyle and cardiovascular risk factors. We found evidence for a dose-response relationship from normal glucose metabolism, to prediabetes, to type 2 diabetes. Prediabetes stages could give a window of opportunity for the early prevention of brain disease.