Supplemental Digital Content is available in the text.

OBJECTIVETo investigate the association between diabetes and brain or hippocampal atrophy in an elderly population. RESEARCH DESIGN AND METHODSA total of 1,238 community-dwelling Japanese subjects aged ‡65 years underwent brain MRI scans and a comprehensive health examination in 2012. Total brain volume (TBV), intracranial volume (ICV), and hippocampal volume (HV) were measured using MRI scans for each subject. We examined the associations between diabetes-related parameters and the ratios of TBV to ICV (an indicator of global brain atrophy), HV to ICV (an indicator of hippocampal atrophy), and HV to TBV (an indicator of hippocampal atrophy beyond global brain atrophy) after adjustment for other potential confounders. RESULTSThe multivariable-adjusted mean values of the TBV-to-ICV, HV-to-ICV, and HV-to-TBV ratios were significantly lower in the subjects with diabetes compared with those without diabetes (77.6% vs. 78.2% for the TBV-to-ICV ratio, 0.513% vs. 0.529% for the HV-to-ICV ratio, and 0.660% vs. 0.676% for the HV-to-TBV ratio; all P < 0.01). These three ratios decreased significantly with elevated 2-h postload glucose (PG) levels (all P for trend <0.05) but not fasting plasma glucose levels. Longer duration of diabetes was significantly associated with lower TBV-to-ICV, HV-to-ICV, and HV-to-TBV ratios. The subjects with diabetes diagnosed in midlife had significantly lower HV-to-ICV and HV-to-TBV ratios than those without and those diagnosed in late life. CONCLUSIONSOur data suggest that a longer duration of diabetes and elevated 2-h PG levels, a marker of postprandial hyperglycemia, are risk factors for brain atrophy, particularly hippocampal atrophy.

Short and long daily sleep duration and hypnotic use are risk factors for dementia and death in Japanese elderly adults.