During a median follow-up of 7 years, 25% of participants were diagnosed with dementia. Among participants who did not have diabetes, risk for developing dementia increased with increasing average glucose levels, after adjustment for potentially confounding variables. For example, in those whose average glucose level was 115 mg/dL, relative risk for dementia was 18% higher than in those whose average glucose level was 100 mg/dL. Among participants who had diabetes, relative risk for dementia was 40% higher in those whose average glucose level was 190 mg/dL compared with 160 mg/dL.

COMMENT

The prospective nature of this study, in which patients screened negative for dementia at baseline, is a strength. However, unmeasured confounders might have influenced the association between glycemia and dementia, and reverse causality is remotely possible (e.g., lifestyle changes in patients with early subclinical dementia might promote higher glucose levels). If higher blood glucose levels within the nondiabetic range do contribute to development of dementia, the mechanism is unclear.