Abstract
Objective
This study identified factors in patients with diabetes associated with risk of developing dementia.
Research Design and Methods
This retrospective, longitudinal study used a national cohort of US Veterans with diabetes 65โyears or older to examine incidence of dementia over 2โyears. A multivariable Cox regression model was used to estimate risk of developing dementia associated with sociodemographic factors, use of diabetes medications, and duration of diabetes.
Results
In all, 377โ838 patients (average age, 75.53โยฑโ6.07โyears) were studied. Over the 2โyear followโup, 14โ580 (3.85%) were diagnosed with dementia. Major risk factors for dementia were age >75โyears (75โ85โyears, hazard ratio [HR] 2.092, 95% confidence interval [CI] 2.017โ2.169; โฅ85 years, HR 3.468, CI 3.274โ3.672), race black versus white, HR 1.218, CI 1.164โ1.274), Southern residence (HRโ=โ1.181, CI 1.133โ1.232), and diabetes duration (HR for 5โyears or more, 1.428, CI 1.357โ1.504). There was a reduced HR for dementia with use of an oral hypoglycemic agent (HR 0.940, CI 0.909โ0.972) and HMGโCoA reductase inhibitors (HR, 0.875, CI 0.846โ0.906). There was no change in HR with insulin use (HR 1.024, CI 0.983โ1.067).
Conclusions
Several important factors were identified that are associated with increased dementia risk, and two factors were identified that are associated with reduced risk. It will be important to ascertain whether riskโfactor modification reduces the HR for dementia in persons with diabetes, and to further examine effects of medication use for comorbid conditions. Copyright ยฉ 2010 John Wiley & Sons, Ltd.