Aims/hypothesis: Diabetes is associated with an increased risk of dementia but the reasons for this association are unclear because there are many potential mechanisms. We explored the relative contribution of diabetes-related variables as predictors of dementia in older individuals with diabetes.
Methods: Survivors, aged > or =70 or more, were recruited from an existing observational cohort study 7.6 +/- 1.0 years after baseline, when they underwent a comprehensive assessment of diabetes, complications and cardiovascular risk factors. Dementia, probable Alzheimer's disease and cognitive impairment without dementia were diagnosed clinically. Logistic regression modelling determined independent predictors of cognitive diagnoses.
Results: Of 302 participants, aged 75.7 +/- 4.6 years, 28 (9.3%) had dementia (16 with probable Alzheimer's disease) and 60 (19.9%) had cognitive impairment without dementia. The major independent longitudinal predictors of dementia were older age (per decade; odds ratio 4.0, 95% CI 1.59-10.10), diabetes duration (for each 5 years; odds ratio 1.69, 95% CI 1.24-2.32), peripheral arterial disease (odds ratio 5.35, 95% CI 2.08-13.72) and exercise (which was protective; odds ratio 0.26, 95% CI 0.09-0.73). For Alzheimer's disease, diabetes duration was an independent predictor in addition to age and diastolic blood pressure. The results of the cross-sectional analyses were similar with respect to diabetes duration and peripheral arterial disease.
Conclusions/interpretation: Peripheral arterial disease is a strong independent risk factor for dementia in diabetes. After adjustment for a wide range of potential risk factors, diabetes duration remains independently associated with dementia and probable Alzheimer's disease, indicating that factors not measured in this study may be important in the pathogenesis of dementia in diabetes.