부산대학교 도서관

Background/Aims. Diabetes might increase the risk of Alzheimer’s disease (AD). For detecting dementia, it is typical to obtain informants’ perceptions of cognitive deficits, but such interviews are usually difficult in routine care. We aimed to develop a model for predicting mild to moderate AD using a self-reported questionnaire and by evaluating vascular risk factors for dementia in elderly subjects with diabetes. Methods. We recruited 286 diabetic and 155 nondiabetic elderly subjects. There were 25 patients with AD and 261 cognitively normal individuals versus 30 with AD and 125 normal subjects, respectively. Each participant answered subjective questions on memory deficits and daily functioning. Information on vascular risk factors was obtained from clinical charts, and multivariate logistic regression was used to develop a model for predicting AD. Results. The predicted probabilities used in screening for AD in diabetic subjects constituted age, education, lower diastolic blood pressure, subjective complaints of memory dysfunction noticeable by others, and impaired medication, shopping, and travel outside a familiar locality. Receiver operating characteristic analysis revealed a satisfactory discrimination for AD specific for diabetic elderly subjects, with 95.2% sensitivity and 90.6% specificity. Conclusion. This is the first useful index that can prescreen for AD in elderly subjects with diabetes.