Based on risk profiles, several approaches for predicting dementia risk have been developed. Predicting the risk of dementia with accuracy is a significant clinical challenge. The goal was to create a modified dementia risk score (MDRS) based on a big sample size. A total of 239,745 participants from UK Biobank were studied (mean follow-up of 8.7 years). The score value of each risk factor was estimated according to the β coefficient in the logistic regression model. The total dementia risk score was the sum of each risk score. Kaplan Meier survival curves and Cox proportional hazards analyses were used to assess the associations between total score and dementia. Among all participants included, 3531 incident cases of all-cause dementia (ACD), 1729 cases of Alzheimer's disease (AD), and 925 cases of vascular dementia (VD) were identified. Several vascular risk factors (physical activity, current smoking status, and glycemic status) and depressive symptoms were found to be significantly related to dementia risk. The modified dementia risk scores predicted dementia well (model 1, area under curve 0.810; model 2, area under curve 0.832). In model 1, the cut-off value for high risk (HR) was 81 or higher, and in model 2 (including the APOE4), it was 98 or higher. According to Kaplan-Meier survival analyses, patients in the HR group had faster clinical progression (p < 0.0001) in either model 1 or 2. Cox regression analyses for HR versus low risk (LR) revealed that the Hazard radio for ACD was 7.541 (6.941 to 8.193) in model 1 and 8.348 (7.727 to 9.019) in model 2. MDRS is appropriate for dementia primary prevention, and may help quickly identify individuals with elevated risk of dementia.
© 2022. The Author(s).