Background: Dementia is a significant cause of disability and dependency. Persons with high dementia risk but intact cognition will benefit from preventive interventions.

Objectives: The aim was to validate dementia risk score Cardiovascular Risk Factors, Aging and Incidence of Dementia (CAIDE) in a national population-based cohort with data on age, education, hypertension, obesity, hyperlipidemia and physical activity. Secondly, we examined if substituting obesity item with Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) would improve predictive value of CAIDE risk score.

Design: Longitudinal, population-based cohort study.

Setting: General population, Finland PARTICIPANTS: Representative sample of Finnish adult population aged over 30 years from Health 2000 Survey (n = 5,806).

Measurements: CAIDE dementia risk score and substituting BMI with HOMA-IR.

Results: Dementia was diagnosed in 571 (9.8 %) participants during the 19 years follow-up. CAIDE risk score predicted dementia well: AUC (area under curve) ROC (receiver-operating characteristic) was 0.78 (95 % CI from 0.76 to 0.79). Secondly, replacing obesity with HOMA-IR in CAIDE risk score generated similar results: ROC AUC 0.78 (95 % CI from 0.76 to 0.80). Adding APOE ε4 status further improved predictive value of risk score: ROC AUC 0.81 (95 % CI from 0.80 to 0.83).

Conclusions: CAIDE dementia risk score predicts dementia well in a national population-based cohort. Adding APOE ε4 genotype improved predictive value of risk score. Insulin resistance measured by HOMA-IR is comparable to obesity as part of CAIDE risk score. These findings imply that CAIDE risk score is applicable for assessing risk of dementia and highlight importance of modifiable risk factors of dementia.