What is the CAIDE score and how is it calculated for middle‑aged patients?

Checked on January 2, 2026
Disclaimer: Factually can make mistakes. Please verify important information or breaking news. Learn more.

This fact-check may be outdated. Consider refreshing it to get the most current information.

Executive summary

The CAIDE (Cardiovascular Risk Factors, Aging, and Dementia) score is a midlife dementia‑risk algorithm that combines demographic, vascular and lifestyle factors to estimate an individual’s probability of developing dementia roughly 20 years later (developed for ages ~40–65) [1] [2]. It is calculated from point‑weights assigned to age, sex, education, systolic blood pressure, body‑mass index (BMI), total cholesterol, physical activity and—in some versions—APOE ε4 genotype, and performs moderately well in validation studies with area‑under‑curve (AUC) values around 0.77–0.78 [2] [3] [4].

1. What CAIDE measures and why it was created

CAIDE was created to map midlife cardiovascular, metabolic and social risk markers onto long‑term dementia risk because observational data showed that midlife vascular and lifestyle factors predict later dementia; the original aim was to identify people in midlife who might benefit from preventive interventions to reduce late‑life dementia incidence [3] [2]. The score is thus framed as a midlife screening and research tool rather than a clinical diagnostic test for current cognitive impairment [3] [1].

2. The components: the exact inputs used in calculation

The CAIDE score uses easily obtainable midlife variables: age, sex, years of education, systolic blood pressure, BMI, total cholesterol, physical activity and, in extended versions, APOE ε4 carrier status; these items are measured or self‑reported depending on the study or app implementation [2] [5] [3] [6]. The CAIDE app and published papers request numeric values (age, BP, cholesterol, BMI) plus categorical inputs for education and activity level to map into the point system [3].

3. How the score is calculated — points, range and versions

CAIDE assigns point weights to categories of each risk factor derived from regression coefficients in the original cohort; summed points produce a total score (reported range in autopsy and cohort studies is 0–18 points in some implementations) [5]. Two common published versions exist: a basic CAIDE excluding APOE and an extended CAIDE that adds APOE ε4 genotype; studies and apps calculate scores at baseline (midlife) and map total points to estimated 20‑year dementia probability using the original nomogram or app conversion [6] [1] [5].

4. How well CAIDE predicts dementia — performance and validation

Validation studies report moderate discrimination: original and app validations show AUCs around 0.769–0.776 for 20‑year dementia prediction in middle‑aged samples, and larger cohort validations report similar AUC ≈0.78; adding APOE ε4 modestly improves discrimination (AUC up to ~0.81 in one analysis) [3] [4] [7]. CAIDE has also been linked to midlife brain imaging changes—higher scores associate with greater brain atrophy and white‑matter disease—supporting biological plausibility [8] [2].

5. Practical use, modifications and limitations

CAIDE is practical because it uses routine clinical measures and has a mobile app, but its performance varies by population and some components (education cutoffs, lab availability, APOE testing) have been modified in alternative or “mCAIDE” versions to improve applicability; for example, mCAIDE replaces some lab measures with self‑report and adapts education categories for US populations [9] [3]. Several studies caution that predictive capacity is uneven across settings and that CAIDE performs differently when age or APOE effects dominate, so it should be used as a risk‑stratification tool, not a definitive prognosis [7] [10].

6. What it means for a middle‑aged patient

For an asymptomatic middle‑aged person, a higher CAIDE score flags elevated long‑term dementia risk driven largely by age, education and vascular/lifestyle factors; because many included items are modifiable (hypertension, obesity, cholesterol, inactivity), the score can identify targets for midlife risk reduction even though the tool does not prove imminent cognitive decline [6] [8]. The literature supports using CAIDE to inform prevention trials and clinical conversations about vascular risk management, while recognizing population‑specific calibration and the potential benefit of adding biomarkers or APOE status when available [4] [7].

Want to dive deeper?
How is the CAIDE score converted into a 20‑year dementia probability (nomogram or app mapping)?
Which modifiable CAIDE components, when improved in midlife, have shown the strongest evidence for reducing later dementia risk?
How does adding APOE ε4 or insulin‑resistance measures change CAIDE’s predictive accuracy in different populations?