Can Dementia Be Predicted 20 Years in Advance?

To develop a model for predicting middle-aged people’s risk for dementia in older age, researchers tracked 1409 participants in the Cardiovascular Risk Factors, Aging, and Dementia (CAIDE) study who were recruited in the 1970s and 1980s. CAIDE participants came from four separate population samples. They completed a baseline questionnaire about health, medical history, and health behaviors. During a mean follow-up of 21 years, dementia was diagnosed via a 3-step cognitive assessment. The authors used logistic-regression analysis to determine the dementia risk score (range, 0–15).

Sixty-one participants (4%) were diagnosed with dementia during follow-up. Dementia was predicted by higher baseline age (≥ 47 years), low education (<10 years), hypertension, hypercholesterolemia, and obesity. Risk for dementia by predictor score was:

• 0 to 5, 1%;

• 6 to 7, 1.9%;

• 8 to 9, 4.2%;

• 10 to 11, 7.4%; and

• 12 to 15, 16.4%.

With a cut-off of 9 points or more, the sensitivity was 0.77, specificity 0.63, and negative predictive value 0.98. The authors conclude that this dementia risk score highlights the role of vascular factors in dementia and that this novel approach can aid in the prediction of later-life dementia risk.