“This new risk index could be very important both for research and for people at risk of developing dementia and their families," says study author Deborah Barnes of the University of California, San Francisco and the San Francisco Veterans Affairs Medical Center. “It could be used to identify people at high risk for dementia for studies on new drugs or prevention methods.”

Barnes says the tool could also identify people who have no signs of dementia but should be monitored closely, allowing them to begin treatment as soon as possible. Such intervention could help these people maintain their thinking and memory skills and quality of life longer.

 

The risk index is a 15-point scale. People who score eight or more points on the scale are at high risk of developing dementia in the next six years. Several of the items on the scale are well-known risk factors for Alzheimer's disease, such as older age, low scores on tests of thinking skills, and having a gene that has been linked to the disease.

 

Other factors predicting dementia risk were less obvious: People who are underweight, do not drink alcohol, have had coronary bypass surgery, or are slow at performing physical tasks such as buttoning a shirt are more likely to develop dementia than people who do not have these risk factors.

 

To develop the index, researchers in the Cardiovascular Health Study examined 3,375 people with an average age of 76 and no evidence of dementia and followed them for six years. During that time, 480 of the people, or 14 percent, developed dementia. The researchers then determined which factors best predicted who would develop dementia and created the point index.

 

A total of 56 percent of those with high scores on the index developed dementia, compared to 23 percent of those with moderate scores and four percent of those with low scores. Overall, the index correctly classified 88 percent of the participants.

 

Barnes says the risk index will need to be validated with other studies, and she and her colleagues are evaluating whether a shorter, more simplified index could be as accurate as this index.