New Risk Index Predicts Dementia in Elderly

Janis Kelly

May 25, 2009

May 25, 2009 — A new 15-point index that includes both conventional and newly identified dementia risk factors was 81% accurate at identifying older adults who would subsequently develop dementia — a finding that may improve patient selection for future intervention studies.

Deborah E. Barnes, PhD, from the University of California, San Francisco, and colleagues describe the late-life dementia risk index in a study published online May 13 in Neurology.

The researchers tested the new scale in 3375 participants (mean age, 76 years) in the Cardiovascular Health Cognition Study. Fifty-nine percent were women and 15% were African American. None had dementia at baseline.

The index items include older age, poor cognitive test performance, body-mass index, 1 or more apolipoprotein E4 alleles, white-matter disease on cerebral magnetic resonance imaging (MRI), ventricular enlargement on cerebral MRI, internal carotid artery thickening on ultrasound, history of bypass surgery, slow physical performance, and lack of alcohol consumption.

Fourteen percent of the patients (n = 480) developed dementia within 6 years of baseline. This included 56% of subjects with high scores (8 or more points), 23% of subjects with moderate scores (4 to 7 points), and 4% of subjects with low scores (0 to 3 points).

Dr. Barnes told Medscape Psychiatry that the index needs to be validated in different study populations but that most of the risk factors had been identified in previous dementia studies, so there were no true surprises.

"We had hypothesized that the index would include a combination of cardiovascular risk factors, such as hypertension and diabetes, but these turned out to not be as predictive as variables such as brain MRI findings or low cognitive test scores, which are likely to be markers that directly reflect the impact of dementia on the brain," Dr. Barnes said.

Too Expensive for Routine Use?

Constantine Lyketsos, MD, from Johns Hopkins Bayview Medical Center, in Baltimore, Maryland, told Medscape Psychiatry that the index is likely to be very helpful for researchers. "If validated, this will enable us to use dementia risk as a research variable in clinical trials," he said.

Dr. Lyketsos, who is also codirector of the division of geriatric psychiatry and neuropsychiatry and director of the Memory and Alzheimer's Treatment Center, noted that to be clinically useful, an index must include variables that are widely available and not too expensive.

He added that several measures included in the index, such as the Modified Mini-Mental State Exam measure of global cognitive function, the Digit Symbol Substitution Test, carotid Doppler ultrasound to measure intima medial thickness, and the particular type of MRI quantification used, are not routine tests.

"This would be an expensive package of tests for the average person, amounting to thousands of dollars," Dr. Lyketsos said, "particularly as we do not yet have an intervention that we would use for someone with a high-risk score."

Dr. Lyketsos called clinicians' attention to 1 interesting finding — that lack of alcohol consumption is a risk factor for dementia. "People think of alcohol as damaging to cognition, but most data show that low to moderate alcohol consumption actually helps you," he said.

Inclusion of African Americans a Strength

Sharon A. Brangman, MD, who heads the geriatrics division at State University of New  York (SUNY) Upstate Medical University, in Syracuse, noted that 1 strong point of this study is that 15% of the subjects were African American.

"Given the increasing racial/cultural diversity of the geriatric population and the higher rates of Alzheimer's disease that are documented in African American and Latino populations (maybe due to the high rates of diabetes and hypertension in these groups), we need to have data that includes these groups and others. I think 15% African American inclusion in this paper is better than most studies," Dr. Brangman said.

Clinically, Dr. Brangman said that this index might help families with long-term planning and physicians in starting therapy early.

"But I am always torn when faced with the dilemma of identifying a terminal condition for which there is no cure," Dr. Brangman said. "Do we increase the emotional burden when we predict a terminal illness that may or may not occur? What are the ethical considerations when using the risk index in the clinical setting if we cannot offer a cure?

"I do think it is important that we are thinking in terms of a risk index, since increasing our understanding of the disease process, especially in its early stages, can only help in better understanding the disease and the eventual identification of beneficial therapies."

Jerson Laks, MD, director of the Alzheimer Center at the Instituto de Psiquiatria in Rio de Janeiro, Brazil, said that the index would potentially be useful in both research and clinical settings. "However, it should prove to have a high sensitivity/specificity, so that the outcome of dementia may be predicted with a low chance of mistake," he warned.

The researchers are currently seeking funding to develop a shorter index that might have greater clinical utility.

The study was supported by the National Heart, Lung and Blood Institute; the National Institute on Aging; the National Institute of Neurological Disorders and Stroke; and the National Institutes of Health. The researchers have disclosed no relevant financial relationships.

Neurology. Published online before print May 13, 2009. Abstract