Are Blood Tests for Alzheimer's Disease Oversold?

Published: Mar 10, 2014 | Updated: Mar 11, 2014

By John Gever, Deputy Managing Editor, MedPage Today
Reviewed by F. Perry Wilson, MD, MSCE; Instructor of Medicine, Perelman School of Medicine at the University of Pennsylvania and Dorothy Caputo, MA, BSN, RN, Nurse Planner
save|AA

http://www.medpagetoday.com/Neurology/Dementia/44688
LINK TO WEB PAGE
For best viewing, click the bottom right corner for full screen.
Action Points
Blood biomarkers in cognitively normal seniors were associated with their 3-year risk of developing mild cognitive impairment or Alzheimer's disease, although the accuracy fell short of what would normally be acceptable for a screening test.
Note that the researchers focused on unconventional markers considered to be components of the blood "metabolome" and "lipidome," metabolites resulting from cellular processes and lipid molecules, including 10 plasma phospholipid molecules.
Be aware that the positive predictive value of a test is highly dependent upon the prevalence of disease in the population studied; large-scale screening of healthy adults would lead to an unacceptable false-positive rate, but targeted screening of high-risk individuals may be more efficacious.
Blood biomarkers in cognitively normal seniors were associated with their 3-year risk of developing mild cognitive impairment or Alzheimer's disease, researchers said, although the accuracy fell short of what would normally be acceptable for a screening test.

Levels of 10 plasma phospholipid molecules -- none of them conventional markers for Alzheimer's disease -- distinguished initially healthy individuals 70 and older who developed cognitive impairments during follow-up from those who remained cognitively normal in a 525-person study, according to Howard J. Federoff, MD, PhD, of Georgetown University in Washington, D.C., and colleagues.

For this distinction, the 10-marker panel had an area under the receiver-operating characteristic curve (AU-ROC) of 0.92 (95% CI 0.87-0.99), with sensitivity and specificity each at 90%, they reported online in Nature Medicine.

"We consider our results a major step toward the commercialization of a preclinical disease biomarker test that could be useful for large-scale screening to identify at-risk individuals," Federoff said in a Georgetown press release.

But the data did not appear to fully support that optimism. If the study cohort's 5% rate of conversion from normal cognition to mild impairment or Alzheimer's disease is representative of a real-world screening population, then the test would have a positive predictive value of just 35%. That is, nearly two-thirds of positive screening results would be false.

In general, a positive predictive value of 90% is considered the minimum for any kind of screening test in normal-risk individuals.

On the other hand, the use of markers unrelated to the APOE gene or beta-amyloid and tau proteins may represent a welcome new direction for Alzheimer's disease risk prediction.

For instance, the test could hold promise as part of clinical trials of new treatments, by enriching patient samples with those progressing quickly to clear impairment, Gisele Wolf-Klein, MD, director of geriatric education at North Shore-LIJ Health System in New Hyde Park, N.Y., told MedPage Today in an email.

"[It] would be a major step in assisting the pharmaceutical industry in producing disease-modifying therapies at both early and preclinical stages of dementia," she said.

Wolf-Klein added, "Further research will enable healthcare practitioners to determine the actual benefits and relevance of this test in clinical practice."

For the study, Federoff and colleagues recruited 525 community-dwelling, physically healthy people who were 70 and older and followed them for up to 5 years with annual examinations.

A total of 74 were determined to have either amnestic mild cognitive impairment or Alzheimer's disease at some point, including 46 at study entry and 28 who converted from normal to impaired during follow-up. These conditions were diagnosed according to standard criteria, based on a suite of 14 neurocognitive tests.

To identify potentially predictive biomarkers, Federoff and colleagues analyzed baseline blood samples from 53 who showed cognitive impairments at the end of year three (including 18 who had converted from normal cognition) and 53 who remained cognitively normal.

The researchers focused on unconventional markers considered to be components of the blood "metabolome" and "lipidome" -- that is, metabolites resulting from cellular processes and lipid molecules. Ten of these were ultimately found to be either significantly increased or decreased in the impaired individuals versus the normal controls. They included amino acids such as proline and lysine, the neurotransmitter serotonin, and others that Federoff and colleagues suggested were indicators of "cell membrane integrity."

"We posit that this 10-phospholipid biomarker panel, consisting of phosphatidylcholine and acylcarnitine species, reveals the breakdown of neural cell membranes in those individuals destined to phenoconvert" from cognitively normal to impaired, they wrote.

Statistical modeling suggested an optimal set of cutoffs for the 10 markers to distinguish impaired individuals from controls. The researchers then applied it to blood samples from a different subset of the study participants, including 10 converters, 20 normal controls, and 11 with impairments at study entry.

The model actually did less well in discriminating the entire group of impaired participants from normal controls (AU-ROC 0.77) than in predicting who would convert from normal to impaired (AU-ROC 0.92).

Incorporating APOE genotype into the model did not improve its performance, the researchers said.

Federoff and colleagues indicated that their biomarker panel's performance in picking up incipient cognitive impairment was better than blood tests for beta-amyloid and tau proteins. Levels of these classical markers of Alzheimer's disease in cerebrospinal fluid have shown better predictive and diagnostic power, but no test requiring lumbar puncture will catch on for large-scale screening of asymptomatic people.

On the other hand, at least two other blood tests for unconventional markers had shown good results for discriminating individuals with current impairment from normal controls and for predicting progression from mild impairment to overt dementia, Federoff and colleagues noted, suggesting that this is a fruitful area to explore.

Keith L. Black, MD, a neurosurgeon and researcher at Cedars-Sinai Medical Center in Los Angeles, told MedPage Today that it would be important to learn more about how these markers connect with neurological dysfunction.

For the biomarkers found in the study, "it's difficult to understand how they can be specifically linked to Alzheimer's disease versus some sort of diffuse neuronal abnormality or connectivity abnormality or synaptic connectivity abnormality," he said.

"The question, I think, becomes for future studies how well do we understand the underlying biology of it, whether it's specific to Alzheimer's disease or whether you might see these same types of changes in these blood biomarkers with other types of diffuse neuronal loss or injury, other types of dementia, or ischemic changes, for example."

In a more restrained vein than Federoff took in the press release, their Nature Medicine report concluded by saying their biomarker panel "requires external validation using similar rigorous clinical classification before further development for clinical use. Such additional validation should be considered in a more diverse demographic group than our initial cohort."

The study was funded by the National Institutes of Health. Study authors declared they had no relevant financial interests.