A panel of 10 plasma protein biomarkers shows the potential to predict progression from mild cognitive impairment to Alzheimer’s disease within a year of blood sampling, but questions about the tool’s predictive value in a real-world setting have drawn some skeptical assessments of the work.
Previous studies have identified various potential Alzheimer’s disease blood biomarkers, and, as recently as March 2014, a possible lipid plasma panel was described, but these new findings represent an important next step for the research, said coauthor of the new study Simon Lovestone, PhD, MRCPsych, from King’s College London, United Kingdom.
“I think that this is one of the most promising sets of markers that both replicate and appear to have sufficient utility value to be worth further development,” said Dr. Lovestone, who is director of the National Institute for Health Research’s Biomedical Research Centre for Mental Health at the South London and Maudsley NHS Trust and the Institute of Psychiatry.
This is especially so, he added, “in relation to the real need, which is for predictors of progression from very early symptomatic pre-disease states such as mild cognitive impairment to Alzheimer’s disease,” he told Medscape Medical News.
Their findings were published online July 7 in Alzheimer’s & Dementia.
For the study, the researchers evaluated blood samples from 3 international cohorts, including 476 patients with Alzheimer’s disease, 220 with mild cognitive impairment, and 452 healthy controls without dementia.
The samples were analyzed for 26 proteins previously identified as being associated with Alzheimer’s disease, and a subgroup of 476 patients across the 3 groups also received brain MRI.
Using a multiplex antibody capture platform, Dr. Lovestone and his team identified 10 of the 26 proteins that correlated with the progression of mild cognitive impairment to Alzheimer’s disease within a year, with an optimal sensitivity of 84.6% and specificity of 88.1%.
Among the key biomarkers identified were plasma clusterin concentration, associated with brain atrophy, disease severity, and rate of progression; transthyretin, an amyloid-binding protein associated with the rate of disease progression; cystatin C; and the APOE genotype.
The primary value of the panel would be in the research setting and used in combination with other modalities, Dr. Lovestone explained.
“This work was predicated on the need for markers to enrich clinical trials of secondary prevention,” he said.
“PET [positron emission tomography] imaging and CSF [cerebrospinal fluid] markers are good ways to do this but have obvious limitations. We seek a test that would be easy to administer that would help recruitment and reduce screen failure in clinical trials.”
“It is likely that a test in blood would in reality be combined with other tests such as PET or CSF.”
With media exposure of the test inevitably speculating on its potential clinical application has come some skepticism, focusing specifically on the positive predictive value.
In an analysis of the study published by NHS Choices, authors with the United Kingdom–based healthcare analysis group Bazian argued that, in the context of a reported 10% to 15% of people with mild cognitive impairment known to convert each year to Alzheimer’s disease, the predictive ability of a positive test drops to only about 50%.
“This means that those who have a positive test have a 50:50 chance of going on to have Alzheimer’s,” they write. “Consequently, on its own, this test is unlikely to be much good for use in clinical practice for the general population.”
Their conclusion, however, is still in line with Dr. Lovestone’s assessment of the test’s research value.
“However, refining this test and combining it with other methods (such as a lipid test) might improve accuracy rates, making it a viable predictive tool in the future,” they conclude.
In response to the analysis, Dr. Lovestone asserted that the critique simply got the figures wrong. The predictive value found in his study, he said, was not based on the denominator of the 10% to 15% rate of conversion rate that is cited in the Health Choices article but on a rate of 23%.
“In our paper the figures are very clear — the relevant denominator or prevalence of conversion is 23%,” he said. “Our critics have made a basic mistake of not paying attention to the denominator and using a figure culled from a website, which is a mean/average.”
Conversion rates in the literature vary from less than 10% among people with mild cognitive impairment in epidemiologic and random screening studies to more than 20% to 30% among patients in memory clinics, he said.
“[The range] is not really surprising because in epidemiological studies [patients] are not complaining and are therefore presumably not worried. Those in memory clinics are often desperately worried or their relatives are. These figures are widely replicated — we find the same and so does Alzheimer’s Disease Neuroimaging Initiative.”
The bottom line, Dr. Lovestone asserted, is that the findings need to be reproduced before any solid conclusions can be drawn.
“It is important to note that replication in far bigger studies is needed,” he said. “We emphasize this in the paper and in all discussions. The value we see is in clinical trials enrichment.”
That’s not to say there isn’t any potential for clinical utility, but that hasn’t been the researchers’ intention, Dr. Lovestone added.
“It would require simply huge studies to get to that point, [and] it would not be used alone — what test ever is?”
However, he said, “there is the potential for the test to have some value in the clinical ‘work up’ of a patient presenting to a memory clinic with mild clinical impairment, just as MRI, PET, CSF have value. Again, if it replicates in further studies.”
Neurologist and Alzheimer’s disease expert Richard J. Casellia, MD, who is medical director for service at the Mayo Clinic in Scottsdale, Arizona, commented that the 10% to 15% conversion rate is not uncommon.
“I think [Dr. Lovestone’s] response is reasonable, although the conversion rate I use is derived from the Mayo Clinic/Olmsted County population, and that is about 10% to 15% — probably closer to 15%,” he told Medscape Medical News.
In offering his take on the plasma panel’s utility, Dr. Casellia suggested that the advantage over other biomarkers remains to be seen.
“Not long ago the Alzheimer’s panel with plasma lipids was in the news,” he said. “It is true that a blood test–based biomarker would be highly desirable, but I am not sure there is an obvious advantage that one panel has over another.”
“A related issue which the study does not address is that of Alzheimer’s disease patients with negative amyloid PET scans, a sizable number of whom are APOE ε4 noncarriers, so about half the clinical Alzheimer’s disease population,” he noted.
That aside, however, Dr. Casellia said the panel nevertheless represents a welcome development with potentially important benefits — even in the clinical setting.
“I do also see potential value of this in the clinical setting and agree with the authors’ call for a replication study for that purpose.”