Proposed New Criteria for Alzheimer's Diagnosis

Susan Jeffrey

July 13, 2007 — On the strength of new developments in the field, an international group of researchers has proposed new diagnostic criteria for Alzheimer's disease (AD) that they hope will provide an earlier firm diagnosis of the disease.
Their proposal is published as a position paper online July 9 and will appear in the August issue of Lancet Neurology.
"The idea is first, to make the diagnosis earlier, and second, to make the diagnosis with a higher specificity than before, and with these criteria you go more quickly and you go better," Bruno Dubois MD, from the Hôpital de la Salpêtrière and the Université Pierre et Marie Curie, in Paris, France, first author on the paper, told Medscape.
Revised Criteria
For the purpose of research, the current standard for the diagnosis of AD is based on criteria set out by the text revision of the Diagnostic and Statistical Manual of Mental Disorders, 4th ed (DSM-IV-TR) and the National Institute of Neurological Disorders and Stroke — Alzheimer Disease and Related Disorders (NINCDS-ADRDA) working group. This latter document, though, was published in 1984 and the former in 2000.
These accepted criteria proscribe a 2-step process that requires the patient to exhibit a dementia syndrome and then for the physician to apply the criteria based on clinical features of the AD phenotype, the authors write.
"It means you can make the diagnosis of Alzheimer's disease only when the patient is demented, indicating that you could not, and you should not, make the diagnosis before," Dr. Dubois said. "The problem is that I don't know any disease for which there is a limitation of the diagnosis until a certain threshold of severity." In the case of AD, leaving the diagnosis to this stage is analogous to delaying the diagnosis of Parkinson's disease until the patient is bedridden, he said.
"Instead, the diagnosis of PD can be made at the earliest tremor of the hand. So, do we have a 'tremor of the hand' for AD? The answer is yes, as long as we use a good memory task that can pick up the specificity of the memory disorders of AD."
Progress in the field has meant that distinctive and reliable biomarkers for AD are now available through the use of technologies such as structural magnetic resonance imaging (MRI), molecular neuroimaging using positron emission tomography (PET), and cerebrospinal fluid (CSF) analysis.
The new proposed criteria, the authors write, "are centered on a clinical core of early and significant episodic memory impairment." The memory task used should help to define the specific amnestic syndrome of the hippocampal type that is characteristic of AD, Dr. Dubois noted.
For a diagnosis of probable AD, patients should in addition have at least 1 or more abnormal biomarkers among several supporting tests: structural neuroimaging on MRI looking for hippocampal atrophy; a specific metabolic pattern on molecular neuroimaging with PET and CSF analysis of amyloid-beta or tau proteins; or the presence of 1 of 3 autosomal dominant mutations related to AD on genetic testing.
Their criteria, though, should be validated in existing and prospective cohorts to optimize their sensitivity, specificity, and accuracy, they add.
"The timeliness of these criteria is highlighted by the many drugs in development that are directed at changing pathogenesis, particularly at the production and clearance of amyloid-beta as well as the hyperphosphorylation state of tau," the authors conclude.
Dr. Dubois pointed out, for example, that phase 3 results with the antiamyloid drug tramiprosate (Alzhemed, Neurochem) are eagerly awaited by the Alzheimer's community. "The day that a treatment will be positive in terms of reducing disease progression, this day, these criteria will be very useful, because we will need to pick up the patients at the early stage of the disease."
"Seize the Opportunity"
In an editorial accompanying the paper, Norman L. Foster, MD, from the Center for Alzheimer's Care, Imaging, and Research at the University of Utah, in Salt Lake City, while supportive of the need for review of current AD diagnostic criteria in light of new technologies, is cautious about whether this particular proposal is ready for clinical use.
"Although presented as criteria, the proposal from Dubois and colleagues instead should be thought of as a crucial framework for future discussion," he writes. "As currently constructed, the criteria would be impossible to implement, because they do not include an operational definition of methods and abnormal results."
How deficits in episodic memory are evaluated is especially important, because this cognitive measure is "crucial" to determine the disease state, he notes, "but the methods used for this evaluation have not been shown to be analytically reliable and diagnostically valid." In addition, previous guidelines for evaluating biomarkers in AD have been "notoriously difficult to meet and have not been widely accepted. Nevertheless," he adds, "adoption of this new framework would provide a focus and stimulus for the validation studies that are needed to advance the field."
Much work remains, Dr. Foster concludes. "However, great progress could be achieved relatively easily if research studies continue to use NINCDS-ARDRA criteria and also implement procedures that allow evaluation of the new criteria."
Ongoing epidemiological studies such as the Alzheimer's Disease Neuroimaging Initiative will provide additional data to support this effort, he writes. "We should seize this opportunity to reopen the discussion of Alzheimer's disease diagnosis. The time is right to use the advanced technology at our disposal to improve the early, accurate diagnosis of dementia and develop more effective treatments."
Defining the Underlying Disease
Dr. Dubois, though, seemed puzzled about Dr. Foster's assertion that the criteria are not ready for use in the clinic. "These criteria are not in fact so difficult," he said. "We have only mixed very usual techniques, adding to memory testing either MRI or lumbar puncture. I'm aware this is not totally insignificant, but when you are faced with a young patient — and in France, we think there are about 30,000 patients under the age of 65 years — then I think it's important to be sure of the diagnosis, and today lumbar puncture and the study of beta amyloid and the hyperphosphorylation state of tau in the CSF is done in most neurological centers."
The point is to take a more biological approach to defining the disease underlying mild cognitive impairment (MCI), he said. He compared a diagnosis of MCI to a diagnosis of fever. "A fever is not the important thing to know. What is important is if the fever is due to meningitis or tonsillitis, because it's not the same prognosis and it's not the same treatment. Similarly, what we need to know is the type of MCI, because it's not the same to have MCI because of depression or because of Alzheimer's disease."
The authors report no conflict of interest.
Lancet Neurol. Published online July 9, 2007.