Redefining Alzheimer's Disease: NIA and Alzheimer's Association Float New Draft Diagnostic Criteria

Susan Jeffrey

July 14, 2010

July 14, 2010 (Honolulu, Hawaii) — Draft reports from 3 workgroups presented here at the Alzheimer's Association International Conference on Alzheimer's Disease 2010 will form the basis for new diagnostic criteria for mild cognitive impairment (MCI) and Alzheimer's disease (AD) — the first update of the current criteria in 25 years — as well as outline a new category of preclinical AD.

The workgroups were convened by the National Institute on Aging and the Alzheimer's Association to update criteria established by the National Institute of Neurological Disorders and Stroke/Alzheimer's Disease and Related Disorders Association — now the Alzheimer's Association — in 1984.

Dr. Reisa Sperling, Dr. Marilyn Albert, and Dr. Guy McKhann

Although it is generally agreed that the earlier diagnostic criteria have stood the test of time, the push for new criteria has come largely from developments in the meantime, said Creighton H. Phelps, PhD, director of the Alzheimer's Disease Centers Program in the Division of Neuroscience at the National Institute on Aging, who led off the discussion here.

There are now 3 dominantly inherited genes related to the disease, for example, he noted. "Imaging has just blossomed, so we now have lots of new imaging techniques, and better neuropsychological techniques and tests," Dr. Phelps said. The concept of MCI was only proposed 10 to 15 years ago, and many biomarkers, including positron emission tomography (PET) imaging for amyloid and other new techniques, have been developed.

"The overall goals of this project are to better define the natural history of AD from the asymptomatic stages to full-blown dementia," Dr. Phelps said of the new diagnostic criteria. "We also want to attempt to relate the clinical symptoms as they emerge to the underlying pathophysiology. We want to use what present knowledge we have to better diagnose the disease, and above all, we want to set the research agenda that's going to fill in the gaps, so we can reach the above goals sooner."

The workgroup reports were presented during a session entitled "Redefining Alzheimer's Disease" here. A Web site was launched immediately after the presentation inviting further input on the proposed criteria.

AD Dementia Workgroup

Guy McKhann, MD, from Johns Hopkins University School of Medicine in Baltimore, Maryland, who chaired the Alzheimer's Disease Dementia Workgroup, was instrumental to the definition of current diagnostic criteria.

Again, although those criteria have been successful, they have some shortcomings, Dr. McKhann noted, most relating to the lack of knowledge at the time they were developed. For example, there was a lack of distinguishing features from other dementing conditions not fully recognized 25 years ago, and no inclusion of biomarkers or genetics, and the criteria greatly emphasized memory as the primary deficit, rather than other presentations.

The aim of this group was to establish criteria that would have wide application and could be used not just by academics but also by people in the community who may not have access to neuropsychological testing or advanced imaging or cerebrospinal fluid (CSF) measures, and also be of use to industry, Dr. McKhann noted.

The report includes proposed criteria for the diagnosis of all-cause dementia, as well as for the diagnosis of AD dementia.

For AD dementia, disease criteria include an insidious onset, with symptoms worsening over months or years. Deficits in 2 or more cognitive domains are established by history and examination. "The most common presentation is the amnestic presentation, but there are others," Dr. McKhann said.

Nonamnestic presentations include language problems often confused with other conditions such as frontotemporal lobe degeneration, primary progressive aphasia, or semantic dementia, he noted. Others start with visual–spatial dysfunction or executive function.

The original criteria dealt in levels of certainty, he said, with definite autopsy-proven AD as the gold standard, followed by probable AD and possible AD. For the new criteria, Dr. McKhann said, "We have essentially retained these 3 categories, but modified them."

The workgroup proposes the following description of AD dementia:

  1. Pathologically proved AD dementia: This diagnosis requires the patient to meet the clinical and cognitive criteria for AD dementia during life and then have proven AD by pathological examination. Dr. McKhann noted that this diagnosis does not include those people who had no symptoms in life but show AD pathology on autopsy. "In other words, we're talking about a clinical diagnosis with a pathological correlation — we're not talking about a primary pathological diagnosis."

  2. Probable AD dementia: These patients meet the clinical and cognitive criteria for dementia and do not have evidence of alternate diagnoses, particularly cerebrovascular disease. Probability can be enhanced by factors including a documented longitudinal decline and positive evidence from biomarkers, or they may be an AD mutation carrier. br Biomarkers include low CSF Aβ42, elevated CSF tau or phospho tau, positive amyloid PET imaging, decreased fludeoxyglucose (FDG) uptake on PET in temporoparietal cortex, disproportionate atrophy on structural MR in medial temporal (especially in the hippocampus), basal and lateral temporal lobe, and medial parietal isocortex.

  3. Possible AD dementia: This group includes patients who have an atypical course; that is, those who meet other clinical or cognitive criteria, but for whom information on the course of progression is lacking or uncertain. It also includes those who meet clinical and cognitive criteria but who are negative for biomarkers. Finally, the group includes those with mixed presentation. "This is an area that we didn't discuss in the initial criteria very much at all, but we're recognizing very much that patients can have concomitant disorders, particularly cerebrovascular disease and Lewy body disease," Dr. McKhann said.

  4. Not AD dementia: These patients clearly do not have AD, he said. "They don't have the clinical criteria, and they have sufficient evidence of alternative diagnoses that we don't think are concomitant with AD," such as HIV dementia or Huntington's disease.

MCI Due to AD Workgroup

The MCI workgroup was chaired by Marilyn Albert, MD, also from Johns Hopkins University School of Medicine, who presented their findings here. Their charge, as they understood it, was to develop criteria for the symptomatic predementia phase of AD.

"They purposely didn't provide a name for this phase, but we very quickly decided amongst ourselves that the overarching concept that we would use was MCI due to AD," Dr. Albert said, building on what she called the "seminal work" of Ron Peterson, MD, from the Mayo Clinic, Rochester, Minnesota.

Because the group was told the criteria would be regularly evaluated and updated, "that encouraged us to go out somewhat on a limb," she noted, and incorporate what is currently known about biomarkers without having to set them in stone.

Four clinical and cognitive criteria for MCI due to AD were developed — similar but slightly different from original criteria for MCI:

  1. Concern regarding a change in cognition: This could be a concern about a change in cognition from the patient's previous level but can also be identified by the patient, an informant, or a skilled clinician. In previous criteria, the cognitive concern had to be expressed by the patient and corroborated by an informant.

  2. Impairment in 1 or more cognitive domains: Performance should be lower than would be expected from the patient's age and education. "We emphasize in the document that episodic memory impairment is clearly the most common impairment seen in the prodromal phase to AD, but we allow for other domains to be impaired," Dr. Albert notes, "and we don't specify cutoffs." Instead, they note that the impairment is typically 1 to 1.5 standard deviations below the mean of the individual, adjusted for age and education. They also acknowledge there may be impairments in more than 1 domain. However, they have not at this point incorporated the terms amnestic or nonamnestic MCI, she noted.

  3. Preservation of independence in functional abilities: They acknowledge that those with MCI may have mild problems with complex tasks such as paying bills, preparing meals, or shopping, taking more time or making more errors. "But the critical thing is they have the ability to maintain independence of function with minimal aids and assistance."

  4. Not demented: The cognitive changes should be sufficiently mild that there is no evidence of impairment in social or occupational function. The changes are intra-individual changes, she notes, which can be challenging to establish when seeing patients for the first time.

"Overall, what we repeatedly say in our criteria as well, not only in the clinical section but in the biomarkers section, is that in making the diagnosis, it really ends up being a matter of clinical judgment, that you can't have certain cutoffs, you can have just rules that everyone is going to follow," Dr. Albert added. "You have to take into account the entire individual in making the final diagnosis of MCI due to AD."

As the AD dementia group did, the MCI due to AD workgroup also provided proposed terminology for classifying these patients with varying levels of certainty, including the proposed use of biomarkers:

  1. MCI of a neurodegenerative etiology: The patient meets the clinical and cognitive criteria for the disorder, and biomarkers may not have been tested, they may have been tested and are ambiguous, or molecular biomarkers may be negative. "We acknowledge that in that instance there would be a very low likelihood of underlying AD pathology, but we still think that that situation should be called MCI of a neurodegenerative etiology, assuming that the patient meets the other clinical and cognitive criteria," she noted.

  2. MCI of the Alzheimer type: The patient meets clinical and cognitive criteria for the disorder, plus has positive findings from 1 of the "downstream" biomarkers of structural or functional change, such as MRI evidence of hippocampal atrophy, or FDG PET alterations. "Consistent with this diagnosis would be that you had not acquired any molecular biomarkers, or that you had evidence from biomarkers that was equivocal."

  3. Prodromal Alzheimer's dementia: The highest level of certainty, in which the individual meets the clinical and cognitive criteria for MCI, plus has biomarker evidence to suggest underlying AD pathology. "Of course in this context, at this moment in time, what we're referring to is low CSF Aβ42, or amyloid accumulation with PET imaging," Dr. Albert noted.

"The most novel aspect of these criteria I think is obvious, which is the incorporation of biomarkers to increase the certainty of the diagnosis, and the fact that we recognize that there's much more that needs to be learned about all of these biomarkers to be able to move them out into community settings," she concluded. "But we believe the time is right, and we urge everyone to move ahead."

Preclinical AD Workgroup

Finally, the third group, chaired by Reisa Sperling, MD, from Harvard Medical School in Boston, was charged with what Dr. Phelps called the most difficult of the tasks — defining a preclinical stage of AD.

"There's converging evidence from both genetic at-risk cohorts and aging cohorts that the pathophysiologic process of AD begins years, probably more than a decade, prior to the time that we make a diagnosis of dementia," Dr. Sperling said, "This long preclinical phase of AD provides a critical opportunity for potential intervention with disease-modifying therapy."

Animal models suggest intervention might be most efficacious in this stage, and economic models show the largest public health benefit may come from intervening early in the process.

The group struggled somewhat with terminology for this phase but settled on preclinical AD, to emphasize the continuum toward MCI and AD, Dr. Sperling noted. This includes individuals who are completely asymptomatic as well as those who have subtle decline but do not yet meet criteria for MCI.

"There's already been some controversy about the idea that we would define a preclinical stage of the disease, and we thought it might be useful to point out that this is actually common in multiple other examples of disease," Dr. Sperling noted.

Carcinoma in situ or asymptomatic coronary artery disease are routinely defined, and symptoms are not required for a diagnosis, she said. Diseases such as kidney disease are detected by blood tests, and treatment can change the course of disease before symptoms emerge. "Importantly, not all individuals who have these risk factors or early stages of these other diseases ever go on to manifest symptoms," Dr. Sperling pointed out. Still, treating high cholesterol has prevented many overt cardiac and stroke events.

"The challenge in AD is that we don't yet have definitive evidence linking the pathologic process of AD to the emergence of clinical symptoms," Dr. Sperling noted. "However, we hypothesized that amyloid-β is an early inciting event that is necessary but may not be sufficient to cause the clinical syndrome of AD, and that there are likely multiple factors which mediate the relationship between AD pathologic changes — both amyloid and other processes — and the clinical manifestations."

Based on currently available, and admittedly limited, data, the workgroup proposes draft operational research criteria for preclinical AD. The criteria are postulates only, she emphasized, based on the hypothesis that the process of AD begins far in advance of symptoms, and that amyloid-β accumulation, "or cerebral amyloidosis, as we decided to call it, is one of the earliest measurable stages of preclinical AD, and that biomarker positivity in combination with other biomarkers may have implications for clinical decline and responsiveness to treatment."

Dr. Sperling added: "These are not intended as clinical diagnostic criteria. They are for research purposes, and importantly, it's very clear that the prognostic utility of these criteria for an individual remains unclear."

The workgroup used stages 1 through 3 for these criteria.

  1. Stage 1 — asymptomatic cerebral amyloidosis: This would include evidence of amyloid-β accumulation, either by low CSF Aβ42 measures or elevated PET amyloid tracer retention, but individuals would be normal on all measures of cognitive function.

  2. Stage 2 — amyloidosis plus evidence of early neurodegeneration: This is evidence of an AD-like pattern of abnormality on other markers — synaptic dysfunction, increased CSF tau or phospho tau, cortical volume loss, gray matter loss, or thinning or hippocampal atrophy. Individuals may still have normal cognition.

  3. Stage 3 — amyloidosis, evidence of neurodegeneration plus subtle cognitive change: This stage includes evidence of subtle decline over time on standard cognitive tests, even though the individual may still perform within the normal range and not meet criteria for MCI.

A Clear, Bright Line

Steven T. DeKosky, MD, from the University of Virginia School of Medicine in Charlottesville, also a member of the MCI working group, acted as discussant during the session.

Dr. Steven DeKosky

He first addressed a question that the workgroups have heard on the need for another set of criteria, particularly as Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, is currently in development, as well as International Classification of Diseases, Eleventh Revision. The different criteria, however, are used for different purposes — some for research and some for clinical use.

He urged audience members, when asked about these new draft criteria, to put a "clear, bright line between those criteria which are intended for general use by clinicians, either specialist or primary care physicians, and those that are intended to be research criteria, which we clearly understand, know how to use, but are not available even to us if we were to go over to the clinic to see patients, as opposed to patients in research protocols."

"I think this is a huge issue with respect to how the world hears this message," he added. "Part of what we want to get out is that we're making progress on the disease, moving earlier and earlier, trying to find ways to diagnose it better and to treat it, but that excitement and enthusiasm which we have ourselves spills over sometimes and leads to public misunderstanding and public expectation that we are not going to be able to meet for a while."

Initial Issues

Although the reception here for the draft diagnostic criteria was generally positive, there were some issues raised immediately — some of them by the workgroups themselves. Each group had prepared the diagnostic criteria in these areas basically in isolation, and some potential problems across the documents became immediately evident.

One of the things that needs to be harmonized is the way that biomarkers are incorporated, Dr. Albert noted. "You will have noticed that the AD dementia group didn't differentiate biomarkers in terms of the levels of certainty that they conferred, and the MCI group spent a lot of time on that, so I think we're going to have to figure out how to make that seem more seamless," she told Medscape Medical News.

There was also a question at the end of the session on mixed presentations, she added, which was discussed at length by the AD dementia group and the preclinical group, but was not discussed by the MCI group. "So we'll have to at least try to be sure that that's addressed."

One thing that the groups tried to avoid was the concept of "conversion" from 1 stage to another, from MCI to AD for example, preferring to emphasize that the process is a continuum rather than a threshold over which patients fall into the next category.

This may have implications for clinical trials, however, as conversion has historically been used as an endpoint to judge treatment effects.

David S. Knopman, MD, from the Mayo Clinic, who was part of the AD Dementia workgroup, said that this change has been a point of discussion with the US Food and Drug Administration and European regulators.

Dr. William Thies

"What I think many in the field would like to see happen is to be able to use continuous measures in patients with MCI, so instead of using dementia as an endpoint in the MCI trial, use change on the cognitive test," Dr. Knopman said. "The regulatory agencies have so far resisted that because the field hasn't given them a clear signal that it's valid.

"But I think what we're saying, and what this meeting is making abundantly clear, is those continuous measures in fact are valid, and in fact do reflect worsening eventually that leads to dementia," he said. "It turns out that using the continuous measures almost certainly is a more efficient strategy for conducting the trials."

The 3 draft diagnostic criteria documents will be posted for comment, and after incorporation of suggested changes, the final document will be prepared for publication in a peer-reviewed journal probably this fall, said William Thies, PhD, chief medical and scientific officer at the Alzheimer's Association in Chicago. "We expect these documents to become relatively living documents, where they are adjusted as new data comes forward," Dr. Thies added.

The working group chairs have disclosed no relevant financial relationships. Full disclosures for workgroup members can be found on the diagnostic criteria Web site.

Alzheimer's Association International Conference on Alzheimer's Disease 2010: Focused Topic Sessions: Redefining Alzheimer's Disease. Presented July 13, 2010.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.