Article Text

Download PDFPDF

New diagnostic criteria for Alzheimer's disease and mild cognitive impairment for the practical neurologist
  1. Andrew E Budson1,2,
  2. Paul R Solomon2,3
  1. 1Center for Translational Cognitive Neuroscience, VA Boston Healthcare System, Boston, Massachusetts, USA, and Boston University Alzheimer's Disease Center, Boston University School of Medicine, Boston, Massachusetts, USA
  2. 2The Memory Clinic, Bennington, Vermont, USA
  3. 3Department of Psychology, Program in Neuroscience, Williams College, Williamstown Massachusetts, USA
  1. Correspondence to Andrew E Budson, Center for Translational Cognitive Neuroscience, VA Boston Healthcare System, 1400 VFW Parkway, West Roxbury MA 02132, USA; abudson{at}


Four articles in the journal Alzheimer's and Dementia in 2011 describe new criteria for Alzheimer's disease (AD) dementia and mild cognitive impairment (MCI) due to the AD pathophysiological process (MCI due to AD), as well as the underlying rationale for them. The new criteria also include preclinical AD criteria but these are intended purely for research purposes. The new criteria emphasise that the AD pathophysiological process starts years and perhaps decades before clinical symptoms, and that biomarkers can detect amyloid β deposition and the effects of neurodegeneration in the brain. The criteria are recommendations based upon consensus meetings and will require future validation. Nonetheless, the authors believe that they are immediately helpful to the practising clinician, providing more accurate and specific guidelines for the diagnosis of AD dementia and MCI due to AD. As new diagnostic tools and treatments for AD become available, diagnoses using these criteria will enable patients with AD dementia, MCI due to AD and eventually preclinical AD to receive the best possible care.

Statistics from

Rationale for new criteria

The National Institute on Aging and the Alzheimer's Association work group on diagnostic guidelines for Alzheimer's disease have updated the 1984 Alzheimer's disease criteria in four articles in the journal Alzheimer's and Dementia in 2011. In these, they describe new clinical criteria for Alzheimer's disease (AD) and mild cognitive impairment (MCI) due to the AD pathophysiological process (MCI due to AD).1 Although these groups were not the first to update the 1984 criteria (see papers by Dubois and colleagues),2 3 they currently have the greatest consensus in the scientific AD community and are likely to have a lasting impact.

The new criteria were developed due to several factors that have changed since 1984:

  • (1) The AD pathophysiological process likely starts years before cognitive changes and decades before onset of clinical dementia4 5 (figure 1). The concept of the ‘AD pathophysiological process’ is thus separated from ‘AD dementia.’

  • (2) Many patients whose cognition is not normal for age do not meet criteria for dementia.

  • (3) Other causes of dementia are more likely mistaken for AD than are thyroid disorders and B12 deficiency.

  • (4) Genetics of AD are better understood.

  • (5) Biomarkers of AD are available in some centres.

  • (6) New criteria are needed for research.

  • (7) Specific treatments for the AD pathophysiological process are being developed; when these treatments are available it will be critical to know if patients have that process.

Figure 1

Postulated temporal lag between the deposition of amyloid β in amyloid plaques from an autopsy series and the development of clinical Alzheimer's disease (AD) dementia based upon three epidemiological studies (from Sperling et al).5

There are three stages of AD:

  • Preclinical AD requires measureable changes in biomarkers and/or poor performance on challenging cognitive tests.

  • MCI due to AD manifests the first clinical changes. Patients and families notice mild changes in memory and other cognitive abilities; these changes can be detected through careful evaluation, but do not interfere with day-to-day activities.

  • Dementia due to AD is characterized by changes in two or more aspects of cognition and behaviour that interfere with function in everyday life.

One model of AD is that many factors combine to cause the accumulation of amyloid β (Aβ) in the brain, which in turn produces synaptic dysfunction, tangle formation, and neuronal death, ultimately leading to cognitive decline (Sperling et al,)5 (figure 2). Because events in this sequence likely take years or decades, there must be a prior ‘preclinical’ stage of AD (figure 3).

Figure 2

Hypothetical model of Alzheimer's disease (AD) pathophysiological cascade sequence (from Sperling et al).5

Figure 3

Model of the clinical course of Alzheimer's disease (from Sperling et al).5 MCI, mild cognitive impairment.

Table 1 presents several currently used biomarkers of Aβ deposition or neurodegeneration. Markers of Aβ deposition include low cerebrospinal fluid (CSF) Aβ42 and positive positron emission tomography (PET) amyloid imaging. Markers of neurodegeneration include elevated CSF tau (both total and hyperphosphorylated tau), decreased metabolism in temporal and parietal cortex on 18flurodeoxyglucose (FDG) PET, and atrophy on MRI in temporal (medial, basal, and lateral) and medial parietal cortex. In clinical practice, one tends to divide the biomarkers by how they are obtained: structural MRI, PET, and CSF studies.

Table 1

Putative biomarkers for the AD pathophysiological process currently being used

Although volumetric MRI analyses are not routinely available, we encourage all clinicians to look for qualitative patterns of atrophy in temporal (medial, basal, and lateral) and medial parietal cortex.6

FDG PET scans show decreased metabolism in temporal and parietal cortex when the AD pathophysiological process has caused neurodegeneration.6 FDG PET scans are available to the clinician now (and are covered by Medicare in the USA). We do not, however, recommend using these scans routinely when the history, physical examination, cognitive testing, and structural imaging are all consistent with AD: it is simply not necessary.6 However, when one suspects an atypical neurodegenerative disease or the patient is younger than 66 years of age (when the prevalence of AD is similar to that of many other aetiologies), an FDG PET scan can help to distinguish AD from another disorder (such as dementia with Lewy bodies or frontotemporal dementia).

Several compounds that can identify Aβ deposition using PET are currently being used for research and/or are under commercial development, including Pittsburgh compound B and florbetapir. Because amyloid accumulation may occur years before clinical symptoms, PET identification of Aβ raises the possibility of identifying and treating patients with AD years before symptoms emerge—once a drug has been proven to be disease-modifying.

Standard CSF biomarkers for AD are Aβ42, total tau, and hyperphosphorylated tau. When all three markers are combined, the accuracy of the diagnosis is the highest, with sensitivity and specificity of 85–90%. Although CSF analysis is already commercially available ( and some clinicians use it to aid diagnosis, we view this test as promising but not yet ready for routine clinical practice.6

Guided by these theoretical underpinnings, the new criteria are presented for all cause dementia, AD dementia, MCI due to AD, and—for research only—preclinical AD. (See figure 4 for an overall flow chart for the evaluation of a patient with cognitive impairment.)

Figure 4

Flow chart for the evaluation of a patient with cognitive impairment leading to the diagnosis of mild cognitive impairment (MCI) and dementia using the new guidelines, with references to the tables (T) and steps (S) in the text.

Criteria for all cause dementia and AD

The new criteria propose four possible classifications of dementia caused by AD: (1) probable AD dementia, (2) probable AD dementia with increased level of certainty, (3) possible AD dementia and (4) probable or possible AD dementia with evidence of AD pathophysiological process.7

The new criteria suggest a four-step approach to diagnosing dementia due to AD (table 2). Step 1 determines that dementia is present, step 2 determines that the dementia is due to AD, step 3 provides an increased level of certainty to the diagnosis and step 4 evaluates the biomarker probability of AD aetiology. The authors do not advocate obtaining biomarkers for routine clinical purposes at present, although they note that they may be used when available and deemed appropriate by the clinician.

Table 2

Clinical and cognitive evaluation for all cause dementia and AD

Biomarkers enable the diagnosis of ‘probable AD dementia with evidence of the AD pathophysiological process’ (table 1). If one of the two biomarker categories is positive, the ‘biomarker probability of AD aetiology’ rises to ‘intermediate,’ and if both categories are positive the probability becomes ‘high’.

‘Possible AD’ is used instead of ‘probable AD’ if the cognitive deficits look like AD but there is an atypical course (either sudden onset or no definite decline) or evidence of a mixed aetiology. Thus, the patient might meet the criteria for probable AD dementia but there is also evidence of significant vascular disease, features of dementia with Lewy bodies or other disease, or condition that could be contributing to the patient's dementia (table 3).

Table 3

Clinical and cognitive evaluation for possible AD

The next category is pathophysiologically proven AD, consisting simply of the unchanged criteria of patients meeting both the clinical and neuropathological criteria for AD. Finally, the authors discuss the criteria for dementia unlikely to be due to AD (table 4).

Table 4

Criteria for dementia unlikely to be due to AD

Criteria for MCI due to AD

MCI due to AD refers to the symptomatic phase of the AD pathophysiological process before the individual develops the functional impairment that defines dementia. The guidelines presented by Albert et al recognise that everyone who eventually develops AD goes through a transitional period of mild but detectable cognitive impairment.8 However, not everyone who is diagnosed with MCI goes on to develop AD. MCI can be due to several disorders, including vascular dementia, frontotemporal dementia, dementia with Lewy bodies, and others. Furthermore, distinguishing between normal cognition and MCI, and between MCI and dementia needs clinical judgment.

The criteria for MCI due to AD include:

  • excluding patients with other causes of MCI, including extensive vascular disease, frontotemporal dementia, and dementia with Lewy bodies

  • including patients with increasing cognitive decline over time

  • including patients with mutations associated with early-onset familial AD (amyloid precursor protein, presenilin 1 or presenilin 2).

First, criteria are presented for the clinical and cognitive syndrome of MCI, and second, criteria are presented regarding the aetiology of the MCI syndrome being consistent with AD (table 5).

Table 5

Clinical and cognitive evaluation for MCI due to AD

Criteria for MCI due to AD incorporating biomarkers (table 1) are next presented. Biomarkers of Aβ protein deposition may help determine aetiology, and markers of neurodegeneration may aid prognosis. If one of these two biomarker categories is positive, the ‘biomarker probability of AD aetiology’ rises to ‘intermediate’; both categories must be positive for the ‘highest’ probability. The ‘lowest’ probability is present if both categories are negative.

Preclinical AD (for research only)

Sperling et al5 discussed three stages of preclinical AD based upon biomarkers and cognitive change (figure 5). Stage 1 is asymptomatic cerebral amyloidosis, determined by the presence of a biomarker of Aβ, without a marker of neurodegeneration or evidence of subtle cognitive change. Stage 2 is asymptomatic cerebral amyloidosis plus neurodegeneration without subtle cognitive change. Stage 3 is amyloidosis plus neurodegeneration plus subtle cognitive or behavioural decline.

Figure 5

Model of how the different stages of Alzheimer's disease may be detected by changes in various biological, cognitive and clinical markers (from Sperling et al).5 CSF, cerebrospinal fluid; FDG, 18flurodeoxyglucose; fMRI, functional Magnetic Resonance Imaging; MCI, mild cognitive impairment; PET, positron emission tomography.

Case examples

Example 1

A 78-year-old man and his wife were both concerned that his memory was not as good as it was last year. He used to be able to remember a short grocery list in his head but now he needed to write it down or he would return with the wrong items. He continued to pay the bills, balance the books and do household projects although these tasks now took him longer to complete. Evaluation of memory testing in the office showed that he scored below normal when repeating a brief story containing 10 details. His laboratory data were unremarkable, and his MRI showed bilateral atrophy in hippocampus (basal temporal lobe), lateral temporal lobe, and parietal lobe.

Because this patient had concerns about a change in cognition, objective impairment in cognition and preservation of functional abilities, he meets the criteria for MCI (table 5 step 1). Because the history and cognitive testing are of memory impairment—the main phenotype of AD—and he has atrophy of the brain in a pattern consistent with the AD pathophysiological process, he meets the criteria for MCI due to AD, with an intermediate biomarker probability of AD aetiology (table 5 step 2).

Example 2

An 84-year-old woman was brought in by her daughter. Although the patient did not believe that anything was wrong, her daughter noted that her mother's thinking and memory had deteriorated slowly over 5 years. Her phone was disconnected because she forgot to pay the bills, and she kept buying the same canned food items whenever she went to the market. There was spoiled food in the refrigerator. When interviewed she had pauses in her speech and her daughter often filled in missing words for her. On a brief office test she was not orientated to the day, date, month, or year; she could not recall any of the few items she was instructed to remember; and she was unable to name even common items. A CT scan of head showed mild small vessel ischaemic disease (average for individuals in their 80s) and atrophy of hippocampi, lateral temporal lobes, and parietal lobes.

Suggestions for further reading

  • Budson & Solomon Memory Loss: A Practical Guide for Clinicians, Philadelphia: Elsevier Inc., 2011.

  • Sperling RA, Aisen PS, Beckett LA, et al. Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011; 7:280-92.

This patient had clear evidence of a decline in prior ability that interfered with function: she has had her phone disconnected, she bought the same food repetitively, and she had spoiled food in the refrigerator. Cognitive impairments in at least two domains (memory and language) are present. Thus, she meets criteria for ‘all cause dementia’ (table 2 step 1). That her memory deteriorated slowly over 5 years suggests a gradual onset; clear cut worsening of memory suggests an amnestic presentation, and the CT scan of head rules out a vascular dementia. There is nothing in the story to suggest another type of dementia or condition that could have a substantial effect on cognition. Thus, this patient's dementia meets criteria for the amnestic presentation of probable AD (table 2 step 2), and the atrophy on the CT suggests an intermediate biomarker probability of AD aetiology (table 2 step 4). There is nothing in the story either to provide an ‘increased level of certainty’ (table 2 step 3) or to suggest ‘possible’ instead of ‘probable’ AD (table 3).

Implications for the practicing clinician

These new guidelines are recommendations based upon consensus meetings and will require validation in the future. They encourage clinicians to:

  • Recognise that AD is the end of a long process, spanning years or perhaps decades.

  • Diagnose (and perhaps treat) AD at the earliest possible stage, at present MCI due to AD, but eventually (with new disease-modifying medications) preclinical AD. (Note that the criteria for preclinical AD are currently intended purely for research purposes.)

  • Consider using biomarkers in the diagnosis of all stages of AD. Our current recommendation is to use biomarkers for those cases that present diagnostic quandaries.6

  • Evaluate patients with cognitive impairment and dementia to determine aetiology, with special attention to amnestic and non-amnestic presentations of AD.

  • Remember that because of the ageing population, numbers of patients with all stages of AD will likely triple in the next 50 years.


This work was supported by National Institute on Ageing grant P30 AG13846. This material is also the result of work supported with resources and the use of facilities at the VA Boston Healthcare System.


View Abstract


  • Funding Funding was provided by the National Institute on Ageing.

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed. We are grateful to Jonathan Schott, London, UK, for reviewing this paper.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Other content recommended for you