Revised criteria for the diagnosis and staging of Alzheimer’s disease

Abstract

Alzheimer’s disease can be treated by targeting amyloid-β plaques and diagnosed in vivo by biomarkers, prompting the revision of criteria for the diagnosis and staging of this disease.

A workgroup convened by the Alzheimer’s Association recently published Revised Criteria for Diagnosis and Staging of Alzheimer’s Disease ¹. Members of this committee (who also co-authored this comment) were selected to provide balanced representation of academia, clinical medicine, industry, the National Institute on Aging, the Food and Drug Administration, and included participants from the USA and Europe. The revised criteria ¹ are an update to prior criteria published in 2011 and 2018 ^{2, 3}. Formulation of these revised criteria ¹ was prompted by two major developments in this field that occurred after the 2018 research framework ³ was published: regulatory approval of the first disease targeted therapeutics for Alzheimer’s disease (AD) ^{4, 5}; and development of accurate blood-based biomarkers of AD ⁶.

The 2024 Revised Criteria for Diagnosis and Staging of AD ¹ are founded on several core principles. AD is defined as a biologic process that is first detectable by abnormal biomarkers of its defining neuropathologic features (β-amyloid plaques and tau neurofibrillary tangles) when an individual is asymptomatic. The disease progresses biologically during the preclinical period and when a sufficient pathologic burden is reached, symptoms appear and then progress. The entire disease course may span up to 30 years.

Categorization of biomarkers

In the revised criteria ¹, biomarkers are categorized into those that are core biomarkers of AD neuropathologic change ⁷ (specific to AD), non-core biomarkers of processes (not specific to AD but still important in AD pathogenesis, such as biomarkers of inflammation or immune activation, and neurodegeneration), and biomarkers of common non-AD copathologies (such as cerebro vascular disease and neuronal α-synuclein disease). This commentary focuses on core AD biomarkers that fall into the following categories: biomarkers of the amyloid-β (Aβ) proteinopathy pathway (labeled ‘A’); biofluid biomarkers of phosphorylated and secreted AD tau, which become abnormal early in the disease process (labeled ‘T₁’); and biomarkers of AD tau neurofibrillary pathology that become abnormal later in the disease process (labeled ‘T₂’). We group biomarkers into Core 1 (A and T₁) and Core 2 (T₂) (Table 1).

Table 1.

Categorization of Core AD imaging biomarkers and biofluid assays

Category	Modalities
Core 1	Amyloid PET and biofluid measures of Aβ proteinopathy (A)
	Biofluid measures of phosphorylated and secreted AD tau (T1)
Core 2	Tau PET and biofluid measures of AD tau proteinopathy (T2)

Core 1 biomarkers define the initial stage of AD that is detectable in vivo. An abnormal Core 1 biomarker is sufficient to establish a diagnosis of AD and to inform clinical decision making (Table 2). Core 2 biomarkers are not typically used as standalone tests for the diagnosis of AD, but when combined with Core 1 may be used to stage biological disease severity and provide prognostic information (Table 2).

Table 2.

Intended uses of Core AD imaging biomarkers and biofluid assays

Category	Intended use	Details
Core 1	Diagnosis	Early detection of AD in asymptomatic individuals
		Confirmation that AD is an underlying pathology in symptomatic individuals
Core 2	Staging and prognosis	Provide information on the likelihood that symptoms are associated with AD
		Inform on the risk of and rate of progression

Diagnosis

An abnormal Core 1 biomarker test is diagnostic of AD. The following are currently considered fit for this purpose: amyloid positron emission tomography (PET); cerebrospinal fluid (CSF) Aβ42/40, CSF p-tau181/Aβ42, CSF t-tau/Aβ42; and accurate plasma assays (defined below). However, the revised criteria outline important caveats around diagnostics, especially using plasma biomarkers. Chief among these is a minimum diagnostic accuracy requirement (90% or greater) with respect to an accepted reference standard in the intended context of use (which includes a description of the intended use population). Amyloid PET ligands and CSF assays listed in Table 1 have received regulatory approval in the USA and EU; the former were validated against autopsy ⁸ and the latter validated against amyloid PET ⁹. While presently no plasma assays have received regulatory approval, this is expected to change soon as some plasma assays (such as p-tau217) have demonstrated diagnostic accuracy equivalent to approved CSF assays ¹⁰. The diagnostic accuracy of other plasma assays, however, do not meet the 90% standard ¹⁰ and would not be considered fit for the purpose of diagnosis.

A second important caveat centers on the role of the clinician. The biologically based diagnosis of AD is meant to assist, rather than supplant, the clinical evaluation of individuals with cognitive impairment. Biomarker testing should only be performed under the supervision of a clinician.

Staging

The revised criteria ¹ describe separate schemes for staging biological AD severity and for staging the severity of clinical symptoms. Biological staging is based on ordering of biomarker events in the natural history of the disease from observational research. A 4-point staging scale (A-D) is outlined that can be accomplished by amyloid and tau PET, by Core 1 biofluids and tau PET, or a forward-looking conceptual staging approach based only on biofluids ^{11, 12}. Staging the severity of clinical symptoms is based on a 6-point numeric scale ranging from asymptomatic (stage 1) to severe dementia (stage 6). Integration of biological and clinical staging is described using an alphanumeric labeling scheme (for example, stage 1A, 4D, and so on).

While clinical stages generally worsen with increasing biological stages, the revised criteria recognize that mismatches between clinical and biological stage commonly occur. This is because AD is only one of several common pathologies that underlie cognitive decline and dementia in older individuals; the other 3 most common are cerebro vascular disease, neuronal α-synuclein disease, and limbic-predominant age-related TDP-43 encephalopathy (LATE) ¹³. Individuals with one or more of these pathologies are likely to have worse clinical stage than expected for their biological AD stage. Conversely, individuals with exceptional resilience or cognitive reserve may have better clinical stage than expected for their biological AD stage.

Use of the name Alzheimer’s disease

Two successive drafts of these criteria were presented at scientific meetings in 2023 and were also posted for public comment. It is our view that some of the comments or critiques offered contradict fundamental positions of the committee and so below we describe these critiques and the rational underlying the committee’s decision to formulate the criteria as done.

The term AD is understood by many in the public and in general medicine to be synonymous with dementia. Many think of AD as a progressive impairment leading to loss of functional independence in older persons, irrespective of etiology. A common critique of the revised criteria ¹ is that to avoid confusion on the part of the public, the term AD should be used to describe all-cause dementia and the distinction between etiology (pathobiology) and symptoms (the result of pathobiology) should be ignored.

The position of the committee is that not only is this scientifically inaccurate, but it is also harmful. The etiology or etiologies underlying clinical symptoms must be understood and accurately diagnosed to enable effective treatments. Although AD is the most common pathologic condition underlying progressive cognitive impairment in elderly persons, non-AD pathologies are also common ¹³. The typical syndrome association with AD (a progressive memory impairment that progresses to other cognitive domains) can also be caused by common non-AD pathologies, most often in combination. For example, in early Aβ immunotherapy trials where eligibility was based on a traditional clinical diagnosis of AD without biological confirmation, 30% of individuals had normal amyloid PET scans ¹⁴, meaning that nearly one third of participants in these trials did not have the disease they were being treated for. The position of the committee is that therapy directed toward a biological target requires confirmation of that biology both in clinical trials and in clinical practice. Thus, AD must be defined by its biology for medical progress to be made; the public is readily able to grasp this concept with the proper educational effort.

Requiring symptoms for the diagnosis of AD

A second critique argues that a diagnosis of AD should require the presence of both abnormal AD biomarkers and the presence of clinical symptoms. The argument is that preclinical AD is not a valid concept and asymptomatic individuals with abnormal biomarkers should be labeled ‘at risk’ for AD. The reasoning is that many asymptomatic individuals with abnormal biomarkers (preclinical AD) will not develop symptoms in their lifetime. The argument continues, labeling such individuals with a disease may cause psychological, financial, or social harm without potential for benefit, since at present, disease targeted treatments are not approved for preclinical AD.

The committee takes a different view ¹. First, diseases are routinely diagnosed in other areas of medicine while patients are asymptomatic. It is axiomatic throughout medicine that treatments are more effective when administered as early in the disease process as possible ². Therefore, conducting clinical trials in persons with preclinical AD is considered by the committee to be a forward-looking approach to identify the most effective means of slowing or preventing the onset of symptoms.

Second, while it is true that many asymptotic individuals with abnormal AD biomarkers will not experience symptoms in their lifetime, this is because of increasing all-cause mortality rates with advancing age, not because AD pathology is benign in the preclinical period. AD is characterized by a preclinical period of 15–20 years or more, which most often begins from age 60 ¹. Over half of individuals newly diagnosed in their mid-70s may not experience symptoms in their lifetime, but nearly half will. The position of the committee is that asymptomatic individuals who may experience symptoms deserve treatment once these are approved for the preclinical population. Mortality from unrelated diseases should not be interpreted to indicate that AD in asymptomatic persons is a benign condition. Individuals with Down syndrome have genetically determined AD (trisomy 21) and 95% of these individuals develop dementia in their lifetime, ¹⁵ because the average age of onset of clinical symptoms is mid-50s, at which age all-cause mortality rates are far lower than in older age. Remaining life expectancy is an important consideration in clinical management, but mortality from unrelated causes should not be a criterion used to define what is and what is not a disease ¹.

Third, the revised criteria have been criticized for advocating screening of the general population with AD biomarkers, which could place asymptomatic persons who test positive in a position of potential harm without benefit. However, the criteria are very clear that since no treatment has yet been approved for asymptomatic persons, biomarker testing should not be performed in this population outside the context of observational or therapeutic research studies ¹. This may change in the future, however, pending results of ongoing trials.

Validation of plasma biomarkers

Criticism was raised that plasma biomarkers are too new to be used clinically and have not been adequately tested in all representative populations. While the committee recognizes that plasma biomarkers are a recent development, we also outline rigorous criteria that plasma biomarkers need to meet to be considered fit for the purpose of diagnosis ¹. The committee strongly endorses the need for testing and prospective evaluation of plasma biomarkers in more representative populations.

Neuropathological underpinnings of diagnosis by biomarkers

Some critics claim that defining AD by an abnormal Core 1 biomarker is a departure from the accepted neuropathological definition which requires the presence of both β-amyloid plaques and tau tangles ⁷. Regulatory approval of amyloid PET ligands was based on their ability to detect moderate to frequent neuritic β-amyloid plaques at autopsy and not both plaques and tangles ⁸, whilst approval of CSF assays was based on their ability to differentiate normal from abnormal amyloid PET scans ⁹. Thus, the argument of these critics is that defining AD by the presence of abnormal Core 1 biomarkers does not fully capture the neuropathologic standard that requires tau tangles in addition to β-amyloid plaques. However, amyloid PET scans (and thus biofluid surrogates) are not able to reliably detect mild levels of AD pathology; it can detect moderate to frequent neuritic plaques but not sparse plaques ⁸. Given that nearly all (over 90%) of individuals with moderate to frequent neuritic plaques at autopsy (and therefore abnormal Core 1 biomarkers) will also have sufficient neurofibrillary tangle pathology to meet criteria for a pathological diagnosis of intermediate to high AD neuropathologic change,⁷ diagnosing AD based on abnormal Core 1 biomarkers will nearly always be consistent with the accepted neuropathologic reference standard for AD.

Future

Many in the AD field are optimistic about the prospects for improving care of patients. Recent advances in the ability to diagnose and treat AD herald a virtuous cycle wherein improvements in diagnostic methods enable more sophisticated treatment approaches which in turn steer advances in diagnostic methods. An unchanging principle, however, is that effective treatment will always rely on the ability to diagnose and stage the biology driving the disease process. The committee recognizes that the feasibility of implementing criteria for biologically-based diagnosis and staging of AD in clinical practise varies across regions, even within high-income countries. We anticipate that the increasing availability and accuracy of blood-based biomarkers will make these criteria more widely deployable.