JAMA Neurology | Special Communication Clinical Review & Education

Alzheimer Disease as a Clinical-Biological Construct — An International Working Group Recommendation

Bruno Dubois, MD, MSc; Nicolas Villain, MD, PhD; Lon Schneider, MD, MSc; Nick Fox, MD, MA; Noll Campbell, PharmD, MSc; Douglas Galasko, MD, MSc; Miia Kivipelto, MD, PhD; Frank Jessen, MD; Bernard Hanseeuw, MD, PhD; Mercè Boada, MD, PhD; Frederik Barkhof, MD, PhD; Agneta Nordberg, MD, PhD; Lutz Frolich, MD, PhD; Gunhild Waldemar, MD, DMSc; Kristian Steen Frederiksen, MD, PhD; Alessandro Padovani, MD, PhD; Vincent Planche, MD, PhD; Christopher Rowe, MD; Alexandre Bejanin, PhD; Agustin Ibanez, PhD; Stefano Cappa, MD; Paulo Caramelli, MD, PhD; Ricardo Nitrini, MD, PhD; Ricardo Allegri, MD, PhD; Andrea Slachevsky, MD, PhD; Leonardo Cruz de Souza, MD, PhD; Andrea Bozoki, MD; Eric Widera, MD; Kaj Blennow, MD, PhD; Craig Ritchie, MD, PhD; Marc Agronin, MD; Francisco Lopera, MD; Lisa Delano-Wood, PhD; Stéphanie Bombois, MD, PhD; Richard Levy, MD, PhD; Madhav Thambisetty, MD, DPhil; Jean Georges, BA; David T. Jones, MD; Helen Lavretsky, MD, MSc; Jonathan Schott, MD, BSc; Jennifer Gatchel, MD, PhD; Sandra Swantek, MD; Paul Newhouse, MD; Howard H. Feldman, MD; Giovanni B. Frisoni, MD

JAMA Neurol. doi:10.1001/jamaneurol.2024.3770 Published online November 1, 2024 · Print: JAMA Neurol. 2024 Dec 01; 81(12):1304–1311 View on JAMA Neurology →

🎯 Context

Since 2018, a movement has emerged to define Alzheimer disease (AD) as a purely biological entity based on biomarker findings. This IWG paper offers a critical response.

Two Competing Views

The Alzheimer Association (AA) 2024 criteria define AD as a purely biological entity — a diagnosis can be given to cognitively normal people based on biomarker findings alone (e.g. CSF Aβ/Tau ratios, plasma phosphoTau217).


The International Working Group (IWG) argues that AD must remain a clinical-biological construct — requiring both measurable cognitive symptoms and positive biomarkers before a diagnosis of AD is made.

The IWG position: Diagnosing AD in cognitively normal individuals based solely on biomarker positivity is clinically unwarranted and potentially harmful — most of these individuals will never develop symptoms.

🔬 The Value of Biomarkers

Biomarkers have transformed Alzheimer research — but their role in the clinic requires careful contextualization.

Validated Tools

Amyloid PET, CSF & plasma biomarkers have been validated against autopsy-derived brain pathology and justify inclusion in diagnostic workups.

Research Revolution

Biomarkers enable in-vivo monitoring of pathological changes, pharmacodynamic evaluation, and surrogate endpoints in clinical trials.

Not Sufficient Alone

"Core 1 biomarkers" are insufficient to account for the clinical expression of the disease. They should be conceptualized as risk markers or markers of pathology, not disease markers.

In Cognitively Impaired Patients

The IWG fully supports the use of biomarkers in patients with cognitive impairment. The combination of a recognized clinical phenotype (amnestic syndrome, logopenic aphasia, posterior cortical atrophy…) plus positive AD biomarkers establishes the clinical-biological diagnosis of AD — from prodromal to dementia stages. This framework was central to regulatory approval of anti-amyloid treatments.

⚠️ The Core Problem

The key concern: applying a biological-only definition of AD to cognitively normal individuals creates serious clinical and societal adverse consequences.

21.9%
Lifetime AD risk for amyloid-positive 65-year-old men — only 1.7× higher than amyloid-negative
17%
Of cognitively normal amyloid-positive individuals progress to MCI over 6 years
Key finding: The majority of biomarker-positive cognitively normal individuals will not become symptomatic along a proximate timeline. The cancer analogy — which implies near-certain progression — does not apply to sporadic AD pathology. More pertinent is the vascular analogy whereby cholesterol and blood pressure increase the risk of adverse outcomes but by no means allow them to be deterministically predicted.

🗂️ The 2024 IWG Framework

The IWG proposes a three-tier lexicon that separates cognitively normal biomarker-positive individuals into distinct, clinically meaningful categories.

At-Risk

Asymptomatic At-Risk

Cognitively normal individuals at increased risk due to a specific biomarker profile: brain amyloidosis (isolated or associated with tauopathy limited to medial temporal regions), or a positive phospho-tau fluid biomarker. Progression is uncertain — they should not be labeled as having AD. Goal: risk communication and risk reduction (pharma and lifestyle).

Presymptomatic AD

Presymptomatic AD

Specific biomarker profiles indicating very high lifetime risk of progression — reserved for: autosomal dominant mutations, Down syndrome, APOE ε4 homozygous + SORL1 loss, or combined amyloid + extensive neocortical tau PET positivity.

Alzheimer Disease

Alzheimer Disease

A recognized clinical phenotype (amnestic syndrome of the hippocampal type, logopenic aphasia, posterior cortical atrophy, or less common variants) plus positive CSF or PET AD biomarkers (amyloid and/or tau). Includes both prodromal (MCI, no loss of function) and dementia (with loss of function) stages. Eligible for diagnosis and treatment according to current guidelines.

The Probabilistic Amyloid Cascade Model

The pathophysiological disease model underlying the IWG criteria posits that the penetrance of Alzheimer's pathology (amyloid and tau) decreases along the spectrum: autosomal dominant (~100%) → APOE ε4 carriers (intermediate) → APOE ε4 non-carriers (lowest). Decreasing penetrance reflects increasing contributions of stochastic factors: non-APOE genes, environment, co-pathologies, and resilience. Frisoni et al., Nat Rev Neurosci 2022 →

⚖️ AA 2024 vs. IWG 2024

Two competing frameworks rest on fundamentally different conceptions of what Alzheimer disease is. The IWG defines AD as a clinical-biological entity; the AA defines it as a biological continuum. This conceptual divide drives every downstream disagreement on diagnosis, communication, and clinical practice.

The conceptual foundation

Where each framework starts from — and what it means in practice for asymptomatic individuals.

International Working Group

Clinical-Biological Entity Dubois et al., Lancet Neurology, 2024

Concept of Alzheimer's disease

Clinical-biological: neuropathological lesions explain the cognitive phenotype.

Definition of Alzheimer's disease

Alzheimer's disease neuropathological lesions are the primary cause of the cognitive phenotype.

Cognitively unimpaired individuals — Risk stratification
  • Presymptomatic AD — risk of developing the cognitive phenotype ≈ 100% (ADAD, widespread neocortical tau)
  • Asymptomatic – at risk for AD — risk significantly lower than 100% (positive PET or CSF biomarkers)
Alzheimer Association

Biological Continuum Jack et al., Alzheimer's & Dementia, 2024

Concept of Alzheimer's disease

Biological: pathophysiological continuum starting from the first neuropathological lesions.

Definition of Alzheimer's disease

Alzheimer's disease neuropathological lesions are Alzheimer's disease.

Cognitively unimpaired individuals — One nosological category
  • Alzheimer's disease — biomarker positivity is sufficient to establish the diagnosis
  • All biomarker-positive individuals have preclinical AD — a stage of the same disease
The root of the disagreement

The dispute is not whether biomarkers are useful — both frameworks agree they are. The dispute is whether biomarker positivity alone should be sufficient to define disease in cognitively unimpaired individuals.

Practical consequences

How the conceptual divide translates into everyday clinical and research decisions.

Dimension AA 2024 IWG 2024
AD Definition Biological only Clinical-biological
Cognitively Normal + Biomarker(+) Can be diagnosed with AD Labeled "at-risk" — NOT AD*
Disclosure to Cognitively Normal, Biomarker-Positive Individuals "You have Alzheimer disease" "You are at-risk for Alzheimer disease"
Cognitively unimpaired individuals One nosological category — all biomarker positives have preclinical AD Risk stratification — presymptomatic AD vs at-risk
Trial Endpoints Biomarkers can be primary endpoints Clinical efficacy required

* Except rare cases meeting presymptomatic AD criteria (autosomal dominant mutations, Down syndrome, very high-risk biomarker profiles).

✅ Key Recommendations

Four core recommendations from the IWG for clinical practice and research.

1

AD = Clinical-Biological Entity

Alzheimer disease requires both documented cognitive impairment and positive AD biomarkers (amyloid and/or tau). A biological finding alone is not sufficient.

2

No Routine Testing in Cognitively Normal Individuals

Biomarkers serve as risk markers — not diagnostic markers — in people without cognitive symptoms. Routine testing in this population is not recommended.

3

At-Risk ≠ Disease

Biomarker-positive cognitively normal individuals should be informed of their risk status — not diagnosed with AD. This should take place in the context of a comprehensive assessment of all known risk factors for cognitive impairment and dementia, including genetic and lifestyle. The communication of risk should likewise be comprehensive.

4

Define Presymptomatic AD — Carefully

Reserve the "presymptomatic AD" label for those with specific biomarker profiles indicating near-certain imminent progression (autosomal dominant mutations, Down syndrome, very high genetic risk or combined amyloid + tau PET positivity). This chapter of clinical science is still being written.

🔭 From Diagnosis to Prevention

The IWG envisions a shift toward structured risk evaluation and individual- and population-level prevention.

Evaluate Risk

Personalized biomarker and lifestyle profiling through dedicated Brain Health Services (dBHS).

Communicate Risk

Amyloid status ≠ AD diagnosis. Structured, nuanced communication of individual risk — not disease labeling.

Reduce Risk

Modifiable risk factors, lifestyle changes, and future preventive treatments — analogous to managing vascular risk to prevent stroke.

Observational Studies

Long follow-up combining lifestyle factors and biomarkers, with emphasis on diverse and underrepresented populations.

Interventional Trials

Test pharmacological and risk-reduction strategies in cognitively normal at-risk individuals.

Equity

Include non-white, ethnic minority, and low/middle-income country (LMIC) populations in future research.

👥 Authors & Affiliations

This recommendation was produced by an international consortium of 46 neurologists, psychiatrists, and researchers across 4 continents.

Lead authors: Bruno Dubois (Paris), Nicolas Villain (Paris) — Senior authors: Howard H. Feldman (San Diego), Giovanni B. Frisoni (Geneva)

AP-HP / Sorbonne Université (Paris, France) · UCL Queen Square Institute of Neurology (London, UK) · Keck School of Medicine USC (Los Angeles, USA) · Karolinska Institutet (Stockholm, Sweden) · University of Cologne / DZNE (Germany) · UCLouvain (Brussels, Belgium) · Ace Alzheimer Center (Barcelona, Spain) · University College London (UK) · University of Gothenburg (Sweden) · University of St Andrews (Scotland, UK) · Mayo Clinic (Rochester, USA) · University of California San Diego · Harvard Medical School · National Institute on Aging (NIH) · Alzheimer Europe (Luxembourg) · University of Geneva / HUG (Switzerland) · and many more institutions across Europe, North America, Latin America, and Australia.