Skip to content

Clinical presentation and staging

TL;DR — The typical presentation is an insidious amnestic syndrome, but a substantial minority — over-represented among young-onset cases — present with visual, language, executive or behavioural syndromes that carry the same underlying pathology (Graff-Radford 2021, PMID 33609479). Progression is slower and more variable than clinic samples suggest: in a population-based incident cohort the mean annual change was −1.53 MMSE points (SD 2.69) and +1.44 CDR-sum-of-boxes points (SD 1.82), and 30–58% of survivors progressed by less than one point per year even 5–7 years after onset (Tschanz 2011, PMID 21606896). Multistate modelling puts total disease duration at about 24 years from age 60 and 15 years from age 80, of which roughly 10 years is preclinical, 4 prodromal and 6 dementia for someone entering preclinical AD at 70 (Vermunt 2019, PMID 31164314). MCI is not a guaranteed pathway: pooled annual conversion to dementia was 9.6% in specialist settings and 4.9% in population studies, and most people with MCI do not progress even after 10 years (Mitchell 2009, PMID 19236314). The staging instruments in use — CDR, MMSE, MoCA, GDS/FAST — measure different constructs at different resolutions, and the anti-amyloid era has made the size of a "meaningful" change on them a contested quantity in its own right (Muir 2024, PMID 38561021).

The continuum and how it is divided

The 2011 NIA-AA workgroups split the disease into three clinical stages — preclinical, MCI due to AD, and AD dementia — each with its own criteria document (Sperling 2011, PMID 21514248; Albert 2011, PMID 21514249; McKhann 2011, PMID 21514250). The 2024 revision replaced this with an integrated biological-and-clinical staging scheme, explicitly built to tolerate the fact that copathology, cognitive reserve and resistance decouple biological from clinical stage (Jack 2024, PMID 38934362). What follows describes the clinical axis; the biological axis and the dispute over it are on diagnostic criteria and the biological definition.

Clinical stage Defining feature Typical duration Key quantities
Preclinical Biomarker abnormality, normal cognition ~10 y from age 70 5-year progression 11% (stage 1) to 56% (stage 3); A+T+ 57% (PMID 24012374; PMID 40522652)
MCI / prodromal Objective cognitive impairment, preserved independence ~4 y Annual conversion 9.6% (clinic) / 4.9% (population) (PMID 19236314)
Mild dementia Impairment interferes with instrumental activities — MCID +2 CDR-SB, −2 MMSE (PMID 38561021)
Moderate dementia Basic ADLs affected, supervision needed — MCID +2 CDR-SB, −1.4 to −3 MMSE (PMID 38561021)
Severe / advanced Loss of ambulation, continence, meaningful speech; eating problems ~6 y total dementia stage 18-month mortality 54.8%; eating problem probability 85.8% (PMID 31164314; PMID 19828530)

Vermunt's multistate model gives the durations and their modifiers: overall AD duration was 24 years at age 60 falling to 15 years at age 80, and male sex, clinical (rather than population) setting, APOE ε4 carriership and abnormal CSF tau each shortened duration, with stage-dependent effects (Vermunt 2019, PMID 31164314). Because the model was fitted across six cohorts including clinic samples, these numbers describe the modelled populations, not an individual prognosis.

Typical amnestic presentation

The 1984 description — insidious onset, progressive impairment of memory and at least one other cognitive domain, no early motor/sensory/coordination deficit, other causes excluded — remains a good clinical summary of the modal case (see criteria). Episodic memory fails first, with impaired encoding and consolidation rather than retrieval: cueing does not rescue recall, distinguishing it from the retrieval-pattern deficits typical of vascular and depressive cognitive impairment. Semantic and executive changes, anomia, and impaired complex-task sequencing follow. Insight declines unevenly, and the discrepancy between self-report and informant report is itself diagnostic information, which is why the CDR requires both interviews (Morris 1997, PMID 9447441).

Atypical phenotypes

Atypical presentations matter operationally because they are the ones that are missed, and because they are enriched in young-onset disease and depleted of APOE ε4 (Graff-Radford 2021, PMID 33609479).

Variant Core syndrome AD-pathology yield Distinguishing quantities
Posterior cortical atrophy (PCA) Early, prominent, progressive visuoperceptual/visuospatial failure with relatively preserved memory early (Crutch 2017, PMID 28259709) Amyloid PET positive in 94% (90–97); tau PET 97% (93–100); AD at autopsy 94% (90–97) Mean age at symptom onset 59.4 y (95% CI 58.9–59.8); 60% women; 80% "pure" syndrome (Chapleau 2024, PMID 38267189)
Logopenic variant PPA (lvPPA) Impaired single-word retrieval and sentence repetition, phonological errors, spared grammar and single-word comprehension Amyloid PET positive in 25/26 (96%) Contrast: 86% of semantic and 90% of nonfluent/agrammatic PPA were amyloid-negative (Santos-Santos 2018, PMID 29309493)
Behavioural/dysexecutive AD (bvAD) Apathy-predominant behavioural change or predominant executive failure Autopsy or biomarker AD by design in the referenced series 52% of bvAD met possible bvFTD criteria; ≥1 APOE ε4 in 60% (bvAD) and 40% (dysexecutive); atrophy temporoparietal-predominant, not frontal (Ossenkoppele 2015, PMID 26141491)

Two corrections to common assumptions come from these data. First, PCA is not a rare curiosity attached to many pathologies — in the largest pooled series (1,092 individuals, 36 centres, 16 countries) it was highly specific for AD, although co-pathology was the rule at autopsy: cerebral amyloid angiopathy 71%, Lewy body disease 44%, cerebrovascular injury 42% (PMID 38267189). Second, "frontal variant AD" is a misnomer: voxel-based morphometry showed temporoparietal-predominant atrophy with only limited frontal involvement, which is why Ossenkoppele proposed the behavioural/dysexecutive terminology (PMID 26141491).

The PPA classification itself is a three-way clinical scheme (nonfluent/agrammatic, semantic, logopenic) with optional imaging-support and definite-pathology qualifiers (Gorno-Tempini 2011, PMID 21325651); its value here is that variant assignment predicts amyloid status strongly enough to change the diagnostic workup.

Neuropathologic subtypes underlie some of the clinical variation

Tangle-density mapping in 889 autopsy cases classified AD into hippocampal-sparing (11%), typical (75%) and limbic-predominant (14%). Hippocampal-sparing cases had higher cortical and lower hippocampal tangle counts, less hippocampal atrophy, younger age at death (mean 72 y, SD 10) and a male predominance (63%); limbic-predominant cases were older (mean 86 y, SD 6) and predominantly women (69%), with MAPT H1H1 more common (70% vs 46%, p=0.011). Clinical presentation, age at onset, duration and rate of decline all differed between subtypes, and the pattern replicated in an independent cohort of 113 (Murray 2011, PMID 21802369). Roughly a quarter of AD is therefore not the Braak-stereotypical case, which has direct consequences for tau imaging and staging.

Survival differs by phenotype

In 2,081 biomarker-confirmed sporadic AD patients from the Amsterdam Dementia Cohort, median survival from first memory-clinic visit was 6.3 years (95% CI 5.8–6.9) for atypical AD versus 7.2 years (7.0–7.5) for typical AD (p=0.02); adding atypical status to a model containing age, sex, education, MMSE and APOE ε4 improved fit and carried HR 1.31 (1.10–1.56). By variant: PCA HR 1.35 (1.05–1.73), lvPPA 1.27 (0.94–1.69), bvAD 1.31 (0.94–1.83) (Bader 2025, PMID 40294367). These are memory-clinic figures with left truncation at first visit, and they are shorter than the ~5.8-year mean survival from diagnosis reported in a broad meta-analysis (Liang 2021, PMID 36097997) only because the populations and index events differ.

Staging and measurement instruments

Instrument Range / direction What it measures Principal limitation
CDR global, CDR-SB 0–3 global; 0–18 sum of boxes, higher = worse Six domains (memory, orientation, judgment/problem-solving, community affairs, home/hobbies, personal care) from patient and informant interview Time-intensive; unsuited to population screening (Morris 1997, PMID 9447441)
MMSE 0–30, lower = worse Orientation, registration, attention, recall, language, construction Ceiling effects in MCI; education and language bias (Folstein 1975, PMID 1202204)
MoCA 0–30, lower = worse Adds executive, abstraction, delayed recall load At a cut-off of 26 detected 90% of MCI versus MMSE's 18%, at specificity 87% versus 100% — a sensitivity/specificity trade, not a free upgrade (Nasreddine 2005, PMID 15817019)
GDS 1–7 stages Global stage from normal to severe dementia Ordinal, coarse; validated against behavioural, neuroanatomic and neurophysiologic measures (Reisberg 1982, PMID 7114305)
FAST 16 ordered items within 7 stages Functional loss, extending resolution into severe disease Designed for the severe end; reliability, validity and ordinality reported separately (Reisberg 1988, PMID 3249767; Sclan 1992, PMID 1504288)

The CDR's value in the current era is that it is the instrument on which anti-amyloid trials report their primary or key secondary outcomes, which makes the size of a clinically meaningful CDR-SB change a live question rather than a psychometric footnote.

How large a change is meaningful?

A rapid systematic review located ten articles reporting MCIDs for AD trial endpoints (Muir 2024, PMID 38561021):

Stage ADAS-Cog CDR-SB iADRS MMSE
MCI +2 to +3 +1 −5 −1 to −2
Mild AD +3 +2 −9 −2
Moderate–severe AD — +2 — −1.4 to −3

The review's own conclusion is the decisive one for this knowledge base: average treatment effects in recent anti-amyloid monoclonal antibody trials are lower than these previously published MCIDs, and MCIDs were derived without systematic patient and care-partner input (PMID 38561021). This is the empirical substrate of the clinical-meaningfulness dispute; the trial numbers are on anti-amyloid immunotherapy.

Rate of decline and its variability

Population-based incident-case data are the least biased estimate available. In 328 Cache County participants with incident AD followed a mean 3.80 years (range 0.07–12.90), mean annual change was −1.53 (SD 2.69) MMSE points, +1.44 (SD 1.82) CDR-SB and +2.55 (SD 5.37) NPI; 30–58% of survivors progressed less than one point per year on these measures even 5–7 years after onset. Women and those with younger onset declined faster on MMSE; APOE ε4 and cholinesterase-inhibitor use were associated with initial MMSE but not with rate of change (Tschanz 2011, PMID 21606896).

Two implications. First, the standard deviations are of the same order as the means, so "expected decline" is nearly uninformative for an individual. Second, clinic-derived progression rates are steeper than population-derived rates, which inflates apparent placebo-arm decline in trials recruited from specialist centres and therefore inflates the absolute difference a treatment can generate.

The severe stage

Advanced dementia has a describable, and largely under-recognised, terminal course. Following 323 nursing-home residents with advanced dementia for 18 months: 54.8% died; probability of pneumonia 41.1%, of a febrile episode 52.6%, of an eating problem 85.8%. Adjusted 6-month mortality after these events was 46.7%, 44.5% and 38.6% respectively. Dyspnoea (46.0%) and pain (39.1%) were common, and 40.7% underwent at least one burdensome intervention (hospitalisation, ED visit, parenteral therapy or tube feeding) in the last three months of life. Where the health-care proxy understood the prognosis and expected complications, the odds of a burdensome intervention were far lower (adjusted OR 0.12, 95% CI 0.04–0.37) (Mitchell 2009, PMID 19828530). This is one of the strongest available arguments that prognostic communication is itself an intervention — see care, caregiving and health systems.

Heterogeneity that a mean slope conceals

Longitudinal MMSE data from 201 Caucasian people with possible or probable AD, followed for as long as 13.5 years, separated into six statistically distinct decline trajectories rather than a single “typical” slope. Baseline MMSE and age helped assign trajectory; psychotic-symptom burden increased the probability of a faster course across age and baseline-severity groups, whereas APOE ε4 did not (Wilkosz 2010, PMID 19781112). This is a reminder that trial-average decline is not a patient-level forecast and that neuropsychiatric phenotype carries prognostic information.

Late-life amnestic decline also has a newly operational competing explanation. Proposed 2025 clinical criteria describe probable LATE as progressive memory loss with substantial hippocampal atrophy and negative amyloid biomarkers, and possible LATE when amyloid is unavailable or present but hippocampal neurodegeneration is out of proportion to expected pure AD. The criteria are explicitly an initial framework awaiting validation; LATE with AD neuropathologic change is associated with a faster course than LATE alone (Wolk 2025, PMID 39807681). This is useful staging content because a positive amyloid test does not prove that amyloid is the dominant driver of the observed slope.

Presentation features that should redirect the diagnosis

Prominent early motor signs, rapid progression over weeks to months, prominent early visual hallucinations or fluctuation, early prominent behavioural disinhibition with frontal-predominant atrophy, or a clinical picture that does not match biomarker status all warrant reconsideration. Variant-specific biomarker yields quantify part of this: in semantic and nonfluent/agrammatic PPA, amyloid positivity was the exception (14% and 10%), and where it occurred with autopsy follow-up, the primary pathology was frontotemporal lobar degeneration with secondary AD (Santos-Santos 2018, PMID 29309493). Detailed mimic and safety material is on red flags and safety concerns.

Open questions

  • What explains the 30–58% of people with AD who progress less than one point per year for years — reserve, subtype, copathology, or measurement floor (Tschanz 2011, PMID 21606896)?
  • Should MCIDs be re-derived with patient and care-partner input, and would the resulting thresholds be larger or smaller than the current ones (Muir 2024, PMID 38561021)?
  • Do hippocampal-sparing and limbic-predominant subtypes require different trial endpoints, given that they differ in age, sex, presentation and decline rate (Murray 2011, PMID 21802369)?
  • Why do atypical variants carry a 31% higher mortality after adjustment for age, sex, education, severity and APOE — is the excess biological or care-related (Bader 2025, PMID 40294367)?
  • Can staging instruments designed before the biomarker era (GDS, FAST, CDR) be mapped onto the 2024 integrated biological-clinical staging scheme without loss (Jack 2024, PMID 38934362)?
  • Would earlier recognition of PCA — mean onset at 59.4 years, typically after long diagnostic delay — change outcomes, and by what mechanism (Chapleau 2024, PMID 38267189)?

References

  1. Sperling RA, et al. Toward defining the preclinical stages of Alzheimer's disease. Alzheimers Dement. 2011;7:280-92. PMID 21514248.
  2. Albert MS, et al. The diagnosis of mild cognitive impairment due to Alzheimer's disease. Alzheimers Dement. 2011;7:270-9. PMID 21514249.
  3. McKhann GM, et al. The diagnosis of dementia due to Alzheimer's disease. Alzheimers Dement. 2011;7:263-9. PMID 21514250.
  4. Jack CR, et al. Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup. Alzheimers Dement. 2024;20:5143-5169. PMID 38934362.
  5. Folstein MF, et al. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-98. PMID 1202204.
  6. Nasreddine ZS, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53:695-9. PMID 15817019.
  7. Reisberg B, et al. The Global Deterioration Scale for assessment of primary degenerative dementia. Am J Psychiatry. 1982;139:1136-9. PMID 7114305.
  8. Reisberg B. Functional assessment staging (FAST). Psychopharmacol Bull. 1988;24:653-9. PMID 3249767.
  9. Sclan SG, et al. Functional assessment staging (FAST) in Alzheimer's disease: reliability, validity, and ordinality. Int Psychogeriatr. 1992;4 Suppl 1:55-69. PMID 1504288.
  10. Morris JC. Clinical dementia rating: a reliable and valid diagnostic and staging measure for dementia of the Alzheimer type. Int Psychogeriatr. 1997;9 Suppl 1:173-6. PMID 9447441.
  11. Crutch SJ, et al. Consensus classification of posterior cortical atrophy. Alzheimers Dement. 2017;13:870-884. PMID 28259709.
  12. Chapleau M, et al. Demographic, clinical, biomarker, and neuropathological correlates of posterior cortical atrophy: an international cohort study and individual participant data meta-analysis. Lancet Neurol. 2024;23:168-177. PMID 38267189.
  13. Gorno-Tempini ML, et al. Classification of primary progressive aphasia and its variants. Neurology. 2011;76:1006-14. PMID 21325651.
  14. Santos-Santos MA, et al. Rates of amyloid imaging positivity in patients with primary progressive aphasia. JAMA Neurol. 2018;75:342-352. PMID 29309493.
  15. Ossenkoppele R, et al. The behavioural/dysexecutive variant of Alzheimer's disease: clinical, neuroimaging and pathological features. Brain. 2015;138:2732-49. PMID 26141491.
  16. Murray ME, et al. Neuropathologically defined subtypes of Alzheimer's disease with distinct clinical characteristics: a retrospective study. Lancet Neurol. 2011;10:785-96. PMID 21802369.
  17. Graff-Radford J, et al. New insights into atypical Alzheimer's disease in the era of biomarkers. Lancet Neurol. 2021;20:222-234. PMID 33609479.
  18. Mitchell AJ, et al. Rate of progression of mild cognitive impairment to dementia — meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand. 2009;119:252-65. PMID 19236314.
  19. Tschanz JT, et al. Progression of cognitive, functional, and neuropsychiatric symptom domains in a population cohort with Alzheimer dementia: the Cache County Dementia Progression Study. Am J Geriatr Psychiatry. 2011;19:532-42. PMID 21606896.
  20. Vermunt L, et al. Duration of preclinical, prodromal, and dementia stages of Alzheimer's disease in relation to age, sex, and APOE genotype. Alzheimers Dement. 2019;15:888-898. PMID 31164314.
  21. Bader I, et al. Survival differences between individuals with typical and atypical phenotypes of Alzheimer disease. Neurology. 2025;104:e213603. PMID 40294367.
  22. Mitchell SL, et al. The clinical course of advanced dementia. N Engl J Med. 2009;361:1529-38. PMID 19828530.
  23. Muir RT, et al. Minimal clinically important difference in Alzheimer's disease: rapid review. Alzheimers Dement. 2024;20:3352-3363. PMID 38561021.
  24. Moscoso A, et al. Frequency and clinical outcomes associated with tau positron emission tomography positivity. JAMA. 2025;334:229-242. PMID 40522652.
  25. Liang CS, et al. Mortality rates in Alzheimer's disease and non-Alzheimer's dementias: a systematic review and meta-analysis. Lancet Healthy Longev. 2021;2:e479-e488. PMID 36097997.
  26. Vos SJ, et al. Preclinical Alzheimer's disease and its outcome: a longitudinal cohort study. Lancet Neurol. 2013;12:957-65. PMID 24012374.
  27. Wilkosz PA, et al. Trajectories of cognitive decline in Alzheimer's disease. Int Psychogeriatr. 2010;22:281-90. PMID 19781112.
  28. Wolk DA, et al. Clinical criteria for limbic-predominant age-related TDP-43 encephalopathy. Alzheimers Dement. 2025;21:e14202. PMID 39807681.