Mild Cognitive Impairment & Dementia Workup
Mild Cognitive Impairment & Dementia Workup
What Do You Need to Know?
- MCI = objective cognitive decline (1–1.5 SD below mean) with preserved functional independence — annual conversion to dementia ~10–15%; amnestic MCI most likely to convert to AD
- MoCA is superior to MMSE for detecting MCI — includes executive/visuospatial tasks; MMSE has ceiling effect and misses early deficits
- Standard dementia workup (AAN): CBC, CMP, TSH, B12, structural neuroimaging — rule out reversible causes first
- MRI atrophy patterns localize the diagnosis: hippocampal = AD, frontal/temporal = FTD, occipital = DLB/PCA — Scheltens MTA scale for grading
- CSF biomarkers: decreased Aβ42 + increased p-tau = AD; Aβ42/40 ratio is more accurate than Aβ42 alone; ATN framework classifies by biomarker profile
- Blood-based biomarkers are emerging: p-tau217 is the most promising screening biomarker for AD pathology
- CDR is the most widely used staging tool — CDR 0.5 = MCI/very mild dementia; CDR-SB (sum of boxes) tracks progression in clinical trials
🚩 Don’t Miss — Test-Day Priorities
- MCI = Petersen/NIA-AA core: patient/informant/clinician cognitive concern + objective impairment (1–1.5 SD) in ≥1 domain + preserved functional independence + does NOT meet dementia criteria — preserved ADLs is the line between MCI and dementia.
- Amnestic MCI (isolated memory loss) has the highest conversion rate to AD; overall MCI → dementia ~10–15%/year in clinic cohorts; ~15–20% revert to normal — always hunt reversible causes (depression, anticholinergics, B12, OSA).
- MoCA > MMSE for MCI — MoCA is a sensitive cognitive SCREEN, not a diagnostic staging test. Commonly cited impairment bands (e.g., ≥26 often called normal, lower scores raising concern; add 1 if ≤12 yr education) must be interpreted with age, education, language/culture, sensory status, and functional history — do NOT diagnose MCI or dementia from MoCA bands alone. MMSE has a ceiling effect and misses executive/visuospatial deficits.
- AAN core dementia workup: CBC, CMP, TSH, B12 + structural neuroimaging (MRI > CT) — rule out reversible causes first; consider RPR, HIV, autoimmune panel based on red flags.
- MRI atrophy patterns localize: hippocampal/MTL → AD; frontal/anterior temporal "knife-edge" → FTD; preserved hippocampi + occipital → DLB; parieto-occipital → PCA; ventriculomegaly out of proportion → NPH.
- FDG-PET signatures: temporoparietal + posterior cingulate hypometabolism → AD; frontal/anterior temporal → FTD; occipital with cingulate island sign → DLB.
- CSF AD signature: low Aβ42 (prefer Aβ42/40 ratio) + high p-tau181/217 + high t-tau; negative amyloid PET effectively rules out AD; positive scan does NOT confirm AD (~30% of normal elderly are amyloid+).
- Plasma p-tau217 is the best emerging blood biomarker for AD (~90% concordance with amyloid PET); ATN framework (A+T+ confirms AD pathology) defines AD biologically across preclinical → MCI due to AD → AD dementia.
- CDR staging: 0 normal, 0.5 = MCI/very mild dementia, 1 mild, 2 moderate, 3 severe; CDR-SB is primary endpoint in lecanemab/donanemab trials; FAST ≥7c qualifies for hospice.
- Anti-amyloid mAb (lecanemab, donanemab) requires amyloid+ biomarker + early symptomatic AD; APOE ε4 genotyping SHOULD be performed before treatment per FDA labels for ARIA risk counseling (ε4 homozygotes highest risk) — recommended, not an absolute prerequisite; if not performed, treatment is not automatically prohibited but ARIA counseling is incomplete. APOE is NOT a diagnostic test.
🔍 Buzzwords & Pathognomonic FindingsClinical / criteria · Bedside cognitive testing · Workup / biomarkers
Clinical / criteria
- Cognitive concern + objective impairment ≥1 domain + preserved ADLs + not dementia → MCI (Petersen / NIA-AA 2011)
- Isolated memory loss with intact function, high AD conversion → Amnestic MCI
- Subjective complaint with NORMAL objective testing → Subjective Cognitive Decline (SCD) — preclinical
- Significant cognitive decline interfering with independence, not delirium → Major neurocognitive disorder (DSM-5 dementia)
- "I don’t know" answers, acute onset, mood history → Pseudodementia (depression)
- Cues do NOT help recall (encoding/storage deficit) → Cortical/AD pattern; cues DO help → subcortical (VaD, PD, NPH)
- A+T+N+ on biomarkers → AD with neurodegeneration (NIA-AA ATN)
Bedside cognitive testing
- MoCA (sensitive cognitive screen, NOT a diagnostic staging test) → Best brief screen for MCI / executive dysfunction; commonly used cutoffs (e.g., ≥26 often called normal; lower scores raising concern) vary with age, education, language/culture, and functional history — do NOT diagnose MCI or dementia from MoCA bands alone
- MMSE ≤23/24 → Dementia screen (ceiling effect — misses MCI)
- 3-word recall + clock drawing → Mini-Cog (primary-care screen)
- Trail Making B, verbal fluency (FAS), Luria fist-edge-palm, Go/no-go → Executive / frontal probes
- Clock drawing, Rey figure copy, pentagons, intersecting circles → Visuospatial / parietal probes
- Boston Naming Test, category fluency (animals), Token Test → Language battery (PPA work-up)
- RAVLT / CVLT / Logical Memory → Episodic memory (hippocampal / AD)
- CDR 0.5 → MCI / very mild dementia; FAST ≥7c → hospice eligibility
Workup / biomarkers
- CBC + CMP + TSH + B12 + MRI → AAN core dementia workup
- MTL/hippocampal atrophy (Scheltens MTA ≥2) → AD; preserved hippocampi + visual hallucinations → DLB
- Anterior temporal "knife-edge" / temporal pole scooped out → svPPA; frontal/insular atrophy → bvFTD/nfvPPA
- Cortical ribboning + caudate/putamen DWI signal → CJD; Evans index >0.3 + DESH → NPH
- Temporoparietal + posterior cingulate FDG hypometabolism → AD; cingulate island sign → DLB
- Low CSF Aβ42 + low Aβ42/40 ratio + high p-tau181/217 → AD-pattern CSF
- RT-QuIC positive (>90% sens, ~99% spec); t-tau/p-tau ratio >20 → CJD
- Plasma p-tau217-based assays → Best-performing blood biomarkers for AD pathology; Fujirebio Lumipulse G pTau217/Aβ42 Plasma Ratio is the first FDA-cleared blood test (cognitively impaired adults in specialty care — not a stand-alone screen); PrecivityAD2 / ALZpath / Quanterix are clinically used and validated as LDTs but are NOT FDA-cleared
- Abnormal DaTscan (reduced striatal uptake) → DLB / PDD (normal in AD)
- APOE ε4 homozygote → Highest ARIA risk before lecanemab/donanemab; C9orf72 hexanucleotide repeat → FTD-ALS; PSEN1 → earliest-onset autosomal dominant AD
Subjective Cognitive Decline (SCD)
- SCD (preclinical): subjective cognitive decline with normal objective testing and preserved function
- ~2× risk of progression to MCI/dementia compared to no SCD
- Particularly concerning ("SCD-plus" features; Jessen 2014 SCD-I) if:
- Persistent over time
- Age >60
- Onset <5 years ago
- Concern about memory specifically (rather than other domains)
- Informant corroboration
- Sits on the cognitive continuum: SCD → MCI → dementia; not all SCD progresses, but it identifies a higher-risk cohort for surveillance
MCI: Definition & Diagnostic Criteria
Core Diagnostic Features (Petersen criteria / NIA-AA 2011)
- Subjective cognitive concern — reported by patient, informant, or clinician
- Objective cognitive impairment — performance 1–1.5 SD below age/education-adjusted norms in ≥1 cognitive domain
- Preserved functional independence — may have mild difficulties with complex IADLs, but fundamentally independent
- Does NOT meet criteria for dementia — impairment does not significantly interfere with daily functioning
- DSM-5 terminology: mild neurocognitive disorder (vs. major neurocognitive disorder = dementia) — a related but not identical construct (1–2 SD threshold; informant or clinician concern; modest cognitive decline)
MCI Subtypes
| Subtype | Domains Affected | Most Likely Etiology |
|---|---|---|
| Amnestic — single domain | Memory only | AD (highest conversion rate) |
| Amnestic — multi-domain | Memory + ≥1 other domain | AD, vascular, mixed |
| Non-amnestic — single domain | One non-memory domain (e.g., executive, language) | FTD, DLB, vascular, psychiatric |
| Non-amnestic — multi-domain | ≥2 non-memory domains | Vascular, DLB, FTD |
MCI Prognosis
- Annual conversion to dementia: ~10–15%/year in specialty clinic / amnestic MCI cohorts; ~5–10%/year in community cohorts; lower for non-amnestic subtypes (vs. 1–2% in age-matched controls)
- ~30–50% of MCI patients progress to dementia within 5 years
- ~15–20% may revert to normal cognition (especially if depression, medication, or sleep disorder was the cause)
- Predictors of conversion: amnestic subtype, APOE ε4, positive amyloid biomarkers, hippocampal atrophy, lower baseline MoCA
AAN 2018 MCI Management Recommendations (Petersen)
- Recommend regular exercise (Level B; ≥2×/week) — the only intervention with a positive evidence-based recommendation for cognitive benefit in MCI
- Optimize vascular risk factors — BP, lipids, glucose, smoking cessation, diet
- Cognitive training may be considered — modest evidence; reasonable to discuss with patients
- NO FDA-approved medications for MCI; cholinesterase inhibitors are NOT routinely recommended (no proven benefit in delaying conversion to dementia)
- Re-evaluate cognition every 6–12 months — track for progression, reversion, or stability
- Discuss prognosis including potential for reversion (~15–20%) and conversion risk; address advance planning while capacity is intact
💎 Board Pearl
- Amnestic MCI has the highest conversion rate to AD — if the board describes isolated memory loss with preserved function, this is the diagnosis
- MCI is a clinical diagnosis — biomarkers add certainty but are not required; preserved functional independence is the key distinguishing feature from dementia
- ~15–20% of MCI patients revert to normal — always look for reversible causes (depression, medications, OSA, B12 deficiency)
Cognitive Domains & Localization
💎 Memory Taxonomy Quick Frame
| Memory Type | Substrate | Bedside Probe |
|---|---|---|
| Episodic (declarative) | Hippocampus, Papez circuit, MTL | "What did you eat yesterday?" / word-list recall (RAVLT, CVLT). Impaired in AD. |
| Semantic | Inferior & anterolateral temporal cortex, temporal poles | Naming, category fluency, picture-meaning matching. Impaired in svPPA / semantic-variant FTD. |
| Working memory | Dorsolateral prefrontal cortex (with Broca/Wernicke phonological loop) | Digit span backward, mental arithmetic. Impaired in frontal/executive dysfunction. |
| Procedural (non-declarative) | Basal ganglia, cerebellum, supplementary motor area | Riding a bicycle, motor sequence learning. Spared in early AD (patient H.M. could learn motor tasks). |
| Cognitive Domain | Key Anatomy | Testing Examples | Impairment Suggests |
|---|---|---|---|
| Memory (episodic) | Hippocampus, medial temporal lobe, Papez circuit | Word list recall, story recall, RAVLT, CVLT | AD, hippocampal sclerosis |
| Executive function | Prefrontal cortex, frontal-subcortical circuits | Trail Making B, Wisconsin Card Sort, Stroop, verbal fluency (FAS) | FTD (bvFTD), vascular dementia, PDD/DLB, normal pressure hydrocephalus |
| Language | Left perisylvian cortex (Broca, Wernicke, arcuate fasciculus) | Boston Naming Test, category fluency (animals), Token Test, repetition | PPA variants (svPPA, nfvPPA, lvPPA), AD (late) |
| Visuospatial | Parietal lobes (right > left), occipitoparietal | Clock drawing, Rey Complex Figure copy, line bisection, Benton Judgment of Line Orientation | DLB, PCA (visual variant AD), CBD |
| Attention / Processing speed | Frontal-subcortical networks, ascending reticular activating system | Digit span, Trail Making A, coding/symbol search | Delirium, vascular dementia, DLB (fluctuations), depression |
| Social cognition / Behavior | Orbitofrontal cortex, anterior temporal lobes, insula, amygdala | Faux pas test, emotion recognition, informant-based behavioral inventories | bvFTD (early, prominent) |
💎 Board Pearl
- Domain-specific impairment predicts etiology: isolated memory loss → AD; executive/behavioral → FTD; visuospatial → DLB/PCA; language → PPA
- Subcortical pattern (slow processing, poor attention, executive dysfunction, retrieval-based memory deficit) vs. cortical pattern (true amnesia, aphasia, apraxia, agnosia) — vascular, PD, NPH, and DLB tend to be subcortical
- In AD, memory deficit is encoding/storage (hippocampal) — cues do NOT help recall; in subcortical dementias, memory deficit is retrieval-based — cues DO help
Cognitive Screening Tools
| Test | Score Range | Cutoff | Domains Tested | Strengths | Limitations |
|---|---|---|---|---|---|
| MMSE | 0–30 | ≤23 or <24 (dementia, traditional Folstein cutoff); not reliably sensitive for MCI | Orientation, registration, attention/calculation, recall, language, construction | Widely validated; quick (~10 min); extensive normative data | Ceiling effect (misses MCI); limited executive/visuospatial testing; copyrighted; education bias |
| MoCA | 0–30 | Sensitive cognitive SCREEN (not a diagnostic staging test). Commonly cited bands: ≥26 often called normal, lower scores raising concern (add 1 point if ≤12 yr education) — interpret with age, education, language/culture, sensory status, and function; do NOT diagnose MCI or dementia from MoCA bands alone | Executive (TMT-B, clock, fluency), visuospatial, naming, memory (delayed recall), attention, language, abstraction, orientation | Better sensitivity for MCI than MMSE; tests executive/visuospatial; free for clinical use | Lower specificity; education/cultural bias; practice effects on serial testing |
| Mini-Cog | 0–5 | ≤2 (positive screen) | 3-word recall + clock drawing | Very brief (~3 min); language-independent; good for primary care screening | Limited sensitivity for MCI; no domain-specific information; binary result |
| SLUMS | 0–30 | HS education: 27–30 normal, 21–26 MCI, ≤20 dementia; <HS: 25–30 normal, 20–24 MCI, ≤19 dementia (Tariq 2006) | Orientation, memory, calculation, animal naming, clock, figure recognition | Free; better MCI detection than MMSE; education-adjusted cutoffs | Less extensive normative data than MMSE/MoCA |
| AD8 | 0–8 | ≥2 (suggests cognitive impairment) | Informant-based: judgment, interest, repetition, tools, date, thinking, appointments, memory | Informant perspective; quick; detects change over time | Depends on informant reliability; does not assess cognition directly |
💎 Board Pearl
- MoCA > MMSE for detecting MCI — MoCA includes executive and visuospatial tasks that MMSE lacks; if a board question asks the best screening tool for early cognitive impairment, the answer is MoCA
- MMSE ceiling effect: highly educated patients can score 28–30 despite having early MCI — the MMSE is NOT sensitive enough for early detection
- Clock drawing tests executive function AND visuospatial ability — abnormal clock with normal word recall suggests non-AD pathology (DLB, FTD, vascular)
Neuropsychological Testing
When to Refer
- Screening test is borderline or discordant with clinical impression
- Need to characterize specific domain deficits (e.g., differentiate AD from FTD or vascular)
- Atypical presentation (young onset, isolated non-memory complaint, psychiatric overlap)
- Medicolegal or disability evaluation; decision-making capacity questions
- Baseline for serial monitoring of progression
- Presurgical evaluation (e.g., DBS for PD, epilepsy surgery)
Standard Battery Components
| Domain | Common Tests |
|---|---|
| Memory | RAVLT, CVLT-3, Logical Memory (WMS), Brief Visuospatial Memory Test (BVMT) |
| Executive | Trail Making Test B, Wisconsin Card Sorting Test, Stroop Color-Word, letter fluency (FAS/CFL) |
| Language | Boston Naming Test, category fluency (animals), Token Test |
| Visuospatial | Rey Complex Figure (copy), Judgment of Line Orientation, Block Design (WAIS) |
| Attention / Speed | Digit Span, Trail Making Test A, Coding (WAIS), Symbol Search |
| Effort / Validity | Test of Memory Malingering (TOMM), Word Memory Test, Green Word Memory Test |
| Mood / Behavior | BDI-2, GDS (Geriatric Depression Scale), NPI (Neuropsychiatric Inventory) |
Key Concepts
- Normative data: scores are compared against age-, education-, and sometimes sex-matched norms; impairment = 1–1.5 SD below mean (~7th–16th percentile)
- Ecological validity: how well test performance predicts real-world functioning (driving, finances, medication management)
- Effort testing: MUST be included — without valid effort, the entire battery is uninterpretable; low effort ≠ malingering (can reflect depression, fatigue, poor motivation)
- Serial testing: repeat at 6–12 month intervals to track progression; use alternate forms to minimize practice effects; decline ≥1 SD on serial testing is clinically significant
Clinical Pearl
- Neuropsychological testing differentiates cortical from subcortical patterns: AD shows rapid forgetting (poor delayed recall, poor recognition) vs. vascular/subcortical shows retrieval deficit (poor free recall but intact recognition)
- A patient who fails effort testing does NOT automatically have a functional disorder — common causes include severe depression, delirium, fatigue, and low education. Interpret in clinical context.
Standard Dementia Workup
AAN Practice Parameter: Core Workup
| Test | Purpose | Key Findings |
|---|---|---|
| CBC | Anemia, infection, malignancy | Macrocytic anemia → B12/folate deficiency |
| CMP (BMP + LFTs) | Metabolic encephalopathy | Hyponatremia, hypercalcemia, hepatic/renal failure |
| TSH | Hypothyroidism | Elevated TSH → treatable cognitive impairment (though rarely sole cause of dementia) |
| Vitamin B12 | B12 deficiency | Low B12 → cognitive impairment, subacute combined degeneration; check methylmalonic acid if borderline |
| Structural neuroimaging | Rule out mass, NPH, SDH; assess atrophy pattern | MRI preferred over CT — see atrophy patterns section |
Extended Workup: When to Add
| Test | When to Order |
|---|---|
| RPR/VDRL | Risk factors for syphilis; rapidly progressive dementia; confirm with FTA-ABS |
| HIV testing | Risk factors; young-onset dementia; subcortical pattern |
| ESR/CRP | Suspected vasculitis, inflammatory, or autoimmune etiology |
| Lumbar puncture | Rapidly progressive dementia (rule out CJD, autoimmune, infection, carcinomatous meningitis); suspected NPH; AD biomarkers |
| EEG | Rapidly progressive dementia (CJD — periodic sharp wave complexes); seizures; transient epileptic amnesia; autoimmune encephalitis |
| Genetic testing | Early-onset (<65 yr), strong family history (autosomal dominant pattern), specific phenotype (e.g., FTD + ALS) |
| Autoimmune panel | Subacute onset, inflammatory CSF, seizures, movement disorder, psychiatric features; send serum + CSF |
| Heavy metals / Toxicology | Occupational exposure history; lead, mercury, arsenic, bismuth |
| Folate, thiamine | Alcohol use disorder, malnutrition; Wernicke encephalopathy |
💎 Board Pearl
- AAN core labs: CBC, CMP, TSH, B12 + structural imaging — this is the minimum standard workup for all dementia evaluations
- Hypothyroidism and B12 deficiency are the classic "reversible" causes tested on boards — but they are rarely the SOLE cause of dementia; more often they are comorbid
- Rapidly progressive dementia (<1–2 years from symptom onset to severe dementia) triggers an expanded workup: LP (14-3-3, RT-QuIC for CJD), EEG, MRI DWI, autoimmune panel, and consideration of brain biopsy
Structural Neuroimaging
MRI vs. CT
- MRI is preferred — superior for detecting atrophy patterns, white matter disease, hippocampal sclerosis, microbleeds
- CT: acceptable if MRI contraindicated; detects mass lesions, hydrocephalus, large-vessel strokes, SDH
- Minimum MRI sequences: T1 (volumetric), T2/FLAIR (white matter), DWI (CJD, acute stroke), T2*/SWI (microbleeds, hemosiderin)
Atrophy Patterns by Dementia Type
| Dementia | Atrophy Pattern | Key MRI Findings |
|---|---|---|
| AD (typical) | Medial temporal (hippocampus, entorhinal cortex) → parietal → diffuse | Hippocampal atrophy (Scheltens MTA score ≥2); posterior parietal atrophy; relative sparing of primary motor/visual cortex |
| bvFTD | Frontal (orbitofrontal, dorsolateral) and anterior temporal | Frontal lobe atrophy (often asymmetric); "knife-edge" gyri; relative sparing of posterior regions |
| svPPA (semantic) | Anterior temporal (left > right) | Asymmetric anterior temporal atrophy; temporal pole "scooped out" |
| nfvPPA (nonfluent) | Left posterior frontal / insular | Left inferior frontal and insular atrophy; perisylvian involvement |
| DLB | Relatively preserved hippocampi; diffuse cortical, occipital | Less hippocampal atrophy than AD (key distinguishing feature); posterior cortical atrophy in some |
| PCA (posterior cortical atrophy) | Occipitoparietal | Posterior-predominant atrophy; usually AD pathology but visuospatial presentation |
| Vascular dementia | Multi-focal or strategic infarcts; white matter disease | Lacunar infarcts (basal ganglia, thalamus); confluent periventricular white matter hyperintensities; microbleeds |
| CJD | Cortical ribboning on DWI; caudate/putamen signal | DWI is most sensitive — cortical ribboning + caudate/putamen high signal; FLAIR less sensitive |
| NPH | Ventriculomegaly out of proportion to sulcal atrophy | Evans index >0.3; disproportionate callosal angle narrowing; tight high-convexity sulci with enlarged Sylvian fissures (DESH) |
Scheltens Medial Temporal Atrophy (MTA) Scale
- Score 0: No atrophy (normal choroid fissure, temporal horn, hippocampus)
- Score 1: Widened choroid fissure only; normal temporal horn and hippocampus
- Score 2: Mild widening of temporal horn; mild hippocampal volume loss
- Score 3: Moderate widening of temporal horn; moderate hippocampal volume loss
- Score 4: Severe hippocampal volume loss; widely open temporal horn
- Age-adjusted abnormal: ≥1.5 if <65 yr; ≥2 if 65–74; ≥2.5 if ≥75
💎 Board Pearl
- DLB has LESS hippocampal atrophy than AD — this is a key differentiating feature on MRI; if a board question shows preserved hippocampi with visual hallucinations and parkinsonism, think DLB
- CJD on MRI: DWI is the most sensitive sequence — look for cortical ribboning + caudate/putamen signal abnormality; FLAIR may be normal early
- NPH triad: gait apraxia (first and most responsive to shunting), urinary incontinence, dementia — Evans index >0.3 on imaging; large-volume LP with gait improvement supports the diagnosis
Functional Neuroimaging
FDG-PET (Fluorodeoxyglucose PET)
- Measures regional cerebral glucose metabolism — hypometabolism reflects synaptic dysfunction/neurodegeneration
- More sensitive than structural MRI for detecting early changes
| Dementia | FDG-PET Pattern |
|---|---|
| AD | Temporoparietal and posterior cingulate/precuneus hypometabolism; frontal involvement in later stages; relative sparing of primary sensorimotor, visual, basal ganglia, cerebellum |
| FTD | Frontal and/or anterior temporal hypometabolism (asymmetric); sparing of posterior regions |
| DLB | Similar to AD (temporoparietal) BUT with occipital hypometabolism (cingulate island sign: preserved posterior cingulate relative to precuneus) |
| PCA | Occipitoparietal hypometabolism |
| CBD | Asymmetric frontoparietal hypometabolism (contralateral to affected limb) + basal ganglia/thalamic asymmetry |
Amyloid PET
- Tracers: florbetapir (Amyvid), florbetaben (Neuraceq), flutemetamol (Vizamyl); Pittsburgh Compound B (PiB) is research-only
- Detects fibrillar Aβ plaques — positive in AD, also positive in ~30% of cognitively normal elderly (≥65 yr)
- Negative amyloid PET effectively rules out AD as the cause of dementia
- CMS coverage (2023 update): the one-per-lifetime limit was removed and amyloid PET is now covered under standard Medicare rules without CED, supporting use for anti-amyloid therapy eligibility and AUC-appropriate indications
Appropriate Use Criteria (AUC)
- Persistent/progressive unexplained MCI
- Atypical clinical presentation or course
- Early-onset dementia (<65 yr)
- Etiologically unclear after standard workup
- NOT appropriate for: asymptomatic individuals, family history alone, severity staging, screening, patients meeting core AD criteria with typical presentation
Tau PET
- Tracer: flortaucipir (Tauvid) — FDA-approved 2020; detects paired helical filament tau (3R/4R)
- Tau PET correlates with Braak staging and cognitive severity better than amyloid PET
- May help differentiate AD from non-AD tauopathies; emerging clinical use
- Useful in research: treatment response monitoring, patient selection for anti-tau trials
DaTscan (Ioflupane SPECT)
- Not strictly a "dementia scan" but critical for differentiating DLB from AD
- Abnormal (reduced striatal uptake) in DLB, PDD — normal in AD
- Indicative biomarker for DLB in the diagnostic criteria (one of three indicative biomarkers with PSG-confirmed RBD and reduced MIBG cardiac uptake per 2017 McKeith criteria)
💎 Board Pearl
- Negative amyloid PET rules out AD — this is the highest-yield amyloid PET fact for boards; a positive scan does NOT confirm AD (30% of normal elderly are amyloid-positive)
- DLB on FDG-PET: occipital hypometabolism distinguishes it from AD; the cingulate island sign (preserved posterior cingulate) is relatively specific for DLB
- DaTscan differentiates DLB from AD — abnormal in DLB (presynaptic dopaminergic degeneration), normal in AD
CSF Biomarkers
Core AD CSF Biomarkers
| Biomarker | What It Measures | Direction in AD | Key Points |
|---|---|---|---|
| Aβ42 | Amyloid-beta 42 peptide | Decreased | Trapped in plaques → less in CSF; decreases early (preclinical stage); affected by total amyloid load |
| Aβ42/40 ratio | Ratio of Aβ42 to Aβ40 | Decreased | More accurate than Aβ42 alone — corrects for inter-individual variation in amyloid production; preferred measure |
| p-tau181 | Phosphorylated tau at threonine 181 | Increased | Reflects tau pathology (tangle formation); relatively specific to AD; does NOT increase in most non-AD dementias |
| p-tau231 | Phosphorylated tau at threonine 231 | Increased | May increase earlier than p-tau181; correlates with amyloid pathology; useful for early detection |
| t-tau | Total tau (reflects neurodegeneration) | Increased | Non-specific marker of neuronal injury/death; elevated in AD, CJD (very high), stroke, TBI; NOT specific to AD |
| NfL (neurofilament light) | Neurofilament light chain | Increased | Non-specific neurodegeneration marker; elevated in FTD, ALS, MS, CJD, AD; highest in CJD and ALS |
ATN Classification Framework (NIA-AA 2018)
- Research framework that defines AD biologically (not clinically)
- A = Amyloid (Aβ42, amyloid PET)
- T = Tau (p-tau, tau PET)
- N = Neurodegeneration (t-tau, NfL, FDG-PET, MRI atrophy)
| ATN Profile | Interpretation |
|---|---|
| A+T+N+ | AD with neurodegeneration (full AD pathology) |
| A+T+N– | AD pathology without neurodegeneration (early AD) |
| A+T–N– | Alzheimer pathologic change (amyloid only; preclinical) |
| A–T+N+ | Non-AD tauopathy with neurodegeneration (e.g., PSP, CBD, FTLD-tau) |
| A–T–N+ | Non-AD neurodegeneration (e.g., FTD-TDP43, DLB, vascular) |
| A–T–N– | Normal AD biomarkers (consider non-neurodegenerative causes) |
CSF in Rapidly Progressive Dementia
- CJD: 14-3-3 protein (sensitivity ~85%, poor specificity); RT-QuIC (real-time quaking-induced conversion) — sensitivity >90%, specificity ~99% — best CSF test for CJD
- CJD: t-tau markedly elevated (commonly >1,150–1,300 pg/mL depending on assay); t-tau/p-tau ratio >20 is suggestive (tau is released but NOT phosphorylated in prion disease)
- Autoimmune encephalitis: mild pleocytosis, elevated protein, oligoclonal bands; send antibody panels (NMDAR, LGI1, CASPR2, GABA-B, AMPA)
💎 Board Pearl
- AD CSF signature: low Aβ42 + high p-tau + high t-tau — this combination has >85% sensitivity and specificity for AD
- Aβ42/40 ratio is more accurate than Aβ42 alone — corrects for individual variation in amyloid production; always prefer the ratio
- RT-QuIC is the best CSF test for CJD — >90% sensitivity, ~99% specificity; has largely replaced 14-3-3 protein
- t-tau/p-tau ratio >20 suggests CJD over AD — in CJD, tau is released massively from dying neurons but is NOT phosphorylated
Blood-Based Biomarkers
Emerging Plasma Biomarkers
| Biomarker | Clinical Significance | Key Points |
|---|---|---|
| p-tau217 | Most promising blood biomarker for AD | High concordance with amyloid PET and CSF biomarkers (~90% accuracy); differentiates AD from non-AD dementias; increases early in preclinical AD. FDA-cleared assay: Fujirebio Lumipulse G pTau217/Aβ42 Plasma Ratio — indicated to aid diagnosis in cognitively impaired adults in specialty care, NOT a stand-alone screening test. PrecivityAD2 / ALZpath / Quanterix are validated as laboratory-developed tests (LDTs) but are NOT FDA-cleared |
| p-tau181 | Good discrimination for AD | Less accurate than p-tau217 but still useful; elevated in AD, not in FTD or DLB; correlates with amyloid and tau PET |
| Aβ42/40 ratio (plasma) | Amyloid status | Decreased in AD; lower accuracy than p-tau markers; better when combined with p-tau; affected by peripheral amyloid sources |
| GFAP (glial fibrillary acidic protein) | Astrocytic activation / neuroinflammation | Elevated in AD (especially preclinical); reflects astrogliosis; may add value in combination panels |
| NfL (plasma) | Non-specific neurodegeneration | Elevated in FTD, ALS, MS, CJD, AD; useful for monitoring neurodegeneration; not disease-specific |
Clinical Role
- Screening: blood biomarkers may reduce need for CSF/PET by identifying amyloid-positive patients non-invasively
- Anti-amyloid therapy eligibility: lecanemab and donanemab require amyloid confirmation — plasma p-tau217 may serve as a gatekeeper before PET/CSF
- Current limitations: not yet standard of care everywhere; BMI, renal function, and comorbidities can affect levels; confirmatory testing (PET or CSF) still recommended for treatment decisions
Clinical Pearl
- p-tau217 is the single best blood biomarker for detecting AD pathology — it outperforms p-tau181 and plasma Aβ42/40 in head-to-head studies and has ~90% concordance with amyloid PET
- Plasma biomarkers are rapidly changing the diagnostic landscape — in the near future, a blood test may be the first step to confirm or rule out AD pathology before proceeding to PET or CSF
Dementia Staging Scales
CDR (Clinical Dementia Rating)
- Gold-standard staging tool; based on structured interview with patient AND informant
- Evaluates 6 domains: memory, orientation, judgment/problem-solving, community affairs, home/hobbies, personal care
- Memory is the primary category — used as the anchor for scoring (Washington University scoring rules)
| CDR Score | Stage | Clinical Description |
|---|---|---|
| 0 | Normal | No cognitive impairment |
| 0.5 | Very mild / Questionable | Consistent slight forgetfulness; MCI or very mild dementia |
| 1 | Mild dementia | Moderate memory loss; moderate difficulty with community affairs; mild but definite impairment at home |
| 2 | Moderate dementia | Severe memory loss; cannot function independently in community; requires assistance with personal care |
| 3 | Severe dementia | Severe memory loss (fragments only); oriented to person only; requires full assistance with personal care |
- CDR-SB (Sum of Boxes): sum of individual domain box scores (range 0–18); more granular than global CDR; primary endpoint in many clinical trials (lecanemab, donanemab); better at detecting change over time
GDS / FAST (Reisberg)
| GDS Stage | FAST Stage | Clinical Description | Approximate MMSE |
|---|---|---|---|
| 1 | 1 | No cognitive decline | 29–30 |
| 2 | 2 | Very mild (subjective complaints only) | 28–29 |
| 3 | 3 | Mild cognitive decline (early confusional; MCI) | 24–28 |
| 4 | 4 | Moderate (late confusional; mild dementia) — needs help with finances, travel | 18–23 |
| 5 | 5 | Moderately severe (early dementia) — needs help choosing clothes | 10–17 |
| 6 | 6a–6e | Severe (middle dementia) — dressing, bathing, toileting, incontinence | 1–9 |
| 7 | 7a–7f | Very severe — limited speech (≤6 words → single word → none), loss of ambulation, inability to sit, loss of smiling | 0 |
- FAST is particularly useful for hospice eligibility: FAST stage ≥7c (unable to ambulate, limited intelligible vocabulary, incontinent) with comorbidities qualifies for hospice
💎 Board Pearl
- CDR 0.5 = MCI or very mild dementia — this is the most commonly tested CDR stage on boards
- CDR-SB is the primary outcome measure in recent anti-amyloid trials (lecanemab, donanemab) — know that it is more sensitive to change than global CDR
- FAST ≥7c is the threshold for hospice eligibility in dementia — boards may ask about end-of-life care planning
- GDS/FAST tracks AD specifically and follows a predictable functional decline; it does NOT apply well to non-AD dementias (FTD, DLB) which have different progression patterns
Genetic Testing Considerations
When to Consider Genetic Testing
- Early-onset dementia (<65 yr) — especially <60 yr with autosomal dominant family history
- Strong family history: ≥3 affected individuals across ≥2 generations (autosomal dominant pattern)
- Specific phenotypes: FTD + ALS (C9orf72), young-onset AD, familial prion disease
- Always with genetic counseling — discuss implications for patient and family members before testing
Diagnostic vs. Predictive Testing
| Type | Setting | Examples | Key Considerations |
|---|---|---|---|
| Diagnostic | Symptomatic patient — confirms etiology | Testing a patient with early-onset AD for PSEN1/PSEN2/APP; FTD patient for C9orf72/GRN/MAPT | Result changes management, prognosis, family counseling; generally less ethically complex |
| Predictive (presymptomatic) | At-risk but asymptomatic family member | Testing child of known PSEN1 carrier; Huntington disease protocol | Requires formal genetic counseling; follow Huntington protocol (pre-test counseling, waiting period, psychological support); insurance/employment discrimination concerns (GINA protections) |
Key Genetic Causes of Dementia
| Gene | Inheritance | Phenotype | Key Points |
|---|---|---|---|
| APP | Autosomal dominant | Early-onset AD (40s–50s) | Chromosome 21; duplication also causes AD; Trisomy 21 (Down syndrome) → near-universal AD neuropathology by age 40; clinical dementia onset typically mid-50s |
| PSEN1 | Autosomal dominant | Earliest-onset familial AD (30s–50s) | Most common cause of autosomal dominant AD; >300 mutations; near-complete penetrance; may present with spastic paraparesis, seizures, myoclonus |
| PSEN2 | Autosomal dominant | Early-onset AD (40s–60s) | Rarer than PSEN1; later onset; incomplete penetrance |
| C9orf72 | Autosomal dominant | FTD, ALS, FTD-ALS | GGGGCC hexanucleotide repeat expansion; most common genetic cause of FTD AND familial ALS; may present with psychosis |
| GRN | Autosomal dominant | FTD (often nfvPPA or bvFTD) | Progranulin; haploinsufficiency; TDP-43 pathology; asymmetric atrophy |
| MAPT | Autosomal dominant | FTD (often bvFTD with parkinsonism) | Tau pathology; chromosome 17; variable phenotype |
APOE Testing Controversy
- APOE ε4: strongest genetic risk factor (not deterministic) for late-onset AD — 1 copy = ~3× risk; 2 copies = ~12× risk (range 8–15× depending on cohort/ancestry)
- APOE ε2: protective — reduces AD risk by ~40%
- NOT recommended for routine clinical testing — it is a risk factor, not diagnostic; many ε4 carriers never develop AD; many AD patients are ε4-negative
- APOE ε4 genotyping should be performed before initiating lecanemab or donanemab per FDA labels — recommended (not an absolute prerequisite) to counsel ARIA risk; ε4 homozygotes have markedly higher ARIA rates (amyloid-related imaging abnormalities) and require shared decision-making. If genotyping is not performed, treatment is not automatically prohibited but ARIA risk counseling is incomplete
- Direct-to-consumer testing (23andMe) reports APOE status — may present in clinic without genetic counseling context
💎 Board Pearl
- PSEN1 is the most common cause of autosomal dominant early-onset AD — onset as early as 30s; near-complete penetrance
- C9orf72 is the most common genetic cause of both FTD and familial ALS — hexanucleotide repeat expansion; may present with psychosis
- APOE ε4 is a risk factor, NOT diagnostic — do NOT order it to diagnose AD; it IS relevant for ARIA risk assessment before anti-amyloid therapy
- Down syndrome (Trisomy 21) → APP gene on chromosome 21 → overproduction of Aβ → near-universal AD neuropathology by age 40; clinical dementia onset typically mid-50s
Driving Evaluation in Cognitive Impairment
- Per AAN 2010 (Iverson) practice parameter:
- CDR ≥1 (mild dementia or greater) generally indicates unsafe driving
- CDR 0.5 (MCI / very mild) requires individualized on-road testing — do not blanket-restrict
- Useful predictors of driving impairment:
- Caregiver concern about driving safety
- History of crashes or citations
- Reduced driving mileage / self-restriction
- Aggressive or impulsive personality
- MMSE ≤24
- Consider formal driving evaluation by occupational therapy (on-road + simulator) when uncertain — the gold standard for capacity assessment
- Know state-specific physician reporting requirements — mandatory reporting in CA, DE, NV, NJ, OR, PA; voluntary or condition-specific elsewhere
- Document the conversation, the recommendation, and any reporting action in the chart
💎 Board Pearl
- CDR ≥1 → unsafe to drive; CDR 0.5 → on-road testing — the most testable point from the AAN 2010 driving guideline
- Caregiver concern is one of the strongest predictors of unsafe driving — weight it heavily even if the patient denies any problems
Putting It All Together: Workup Algorithm
Step-by-Step Approach
- Step 1 — Confirm cognitive impairment: screening (MoCA preferred) + informant history + functional assessment
- Step 2 — Rule out reversible causes: CBC, CMP, TSH, B12; medication review (anticholinergics, benzodiazepines, opioids); screen for depression (PHQ-9/GDS), sleep disorders (OSA), delirium
- Step 3 — Structural imaging: MRI brain (CT if MRI contraindicated) — assess atrophy pattern, rule out mass/NPH/SDH/vascular disease
- Step 4 — Neuropsychological testing: when diagnosis is uncertain, atypical presentation, need for domain-specific characterization, or baseline for monitoring
- Step 5 — Biomarkers (if indicated): CSF (Aβ42/40, p-tau, t-tau) or amyloid PET; blood biomarkers (p-tau217) emerging as screening step
- Step 6 — Expanded workup (if indicated): RPR, HIV, LP, EEG, genetic testing, autoimmune panel — based on clinical features and red flags
- Step 7 — Stage severity: CDR or GDS/FAST; discuss prognosis, advance directives, caregiver support
Standard Workup Checklist
| Category | All Patients | If Indicated |
|---|---|---|
| Cognitive screen | MoCA (or MMSE, Mini-Cog) | Formal neuropsychological testing |
| Labs | CBC, CMP, TSH, B12 | RPR, HIV, ESR, folate, thiamine, heavy metals, methylmalonic acid |
| Imaging | MRI brain (or CT) | FDG-PET, amyloid PET, tau PET, DaTscan |
| CSF | — | AD biomarkers (Aβ42/40, p-tau, t-tau); CJD (14-3-3, RT-QuIC); autoimmune antibodies; cytology |
| Blood biomarkers | — | p-tau217, NfL, GFAP (emerging; pre-treatment screening) |
| EEG | — | Rapidly progressive dementia, seizures, fluctuating cognition |
| Genetic testing | — | Early-onset, autosomal dominant pattern, FTD-ALS phenotype; APOE for ARIA risk (anti-amyloid therapy) |
| Mood / Sleep | Depression screen (PHQ-9, GDS) | Sleep study (OSA); substance use screen |
Clinical Pearl
- Medication review is critical — anticholinergics (diphenhydramine, oxybutynin, TCAs), benzodiazepines, and opioids are the most common iatrogenic causes of cognitive impairment in the elderly. Always check the Beers Criteria before attributing symptoms to neurodegeneration.
- Depression mimics dementia ("pseudodementia") — key clues: acute onset, subjective complaints disproportionate to objective deficits, "I don't know" answers (vs. confabulation in AD), history of mood disorder, and response to antidepressants. Treat depression first, then re-evaluate cognition.
References
- Petersen RC, Lopez O, Armstrong MJ, et al. Practice guideline update summary: mild cognitive impairment. Neurology. 2018;90(3):126–135.
- Jack CR Jr, Bennett DA, Blennow K, et al. NIA-AA research framework: toward a biological definition of Alzheimer's disease. Alzheimers Dement. 2018;14(4):535–562.
- Knopman DS, DeKosky ST, Cummings JL, et al. Practice parameter: diagnosis of dementia (an evidence-based review). Neurology. 2001;56(9):1143–1153.
- Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699.
- Johnson KA, Minoshima S, Bohnen NI, et al. Appropriate use criteria for amyloid PET: a report of the Amyloid Imaging Task Force. Alzheimers Dement. 2013;9(1):e1–e16.
- Hansson O, Edelmayer RM, Boxer AL, et al. The Alzheimer's Association appropriate use recommendations for blood biomarkers in Alzheimer's disease. Alzheimers Dement. 2022;18(12):2669–2686.
- Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology. 1993;43(11):2412–2414.
- Reisberg B, Ferris SH, de Leon MJ, Crook T. The Global Deterioration Scale for assessment of primary degenerative dementia. Am J Psychiatry. 1982;139(9):1136–1139.
- Scheltens P, Leys D, Barkhof F, et al. Atrophy of medial temporal lobes on MRI in "probable" Alzheimer's disease and normal ageing. J Neurol Neurosurg Psychiatry. 1992;55(10):967–972.
- Palmqvist S, Janelidze S, Quiroz YT, et al. Discriminative accuracy of plasma phospho-tau217 for Alzheimer disease vs other neurodegenerative disorders. JAMA. 2020;324(8):772–781.
Continue reading — sign in
The full note has more clinical pearls, tables, and board-focused tips. Free account, no fee.