Yes. You can have Alzheimer's disease without dementia, because the two terms describe different things: Alzheimer's is a biological process of amyloid plaques and tau tangles, while dementia is a clinical syndrome of cognitive decline severe enough to disrupt daily life. In the preclinical stage, that pathology is measurable on PET scans or in cerebrospinal fluid and blood while memory and function remain normal. Pooled amyloid PET data show roughly 25 to 35 percent of cognitively normal people in their 70s carry the pathology. Most never know it.
Why Alzheimer's Disease and Dementia Describe Different Things
Alzheimer's disease is defined by proteins, not by whether a person can manage a checkbook. Amyloid-beta 42 clumps into plaques outside neurons, and hyperphosphorylated tau forms tangles inside them. The tangles follow a route mapped by Braak stages I through VI. They begin in the entorhinal cortex and reach the neocortex only at stages V and VI, which is roughly where most people cross into dementia.
Dementia is a syndrome rather than a disease. It means acquired cognitive decline in at least one domain that interferes with independence. Roughly 60 to 80 percent of dementia cases involve Alzheimer's pathology, but more than half of autopsied brains carry a second contributor. Vascular injury, Lewy bodies, and TDP-43 are the usual partners. In plain terms: Alzheimer's is one way a brain can be damaged; dementia is what that damage does to daily life.
That separation makes a specific combination possible. A person can carry enough plaques and tangles to satisfy neuropathologic criteria for Alzheimer's disease and still score normally on cognitive testing. The reverse is equally real: a person can have dementia with no Alzheimer's pathology whatsoever.
The scale is large. About 7 million Americans aged 65 and older live with Alzheimer's dementia, roughly 1 in 9 people in that age band. The number who carry the pathology without the syndrome is larger still, and the great majority have never been tested.
How Can You Have Alzheimer's Without Dementia? The Preclinical Window
The idea that Alzheimer's can exist without dementia is not new, but it only became testable around 2011. That year, the National Institute on Aging and Alzheimer's Association workgroup, led by Sperling, split the disease into three clinical stages: preclinical, mild cognitive impairment, and dementia. Preclinical disease was further divided into three substages. Stage 1 meant amyloid abnormality alone, stage 2 added neurodegeneration, and stage 3 added measurable but subtle cognitive decline.
The 2018 research framework from the same groups, led by Jack, went further and removed symptoms from the definition entirely. It introduced the AT(N) scheme: amyloid, tau, and neurodegeneration, each scored positive or negative. Under that framework, an amyloid-positive person with no tau is labeled "Alzheimer's pathologic change," not Alzheimer's disease. Full disease requires both A and T to be positive.
The 2024 revision kept the biological stance and added blood-based core biomarkers, chiefly the plasma p-tau217 ratio. It also introduced a staging ladder from 1a through 6. A person at stage 1a has abnormal amyloid and no detectable tau; a person at stage 6 has widespread tau, atrophy, and severe symptoms.
Why do symptoms lag so far behind the biology? Three mechanisms matter. Amyloid deposition is slow and largely silent at the synapse level for years. The brain recruits reserve capacity and redundant networks to compensate. Symptoms appear only when tau spreads beyond the medial temporal lobe and synapse loss exceeds what those networks can absorb. In plain terms: the brain hides the damage until it runs out of workarounds.
How Often Does Alzheimer's Exist Without Dementia?
The prevalence depends entirely on age. Pooled amyloid PET data reported by Jansen and colleagues in 2015 showed abnormal amyloid in about 7 percent of cognitively normal people in their 50s. The figure rose to roughly 15 percent in the 60s, about 25 percent in the 70s, and roughly 35 percent in the 80s. Above age 90, it approached 50 percent.
Trial screening data agree. The A4 prevention study screened 4,486 cognitively normal adults aged 65 to 85 and found that about 1 in 3 carried abnormal amyloid. Autopsy series point the same way: roughly 30 percent of older adults who died with normal cognition meet neuropathologic criteria for intermediate or high Alzheimer's disease.
Those numbers reconcile once age weighting is made explicit. The 25 to 35 percent band for the 70s, the roughly 30 percent autopsy figure, and the 1-in-3 A4 screening result all describe the same underlying curve viewed at slightly different ages.
Translate that into people. Applying a 25 to 35 percent range to the roughly 61 million Americans aged 65 and older yields something on the order of 15 to 20 million adults with amyloid in the brain and no dementia. Progression is real but far from universal. Amyloid-positive cognitively normal adults convert to mild cognitive impairment or dementia at about 2 to 3 percent per year, against under 1 percent per year for amyloid-negative peers. Over five years, roughly 15 to 20 percent of the positive group will have moved to a symptomatic stage.
The Silent Timeline: Two Decades Before the First Symptom
The sequence of biomarker change is one of the most reproducible findings in the field. Amyloid moves first, tau follows, neurodegeneration follows tau, and cognition fails last.
| Time before symptom onset | What is happening | What a test can show | Clinical status |
|---|---|---|---|
| 20 to 25 years | Amyloid-beta 42 begins depositing in the neocortex | Cerebrospinal Aβ42/40 ratio falls; amyloid PET turns positive | Normal |
| 15 to 20 years | Amyloid plateaus; p-tau217 starts to rise | Plasma p-tau217 abnormal; tau PET still negative | Normal |
| 10 to 15 years | Tangles spread through entorhinal cortex and hippocampus | Tau PET in medial temporal lobe; hippocampal volume at low end | Normal, occasional complaints |
| 5 to 10 years | Neurodegeneration accelerates; network redundancy shrinks | FDG-PET hypometabolism; cortical thinning on MRI | Subtle decline on sensitive tests, still independent |
| 2 to 5 years | Tau reaches association cortex | Tau PET spreads beyond the temporal lobe | Mild cognitive impairment |
| 0 | Compensation is exhausted | No single threshold test | Dementia: daily function impaired |
| 4 to 8 years after | Pathology continues to spread | Global atrophy on MRI | Moderate to severe dementia |
Two features of this table surprise people. First, the amyloid phase lasts roughly 15 to 20 years on its own, so a positive scan in a healthy 60-year-old does not mean symptoms next year. Second, tau position, not amyloid load, tracks symptom severity. A person can have a heavy amyloid burden and mild tau and remain cognitively normal for a decade.
Preclinical vs. MCI vs. Dementia: Side-by-Side
The three stages differ in definition, biomarker profile, and what treatment is available. The last column is where the practical stakes sit.
| Feature | Preclinical Alzheimer's | MCI due to Alzheimer's | Dementia due to Alzheimer's |
|---|---|---|---|
| Definition | Biomarkers abnormal, cognition normal | Objective impairment, daily function preserved | Impairment plus loss of independence |
| Core biomarkers | Amyloid positive, tau often negative | Amyloid positive, tau usually positive | Amyloid positive, tau positive, atrophy present |
| How common | Roughly 25 to 35 percent of cognitively normal people in their 70s | 10 to 20 percent of adults over 65 have MCI from any cause; a minority is amyloid positive | About 7 million Americans aged 65 and older |
| Annual progression to dementia | 2 to 3 percent | 5 to 10 percent in community samples, about 15 percent in specialty clinics | Not applicable |
| Anti-amyloid eligibility | No, outside a clinical trial | Yes, if amyloid is confirmed | Yes, but only in the mild dementia stage |
| Usual route to detection | Research study or prevention trial | Clinical evaluation after a complaint | Clinical evaluation after functional decline |
The eligibility row carries the most weight for patients. Lecanemab, approved by the FDA in 2023, and donanemab, approved in 2024, are indicated for mild cognitive impairment or mild dementia with confirmed amyloid. Neither is approved for people who are cognitively normal. That single line explains why so much current research funding goes toward prevention trials rather than treatment trials.
Misconceptions and Edge Cases
The most common misconception is that amyloid positivity is a countdown timer. It is not. In the Nun Study cohort followed from the 1990s, a substantial fraction of women who died with normal cognition met neuropathologic criteria for Alzheimer's disease at autopsy. Education, brain size, vascular health, and lifetime cognitive engagement all shifted the point at which pathology became symptomatic. Some brains tolerate a heavy load indefinitely.
A second misconception is that "amyloid positive" means Alzheimer's disease. Under the 2018 and 2024 frameworks, amyloid alone without tau is Alzheimer's pathologic change, a precursor state. Most cognitively normal amyloid-positive people fall into exactly that category, which is one reason their near-term risk is measured in single-digit percentages per year.
A third issue is diagnostic mimicry. Limbic-predominant age-related TDP-43 encephalopathy, or LATE, is found in a large share of autopsied brains over age 80 and produces a memory syndrome that closely resembles Alzheimer's. Many people diagnosed clinically with Alzheimer's dementia turn out to have LATE alone or in combination. This matters because LATE does not respond to anti-amyloid drugs.
Finally, removing amyloid is not the same as restoring cognition. In the A4 prevention trial reported by Sperling and colleagues in 2023, solanezumab cleared peripheral amyloid but did not slow cognitive decline in preclinical disease. In symptomatic disease, lecanemab slowed decline by about 0.45 points on an 18-point clinical scale over 18 months, roughly 27 percent. That is a real but modest effect. The honest summary is that amyloid is necessary for the diagnosis and insufficient as an explanation.
What to Actually Do
The practical question is what to do with a memory complaint, either your own or a relative's. Red flags that justify a formal evaluation rather than watchful waiting:
- Repeating the same question within minutes and not recalling the answer
- Missing medications, bills, or appointments in a way that is new for the person
- Getting disoriented in familiar places or on familiar routes
- Word-finding pauses that now interfere with conversation, not just naming
- Personality or judgment changes noticed by others before the person notices them
- A rapid decline over weeks rather than months, which suggests another cause
For evaluation, start with a structured cognitive screen such as the MoCA, a 30-point test where scores below 26 warrant further testing. Add blood work for vitamin B12, thyroid function, and metabolic causes, plus a medication review and an MRI. Treat sleep apnea and hearing loss, because both are common, treatable, and independently associated with cognitive decline; population models attribute roughly 7 percent of dementia cases to untreated hearing loss alone. Formal neuropsychological testing takes two to four hours and separates true impairment from anxiety, depression, and low education.
Biomarker testing is where restraint matters. Current appropriate-use guidance limits amyloid testing to people with objective cognitive symptoms under specialist evaluation, or to participants in research. It does not support screening asymptomatic adults in primary care, because a positive result in a cognitively normal 65-year-old changes management very little and can complicate insurance, employment, and family decisions. If a blood p-tau217 test is offered, ask what will change based on the result.
Several limits on this evidence deserve stated plainly. The biological definition of Alzheimer's disease is a research construct that clinicians still debate, and labeling asymptomatic people with a disease name carries psychological and social costs that trials do not measure. Blood biomarkers are accurate in research cohorts but less validated in primary care populations with mixed pathology. Prevention trials such as the ongoing AHEAD 3-45 study are testing whether anti-amyloid treatment in preclinical disease delays symptoms, and until those results read out, the claim that early treatment prevents dementia remains a hypothesis rather than an established fact.