Alzheimer's Disease and Dementia

TL;DR. Dementia is not one disease and it is not normal ageing. It is a syndrome of progressive loss of memory, thinking, language, and judgement severe enough to interfere with daily life, and it has several causes. Alzheimer's disease, the commonest, involves two abnormal proteins: amyloid-beta, which accumulates outside neurons in plaques, and tau, which tangles inside them. The disease begins in the brain twenty years before the first symptom. Nothing available today reverses it. But roughly 45 percent of dementia cases worldwide are associated with fourteen modifiable risk factors, which makes prevention the most promising part of this chapter, and the first drugs that alter the disease process rather than just the symptoms arrived in 2023.

Key takeaways

• About 57 million people live with dementia, with nearly 10 million new cases a year, and over 60 percent live in low- and middle-income countries.
• Alzheimer's accounts for 60 to 70 percent; the rest is vascular, Lewy body, frontotemporal, and mixed, and mixed pathology is the norm in the very old.
• The disease process starts around 20 years before symptoms, which is why treatment attempts at the symptomatic stage have mostly failed and why prevention gets a whole section here.
• Age is the dominant risk factor, roughly doubling risk every five years after 65, but dementia is not an inevitable consequence of ageing. Age-specific incidence has actually fallen in several high-income countries.
• The Lancet Commission (2024) lists 14 modifiable factors associated with about 45 percent of cases: hearing loss, high LDL cholesterol, low education, depression, traumatic brain injury, physical inactivity, diabetes, smoking, hypertension, obesity, excess alcohol, social isolation, air pollution, and untreated vision loss.
• Dementia is not contagious.

What it is

In short: Dementia is a syndrome with several causes, and separating them matters because they behave and respond differently.

Dementia (increasingly called major neurocognitive disorder) is acquired, progressive impairment in two or more cognitive domains, severe enough to interfere with independence. Mild cognitive impairment is measurable decline that has not yet crossed that line; roughly 10 to 15 percent of people with it progress to dementia per year, and some revert to normal.

Reversible or treatable mimics must be excluded and are missed often enough to matter: B12 and thyroid deficiency, depression (which can present as cognitive impairment), normal-pressure hydrocephalus, subdural haematoma, medication effects (anticholinergics, sedatives), alcohol, and delirium.

Don't be confused: delirium is not dementia. Delirium is acute (hours to days), fluctuating, usually with altered consciousness and attention, and it is caused by something else: infection, drugs, dehydration, pain, or surgery. It is a medical emergency and it is largely reversible. Dementia is chronic and progressive. The two frequently coexist, and an older person with sudden confusion should be assumed to have delirium and investigated, not written off as demented.

The history

In short: A single 1901 patient, then a century of dead ends, one integrity scandal, and finally two drugs that modestly slow the disease.

In 1901, Alois Alzheimer examined a 51-year-old woman named Auguste Deter at the Frankfurt asylum. She had memory loss, disorientation, jealousy toward her husband, and unpredictable behaviour. Asked to write her name, she said, "I have lost myself, so to speak." She died in 1906, and Alzheimer examined her brain, using new silver staining methods, and described the two hallmarks: dense plaques between cells and neurofibrillary tangles within them.

For decades the disease was thought to be a rare condition of the middle-aged, while the same changes in older people were called senility and considered normal ageing. The recognition in the 1970s that they were the same disease reframed dementia as a medical condition and created the field.

That sequence is worth reading carefully because it is a good example of how science actually proceeds: a dominant hypothesis, decades of failure, a serious integrity scandal in one strand of it, and then partial vindication with an effect size smaller than anyone hoped.

What actually goes wrong

In short: Amyloid accumulates outside neurons and tau tangles inside them, and it is tau, not amyloid, that tracks the symptoms.

Amyloid. A normal membrane protein, amyloid precursor protein (APP), is cut by enzymes. Cut one way it produces harmless fragments; cut by beta-secretase and then gamma-secretase it produces amyloid-beta 42, a sticky peptide that aggregates into oligomers and then into extracellular plaques. Clearance declines with age, so accumulation is a balance problem rather than purely an overproduction one.

The genetic evidence for amyloid's causal role is strong. Mutations in APP, and in the presenilin genes that form part of gamma-secretase, cause early-onset familial Alzheimer's with near-complete penetrance. People with Down syndrome, who carry three copies of chromosome 21 and therefore an extra copy of the APP gene, develop Alzheimer pathology almost universally by their forties, and a majority develop dementia. A rare Icelandic APP variant that reduces amyloid production protects against Alzheimer's and against cognitive decline in general.

Tau. Inside neurons, tau normally stabilises the microtubule tracks along which cargo moves. In Alzheimer's it becomes hyperphosphorylated, detaches, and aggregates into neurofibrillary tangles. Transport fails and the neuron dies. Critically, tau spread correlates with symptoms far better than amyloid does: the amount and distribution of tangles matches the clinical picture, while plaque load does not.

The spread pattern (Braak staging) is stereotyped: tau pathology begins in the entorhinal cortex, moves to the hippocampus (hence memory as the first casualty), then out into the association cortex, then everywhere. The pathology appears to propagate along connected circuits, which has led to the idea that misfolded tau templates the misfolding of adjacent normal tau, prion-like in mechanism though not in transmissibility.

Everything else. Synapse loss is the change that correlates best with cognitive decline. Neuroinflammation, driven by microglia, is an active contributor rather than a bystander, and several risk genes (TREM2, CD33) are microglial. Loss of acetylcholine-producing neurons in the basal forebrain underpins the only symptomatic drugs available. And vascular damage contributes in most patients, which is why blood pressure control matters here.

APOE. The apolipoprotein E gene comes in three common versions. APOE4 raises risk roughly 2 to 3 fold with one copy and around 8 to 12 fold with two copies, and lowers the age of onset. APOE2 is protective. APOE4 is common (roughly 15 to 25 percent of people carry at least one copy) and is a risk factor, not a diagnosis: many carriers never develop dementia and many people with Alzheimer's carry no copy.

What it does to the body

In short: Memory first, then language and orientation, then independence, and finally the ability to swallow safely, which is usually what kills.

Early: difficulty forming new memories (repeating questions, losing track of recent conversations, misplacing objects), word-finding trouble, disorientation in unfamiliar places, withdrawal from complex tasks. Insight is often partly preserved, and the awareness of slipping is itself distressing.

Middle: established memory loss extending backward in time, disorientation to time and place, difficulty with familiar tasks, wandering, sleep disturbance and evening agitation ("sundowning"), behavioural and psychological symptoms including suspicion, agitation, apathy, and sometimes aggression. Independence in daily activities is lost progressively: finances, then medication, then cooking, then washing and dressing.

Late: minimal speech, immobility, incontinence, failure to recognise close family, and loss of the ability to swallow safely, which leads to aspiration pneumonia, the most common immediate cause of death.

Vascular dementia typically adds physical signs (weakness, gait disturbance). Lewy body dementia brings falls, hallucinations, and severe reactions to antipsychotics. Frontotemporal dementia presents in a way families often interpret as a personality change or a psychiatric illness, sometimes for years before diagnosis, and it strikes at younger ages, which compounds the social damage.

Is it deadly?

Yes. Dementia is a terminal illness, and describing it that way changes care for the better, because it prompts advance planning and comfort-focused decisions rather than repeated aggressive interventions.

• Median survival after diagnosis of Alzheimer's is roughly 4 to 8 years, with wide variation; younger and fitter patients can live 15 to 20 years.
• Dementia is among the leading causes of death globally and the leading cause of death in women in several high-income countries.
• Death is usually from pneumonia, from complications of immobility, or from failure to eat and drink.
• Dementia is the leading cause of dependency in older people, and the cost of care, mostly borne by families, exceeds a trillion US dollars a year globally.

Is it contagious?

No. You cannot catch dementia. Living with, caring for, or being related to someone with Alzheimer's does not transmit it.

Two adjacent facts get misreported into a scare and deserve accurate statement. Prion diseases (Creutzfeldt-Jakob disease and variant CJD) are genuinely transmissible through contaminated neurosurgical instruments, cadaveric tissue, and, for variant CJD, BSE- contaminated beef. They are rare and clinically distinct, and they are a separate category from Alzheimer's. Separately, a 2024 report described early-onset Alzheimer-type disease in a handful of people who had received cadaveric human growth hormone decades earlier, batches of which were contaminated with amyloid-beta, suggesting amyloid seeding can be transmitted iatrogenically. This is a historical medical exposure that stopped in 1985. It says something interesting about amyloid biology and nothing about ordinary contact.

Who gets it

In short: Age dominates, two-thirds are women, and fourteen modifiable factors are associated with about 45 percent of cases.

Age. The strongest risk factor, roughly doubling every five years after 65. But age-specific incidence has fallen by roughly 13 percent per decade in several high-income countries over the last thirty years, most plausibly because of better education, cardiovascular treatment, and less smoking. The absolute number of cases still rises because populations are ageing.

Sex. About two-thirds of people with Alzheimer's are women, only partly explained by longer life expectancy; hormonal and genetic factors, including a stronger APOE4 effect in women, are under investigation.

Genetics. Early-onset familial Alzheimer's (APP, PSEN1, PSEN2 mutations) accounts for under 1 percent of cases and is autosomal dominant. Late-onset disease is polygenic with APOE dominating.

The 14 modifiable factors, from the Lancet Commission's 2024 update, grouped by life stage:

Together these are associated with about 45 percent of dementia cases worldwide. That is a population-attributable estimate from observational data, not a guarantee that addressing them prevents that share, and the causal evidence is stronger for some (hypertension, smoking, hearing) than others. It is still the most actionable list in this chapter.

Hearing loss deserves emphasis because it is the largest single midlife factor and the most fixable. The ACHIEVE randomised trial found that hearing aids slowed cognitive decline substantially in older adults at higher risk, though not in the lower-risk group, which is consistent with hearing loss contributing through reduced cognitive stimulation and social isolation.

Treatment, and how it works

In short: Two old drug classes that help symptoms modestly, and two new antibodies that slow decline by roughly a quarter at the cost of brain swelling in a minority.

Symptomatic drugs

Cholinesterase inhibitors (donepezil, rivastigmine, galantamine) block the enzyme that breaks down acetylcholine, raising levels of a neurotransmitter depleted by the loss of basal forebrain neurons. They produce modest improvements in cognition and function, on the order of delaying decline by several months, and they do not alter the disease course.

Memantine blocks NMDA glutamate receptors, reducing excitotoxic damage from chronic glutamate signalling. It is used in moderate to severe disease, often alongside a cholinesterase inhibitor, with modest benefit.

Disease-modifying drugs

Lecanemab and donanemab are monoclonal antibodies that bind aggregated amyloid, mark it for removal by microglia, and clear plaques from the brain, which is visible on PET imaging. In phase 3 trials in early Alzheimer's disease with confirmed amyloid, both slowed clinical decline by roughly 25 to 35 percent over 18 months relative to placebo.

Reading that honestly requires two statements at once. It is the first time any drug has altered the trajectory of the disease, which validates decades of work and confirms amyloid is on the causal path. And the effect is small: it slows decline rather than stopping or reversing it, the difference is at or near the threshold of what a family would notice, and it requires fortnightly or monthly infusions, amyloid confirmation by PET or spinal fluid, APOE genotyping, and regular MRI monitoring.

ARIA (amyloid-related imaging abnormalities) is the specific risk: brain swelling (ARIA-E) or microbleeds (ARIA-H), occurring in a substantial minority of treated patients, usually asymptomatic and detected on surveillance MRI, occasionally serious and rarely fatal. Risk is much higher in APOE4 homozygotes, which is why genotyping is done before treatment. People on anticoagulants are generally excluded.

Managing symptoms and behaviour

Non-drug approaches come first for agitation, aggression, and distress: identify the trigger (pain, infection, constipation, need for the toilet, overstimulation, fear), structure the environment, keep routines, use music and reminiscence, and support the caregiver.

Antipsychotics are frequently used and carry a clear warning: they increase mortality and stroke risk in older people with dementia, and in Lewy body dementia they can cause severe, sometimes life-threatening reactions. They have a legitimate but narrow role in severe distress or danger, at low dose, for a short time, with review.

Depression, pain, constipation, poor sleep, and sensory impairment are all common, all treatable, and all frequently missed as causes of "behavioural" symptoms.

What treatment costs

• Cholinesterase inhibitors: nausea, diarrhoea, loss of appetite and weight, vivid dreams, slow heart rate and falls.
• Memantine: dizziness, headache, confusion, generally well tolerated.
• Anti-amyloid antibodies: ARIA as described, infusion reactions, brain volume loss of uncertain significance, high cost, and a substantial monitoring burden. Several health systems have declined to fund them on cost-effectiveness grounds, which is a defensible reading of the same data.
• Antipsychotics: increased mortality, stroke, falls, sedation, parkinsonism.

What the person can do

In short: Hearing aids, blood pressure, exercise, and social engagement, and the prevention list is almost identical to the cardiovascular one.

Prevention is where the largest gains are, and the actions overlap almost entirely with cardiovascular health, which is not a coincidence.

• Treat hearing loss. Get hearing tested and use aids if needed.
• Control blood pressure from midlife. The SPRINT MIND trial found intensive blood pressure lowering reduced mild cognitive impairment.
• Treat diabetes, high cholesterol, and obesity; stop smoking; keep alcohol low.
• Stay physically active. Exercise has the most consistent observational evidence of any behaviour, and plausible mechanisms including vascular health and neurotrophic signalling.
• Stay cognitively and socially engaged. Education and cognitively demanding work build cognitive reserve, the capacity to sustain pathology while still functioning. Social isolation is an independent risk factor, and loneliness in older adults is a health problem in its own right.
• Protect your head: helmets, fall prevention, and taking repeated head impacts in sport seriously.
• Treat depression and poor sleep.
• Correct vision. Cataract surgery is associated with lower dementia incidence in cohort studies.

The FINGER trial in Finland randomised at-risk older adults to a multi-component programme (diet, exercise, cognitive training, vascular risk monitoring) and found a small but significant benefit on cognition, which is the best randomised evidence that multi-domain prevention does something. Larger, longer trials are running.

Single supplements, brain-training apps sold as prevention, and coconut oil have no convincing evidence.

Living with it

In short: The person most affected long term is often the unpaid family carer, and supporting them is a treatment rather than a courtesy.

Dementia is unusual in this book because the person most affected in the long run is often not the patient. Family caregivers, mostly women, provide the majority of care worldwide, at substantial cost to their own physical health, mental health, employment, and finances. Caregiver support is a treatment: structured education and support programmes measurably delay institutionalisation and improve caregiver wellbeing.

Practical matters that should be addressed early, while capacity permits: legal and financial powers of attorney, advance care planning including decisions about hospitalisation and feeding tubes, driving (which usually has to stop, and which is one of the hardest conversations), and safety at home.

Feeding tubes in advanced dementia deserve a specific mention because they are commonly requested and the evidence does not support them: they do not prolong life, prevent aspiration, or improve comfort in advanced dementia, and careful hand feeding is preferred.

And a point about language: people with dementia retain emotional memory and awareness of tone long after they lose factual memory. Talking about someone as though they are absent while they are in the room is both unkind and noticed.

What's next

• Blood biomarkers. Plasma p-tau217 tests now detect Alzheimer pathology with accuracy approaching PET and spinal fluid analysis. This changes diagnosis from an expensive specialist procedure into a blood test, and it is arguably the most consequential recent development in the field, since it makes both earlier treatment and large prevention trials practical.
• Tau-targeting drugs, on the reasoning that tau tracks symptoms better than amyloid.
• Prevention trials in asymptomatic people with amyloid on scans, testing whether treating 20 years earlier does what treating late cannot.
• Metabolic and inflammatory approaches, including GLP-1 receptor agonists, currently in large trials.
• Combination therapy, on the model of HIV and cancer, rather than single agents.

Sources and notes

Prevalence figures: WHO dementia fact sheet (57 million people with dementia, nearly 10 million new cases annually, over 60 percent in low- and middle-income countries). Modifiable risk factors and the 45 percent figure: Livingston et al., Lancet Commission on dementia prevention, intervention, and care, 2024 update. Auguste Deter's case: Alzheimer's 1906 presentation and 1907 paper; her records were rediscovered in 1996. APOE4: Corder et al., Science, 1993; effect sizes vary by ancestry and sex. Lecanemab: CLARITY-AD, NEJM, 2023. Donanemab: TRAILBLAZER-ALZ 2, JAMA, 2023. ARIA rates from those trials. ACHIEVE hearing trial: Lin et al., The Lancet, 2023. FINGER: Ngandu et al., The Lancet, 2015. SPRINT MIND: JAMA, 2019. Declining age-specific incidence: Wolters et al., Neurology, 2020, pooled cohorts. Image manipulation investigation: Piller, Science, 2022. Iatrogenic amyloid-beta transmission: Banerjee et al., Nature Medicine, 2024. Feeding tubes in advanced dementia: Cochrane review and American Geriatrics Society position statement.

Open questions. Whether removing amyloid early enough prevents dementia is the field's central unanswered question. Why women are disproportionately affected is unresolved. The mechanism connecting hearing loss to dementia is not established.

Next: the other great neurodegenerative disease, and the one where a single missing chemical explains most of the symptoms. 👉