Dementia

Two things are true at once and easily confused. The number of people living with dementia is rising in every ageing country, because more people reach the ages at which it is common. The risk at any given age has fallen in several high-income cohorts studied over the last three decades — in the United States, England, the Netherlands and Sweden, by something like a tenth to a quarter per decade — plausibly from better education, blood pressure control and less smoking. Whether that fall continues, has stopped, or reverses with obesity and diabetes is the question, and the evidence has stopped agreeing: the newest cohorts show a flattening in some countries and continued decline in others.

Earlier detection makes the count harder to read. Diagnosis rates rose sharply after 2010 in England when doctors were encouraged to record it, coding changes moved deaths into dementia in a single year, and blood tests for the proteins of Alzheimer's disease are now accurate enough to enter clinical use, which will find cases years earlier and raise recorded incidence without changing the underlying risk. The cohort studies that measure incidence directly are small and slow.

Treatment has arrived and its effect is modest. The first antibodies that clear amyloid slow decline by a few months over eighteen months in early disease, at a cost and with a risk of side effects that has led some health systems to license them and decline to fund them. Prevention rests on the risk factors: the latest estimates attribute a substantial minority of cases to modifiable exposures, hearing loss and hypertension prominent among them, and the first trials of treating those exposures have shown effects on cognition, not yet on dementia.

The deaths say dementia is now the leading recorded cause in England and Wales and rising in most countries' statistics, but a large part of that rise is certification: doctors record it more, coding rules changed, and people die with it more than of it. So the prevalence trend is certain, the incidence trend is contested, and the two are routinely reported as though they were the same thing.

Status

contested

Direction

mixed

Would change our view

A new wave of any of the long-running cohorts showing incidence clearly falling or clearly rising; a registry-based incidence series large enough to settle it in one country; or population-level outcome data on the anti-amyloid drugs.

Trackers

Gaps

Incidence data outside North America and Northern Europe; anything comparable from Japan or Korea; the effect of diagnosis-rate campaigns and biomarker testing separated from true change.