Smoking, cognitive decline and dementia: what is actually being measured?

01

Three outcomes that should not share one label

Dementia results from different diseases that damage the brain and affect memory, thinking and the ability to manage daily activities. Cognitive decline means change in abilities such as memory or thinking, often measured by repeated tests. For example, the EURODEM study compared changes in cognitive test scores in older adults without dementia; its score outcome was not a new dementia diagnosis. The analysis also excluded people newly diagnosed with dementia during follow-up, so it cannot supply an estimate of new dementia diagnoses.

All-cause dementia groups together diagnoses from different causes. Alzheimer disease and vascular dementia are more specific outcomes, and mixed disease can occur. A result labelled “all dementia” cannot be described as an Alzheimer-only result.

[1][4]

02

What comparisons of smoking groups show

A 2015 review of prospective cohorts found higher rates of all-cause dementia, Alzheimer disease and vascular dementia among current smokers compared with never smokers. Each outcome was analysed separately. Former smokers are another group, with different histories and reasons for stopping.

An association in a cohort is not a test that smoking caused one person’s disease. Age, education, vascular health and other circumstances can differ between smoking groups; adjustment addresses only the factors measured well enough.

[3]

03

Why an older surviving group can look misleadingly favourable

Some people who smoke die before recruitment or before a dementia diagnosis. A study that only includes surviving participants can miss those people; death during follow-up also prevents a later dementia diagnosis. A weaker association at very old ages is therefore not proof that smoking protects the brain.

As an illustration of the selection problem, imagine comparing only people still attending a late-life assessment. Those absent because they died cannot contribute a dementia diagnosis, even though their overall health outcome was not favourable. This is a study-reading issue, not a forecast of any participant’s lifespan.

[2][3]

04

Strong guidance and low-certainty evidence can coexist

The WHO’s 2026 second edition validates a strong recommendation to offer tobacco-cessation interventions for cognitive decline and dementia risk reduction, with low certainty of evidence. Its updated review describes observational cessation studies and found no systematic reviews or meta-analyses of intervention studies.

The strength of a public-health recommendation is not the same as certainty about a dementia-specific effect for one person. Different cohort timelines cannot be turned into a date when an individual becomes protected, nor into a promise that existing dementia will reverse.

[2]

05

Keep memory concerns and stopping support as separate requests

Bring concerns about memory or daily functioning to a qualified local clinician rather than attributing them to smoking or age alone. A useful question is: “What needs assessment, and where can I obtain smoking support separately?” A family member’s observation is useful context, not a home diagnosis.

In England, NHS Better Health directs adults to local Stop Smoking Services; adults elsewhere can use their local health system. The service helps with stopping smoking and does not certify future dementia-free years.

[1][5]

06

What to keep in mind

• Cognitive test change and dementia diagnosis are different endpoints.
• Survival and deaths can distort comparisons in older groups.
• A strong recommendation does not guarantee an individual outcome.