News · Fitness & Exercise
Walking tracked with less dementia, nearly as much as running
A Lancet Public Health study followed 107,036 US health professionals for over 30 years and found walking tracked with lower dementia risk almost as strongly as vigorous exercise.
Based on a peer-reviewed cohort study in The Lancet Public Health
- Researchers used the Nurses' Health Study and the Health Professionals Follow-Up Study, publishing in The Lancet Public Health.
- The dementia analysis covered 63,596 women and 43,440 men, followed from 1990 to 2023.
- Activity was reported every 2 to 4 years from 1986, so this measures a habit over decades rather than a snapshot.
- The most active quarter had 28% lower dementia risk than the least active (HR 0.72).
- Walking alone tracked with 24% lower risk; vigorous aerobic exercise with 11% lower.
- The same pattern held for self-reported cognitive decline.
- For measured cognition, vigorous exercise showed a benefit where walking did not.
- Observational and self-reported. A 4-year lag was used to limit reverse causation, but people who walk more differ in many ways.
Exercise advice for the brain usually implies effort: get your heart rate up, work hard enough to notice. A study in The Lancet Public Health tracked over a hundred thousand people for three decades and found walking doing most of the work.
Being in the highest versus the lowest tertile for walking was associated with lower risk for dementia, and by a wider margin than vigorous exercise managed.
Why the earlier evidence was weak
The question is old and the answers have been thin. Previous cohort studies of physical activity and cognitive health have often been limited by small sample sizes, short follow-up durations, absence of long-term assessments of physical activity, and potential reverse causation.
That last one is the difficult one. Dementia dampens activity years before anyone diagnoses it, so a single snapshot of exercise can capture an early symptom and mislabel it a cause.
Three decades of the same people
The study used two ongoing US prospective cohorts: the Nurses’ Health Study and the Health Professionals Follow-Up Study, covering registered nurses and male health professionals recruited in 1976 and 1986.
Activity was assessed approximately every 2-4 years using self-report questionnaires, beginning in 1986 for both cohorts - not once, but repeatedly across most of an adult lifetime.
The dementia analysis drew on 63,596 Nurses’ Health Study participants and 43,440 Health Professionals Follow-Up Study participants, followed to 2023. A four-year lag was applied specifically to blunt the reverse-causation problem.
Where walking landed
The headline comparison favored activity in general. Those in the highest quarter for total physical activity had a lower risk of dementia, about 28% lower.
Split by type, the ordering is the surprise. Walking came out around 24% lower; vigorous aerobic exercise around 11% lower. Both lowered risk. The gentler one lowered it more.
Self-reported cognitive decline followed the same shape, with walking again ahead of vigorous exercise.
The outcome where intensity won
Measured cognition told a different story, and the paper does not hide it.
Being in the highest quarter of total physical activity was associated with fewer years of age-related cognitive differences in global cognition and verbal memory. Similar decreases were observed for vigorous aerobic exercise but not for walking.
So intensity did something walking did not, on the outcome measured by a test rather than a diagnosis. Two outcomes, two answers, and the study reports both rather than picking the tidier one.
What self-reported activity cannot show
Nobody was assigned to walk. People who walk a great deal are different from people who do not, in health, income, mobility and much else that adjustment reaches only partly.
The exposure is also self-reported, and people are famously generous about their own exercise. Repeating the questionnaire every few years for thirty years helps; it does not make it a measurement.
The authors keep the claim general: higher levels of physical activity were associated with a lower risk of dementia and better cognitive performance.
For a reader, the practical content is that the version of exercise most people can actually sustain did not come second here.
People also ask
Does walking really work as well as harder exercise?
For the dementia endpoint in this study, walking looked slightly better, not worse. But that comparison deserves care: people who report a lot of walking and people who report a lot of vigorous exercise differ in age, health and what else they do. What the study supports is that walking is not a lesser substitute on this outcome, which is more useful to most people than a ranking.
Why did vigorous exercise win on cognitive tests but not on dementia?
The two outcomes measure different things. Dementia is a clinical diagnosis over decades; the cognitive tests are performance on a telephone assessment at one point. Vigorous exercise showed fewer years of age-related difference in global cognition and verbal memory, while walking did not. That could be a real difference in what intensity does, or an artifact of who does which.
What is a 4-year lag and why does it matter?
Early dementia makes people less active years before diagnosis, which would make inactivity look like a cause when it is an early symptom. Lagging the exposure by four years means activity is measured well before the outcome window, which weakens that explanation. It does not eliminate it, because the process can begin more than four years out.
How much walking are we talking about?
The study expressed activity in metabolic equivalent hours per week rather than in minutes, and compared people by tertile and quartile rather than naming a threshold. So it supports 'more walking tracked with less dementia' rather than a specific prescription. Existing physical activity guidance remains the practical reference point.
How much should self-reported activity be trusted?
It is the main weakness, and the study's own framing acknowledges that earlier work suffered from short follow-up and single measurements. Repeating the questionnaire every 2 to 4 years across three decades is a substantial improvement on one snapshot, but people still misjudge and over-report their own exercise.