Exercise, cognition and dementia risk

A 68-year-old retired teacher comes to the appointment worried. Her mother had Alzheimer's disease, and she read that walking "protects the brain". She wants to know how much she needs to walk to avoid the same fate. She does little physical activity, has controlled hypertension and sleeps poorly.

The question is honest and deserves an honest answer. Physical activity appears among the 14 modifiable risk factors in the Lancet commission on dementia, which together account for almost half of cases that in theory could be prevented (Livingston, 2024). At the same time, randomized trials of exercise on cognition show modest or null results (Sink, 2015).

This text explains what cohorts and trials show, why reverse causality confuses this field so much and how to counsel the patient without promising what the evidence does not yet support.

Key points

• The 2024 Lancet commission lists physical inactivity among 14 modifiable risk factors for dementia (Livingston, 2024).
• In a meta-analysis of 58 studies, physical activity was associated with a lower risk of dementia, with a pooled relative risk of 0.80 (Iso-Markku, 2022).
• When activity was measured 10 years or more before diagnosis, the association disappeared in a meta-analysis of 404,840 people (Kivimäki, 2019).
• In the Whitehall II cohort, physical activity began to decline up to 9 years before the diagnosis of dementia (Sabia, 2017).
• Trials in older adults without cognitive impairment and in people with mild to moderate dementia showed no clear cognitive gain (Sink, 2015).
• In people with dementia, exercise may improve activities of daily living, with very low quality evidence (Forbes, 2015).

Summary

• Why the question matters
• What the cohorts show
• The problem of reverse causality
• What the randomized trials show
• Exercise in people who already have dementia
• How to read the evidence as a whole
• How to counsel the patient in practice
• How NEXO helps train this reasoning
• Frequently asked questions
• Take-home messages

Why the question matters

Dementia is one of the major causes of disability in old age, and the number of people affected grows as the population ages. The Lancet commission estimates 57 million people with dementia worldwide in 2019 and projects 153 million in 2050 (Livingston, 2024).

Modifiable risk factors

The 2024 commission added vision loss and high cholesterol to the 12 factors in the 2020 report (Livingston, 2024). The list includes lower education, hearing loss, hypertension, smoking, obesity, depression, physical inactivity, diabetes, high alcohol consumption, traumatic brain injury, air pollution and social isolation (Livingston, 2024).

According to the commission, the 14 factors together account for almost half of dementias, which in theory could be prevented if these risks were eliminated (Livingston, 2024). The word "theory" weighs heavily here: the calculation assumes that each factor is a cause and that removing it has a full effect.

One factor among several

Physical inactivity is one factor among many, and it intertwines with others. It increases the risk of diabetes, coronary disease and stroke, which in turn increase the risk of dementia (Kivimäki, 2019).

For this reason, exercise also plays a part in prevention through indirect pathways, such as blood pressure control. The effect of exercise on blood pressure is covered in detail in the text on how much exercise lowers blood pressure.

What the cohorts show

Most of the evidence in favor of exercise comes from observational studies, which follow active and inactive people over time.

The 2022 meta-analysis

The Iso-Markku meta-analysis included 58 studies with follow-up of at least one year (Iso-Markku, 2022). Physical activity was associated with a lower risk of all-cause dementia, with a pooled relative risk of 0.80, in an analysis of 257,983 people (Iso-Markku, 2022).

The association also appeared for Alzheimer's disease, with a relative risk of 0.86, and for vascular dementia, with 0.79 (Iso-Markku, 2022). The authors observed a graded relationship between the amount of activity and the risk of all-cause dementia (Iso-Markku, 2022).

Long follow-ups

A strength of this meta-analysis was separating the studies with follow-up of 20 years or more (Iso-Markku, 2022). The association with lower risk of all-cause dementia and of Alzheimer's persisted even in these long follow-ups (Iso-Markku, 2022).

The authors concluded that physical activity is a modifiable protective factor, even after reducing the effects of reverse causality (Iso-Markku, 2022). The Lancet commission cited this result and added that the greatest risk reduction occurred in the move from extreme sedentariness to some activity (Livingston, 2024).

Maintaining and increasing activity

The commission describes a cohort of 29,826 people followed for a median of 24.5 years, with two measures of activity taken 10 years apart (Livingston, 2024). Those who maintained an adequate level of activity had lower risk of dementia than those who remained inactive, with a hazard ratio of 0.75 (Livingston, 2024). Those who increased activity up to that level also had lower risk, with a hazard ratio of 0.83 (Livingston, 2024).

The problem of reverse causality

Cohorts have a known trap. Dementia begins in the brain many years before diagnosis, and this preclinical phase can already reduce physical activity.

When the disease reduces movement

If a person moves less because the disease has already begun, inactivity becomes a consequence, not a cause. Studies with short follow-up, generally less than 10 years, have difficulty separating the two effects (Sabia, 2017).

The Lancet commission acknowledges this point and describes the relationship between exercise and dementia as probably bidirectional (Livingston, 2024).

What the Whitehall II cohort showed

The Whitehall II cohort followed 10,308 British civil servants who were between 35 and 55 years old at the start, with seven measures of physical activity between 1985 and 2013 (Sabia, 2017). Over a mean follow-up of 27 years, 329 participants were diagnosed with dementia (Sabia, 2017).

Physical activity was not associated with cognitive decline or with risk of dementia, with a hazard ratio of 1.00 in the group that met the recommendation (Sabia, 2017). Among those who developed dementia, physical activity began to fall up to 9 years before diagnosis (Sabia, 2017).

The meta-analysis with individual data

Kivimäki gathered individual data from 19 cohorts, with 404,840 people initially free of dementia (Kivimäki, 2019). When activity was measured less than 10 years before diagnosis, inactivity was associated with higher risk of dementia, with a hazard ratio of 1.40 (Kivimäki, 2019).

When the measure was taken 10 years or more before the onset of dementia, the difference between active and inactive people disappeared, with a hazard ratio of 1.01 (Kivimäki, 2019). The authors saw a sign of higher risk only in inactive people who developed cardiometabolic disease (Kivimäki, 2019).

Two readings of the same problem

The two meta-analyses reached different conclusions. Iso-Markku found an association in follow-ups of 20 years or more (Iso-Markku, 2022); Kivimäki and the Whitehall II cohort, with another design, did not (Kivimäki, 2019). The difference is a reminder that observational evidence calls for caution in interpretation.

This reasoning about sequence in time and cause appears in other fields. The text on how sequence in time misleads discusses the same error in another context.

What randomized trials show

Randomized trials solve the problem of reverse causality, but they have limits of their own: short follow-up, subtle cognitive outcomes and participants healthier than average.

Older adults without cognitive impairment

The Cochrane review by Young on aerobic exercise in older adults without known cognitive impairment included 12 trials with 754 participants (Young, 2015). The authors concluded that the available data showed no cognitive benefit (Young, 2015).

The most curious finding was that, even in trials where cardiorespiratory fitness improved, this did not coincide with improvement in any cognitive domain assessed (Young, 2015).

The LIFE study

The LIFE study randomized 1,635 sedentary older adults aged 70 to 89 to a moderate-intensity physical activity program or to health education (Sink, 2015). The program included walking, resistance training and flexibility (Sink, 2015).

After 24 months, cognitive scores were the same in both groups (Sink, 2015). The authors concluded that the program did not improve global cognitive function or that of specific domains (Sink, 2015). There were signs of improvement in executive function in participants aged 80 or older and in those with worse physical performance, as a secondary analysis (Sink, 2015).

The hippocampus trial

A trial with 120 older adults compared aerobic walking three times a week with stretching and toning for one year (Erickson, 2011). The aerobic group had an increase in hippocampal volume, which the authors described as reversal of the age-related loss of 1 to 2 years (Erickson, 2011).

The result is biologically interesting, but it is an imaging outcome in a small trial, not a dementia outcome (Erickson, 2011).

What the commission says

The Lancet commission cites a trial with 945 participants with a mean age of 78, randomized to control, moderate continuous training or high-intensity interval training, twice a week, for 5 years (Livingston, 2024). With no significant difference in cognition between the groups, the commission considers the findings consistent with the small cognitive benefit seen in an umbrella review of trials (Livingston, 2024).

Proposed mechanisms

The Lancet commission summarizes the most discussed mechanisms. Exercise appeared useful for cognition at any age, possibly through changes in cerebral blood flow, lower hypertension and more nitric oxide, reflected in brain plasticity and less neuroinflammation (Livingston, 2024).

People who do moderate to vigorous exercise on more days of the week have relatively larger brain volumes than those who do little or no exercise (Livingston, 2024). The commission also cites mouse studies on irisin, a myokine released during exercise, as a possible neuroprotective factor (Livingston, 2024).

These mechanisms make the hypothesis plausible, but biological plausibility is different from clinical proof. An imaging outcome or a blood marker can improve without the risk of dementia changing, and that is why the clinical outcomes of trials weigh more.

Why the trials are difficult

Testing exercise against dementia requires thousands of people followed for many years, because the disease develops slowly. The available trials use intermediate outcomes, such as scores on cognitive tests, which change little in healthy people.

It is also difficult to maintain the difference between the groups. In the 5-year trial cited by the commission, the vast majority of the control group followed national physical activity recommendations, which brought the groups closer together (Livingston, 2024). When the control group also moves, the trial loses the ability to show an effect.

Outcomes that matter to the patient

For the patient and the family, what weighs most is maintaining independence. On this point, the Cochrane review in dementia showed a sign of gain in activities of daily living (Forbes, 2015). In one of the trials, the burden on caregivers who supervised the exercise at home was lower (Forbes, 2015).

Exercise in those who already have dementia

For those who already have dementia, the question changes. The target is no longer prevention and becomes maintaining function, autonomy and quality of life.

The DAPA trial

The DAPA trial randomized 494 people with mild to moderate dementia, with a mean age of 77, to usual care plus four months of supervised aerobic and strength exercise or to usual care alone (Lamb, 2018).

At 12 months, the ADAS-cog cognitive score worsened slightly more in the exercise group, with an adjusted difference of 1.4 points (Lamb, 2018). The conclusion was direct: the moderate- to high-intensity program did not slow cognitive decline (Lamb, 2018).

The Cochrane review in dementia

The Cochrane review by Forbes included 17 trials with 1,067 participants with dementia (Forbes, 2015). For cognition, the evidence of benefit was inconclusive, with a standardized mean difference of 0.43 and a confidence interval that crossed zero (Forbes, 2015).

For activities of daily living, exercise showed benefit in six trials with 289 participants, with a standardized mean difference of 0.68 (Forbes, 2015). The quality was very low and heterogeneity was high, which calls for caution (Forbes, 2015).

What to expect

In people who already have dementia, exercise makes sense for mobility, fall prevention and autonomy, not as a treatment for cognitive decline (Forbes, 2015). This distinction needs to be clear in the conversation with the family. The conversation about treatments that act on the disease is a different one, as the text on anti-amyloid antibodies and who can receive them shows.

How to read the evidence as a whole

The table summarizes the main studies and what each design allows us to conclude.

Study	Design and population	Main result
Iso-Markku, 2022	Meta-analysis of 58 observational studies (Iso-Markku, 2022)	Relative risk of 0.80 for all-cause dementia (Iso-Markku, 2022)
Kivimäki, 2019	Individual data from 19 cohorts, 404,840 people (Kivimäki, 2019)	Hazard ratio of 1.01 with activity measured 10 or more years earlier (Kivimäki, 2019)
Whitehall II, 2017	Cohort of 10,308 adults, mean follow-up of 27 years (Sabia, 2017)	Hazard ratio of 1.00; activity declines up to 9 years before diagnosis (Sabia, 2017)
Young, 2015	Cochrane review of 12 trials, 754 older adults without impairment (Young, 2015)	No cognitive benefit in the available data (Young, 2015)
LIFE, 2015	Trial with 1,635 sedentary older adults for 24 months (Sink, 2015)	Cognitive scores equal between groups (Sink, 2015)
DAPA, 2018	Trial with 494 people with mild to moderate dementia (Lamb, 2018)	Cognitive decline equal or slightly worse with exercise (Lamb, 2018)
Forbes, 2015	Cochrane review of 17 trials, 1,067 people with dementia (Forbes, 2015)	Possible gain in activities of daily living, with very low quality (Forbes, 2015)

Why the results diverge

Cohorts measure habits over decades, and trials test programs lasting months or a few years. A two-year program may be too short to change a process that sets in over decades.

In addition, trials with a cognitive outcome tend to recruit people who are more motivated and healthier, and control groups often also become more active. This reduces the difference between groups.

Benefits that do not depend on cognition

Even with uncertainty about dementia, the benefits of exercise for cardiovascular, metabolic, mental and functional health are well established (Bull, 2020). The World Health Organization guideline lists cognitive health among the outcomes favored by physical activity, alongside mortality, hypertension, diabetes and mental health (Bull, 2020).

The effect on depressive symptoms, another risk factor listed by the commission, is discussed in the text on what trials show for exercise in depression.

How to guide the patient in practice

The guidance starts from the World Health Organization guideline, which applies to general health and does not depend on proving an effect on dementia (Bull, 2020).

How much to move

Adults should do 150 to 300 minutes per week of moderate aerobic activity, or 75 to 150 minutes of vigorous activity, or an equivalent combination (Bull, 2020). The guideline stresses that some physical activity is better than none (Bull, 2020).

What to add for older adults

Older adults should include multicomponent activity, with emphasis on functional balance and strength, at moderate or greater intensity, on 3 or more days per week (Bull, 2020). The stated goal is to improve functional capacity and prevent falls (Bull, 2020).

Reduce sitting time

The guideline also recommends reducing sedentary behavior (Bull, 2020). Replacing sitting time with activity of any intensity brings benefit (Bull, 2020).

Managing the other factors

Since inactivity is one factor among 14, the guidance should include the others: blood pressure, hearing, vision, smoking, alcohol, diabetes, cholesterol, depression and social contact (Livingston, 2024). Exercise helps with several of them at once (Kivimäki, 2019).

Talking without promising

How it is presented matters. Saying that exercise "prevents Alzheimer's" creates an expectation that the evidence does not yet support and can cause guilt if the disease appears. Techniques that explore the patient's own motivation have evidence of their own and work better than fear.

It is also worth avoiding excessive investigation in someone who just wants to know whether they should walk. The concept of cascades of care reminds us that tests without a clear indication can do more harm than good.

When the complaint is memory

If the patient reports new forgetfulness, the reasoning changes: cognition must be assessed and treatable causes sought. In the hospitalized older adult, acute confusion has another name and another urgency, as the text on how to screen for delirium with 4AT and CAM shows.

Back to the case

The 68-year-old teacher can hear a clear answer. Staying active brings certain benefits for blood pressure, mood, sleep, balance and functional capacity (Bull, 2020). On dementia, the cohorts suggest an association, but reverse causality and the modest trials prevent promising prevention (Kivimäki, 2019). The practical goal is the one in the guideline: 150 to 300 minutes of moderate activity per week, with strength and balance on 3 or more days (Bull, 2020).

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by looking for the ICD-10 code.

The limits of using the game are described on the NEXO economy page. If you research clinical reasoning teaching, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and under creators, and institutions can learn about the proposal on the partnerships page. For the most frequent reasons for consultation, see the most common codes in primary care.

Frequently asked questions

Does exercise prevent dementia?

The cohorts show an association between physical activity and lower dementia risk, with a relative risk of 0.80 (Iso-Markku, 2022). The randomized trials, however, show a small or null cognitive effect, and prevention still lacks proof (Sink, 2015).

What is reverse causality in this topic?

It is the possibility that early-stage dementia reduces physical activity, and not the other way around. In the Whitehall II cohort, activity began to fall up to 9 years before diagnosis (Sabia, 2017).

How much exercise does the guideline recommend?

From 150 to 300 minutes per week of moderate activity, or from 75 to 150 minutes of vigorous activity (Bull, 2020). Older adults should add strength and balance on 3 or more days per week (Bull, 2020).

Does aerobic exercise improve memory in healthy older adults?

The Cochrane review of 12 trials concluded that the available data did not show a cognitive benefit (Young, 2015). A smaller trial saw an increase in hippocampal volume after a year of walking (Erickson, 2011).

Does exercise treat dementia that is already established?

The DAPA trial showed that moderate to high intensity exercise did not slow cognitive decline (Lamb, 2018). There may be a gain in activities of daily living, with very low quality evidence (Forbes, 2015).

Which type of exercise should be chosen?

The one the person can keep up, adding aerobic activity to strength and balance, as the World Health Organization guideline recommends (Bull, 2020).

Is inactivity a modifiable risk factor?

The Lancet commission includes it among 14 modifiable risk factors for dementia (Livingston, 2024). The strength of this causal relationship is still debated (Kivimäki, 2019).

Take-home messages

• Physical inactivity is among the 14 modifiable risk factors of the Lancet commission.
• The cohorts show an association between activity and lower dementia risk.
• Reverse causality explains part of this association, because activity declines years before the diagnosis.
• Randomized trials show a small or null cognitive effect.
• In people who already have dementia, exercise serves function and autonomy, not slowing cognitive decline.
• The benefits of exercise for general health justify the recommendation on their own.
• Guide by the WHO guideline, with strength and balance in older adults.
• Talk without promising dementia prevention.

Play today's case on NEXO.

References

1. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024. PMID 39096926. DOI 10.1016/S0140-6736(24)01296-0.
2. Iso-Markku P, Kujala UM, Knittle K, et al. Physical activity as a protective factor for dementia and Alzheimer's disease: systematic review, meta-analysis and quality assessment of cohort and case-control studies. Br J Sports Med. 2022. PMID 35301183. DOI 10.1136/bjsports-2021-104981.
3. Kivimäki M, Singh-Manoux A, Pentti J, et al. Physical inactivity, cardiometabolic disease, and risk of dementia: an individual-participant meta-analysis. BMJ. 2019. PMID 30995986. DOI 10.1136/bmj.l1495.
4. Sabia S, Dugravot A, Dartigues JF, et al. Physical activity, cognitive decline, and risk of dementia: 28 year follow-up of Whitehall II cohort study. BMJ. 2017. PMID 28642251. DOI 10.1136/bmj.j2709.
5. Young J, Angevaren M, Rusted J, Tabet N. Aerobic exercise to improve cognitive function in older people without known cognitive impairment. Cochrane Database Syst Rev. 2015. PMID 25900537. DOI 10.1002/14651858.CD005381.pub4.
6. Sink KM, Espeland MA, Castro CM, et al. Effect of a 24-Month Physical Activity Intervention vs Health Education on Cognitive Outcomes in Sedentary Older Adults: The LIFE Randomized Trial. JAMA. 2015. PMID 26305648. DOI 10.1001/jama.2015.9617.
7. Erickson KI, Voss MW, Prakash RS, et al. Exercise training increases size of hippocampus and improves memory. Proc Natl Acad Sci U S A. 2011. PMID 21282661. DOI 10.1073/pnas.1015950108.
8. Lamb SE, Sheehan B, Atherton N, et al. Dementia And Physical Activity (DAPA) trial of moderate to high intensity exercise training for people with dementia: randomised controlled trial. BMJ. 2018. PMID 29769247. DOI 10.1136/bmj.k1675.
9. Forbes D, Forbes SC, Blake CM, Thiessen EJ, Forbes S. Exercise programs for people with dementia. Cochrane Database Syst Rev. 2015. PMID 25874613. DOI 10.1002/14651858.CD006489.pub4.
10. Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020. PMID 33239350. DOI 10.1136/bjsports-2020-102955.