How many steps a day? What the newest meta-analysis says
Step counters made a round number famous. A 2025 meta-analysis asks a plainer question: how the risk in the pooled studies changes as the step count rises. This post does not retell where any round target came from.
Short answer
- Ding and colleagues pooled 57 studies in The Lancet Public Health (202500164-1/fulltext)).
- About 7,000 steps a day was linked to a 47% lower risk of death than 2,000 steps.
- For several outcomes the curve flattened around 5,000 to 7,000 steps.
- The studies are observational, with few studies per outcome, and no split by age.
What was pooled
The review looked at studies that counted steps with a device, not a guess. Thirty-one of the 57 studies went into the meta-analyses. For all-cause mortality the comparison is explicit: 7,000 versus 2,000 steps a day, hazard ratio 0.53, which is about a 47% lower risk in the model. That is an association across cohorts. It is not an experiment that assigned people a step goal.
Where the curve flattens
For all-cause mortality, cardiovascular disease incidence, dementia and falls, the dose-response was a curve that bent, with inflection points around 5,000 to 7,000 steps. The sketch on this page is a shape, not a copy of the paper's figure. Risk in the model kept edging down a little at higher counts for some outcomes, and the authors say the evidence is thinner there. They also note few studies for most outcomes and no age-specific analysis.
An earlier meta-analysis by Paluch and colleagues, in the same journal (2022), had already described a curve that leveled off as steps rose. The 2025 paper is a wider update, not the first time the line bent.
Interval walking, a different question about how a walk is structured, is in [the Japanese interval walking study](../japanese-interval-walking-study/).
Limits
Device cohorts are not the whole public. People who walk more often differ in health, work and neighbourhood. Reverse causation is plausible: illness can cut steps before anyone counts them. A lower risk in a model is not a guarantee for one person who adds a loop around the block.
This post is general information, not medical or dietary advice.
Sources
- Ding D. et al. (2025), daily steps and health outcomes, systematic review and dose-response meta-analysis, The Lancet Public Health. thelancet.com00164-1/fulltext) · PubMed
- Paluch A. E. et al. (2022), daily steps and all-cause mortality, The Lancet Public Health. PubMed
Video transcript
How many steps a day? What does the data say?
A 2025 meta-analysis combined fifty-seven studies on daily steps and health.
Around seven thousand steps a day was linked to forty-seven percent lower risk of death than two thousand. For several outcomes, the benefit started to flatten around five to seven thousand.
Observational, with few studies per outcome, and no split by age.
PastePlate-AI on your phone
The logbook stays on your device. You bring the chat AI you already use.
This blog does not give medical advice. Health disclaimer.