Not necessarily. In a pooled analysis of 47,471 adults, the risk of death during follow-up stopped falling at about 6,000 to 8,000 steps a day for people aged 60 and over, and at about 8,000 to 10,000 for younger adults1. Ten thousand is not a threshold the data drew; it is a round number.
The data do show that more steps went with better outcomes, up to a point, and that the biggest difference came at the low end. This article walks through what was measured, what the curve looks like, and why it is an association rather than a cause.
Where 10,000 came from
The number predates the research. The authors of the largest analysis wrote that the 10,000-step goal "is not based on evidence, but instead originates from a marketing campaign in Japan"1. It became a default on step counters and stuck.
What the 2022 analysis measured
The Steps for Health Collaborative pooled 15 cohorts: groups of people measured once and then followed over the years1. Eight were in the US, four in Europe, one in Japan, one in Australia and one spanned 40 countries. Each team analysed its own participant data with the same protocol, and the results were combined in a meta-analysis (one analysis that pools many studies).
| What | Number |
|---|---|
| Adults | 47,471 |
| Average age | 65 years (SD 12.4) |
| Women | 68% |
| How steps were counted | A step-counting device worn for about one week, mostly at the hip |
| Median follow-up | 7.1 years |
| Deaths recorded | 3,013 |
Participants were split into quartiles, four equal groups from fewest to most steps. The analysis compared each group's risk of death during follow-up with the lowest group, after adjusting for age, sex, education or income, body-mass index, chronic disease, self-rated health, smoking, alcohol and other factors1. The comparison is a hazard ratio: 1.0 means no difference, and 0.60 means deaths occurred at about 60% of the rate of the reference group.
In plain terms: tens of thousands of adults wore a step counter for a week, and researchers checked who died over the next seven years or so.
The results
More steps went with a smaller risk, and the first steps mattered most. The four groups' median step counts were 3,553, 5,801, 7,842 and 10,901 a day1. Against the lowest group, the adjusted hazard ratios were 0.60 for the second group, 0.55 for the third and 0.47 for the highest.
Risk of death during follow-up by daily steps, vs the lowest group · hazard ratio
The jump from the lowest group to the second, about 2,250 more steps a day, carried most of the difference. Each group after that added less.
The curve levelled off, and where depended on age. Adults 60 and over showed no further difference beyond about 6,000 to 8,000 steps a day. Adults under 60 levelled off at about 8,000 to 10,0001. Men and women did not differ. And the analysis found no higher risk at very high step counts.
Pace was mixed. Before total steps were accounted for, faster stepping went with a smaller risk. After adjusting for total steps, two pace measures (the fastest 30 and 60 minutes of the day) stayed linked, with a hazard ratio of 0.67 each, while time spent above 40 or 100 steps a minute did not. The authors called the pace findings "inconclusive"1.
In plain terms: in these adults, going from very few steps to a moderate number went with the biggest difference, and the benefit tailed off well before 10,000 for older adults.
Other analyses point the same way
2023, 12 studies, 111,309 people. Compared with 2,000 steps a day, a significant difference in the risk of death appeared by about 2,500 steps, and the curve reached its lowest point at about 8,763 steps (hazard ratio 0.40)2. The association was stronger in studies that used hip-worn devices than pedometers or wrist-worn devices.
2023, 17 cohorts, 226,889 people. Each extra 1,000 steps a day went with a hazard ratio of 0.85 for death from any cause3. The curve was non-linear, with the risk still falling at higher step counts in this analysis.
2025, 57 studies. The newest review pooled 31 studies across several outcomes. For death from any cause, the curve bent at about 5,000 to 7,000 steps a day, and 7,000 steps went with a hazard ratio of 0.53 compared with 2,0004. For some other outcomes, including cardiovascular death, cancer and type 2 diabetes, the association ran in a straight line with no bend, so more steps kept going with lower risk. The authors concluded that 10,000 "can still be a viable target for those who are more active", while 7,000 "might be a more realistic and achievable target for some".
Where the step curve flattens, by analysis
- 202215 cohorts, 47,471 adultsLevelled off at about 6,000–8,000 steps (60 and over) and 8,000–10,000 (under 60)
- 202312 studies, 111,309 peopleSignificant from about 2,500 steps; lowest point about 8,763 steps
- 202317 cohorts, 226,889 peopleHazard ratio 0.85 per extra 1,000 steps; still falling at higher counts
- 202557 studies, 31 pooledCurve bent at about 5,000–7,000 steps; 7,000 vs 2,000 steps hazard ratio 0.53
The analyses disagree on the exact point where the curve flattens, from about 5,000 to about 10,000 steps depending on the data, the devices and the ages studied. They agree that the steepest part is at the low end.
What this does and does not mean
It is an association, not a cause. These are observational cohorts. The authors state that "causal inferences cannot be made"1. People with an illness tend to walk less and are also more likely to die, which can make steps look more protective than they are. When deaths in the first two years were excluded to test this, the link weakened but remained.
It is about death from any cause. The 2022 analysis did not look at fitness, strength or body composition. The 2025 review covered other outcomes too, with evidence certainty rated moderate for most and low or very low for some4.
Your device may count differently. Research and consumer devices can differ by 20% or more, and most of the 2022 data came from hip-worn devices1. A wrist tracker's 8,000 may not equal a hip device's 8,000.
Not a prescription. None of these studies tested telling people to walk more. They describe what step counts went with in large groups. For individual questions about activity and health, a doctor is the right person to ask.
Steps and energy
Steps also carry energy. Everyday walking is part of NEAT, the energy of non-exercise movement (see what is NEAT). Moving from a desk-bound day to an active one is also why picking an activity level matters for a nutrition target (see which activity level to pick). The step studies above did not measure energy; that is a separate question.
What the studies found
| Study | Who was studied | Key result | Year |
|---|---|---|---|
| Paluch et al., *Lancet Public Health*1 | 47,471 adults in 15 cohorts, median follow-up 7.1 years | Hazard ratio 0.47 in the highest step group vs lowest; levelled off at 6,000–8,000 (60+) and 8,000–10,000 (under 60) | 2022 |
| Stens et al., *J Am Coll Cardiol*2 | 111,309 people in 12 studies | Significant from about 2,500 steps; lowest point about 8,763 steps | 2023 |
| Banach et al., *Eur J Prev Cardiol*3 | 226,889 people in 17 cohorts | Hazard ratio 0.85 per extra 1,000 steps a day | 2023 |
| Ding et al., *Lancet Public Health*4 | 57 studies from 35 cohorts; 31 in meta-analyses | Curve bent at about 5,000–7,000 steps; 7,000 vs 2,000 steps hazard ratio 0.53 | 2025 |
Limits worth knowing
- Observational. Residual confounding and reverse causation are possible; none of these studies shows cause1.
- One week of steps. Steps were counted once, at the start, and habits may have changed over the follow-up.
- Who was studied. The 2022 sample averaged 65 years old, was 68% women and more than 70% White, from high-income countries.
- Devices differ. Counts can vary by 20% or more between devices1, and one 2023 analysis found stronger associations with hip-worn devices than with pedometers or wrist-worn ones2.
- Abstracts only for the 2023 analyses and the 2025 review. The 2022 analysis was read in full, apart from its online appendix.
- Funding. The 2022 analysis was funded by the US Centers for Disease Control and Prevention1.
Latest research
The newest analysis we opened is the 2025 review in *Lancet Public Health*4. It covered studies published up to February 2025, found the association with death from any cause bending at about 5,000 to 7,000 steps a day, and framed 7,000 as a more realistic target for some people while keeping 10,000 as viable for the more active. It also notes that most outcomes rested on few studies and lacked age-specific analysis.
The short version
- 10,000 steps came from marketing, not from research.
- In large cohorts, more steps went with a smaller risk of death, with the biggest difference at the low end.
- The association levelled off at about 6,000–8,000 steps for adults 60 and over and 8,000–10,000 for younger adults.
- These are associations in groups, not proof that steps cause the difference.
Educational overview only — not medical advice.
