Tip · Movement

6,000 to 10,000 steps a day

Q: Do I really have to walk 10,000 steps a day?

Not for mortality. In the largest pooled analysis to date, of 15 cohorts with 47,471 adults, the benefit leveled off between 6,000 and 8,000 steps in people aged 60 and over and between 8,000 and 10,000 in younger people. A cohort of 16,741 older women found the plateau at around 7,500. Only for the endpoint dementia was the optimum, at 9,826 steps in a UK Biobank analysis, close to the familiar number.

Q: Where does the 10,000 come from in the first place?

From none of the large studies. The authors of the 2022 meta-analysis write explicitly that little evidence exists for this recommendation, and the authors of the 2019 women’s cohort that the goal has a limited scientific basis. The widespread story of a Japanese pedometer advertisement from the 1960s could not be verified in the scientific databases.

Q: What do I gain if I go from 3,000 to 6,000 steps?

In the pooled meta-analysis, exactly that was the largest single jump. From the lowest quartile with a median of 3,553 steps to the second with 5,801, the hazard ratio fell to 0.60. The further step from 7,842 to 10,901 steps brought only a shift from 0.55 to 0.47. So the benefit lies overwhelmingly at the lower end.

Q: Does brisk walking count more than slow walking?

Hardly, for mortality. In both large analyses, step intensity largely lost its significance once adjusted for total step count. For dementia it is different: there, a peak cadence of 112 steps per minute over 30 minutes had the strongest single association, with a hazard ratio of 0.38.

Q: Does a standing desk do the same as walking?

No. In a meta-analysis of seven crossover studies, standing as an interruption of prolonged sitting did lower the post-meal blood glucose response significantly, but only about half as much as light walking, and standing had no significant effect on insulin and blood pressure. In direct comparison, walking was superior to standing. Standing is an intermediate step, not a substitute.

Q: Are these numbers proof that walking prolongs life?

No. All three large analyses are observational studies. They show an association between step count and later events, not a measured effect. People who walk little are often less healthy to begin with, and this can be statistically mitigated but not ruled out. In addition, step count was usually measured only once, over seven days.

The association with mortality is clear in large cohorts – but most of the benefit is already reached at 4,000 to 8,000 steps.

Step count is not an active factor in its own right but the most easily measured proxy for overall everyday movement — and that is exactly why unusually large datasets exist on it. This page covers the daily amount: what the cohorts show, where the benefit levels off and where the round number comes from.

In short

The association is clear: in the pooled analysis of 15 cohorts with 47,471 adults, mortality in the second quartile was 40 % lower than in the lowest (hazard ratio 0.60), 0.55 in the third and 0.47 in the fourth. But the benefit levels off well before 10,000 steps — from age 60 between 6,000 and 8,000 steps per day, under age 60 between 8,000 and 10,000. In older women the plateau was at around 7,500. Only for the endpoint dementia is the optimum, at 9,826 steps, close to the familiar number. The catch: all of these analyses are observational studies, not interventions.

What is behind it

In the short term, everyday movement acts through glucose uptake by the muscles. Even light walking as an interruption of prolonged sitting lowers post-meal glucose (Δ = −0.72, 95 % CI −1.03 to −0.41) and insulin (Δ = −0.83, −1.18 to −0.48) compared with uninterrupted sitting.

Over years, this becomes an association with mortality that is consistent across several large accelerometer cohorts and reaches its plateau between 6,000 and 10,000 steps depending on age. What does not hold up is noteworthy: after adjustment for daily step count, the association with walking speed largely disappears. It is the amount, not the pace.

Where the benefit really lies

The largest single jump is at the lower end. From the lowest quartile with a median of 3,553 steps to the second with 5,801, the hazard ratio fell to 0.60. The further step from 7,842 to 10,901 steps brought only a shift from 0.55 to 0.47. Those who walk little gain the most, and they gain it early.

For the endpoint dementia, the curve looks different. In a UK Biobank analysis with 78,430 adults and 866 dementia cases, the optimal dose was 9,826 steps (hazard ratio 0.49, 95 % CI 0.39–0.62), and the minimal dose with half the maximum risk reduction was 3,826 steps (0.75, 0.67–0.83). This is the only robust finding that supports the round number.

Pace, standing, distribution

On pace, the endpoints diverge. For mortality, the intensity measures lost almost all of their explanatory power after adjustment for step count — peak 1-minute cadence had a hazard ratio of 0.87 (95 % CI 0.68–1.11), peak 30-minute cadence 0.86 (0.65–1.13). For dementia, by contrast, a peak 30-minute cadence of 112 steps per minute was the strongest single association, at 0.38 (0.24–0.60).

A standing desk is an intermediate step, not a substitute. Standing did lower post-meal glucose significantly (Δ = −0.31, 95 % CI −0.60 to −0.03), but less than half as much as walking, and it had no significant effect on either insulin or systolic blood pressure. In direct comparison, walking was superior to standing (Δ = −0.30, −0.52 to −0.08; p < 0.009).

What is well supported

The most robust evidence is the pooled participant-level meta-analysis: 15 cohorts, 47,471 adults, 3,013 deaths, 297,837 person-years. All study teams processed their individual data according to a common protocol, which is considerably stronger than a meta-analysis of published estimates. The median step counts of the quartiles were 3,553, 5,801, 7,842 and 10,901, and the adjusted hazard ratios compared with the lowest quartile were 0.60 (95 % CI 0.51–0.71), 0.55 (0.49–0.62) and 0.47 (0.39–0.57). The second large cohort, with 16,741 older women analyzed, shows the same pattern: hazard ratios of 0.59, 0.54 and 0.42 across the quartiles.

What the studies show

Paluch 2022: the largest pooled analysis

Participant-level meta-analysis across 15 studies, with study start between 1999 and 2018. 47,471 adults, 3,013 deaths, corresponding to 10.1 per 1,000 participant-years, median follow-up 7.1 years. The primary endpoint was all-cause mortality from death certificates and national registries. Restricted cubic splines showed a progressively declining risk in adults aged 60 and over up to 6,000 to 8,000 steps, and in younger adults up to 8,000 to 10,000 steps per day.

Lee 2019: the cohort that first showed the plateau

18,289 participants in the Women’s Health Study wore an accelerometer for seven days between 2011 and 2015; 16,741 met the wear criteria. Mean age 72.0 years (SD 5.7), mean step count 5,499 per day, 504 deaths over an average of 4.3 years. The hazard ratios fell across the quartiles to 0.42 (95 % CI 0.30–0.60), and in the spline analysis up to about 7,500 steps; after that the curve stayed flat.

Buffey 2022: the experimental cross-check

Meta-analysis of seven acute, one-day randomized crossover studies in predominantly overweight or obese adults, in which prolonged sitting was interrupted with short bouts of standing or walking. Walking beats standing, standing beats sitting — but only for glucose. Blood pressure did not change significantly in any condition.

Where the number comes from — and where it does not

In none of these papers is 10,000 an optimum for mortality. The large meta-analysis describes the range 6,000 to 8,000 from age 60 and 8,000 to 10,000 below that; the women’s cohort describes a plateau at around 7,500. The authors themselves write that 10,000 steps is widely promoted but that there is little evidence for it; the women’s cohort states that the goal has a limited scientific basis. The widespread story that the number comes from advertising for a Japanese pedometer in the 1960s could not be confirmed by a peer-reviewed paper in the scientific databases — it is popular, but not verifiable here.

How to do it

Target values from the studies, by age and endpoint: for all-cause mortality, the benefit levels off from age 60 between 6,000 and 8,000 steps per day, under age 60 between 8,000 and 10,000. In older women the plateau is at around 7,500. For dementia, the optimum is 9,826 steps, and half the maximum risk reduction is reached at 3,826. The largest single jump is between the first and second quartile, that is, between about 3,553 and 5,801 steps. For distribution over the day, a meta-analysis of 17 studies with 261 participants with obesity found a reduction in glucose iAUC (SMD −0.49, 95 % CI −0.85 to −0.14) and insulin iAUC (−0.26, −0.50 to −0.03), with exploratory indications of larger effects with interruptions every 30 minutes or more often with 2 to 5 minutes of light walking. For hard endpoints, this comparison does not exist.

Safety

Walking is one of the safest forms of exercise there is, and for healthy adults there are no contraindications. With a risk of falls, however, more walking is not automatically better: the German DVO S3 guideline on fracture prevention recommends a targeted program with around three sessions per week for fall prevention (recommendation grade A, evidence level 1a) and provides for supervision and partner work for participants with low performance. For bone, walking is not enough: the same guideline calls for weight-bearing loading and/or dynamic strength training. After a vertebral fracture, exercises with impact loading of the spine are contraindicated — normal walking does not fall under this, jogging and jumping do; the same applies to knee and hip replacements. This is information, not a treatment recommendation.

BK-Score Well supported, heavily overhyped

Human evidence		8
Mechanism		7
Safety data		10
Hype gap		5
Track record of use		10

Large cohorts with step counters show a clear association between step count and mortality – with most of the benefit reached, depending on age, between 6,000 and 10,000 steps. There is little evidence for 10,000 as a goal, as the cohort papers themselves note; for dementia, however, the optimum is almost exactly 10,000 steps.

The score rates the state of knowledge, not the effect. “Safety data 9” means well studied – not harmless.
Subjective assessment by Biohacking Kompakt based on published scoring rules – not a scientific rating and not a medical recommendation. Rules and all ratings (German)

Frequently asked questions about 6,000 to 10,000 steps a day

Do I really have to walk 10,000 steps a day?