The most influential number in modern walking did not come out of a clinical trial. It came out of a Tokyo marketing department. In 1965, a small company called Yamasa Tokei Keiki launched a pocket pedometer for the post-Olympic fitness market and named it the manpo-kei — literally, "ten-thousand-step meter" [1]. The number was chosen because the Japanese character for ten thousand resembles a striding figure, and because round numbers sell pedometers. Six decades later, that marketing slogan sits inside the activity goal of every major fitness wearable, the default target of nearly every NHS walking campaign, and the back-of-mind benchmark of every adult who has ever opened the Health app on their phone. It has, until this week, been the single most quoted figure in walking medicine.
It is wrong, or at least, it is the wrong threshold. The right threshold, according to a meta-analysis published on 12 May in the American Journal of Health Promotion, is closer to 5,700 [2].
What the meta-analysis actually says
The paper pools forty-six prospective cohort studies of older adults aged sixty-five and over, comprising 367,843 participants across twenty-eight distinct health outcomes. The headline analysis is a random-effects meta-analysis of daily step counts against all-cause mortality, drawn from the subset of cohorts where the average daily step volume could be reliably extracted — some 19,206 older adults in total. The result is the kind of clean finding that makes editors phone the authors at home. Older adults whose habitual walking averaged 5,694 steps a day had a thirteen per cent lower risk of dying from any cause per additional thousand steps walked, with a hazard ratio of 0.87 (95% CI 0.80–0.95).
The pooled effect was statistically robust and, more importantly, replicable across the included cohorts. Benefits, the authors note, were observed not only for mortality but for incident disease, hospitalisation, cognitive decline, psychological wellbeing, and functional capacity. The same per-thousand-step gradient appeared, at different magnitudes, in every outcome the meta-analysis touched. Walking, in the population of adults most likely to be told to take it easy, was associated with a lower risk of essentially everything that worries the geriatrician.
Where the 10,000 number actually came from
The Yamasa pedometer story is one of those rare cases where the marketing origin is uncontested. A 2016 Sports Medicine commentary by Catrine Tudor-Locke traced the 10,000-step number from the manpo-kei brochure of 1965 through a 1980s Japanese health-promotion campaign and into the World Health Organisation's published activity guidance of the early 2000s [3]. Subsequent research by Yoshiro Hatano in Kyushu lent the figure a thin veneer of rationale — 10,000 steps approximated the additional calorie expenditure thought to offset sedentary Japanese office work — but the number itself preceded the science. It was a target chosen for typographical reasons and validated, in retrospect, by approximations.
The number was chosen because the Japanese character for ten thousand resembles a striding figure, and because round numbers sell pedometers.
What the actual research finds, in 2026, is that 10,000 is closer to a ceiling than a target. A 2022 Lancet Public Health meta-analysis of fifteen cohorts and 47,471 adults reported that the mortality benefit of additional steps plateaued at roughly 6,000–8,000 daily steps in older adults [4]. The Saint-Maurice cohort in JAMA 2020, an accelerometer-measured sample of 4,840 US adults, found mortality risk halved at 8,000 steps and gained nothing further past 12,000 [5]. A 2023 European Journal of Preventive Cardiology meta-analysis of 226,889 adults found every additional thousand steps reduced all-cause mortality by fifteen per cent [6]. The new American Journal of Health Promotion meta-analysis sits inside this body of work; it is not an outlier but a refinement. The number is lower than the wearable says.
What this means in plain Tuesday terms
A sedentary office worker in the United Kingdom averages somewhere between 3,500 and 5,000 daily steps without trying [7]. A thirty-minute purposeful walk at a moderate pace adds another 2,500 to 3,500 steps to that baseline. The arithmetic resolves itself: the meta-analytic threshold for a thirteen per cent mortality reduction per additional thousand is achievable in one short walk a day, on top of whatever incidental walking the day already contains. The implication is not that walking less is better. The implication is that the bar to start seeing benefit is lower than the wearable target has been telling adults over sixty for a decade.
There is a particular cruelty to a goal that has been quietly unreachable. Adults who hit 6,200 steps in their busiest day, see the wearable show 62 per cent of goal, and conclude they failed have, in fact, just done something the published literature says is associated with measurably longer life. The 10,000-step ring on the phone face shows them red. The meta-analysis shows them green.
What the meta-analysis does not say
The Walking and Health Outcomes paper studies habitual walking — the regular, non-prescribed, daily-life movement of older adults across decades of cohort follow-up. It does not study structured walking, the kind that alternates brisk and easy blocks and asks the cardiovascular system to adapt. The authors note that walking pace did not meaningfully modify the mortality association in their analysis, but that finding describes the variation present in the underlying cohorts, not the upper limit of what structured intervention can produce. The Shinshu Japanese Walking Method trial found that interval walking, performed four times a week for five months, lowered systolic blood pressure by nine millimetres of mercury, raised peak aerobic capacity by nine per cent, and added fifteen per cent of leg strength [8]. None of that disappears because the threshold-step meta-analysis came in lower than expected. The two findings are stacked, not competing. The first identifies the minimum volume associated with mortality benefit. The second identifies the protocol that compounds it.
For an adult over sixty, the practical reading is straightforward enough. The daily floor sits at roughly 5,700 steps — lower than the wearable, achievable inside one purposeful walk — and that floor delivers the mortality reduction the meta-analysis describes. Layering thirty minutes of brisk-and-easy interval walking, four times a week, on top of that floor delivers the cardiovascular adaptations the Shinshu trial described. The first number is permission. The second is the programme.
The number on the phone, recalibrated
The 10,000-step goal will not disappear from wearables, because it sells wearables. What can change is the adult walker's response to the number. A wearable that shows 6,000 steps at the end of an ordinary day is showing a number that the new meta-analysis associates with mortality reduction. A wearable that shows 4,500 is showing a number that is one short walk away from doing the same. The data, on this point, is now firm enough that a primary-care doctor speaking to a sixty-five-year-old can, in good conscience, point at the screen and say: that is enough to matter; now let us add a half-hour walk most days, and let us make that half-hour interval-paced rather than steady. That second sentence is where the library begins.
