The evidence is very uncertain about the effect of step counts and risk of cardiovascular disease mortality, risk of cancer incidence, type 2 diabetes, and risk of falls.
Health risks generally decreased with every 1000 steps per day increment across most outcomes, up to the highest analysed category of 12 000 steps per day. The incremental improvement beyond 7000 steps per day was small, and there was no statistical difference between 7000 steps per day and a higher step count for all the other outcomes.
The evidence is based on cohort studies. For most outcomes a dose-response was observed. The findings of the study should be interpreted in light of limitations, such as the small number of studies available for most outcomes, a lack of age-specific analysis, and biases at the individual study level, including residual confounding.
The evidence is applicable to healthy adults and limited to patients with chronic conditions, poor health, or mobility limitations.
| Reference | Study type | Population | Exposure (E) and comparison (C) | Outcomes | Risk of bias |
|---|---|---|---|---|---|
| SR=systematic review; MA=meta-analysis | |||||
| «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | SR+MA | Adults aged 18 years and older, both apparently healthy or those living with a chronic
condition or disability. USA, UK, Japan, Australia, Taiwan, Brazil, Estonia, Norway, Spain, Sweden, Thailand |
E: Higher device-measured daily step counts measured with accelerometer, pedometer,
smartwatch, or other step-counting device in free-living conditions. C: Lower daily step counts, expressed as dose categories or per-increment comparisons, with a reference point of 2000 steps per day used in the meta-analyses. |
All-cause mortality Cardiovascular disease incidence and mortality Cancer incidence and mortality Type 2 diabetes incidence Dementia Depressive symptoms Falls |
Moderate |
| Reference | Comments |
|---|---|
| «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | Studies were eligible if they (1) had a prospective design where the exposure was
ascertained before the outcome (eg, cohort or intervention studies); (2) examined
the association between device-measured step counts and at least one of the relevant
health outcomes; and (3) were conducted among adults aged 18 years and older. Studies
of any language, peer reviewed or grey literature (including preprints), were considered.
Studies were excluded if done in a non-free-living context (eg, a laboratory setting). There is a moderate risk of bias due to adequate control of confounding. |
Results
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: moderate The quality of evidence is upgraded due to dose-response (non-linear). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 14 studies, 161176 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 29 months to 17 years (mean) | NR | NR | 0.77 (0.71–0.83) |
| 5000 steps/day vs. 2000 steps/day | 0.57 (0.5–0.66) |
||||
| 7000 steps/day vs. 2000 steps/day | 0.53 (0.46–0.60) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.52 (0.45–0.59) |
||||
| 12 000 steps/day vs. 2000 steps/day | 0.45 (0.39–0.53) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: moderate The quality of evidence is upgraded due to dose-response (non-linear). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 6 studies, 111 349 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs 2000 steps/day | From 5 years (mean) up to 20 years | NR | NR | 0.93 (0.90–0.96) |
| 5000 steps/day vs. 2000 steps/day | 0.82 (0.75–0.90) |
||||
| 7000 steps/day vs. 2000 steps/day | 0.75 (0.67–0.85) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.70 (0.62–0.79) |
||||
| 12 000 steps/day vs. 2000 steps/day | 0.69 (0.62–0.78) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: very low The quality of evidence is downgraded due to inconsistency (I2=78.2%) E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 3 studies, 120 758 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 6.8 (median) to 17 years (mean) | NR | NR | 0.74 (0.60–0.91) |
| 5000 steps/day vs. 2000 steps/day | 0.55 (0.37–0.80) |
||||
| 7000 steps/day vs. 2000 steps/day | 0.53 (0.37–0.77) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.56 (0.40–0.79) |
||||
| 12 000 steps/day vs. 2000 steps/day | 0.48 (0.33–0.71) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: very low The quality of evidence is downgraded due to imprecision and inconsistency (I2=73.7%). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 2 studies, 100505 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 6.9 years (median) to 10.1 years (mean) | NR | NR | 0.99 (0.97–1.00) |
| 5000 steps/day vs. 2000 steps/day | 0.96 (0.92–1.00) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.94 (0.87–1.01) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.90 (0.81–1.01) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.88 (0.76–1.01) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: moderate The quality of evidence is upgraded due to dose-response (non-linear). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 3 studies, 105660 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 6.9 years (median) to 10.1 years (mean) | NR | NR | 0.82 (0.74–0.9) |
| 5000 steps/day vs. 2000 steps/day | 0.66 (0.58–0.76) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.63 (0.55–0.72) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.6 (0.51–0.70) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.50 (0.41–0.60) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: very low The quality of evidence is downgraded due to imprecision. E= exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR= not reported |
|||||
| 4 studies, 61595 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 2.6 to 6 years (mean) | NR | NR | 0.97 (0.94–1.00) |
| 5000 steps/day vs. 2000 steps/day | 0.91 (0.83–1.00) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.86 (0.74–1.00) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.78 (0.61–0.99) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.73 (0.54–0.99) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: moderate The quality of evidence is upgraded due to dose-response (non-linear). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 2 studies, 79699 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 4.2 to 6.9 years (median) | NR | NR | 0.9 (0.86–0.93) |
| 5000 steps/day vs. 2000 steps/day | 0.73 (0.66–0.81) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.62 (0.53–0.73) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.55 (0.46–0.66) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.58 (0.49–0.70) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: moderate The quality of evidence is upgraded due to clear dose-response (linear). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 3 studies, 77565 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 2 years to 7.4 years (median) | NR | NR | 0.95 (0.94–0.96) |
| 5000 steps/day vs. 2000 steps/day | 0.86 (0.83–0.90) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.78 (0.73–0.83) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.67 (0.61–0.75) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.61 (0.53–0.70) | ||||
| Number of studies and number of patients [ref.] | Comparison | Follow-up time | Absolute risk E | Absolute risk C | Hazard ratio (95% CI) |
|---|---|---|---|---|---|
| Level of evidence: very low The quality of evidence is downgraded due to inconsistency of results (different directions of associations across studies). E=exposure; C=comparison; CI=confidence interval; HR=hazard ratio; NR=not reported |
|||||
| 4 studies, 94901 participants «Ding D, Nguyen B, Nau T, et al. Daily steps and he...»1 | 3000 steps/day vs. 2000 steps/day | From 6 months to 7 years (mean) | NR | NR | 0.93 (0.90–0.95) |
| 5000 steps/day vs. 2000 steps/day | 0.8 (0.74–0.87) | ||||
| 7000 steps/day vs. 2000 steps/day | 0.72 (0.65–0.81) |
||||
| 10000 steps/day vs. 2000 steps/day | 0.70 (0.63–0.79) | ||||
| 12 000 steps/day vs. 2000 steps/day | 0.78 (0.67–0.92) | ||||