Paluch · Circulation 2023 · Harmonized meta-analysis of prospective cohort studies · n=20152

Prospective Association of Daily Steps With Cardiovascular Disease: A Harmonized Meta-Analysis.

Cited 117 times in the scientific literature.

Level 3 - non-randomized controlled study

Meta-analysis of prospective cohort studies

PubMed 36537288 · doi:10.1161/CIRCULATIONAHA.122.061288 · record verified 2026-08-29

What was done

A harmonized meta-analysis of 8 prospective cohort studies examined 20,152 adults aged ≥18 years (mean age 63.2 ± 12.4 years, 52% women) with device-measured daily steps. Participants were followed for incident fatal and nonfatal cardiovascular disease (CVD) events (coronary heart disease, stroke, and heart failure) over a mean follow-up of 6.2 years (123,209 person-years). Cox proportional hazards regression models evaluated study-specific step quartiles, pooled via inverse-variance-weighted random-effects models and stratified by age (<60 vs. ≥60 years).

What was found

Across 1,523 CVD events (12.4 per 1,000 participant-years), step-risk associations differed significantly by age. For older adults (≥60 years), higher daily steps progressively decreased CVD risk compared with quartile 1: HR 0.80 (95% CI, 0.69 to 0.93) for quartile 2, HR 0.62 (95% CI, 0.52 to 0.74) for quartile 3, and HR 0.51 (95% CI, 0.41 to 0.63) for quartile 4, with restricted cubic splines confirming a nonlinear association. For younger adults (<60 years), estimates were not statistically significant compared with quartile 1: HR 0.79 (95% CI, 0.46 to 1.35) for quartile 2, HR 0.90 (95% CI, 0.64 to 1.25) for quartile 3, and HR 0.95 (95% CI, 0.61 to 1.48) for quartile 4.

Why it matters

This study provides prospective evidence that higher daily step volume is linked to substantially reduced CVD risk in older adults, supporting step counting as an accessible clinical and public health metric.

Limits

Observational designs cannot establish causality or eliminate residual confounding and reverse causality. The null findings in younger adults may stem from low event rates and wide confidence intervals over 6.2 years of follow-up rather than a true absence of benefit. Absolute step count values corresponding to the quartiles are not reported in the abstract.

Cited by