Rezende · British journal of sports medicine 2018 · umbrella review of systematic reviews and meta-analyses · n=19 reviews (541 original studies)

Physical activity and cancer: an umbrella review of the literature including 22 major anatomical sites and 770 000 cancer cases.

Cited 275 times in the scientific literature.

Level 1 - systematic review of randomized trials

Umbrella review of systematic reviews and meta-analyses

PubMed 29146752 · doi:10.1136/bjsports-2017-098391 · record verified 2026-08-30

What was done

Umbrella review searching Medline, Embase, Cochrane Database, and Web of Science for systematic reviews and meta-analyses examining physical activity in relation to cancer incidence and mortality across different anatomical sites in the general population. The authors assessed evidence strength, between-study heterogeneity, 95% prediction intervals, small-study effects, and excess significance bias.

What was found

The review included 19 reviews, 26 exposure-outcome meta-analyses, 541 original studies, and 770,000 cancer cases across 22 anatomical sites. Physical activity was associated with reduced risk for 7 cancer sites: colon, breast, endometrial, lung, oesophageal, pancreas, and meningioma. Only colon cancer (protective with recreational activity) and breast cancer (protective with overall activity) were supported by strong or highly suggestive evidence, respectively. The other 5 sites showed uncertainty and bias, such as P values not reaching < 10^-6, substantial heterogeneity, or 95% prediction intervals not demonstrating a definite effect direction. Four meta-analyses exhibited small-study effects and four showed excess significance. Numerical summary effect sizes were not reported in the abstract.

Why it matters

This synthesis demonstrates that while physical activity is linked to lower risks for several cancers, only colon and breast cancer associations meet stringent epidemiological criteria for strong or highly suggestive evidence.

Limits

The underlying evidence comprises observational literature susceptible to residual confounding and exposure misclassification. The abstract omits precise pooled effect estimates (such as hazard or odds ratios), and several cancer site analyses were limited by high heterogeneity, small-study effects, and excess significance bias.

Cited by