Exercise & Cancer Risk

Physical activity is one of the few behaviors consistently associated with lower risk of several cancers — but the evidence is almost entirely observational, the effect sizes are modest, and they vary by cancer site, sex, and population. This hub maps what the large pooled cohorts actually found, at what dose, and where the estimates carry qualifiers.

The Landmark: 1.44 Million Adults, 26 Cancers

The reference point for this entire territory is Moore et al.'s pooled analysis of 12 US and European cohorts (JAMA Internal Medicine, 2016): 1.44 million adults, median age 59, followed for a median of 11 years, during which 186,932 cancers were diagnosed. Comparing people at the 90th percentile of leisure-time activity against those at the 10th percentile, higher activity was associated with lower risk for 13 of the 26 cancer types examined, with reductions of 20% or more for seven of them.

Two honest footnotes belong next to that chart. First, the biggest relative estimate (esophageal adenocarcinoma, roughly 40% lower) attaches to a comparatively rare cancer; the absolute numbers matter more for common sites like colon and breast. Second, the same analysis found activity associated with higher rates of malignant melanoma (HR 1.27) and prostate cancer (HR 1.05) — plausibly sun exposure and screening detection, not biology, but a reminder that the associations cut in more than one direction.

Site by Site, With the Qualifiers Attached

The sex and population qualifiers are not fine print — they are the finding. Colon estimates are partly driven by male cohorts; breast estimates mostly reflect postmenopausal women; and the pooled cohorts were predominantly white, US and European populations, so transportability elsewhere is an assumption, not a demonstrated fact.

How Much Activity, and How Hard

The dose question has a direct answer in Matthews et al.'s pooled analysis of nine cohorts (more than 750,000 adults; Journal of Clinical Oncology, 2020). Activity in the range of 7.5–15 MET-hours per week — roughly 2.5 to 5 hours of brisk walking, or 1.25 to 2.5 hours of vigorous work — was associated with significantly lower risk of seven of 15 cancer types studied: colon (in men), breast, endometrial, kidney, myeloma, liver, and non-Hodgkin lymphoma (in women). Pushing from the lower to the upper end of that band was associated with further reductions — for example, colon cancer in men went from about 8% to about 14% lower, and endometrial from about 10% to about 18% lower.

That dose range is deliberately unglamorous: it is the same range the Zone 2 topic treats as an aerobic base and the Walking topic treats as ordinary infrastructure. The World Cancer Research Fund / American Institute for Cancer Research recommendation sits in the same place — be physically active as part of everyday life, at guideline-level amounts. The deeper dose-response curve, and how much vigorous intensity adds, is its own subtopic.

Why These Are Associations, Not Promises

Every number above comes from observational cohorts, and the honest reading requires knowing what that does and does not carry:

• 🔀 Confounding: highly active people tend to be leaner, smoke less, drink less, and screen differently — models adjust for what was measured, never for everything.
• ⚖️ Reverse causation: early, undiagnosed illness makes people less active; most analyses exclude early follow-up years, but the problem never fully disappears.
• 📏 Self-reported activity: the exposure is questionnaires, not accelerometry — misclassification that typically blurs estimates toward the null.
• 🚫 No randomized trials: exercise randomized to cancer incidence has not been tested at scale and probably never will be; plausible mechanisms (below) are supportive, not confirmatory.

The mechanism list is where the biology is genuinely plausible even though the causal claim stays unproven. Candidate pathways recur across reviews:

• 🧬 Sex hormones: activity lowers circulating estrogens and androgens — the most cited route for breast and endometrial associations.
• 🩸 Insulin pathways: lower fasting insulin and better glucose handling reduce a growth-signal environment several cancers exploit — the metabolic chain the Insulin Resistance topic covers from the other end.
• 🔥 Inflammation: chronic low-grade inflammation is a tumor-promoting background; habitual activity tends to lower inflammatory markers.
• 🦠 Immune surveillance: exercise immunology — natural killer cell activity in particular — provides a plausible link to cancer control, though mostly from small short-term studies.
• ⏱️ Gut transit: faster colonic transit is the classic proposal for the colon association — less contact time between carcinogens and the mucosa.

None of these mechanisms is a demonstrated explanation for the cohort numbers; together they make the associations biologically credible, which is exactly as far as the evidence licenses.

Incidence and Mortality Are Different Questions

Most quoted numbers — including everything above — concern getting cancer. Whether exercise after a diagnosis changes survival is a separate literature, overwhelmingly observational, and more vulnerable to confounding: healthier patients exercise more, and sicker patients cannot. The associations in breast and colon cancer survivorship are suggestive and directionally consistent, but this hub deliberately does not borrow them as prevention evidence. The distinction, and the survivor-side territory, sits inside the site-specific subtopics below.

Where the Evidence Thins

Beyond the headline sites, the picture gets patchy fast:

• 🫁 Common but weakly linked: for some common cancers the pooled estimates are modest or absent — prostate is slightly positive, ovarian largely null.
• 🔢 Small numbers: rarer sites (pancreas, thyroid, brain) carry wide confidence intervals that overlap no-effect.
• 🌍 Narrow populations: the pooled cohorts are mostly white and Western; whether the same magnitudes transfer globally is understudied.
• 🚶 Unmeasured domains: occupational and household activity were mostly excluded — "leisure-time" is only one slice of total movement.

The full tour of that long tail — which cancers have credible evidence and which are folklore — is the other-cancers subtopic's territory.

Questions, Answered Briefly

• Which site has the strongest association? In relative terms, esophageal adenocarcinoma (HR 0.58) — but it is uncommon; in terms of cases potentially influenced, colon and breast dominate.
• Does plain walking count? Yes — moderate intensity is the reference exposure in most cohorts, and 2.5–5 weekly hours of it is the range Matthews et al. tied to lower risk of seven cancers.
• Does intensity add anything? Vigorous activity appears to deepen the association for several sites, but on a per-MET basis the difference is debated — the dose subtopic owns this.
• I already had cancer — does this apply? Different question, different evidence base; start with the site-specific subtopics rather than extrapolating from prevention cohorts.

Go Deeper: Subtopics

• 🔎 Colon cancer and physical activity — the most consistently associated major site: the pooled estimates, the male-cohort weighting, and the mechanisms (transit time, insulin, inflammation) under discussion. Read it →
• 🔎 Breast cancer, menopause, and activity — why the pooled HR ≈ 0.90 is mostly a postmenopausal story, and how timing of activity across life stages enters the evidence. Read it →
• 🔎 Endometrial cancer: activity and adiposity — the site where the largest share of the association travels through body fat, and what survives BMI adjustment. Read it →
• 🔎 The activity dose–cancer risk curve — 7.5–15 MET-hours and beyond: the shape of the dose-response, what vigorous intensity adds, and where returns flatten. Read it →
• 🔎 Other cancers: the evidence tour — the long tail beyond the headline sites: which associations are credible, which are null, and which are folklore. Read it →

Related Topics