Cancer and plant-based diets

Cancer is the second leading cause of death worldwide, and diet is one of the most studied modifiable exposures. The evidence base for plant-based eating and cancer is less dramatic than the cardiovascular literature — effect sizes are generally smaller and site-specific rather than across-the-board — but it is coherent. The strongest signal is for colorectal cancer, where processed and red meat exposure is firmly classified as carcinogenic, and where fibre, whole grains, and legumes show protective associations.

The picture is one of modest, targeted benefit, not miracle cure.

What the major reviews concluded

The most authoritative synthesis is the WCRF/AICR Third Expert Report (2018), which pooled the global prospective literature and graded evidence by strength. Its judgments relevant to plant-based diets include: convincing evidence that processed meat causes colorectal cancer; probable evidence that red meat causes colorectal cancer; convincing evidence that whole grains and fibre-containing foods protect against colorectal cancer; and probable evidence that non-starchy vegetables and fruit protect against several aerodigestive cancers (WCRF/AICR, 2018). The report’s headline recommendation — limit red meat to no more than three portions a week and avoid processed meat — follows directly.

The International Agency for Research on Cancer reached parallel conclusions a few years earlier. Its Monograph Working Group (Bouvard et al., 2015) classified processed meat as Group 1 — carcinogenic to humans based on sufficient evidence for colorectal cancer, and red meat as Group 2A — probably carcinogenic to humans based on limited evidence for colorectal cancer and supportive mechanistic data. The Group 1 classification places processed meat in the same evidentiary tier as tobacco smoking and asbestos, though the magnitude of risk is far smaller: IARC estimated roughly an 18% increase in colorectal cancer risk per 50 g per day of processed meat.

Classification reflects strength of evidence, not potency.

What the plant-based cohorts show

EPIC-Oxford followed around 61,000 British adults including roughly 32,000 meat eaters, 8,600 fish eaters, 18,000 vegetarians, and 2,200 vegans. Over an average 14.9 years, vegetarians had an 11% lower risk of all cancers combined compared with meat eaters, with site-specific reductions for stomach cancer, cancers of the lymphatic and haematopoietic tissues, and postmenopausal breast cancer (Key et al., 2014). Vegans showed a 19% lower all-cancer incidence, though with wide confidence intervals reflecting smaller numbers.

Adventist Health Study-2 (Tantamango-Bartley et al., 2013) tracked about 69,000 participants and reported that vegetarian diets overall were associated with a 8% lower all-cancer incidence, with vegans showing the strongest signal — a 16% reduction — and a particularly notable 34% lower risk of female-specific cancers. A separate AHS-2 analysis focused on colorectal cancer (Orlich et al., 2015) found that vegetarian patterns were associated with a 22% lower risk, with pesco-vegetarians showing the largest reduction at 43%.

UK Biobank results have been more mixed. Watling et al. (2022) reported that compared with regular meat eaters, low meat eaters had a 2% lower all-cancer risk, fish eaters 10% lower, and vegetarians 14% lower, with specific reductions in colorectal and postmenopausal breast cancer among vegetarian women. The breast cancer signal attenuated after adjustment for body mass index. Bradbury et al. (2020) found that low meat eaters had an 11% lower risk of colorectal cancer compared with regular meat eaters, with a clear dose-response to red and processed meat intake.

Taken together, these cohorts point to modest reductions — roughly 8 to 15% for all cancers combined, with larger effects for colorectal cancer specifically.

Mechanisms: why meat, why fibre, why IGF-1

Four mechanisms carry most of the biological argument.

Heme iron. Red and processed meats are rich in heme iron, which catalyses the formation of reactive oxygen species in the colonic lumen and promotes the endogenous nitrosation of dietary amines. In animal models heme iron promotes aberrant crypt foci and tumour formation in a dose-dependent manner. Plant-based diets deliver predominantly non-heme iron, which is less bioavailable and does not carry the same luminal oxidative burden.

N-nitroso compounds. Processed meats are cured with nitrites, which react with amines and amides in the stomach and colon to form N-nitroso compounds — a class including several potent DNA-alkylating carcinogens. High-temperature cooking of red meat separately generates heterocyclic amines and polycyclic aromatic hydrocarbons. This chemistry underlies the IARC processed meat classification (Bouvard et al., 2015).

IGF-1. Animal protein intake — particularly dairy — raises circulating insulin-like growth factor 1, which promotes cell proliferation and inhibits apoptosis in a range of tissues. Bradbury et al. (2017) found that vegans in EPIC had markedly lower serum IGF-1 than meat eaters, with vegetarians intermediate. Higher IGF-1 has been associated with increased risk of prostate, breast, and colorectal cancer in meta-analyses, making this a plausible pathway for the cohort findings.

Fibre and the microbiome. Soluble and insoluble fibre shorten colonic transit time, dilute luminal carcinogens, and feed colonic bacteria that produce butyrate — a short-chain fatty acid with anti-inflammatory and pro-apoptotic effects on colonocytes. The WCRF/AICR (2018) convincing judgement on whole grains and fibre rests on this combined epidemiologic and mechanistic base.

What the evidence does not show

Plant-based diets are not a cure for established cancer, and the protective effects for most cancer sites outside the colon are either modest or inconsistent. Breast cancer associations are weaker than for colorectal: Bradbury et al. (2020, BMC Med, EPIC-Oxford breast cancer analysis) found no significant overall difference in breast cancer incidence between vegetarians and meat eaters after adjustment, though some subgroup signals persisted. Prostate cancer results across AHS-2 and EPIC have been heterogeneous.

The Fung et al. (2010) analysis of the Nurses’ Health Study and Health Professionals Follow-up Study cohorts added an important nuance: a low-carbohydrate dietary pattern high in animal protein and fat was associated with higher all-cause and cancer mortality, while a low-carbohydrate pattern high in vegetable protein and fat was associated with lower mortality. The composition of a diet — not just what it excludes — drives outcomes.

Caveats

Cohort participants who adopt vegetarian or vegan diets in the West tend to be leaner, smoke less, and exercise more, which introduces residual confounding. Adventist and EPIC-Oxford cohorts partially address this through comparators with similar lifestyle baselines. Ascertainment bias also affects cancer incidence data, since health-engaged participants are screened more often. And all of this evidence concerns population-level associations, not individual prognosis.

The bottom line

For colorectal cancer, the case is strong and mechanistically coherent: processed meat is a Group 1 carcinogen, red meat is probably carcinogenic, and fibre-rich whole-food plant-based patterns are protective. For all cancers combined, plant-based cohorts suggest a modest 10 to 15% reduction in incidence. For most other specific sites the picture is mixed but directionally favourable. The WCRF/AICR recommendations — eat mostly plants, limit red meat, avoid processed meat, prioritise fibre and whole grains — summarise the translatable science, and a whole-food plant-based diet meets them by default.