Cancer Prevention: What the Science Actually Says (2025)

Cancer will be diagnosed in approximately 1 in 2 men and 1 in 3 women over a lifetime in high-income countries. Yet a landmark study published in Nature Medicine in February 2026 — the most comprehensive global analysis to date, led by IARC/WHO — found that 37.8% of new cancer cases globally are attributable to modifiable risk factors. That means more than one in three cancer diagnoses could, in principle, have been prevented. Here is what the evidence actually shows, and what is still uncertain.

Note: This article is educational and does not constitute medical advice. Consult a qualified healthcare provider about personal cancer screening or risk management.

Key takeaways

• Tobacco remains the single largest preventable cause — responsible for 15.1% of all global cancer cases and approximately 30% of cancer deaths.
• Obesity is linked to 13 cancer types and is the second leading preventable cause in high-income countries.
• Alcohol causes at least 7 cancer types; there is no established "safe" level for cancer risk.
• Infections (HPV, hepatitis B, H. pylori) account for 10.2% of global cancer cases — and several are vaccine-preventable.
• Genetics accounts for approximately 5–10% of cancers; the vast majority are driven by environmental and lifestyle exposures.

What cancer actually is — and what makes it preventable

Cancer is not a single disease. It is a broad category of over 100 distinct conditions, all defined by the same core mechanism: uncontrolled division of abnormal cells that invade surrounding tissue and, in advanced stages, spread to distant organs (metastasis).

Every time a cell divides, its DNA is copied — approximately 6 billion base pairs replicated with extraordinary fidelity. Errors occur roughly once per billion bases copied, meaning about six errors per cell division. Most errors are corrected by DNA repair mechanisms; most that slip through are in non-critical regions or are corrected by cellular self-destruction (apoptosis). Cancer arises when errors accumulate in the specific genes that regulate cell growth — proto-oncogenes and tumour suppressor genes — and the cell's quality control mechanisms fail to catch them.

Carcinogens accelerate this process. Tobacco smoke contains over 70 identified carcinogens that directly damage DNA. Ultraviolet radiation causes specific DNA lesions (cyclobutane pyrimidine dimers) in skin cells. Alcohol metabolises to acetaldehyde, which crosslinks DNA strands. Obesity creates chronic low-level inflammation and hormone dysregulation that increase cell turnover rates and therefore the number of opportunities for copy errors.

Think of cancer risk like compound interest on DNA damage: each additional carcinogen exposure adds more "interest" to the account. Prevention is about reducing the principal balance — the total carcinogen load your cells are exposed to over decades.

Source: IARC World Cancer Report 2025; Hanahan and Weinberg, "Hallmarks of Cancer: The Next Generation," Cell 2011 (landmark review, 62,000+ citations).

The 2026 IARC study: what 30 modifiable risk factors actually account for

The most comprehensive cancer attribution study to date, published in Nature Medicine in February 2026, examined 30 modifiable risk factors — including 9 infectious agents and 13 occupational exposures — across 36 cancer types in 185 countries using 2022 cancer incidence data.

Key findings:

• 37.8% of all new cancer cases globally in 2022 were attributable to at least one modifiable risk factor.
• The burden was substantially higher in men (45.4% of male cancers attributable) than women (29.7%).
• The three largest contributors globally: smoking (15.1%), infections (10.2%), alcohol (3.2%).
• High body mass index, physical inactivity, and occupational exposures each contributed additional significant fractions.

The study's inclusion of infectious causes alongside behavioural factors for the first time in a comprehensive global analysis is significant — it demonstrates that vaccines (for HPV and hepatitis B) and antimicrobial treatment (for H. pylori) are among the most scalable cancer prevention tools available.

The figure of 37.8% is a population-attributable fraction — it estimates what proportion of cancers would not have occurred in the absence of these risk factors, under certain assumptions. It is not a precise ceiling on preventability; some cancers have multiple contributing causes, and removing one factor does not guarantee the cancer would not have developed via another pathway.

Source: Global Burden of Disease Cancer Collaboration, "Global and Regional Cancer Burden Attributable to Modifiable Risk Factors," Nature Medicine, February 2026; IARC/WHO summary, February 2026.

What causes cancer — risk factors by type

Cancer risk factors vary significantly by cancer type. The table below covers the leading modifiable risks for the major cancer types.

Cancer type	Primary modifiable risk factors
Lung	Smoking (80–90% of cases), radon gas, air pollution, asbestos
Colorectal	Red/processed meat, obesity, alcohol, sedentary lifestyle, low fibre intake
Breast	Alcohol, obesity (post-menopause), hormone replacement therapy, reproductive factors
Liver	Hepatitis B/C (infections), alcohol, obesity/NAFLD, aflatoxin (a mould toxin)
Cervical	HPV infection (virtually 100% of cases)
Skin (melanoma)	UV radiation (sunburn especially), tanning beds
Bladder	Smoking, occupational aromatic amine exposure
Oesophageal	Alcohol, tobacco, obesity (via reflux), very hot beverages (IARC Group 2A)
Stomach	H. pylori infection (~90% of cases), salt-preserved foods, smoking
Pancreatic	Smoking, obesity, type 2 diabetes, chronic heavy alcohol use
Oral cavity	Tobacco (including smokeless), alcohol, HPV
Endometrial	Obesity (strongest single risk factor), hormone therapy, diabetes

Source: IARC Monographs on the Identification of Carcinogenic Hazards to Humans; WCRF/AICR Cancer Prevention Recommendations 2018 (updated 2024).

1. Tobacco: the clearest causal case in medicine

The evidence linking tobacco to cancer is among the most robust in all of medicine — the product of 70 years of epidemiological, mechanistic, and interventional research.

The scope: Tobacco causes approximately 22% of global cancer deaths (WHO 2024). In lung cancer, smoking accounts for 80–90% of cases. The IARC has identified over 70 carcinogens in tobacco smoke, including polycyclic aromatic hydrocarbons, nitrosamines, benzene, formaldehyde, and radioactive polonium-210. These carcinogens damage DNA in lung cells directly and act on tissues throughout the body wherever tobacco smoke or its metabolites reach — which explains why smoking causes cancers not just in the lung but in the mouth, throat, larynx, oesophagus, stomach, pancreas, kidney, bladder, cervix, and blood cells (leukaemia).

Secondhand smoke: The IARC classifies secondhand smoke as a Group 1 carcinogen (definite human carcinogen). Non-smokers regularly exposed to secondhand smoke have approximately 25% higher lung cancer risk than non-exposed non-smokers.

The dose-response relationship is clear: Heavy smokers (40+ cigarettes/day for decades) have approximately 20–30 times the lung cancer risk of lifelong non-smokers. The relationship is not just threshold-based — each additional cigarette represents additional risk, not a safe zone below some threshold.

E-cigarettes: Current evidence suggests vaping carries substantially lower cancer risk than combustible cigarettes (primary cancer risk from smoking comes from combustion products, not nicotine). However, long-term data are unavailable — e-cigarettes are too new for 20-year cancer outcome studies. Some flavouring chemicals used in vaping liquids have been found to be carcinogenic or cytotoxic. The absence of evidence of harm is not evidence of absence of harm.

Quitting works substantially: Ten years after smoking cessation, lung cancer risk falls to approximately half that of a current smoker. The benefit begins within months of quitting and compounds over decades. No pharmacological intervention comes close to smoking cessation in terms of reduction in cancer risk.

Source: IARC Monograph 100E (Tobacco Smoking); WHO Tobacco Fact Sheet 2024; WCRF/AICR Diet, Nutrition, Physical Activity and Cancer: A Global Perspective 2018 continuous update.

2. Obesity and metabolic health

Obesity's cancer link is mechanistically well understood — perhaps better understood than the tobacco link was at the same stage of research. Yet public awareness of obesity as a cancer risk factor remains far lower than awareness of the tobacco link.

13 cancer types: The IARC identifies obesity (excess adiposity) as a cause of 13 cancers: colorectal, breast (post-menopausal), endometrial, kidney (renal cell carcinoma), pancreatic, oesophageal (adenocarcinoma), liver, gallbladder, ovarian, thyroid, meningioma (brain), and multiple myeloma, as well as stomach cardia cancer.

The mechanisms are multiple and reinforcing:

Excess adipose tissue is not metabolically inert — it is an active endocrine organ. In obesity, adipose tissue produces excess oestrogen (synthesised from androgens by the aromatase enzyme), which drives hormone-sensitive cancers including breast and endometrial cancer. This is why the obesity-breast cancer link is most pronounced in post-menopausal women, when ovarian oestrogen production has ceased and adipose tissue becomes the dominant oestrogen source.

Obesity also elevates fasting insulin and insulin-like growth factor 1 (IGF-1) — both of which stimulate cell growth signalling pathways (the PI3K/AKT/mTOR pathway) that drive proliferation in multiple cell types. Hyperinsulinaemia is strongly associated with colorectal and endometrial cancer risk in observational data.

Chronic low-grade inflammation is a third mechanism: adipose tissue, especially visceral fat, produces inflammatory cytokines (TNF-alpha, IL-6, leptin) that create a microenvironment promoting cell survival, angiogenesis (blood vessel growth), and resistance to apoptosis — all features of tumour growth.

The communication gap: Surveys consistently show under 50% of the public knows that obesity significantly raises cancer risk. Cancer charities and public health bodies have been slow to communicate this connection as clearly as they communicate the tobacco link — partly for fear of stigmatising individuals with obesity and partly because the relationship (while robust) is more complex than the tobacco-lung cancer connection.

Source: IARC Weight Control and Physical Activity Monograph 2002 (updated in subsequent cancer site-specific monographs); WCRF/AICR 2018 Cancer Prevention Recommendations; Lauby-Secretan B et al., "Body Fatness and Cancer," NEJM 2016.

3. Alcohol: the underestimated carcinogen

Despite cultural normalisation in most high-income countries, alcohol's status as a Group 1 carcinogen — a definite cause of cancer in humans — is established with the same certainty as tobacco.

Seven cancer types: The IARC classifies alcohol as a Group 1 carcinogen for cancers of the mouth, pharynx, larynx, oesophagus, liver, colorectum, and breast. The dose-response relationship is continuous — more alcohol consumed over a lifetime correlates with higher risk, with no safe threshold below which cancer risk is zero.

The no-safe-level finding: This is often misunderstood by the public, who have absorbed decades of messaging about the cardiovascular benefits of moderate red wine consumption. Cancer risk exists even at one drink per day. The WCRF/AICR and now Health Canada explicitly state that even moderate alcohol consumption increases cancer risk, and that for cancer prevention the safest option is not drinking at all. Individual risk decisions appropriately consider the whole risk-benefit picture, but the absence of a "safe" level for cancer is not contested by the evidence.

The acetaldehyde mechanism: Alcohol (ethanol) is oxidised in the body to acetaldehyde by alcohol dehydrogenase enzymes. Acetaldehyde is directly genotoxic — it forms DNA adducts (crosslinks and strand breaks) that, if not repaired, can lead to mutations. Acetaldehyde is classified as a Group 1 carcinogen in its own right. People with genetic variants that slow acetaldehyde metabolism — common in East Asian populations (variants in ALDH2) — face dramatically higher cancer risk from alcohol because acetaldehyde accumulates in tissues.

Breast cancer specifically: The breast cancer-alcohol link is among the best-quantified in cancer epidemiology. Meta-analyses consistently find a 7–10% increase in relative breast cancer risk per 10g of alcohol per day (roughly one standard drink). For a woman who drinks two standard drinks per day lifelong, this represents a material increase in absolute risk.

Source: IARC Monograph 100E (Alcohol); World Cancer Research Fund/AICR Alcoholic Drinks and the Risk of Cancer (continuous update 2021); Cao Y et al., "Light to Moderate Intake of Alcohol, Drinking Patterns, and Risk of Cancer," BMJ 2015.

4. Diet, physical activity, and the evidence

The evidence linking diet and physical activity to cancer is real but more nuanced than the tobacco or alcohol cases.

Processed meat — the clearest dietary carcinogen: Processed meat (bacon, sausage, hot dogs, deli meat, salami, pepperoni) is classified as a Group 1 carcinogen for colorectal cancer. This classification, made by IARC in 2015, caused significant media controversy — but the Group 1 designation means "definite evidence of carcinogenicity," not "as dangerous as tobacco." The absolute risk increase from occasional consumption is modest; the risk from regular, high consumption across a lifetime is clinically significant. A diet high in processed meat adds approximately 18% relative increase in colorectal cancer risk per 50g/day.

Red meat — probable: Unprocessed red meat (beef, pork, lamb, veal) is classified Group 2A (probably carcinogenic to humans) for colorectal cancer. The evidence is weaker and more confounded than for processed meat — red meat consumption correlates with other lifestyle factors that independently affect colorectal cancer risk.

Fibre — protective: High fibre intake from whole grains, legumes, and vegetables is consistently associated with 15–25% lower colorectal cancer risk across multiple large cohort studies. Mechanisms include reduced intestinal transit time (less contact between potential carcinogens and colonic epithelium), altered gut microbiome composition, and butyrate production by gut bacteria — butyrate being a short-chain fatty acid that promotes colonocyte differentiation and apoptosis.

Physical activity — independently protective: Regular moderate-to-vigorous physical activity is associated with approximately 20–30% reduced risk for colon, breast, and endometrial cancer, independent of body weight. This means the benefit is not simply mediated through weight loss — exercise has direct protective effects via reduced inflammation, improved insulin sensitivity, altered sex hormone levels, and possibly enhanced immune surveillance.

Ultra-processed foods: Emerging evidence from large prospective cohorts (including the French NutriNet-Santé study with 105,000 participants) suggests ultra-processed food consumption is independently associated with cancer risk beyond the contributions of calories, fat, and sugar individually. Mechanisms under investigation include food additives, packaging chemicals (PFAS, bisphenols), and the disruption of food matrix — but causality has not been established in randomised trials.

Source: IARC Monograph 114 (Red Meat and Processed Meat) 2015; WCRF/AICR Continuous Update Project 2023; Fiolet T et al., "Consumption of Ultra-Processed Foods and Cancer Risk," BMJ 2018; Friedenreich CM et al., "Physical Activity and Cancer Risk," Cancer Epidemiology, Biomarkers & Prevention 2021.

5. Infection and vaccination: the overlooked prevention route

Infectious agents cause approximately 10.2% of global cancer cases — the second-largest attributable factor after tobacco in the 2026 IARC study. Several are vaccine-preventable, making this perhaps the highest-leverage cancer prevention tool that does not require individual behaviour change.

HPV and cervical cancer: Human papillomavirus causes virtually 100% of cervical cancers and a rapidly growing proportion of throat (oropharyngeal) cancers — particularly in men, where HPV-positive oropharyngeal cancer has increased dramatically since the 1980s. The HPV vaccine (Gardasil-9, protecting against 9 strains including the highest-risk types 16 and 18) has efficacy above 97% against vaccine-type HPV infection. Countries with high vaccination coverage are already seeing dramatic drops in cervical cancer rates: Australia, which has had a school-based HPV vaccination programme since 2007, is on track to eliminate cervical cancer as a public health problem (defined as fewer than 4 cases per 100,000 women per year) by the 2030s.

Hepatitis B and liver cancer: The hepatitis B vaccine, given at birth in most national immunisation programmes, prevents chronic hepatitis B infection — the leading cause of hepatocellular carcinoma (liver cancer) globally, responsible for approximately 50–80% of cases. Countries with high neonatal HBV vaccination coverage (including most of Asia, where HBV is endemic) are seeing significant declines in liver cancer incidence in younger cohorts.

H. pylori and stomach cancer: The bacterium Helicobacter pylori causes approximately 90% of stomach cancers (gastric adenocarcinoma). It is treatable with a 1–2 week course of antibiotics (standard triple therapy: amoxicillin, clarithromycin, proton pump inhibitor). Mass H. pylori screening and treatment programmes in high-incidence countries — Japan, South Korea, Taiwan, China — have produced measurable declines in stomach cancer rates. Japan's National Cancer Research Center estimated that universal H. pylori eradication could prevent approximately 60,000 stomach cancer cases annually in Japan alone.

Source: WHO/IARC World Cancer Report 2025; Australian Institute of Health and Welfare Cervical Cancer Report 2024; Sugano K, "Strategies for Prevention of Gastric Cancer," Journal of Gastroenterology 2021.

6. Screening: what the evidence supports

Screening identifies cancers or precancerous changes before symptoms develop, when treatment is more effective or curative. Not all screening is equally supported by evidence; some screening causes more harm than benefit through false positives, overdiagnosis, and treatment of conditions that would never have caused harm.

Evidence-supported population screening:

Cervical: HPV test (preferred over Pap smear in most updated guidelines) every 5 years for vaccinated women, or every 3 years (Pap smear) for unvaccinated, from ages 25–65. Cervical screening has reduced cervical cancer mortality by over 70% in populations with high screening uptake.

Colorectal: Colonoscopy every 10 years or annual faecal immunochemical test (FIT) from ages 45–75. Colonoscopy detects and removes precancerous polyps before they become malignant — making it genuinely preventive, not just diagnostic.

Breast: Mammography every 1–2 years for women aged 40/50–74 (age range varies by country guideline). Meta-analyses estimate that mammography screening reduces breast cancer mortality by approximately 15–20% in invited populations. Overdiagnosis (detection of cancers that would never have caused symptoms) is a real cost of breast screening — estimates range widely from 1% to 30%, and remains the subject of ongoing scientific debate.

Lung: Annual low-dose computed tomography (LDCT) for heavy smokers aged 50–80 with 20+ pack-year smoking history. The US NLST trial found a 20% reduction in lung cancer mortality with LDCT screening in this high-risk group. Guidelines from the US Preventive Services Task Force (updated 2021) recommend this screening.

Not evidence-supported for general population screening:

PSA testing for prostate cancer: The PSA test has a high false-positive rate and leads to significant overdiagnosis and overtreatment (surgery, radiation) of slow-growing cancers that would never have become symptomatic. Major guidelines (USPSTF, European Association of Urology) support shared decision-making for men aged 55–69 rather than routine screening.

Whole-body CT scans, thermography, and most "liquid biopsy" cancer screening tests marketed directly to consumers: No randomised trial evidence supports any of these for population screening. Some are actively harmful via radiation exposure or false reassurance.

Source: US Preventive Services Task Force Screening Recommendations (updated 2021–2024); NEJM Evidence review of cancer screening; European Code Against Cancer (4th edition).

What the evidence does not support

Several widely believed cancer prevention claims are not supported by current evidence:

Antioxidant supplements: No randomised trial has shown benefit for cancer prevention from vitamin C, vitamin E, or antioxidant supplements in well-nourished populations. Two large trials (ATBC and CARET) were halted early because beta-carotene supplements were found to increase lung cancer risk in smokers and former smokers. The hypothesis that dietary antioxidants could be administered as supplements to prevent cancer has been repeatedly refuted.

Alkaline diet: The claim that cancer "grows in an acidic environment" and can be prevented by eating alkaline foods misunderstands tumour biology. Blood pH is tightly regulated by respiratory and renal systems to 7.35–7.45 regardless of food intake. Consuming alkaline foods does not alter blood pH. Tumours do create acidic microenvironments — but this is a consequence of their high metabolic activity, not a cause that can be manipulated by diet.

Organic food: No randomised trial evidence supports organic food consumption as cancer-preventive. Pesticide residues on conventionally grown produce are typically far below levels associated with human toxicity, as assessed by major food safety agencies. The cancer risk from not eating fruits and vegetables due to the higher cost of organic options is far greater than any theoretical pesticide risk from conventional produce.

Stress directly causes cancer: Psychological stress does not appear to directly cause cancer through a proven biological mechanism, based on current evidence. Stress-related behaviours (smoking, excess alcohol, poor sleep, physical inactivity, overeating) increase cancer risk, but epidemiological studies controlling for these behaviours do not find a robust independent effect of perceived stress on cancer incidence.

Source: Cochrane Review of Antioxidant Supplements for Prevention of Mortality 2012; WCRF/AICR Food, Nutrition, Physical Activity and Cancer Continuous Update 2024; Cancer Research UK position on alkaline diet.

Frequently asked questions

Q: Can eating sugar "feed" cancer and should I avoid sugar? A: Cancer cells do consume more glucose than normal cells (the Warburg effect — they preferentially use glycolysis even in the presence of oxygen). But this does not mean that eating sugar directly feeds tumours in a way that avoiding dietary sugar can prevent or slow. Blood glucose is tightly regulated by insulin; starving cancer cells of sugar would require starving all cells of glucose, which is incompatible with normal physiology. There is no clinical evidence that dietary sugar restriction prevents cancer development or slows established tumours outside of extreme conditions. The cancer risk associated with high sugar consumption is mediated through obesity and insulin resistance, not through direct glucose supply to tumours.

Q: Does sunscreen actually prevent skin cancer? A: Yes, with good evidence. UV radiation causes melanoma and non-melanoma skin cancers via specific DNA damage mechanisms. Randomised trial evidence (the Queensland Melanoma Study) and large observational data support regular sunscreen use (SPF 30+, broad-spectrum UVA/UVB protection) reducing skin cancer risk in high-UV environments. Avoiding tanning beds is strongly supported — the IARC classifies tanning bed use as a Group 1 carcinogen, with a 75% increased melanoma risk for users who first use tanning beds before age 35.

Q: How much does family history of cancer increase risk? A: It depends heavily on which cancer and which specific family history. Having one first-degree relative (parent, sibling, child) with colorectal cancer approximately doubles personal risk; having two first-degree relatives raises it 3–4 fold. For breast cancer, carrying a BRCA1 or BRCA2 pathogenic variant raises lifetime breast cancer risk to 50–85% (vs. ~12% population baseline). However, most people with a family history of cancer do not carry a high-penetrance gene variant — the increased risk reflects shared environmental and lifestyle factors as well as lower-penetrance genetic variants. Genetic testing and counselling are appropriate for people with strong family histories of specific cancers (breast/ovarian, colorectal Lynch syndrome).

Q: Does eating burnt or charred food cause cancer? A: Charring or grilling meat at high temperatures produces heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs) — compounds that are carcinogenic in high-dose animal studies and are classified as probable human carcinogens (Group 2A) by the IARC. The evidence in humans is less clear than animal data would suggest, because typical human exposure from cooked meat is much lower than doses used in animal studies. It is reasonable to moderate consumption of heavily charred meat, but occasional grilling is unlikely to represent a material cancer risk in the context of overall diet.

Q: What is the most impactful single thing a person can do to reduce cancer risk? A: Not smoking — or if currently smoking, quitting. Tobacco accounts for 15.1% of all global cancer cases and 30% of cancer deaths. No other single intervention comes close in magnitude. For non-smokers, the next most impactful factors are maintaining a healthy weight, limiting alcohol, and completing recommended cancer screening for age and risk group. Physical activity and a diet with high fibre and limited processed meat provide additional risk reduction.

The takeaway

Cancer prevention evidence points consistently in a few directions: don't smoke, maintain a healthy weight, limit alcohol, exercise regularly, eat abundant vegetables and fibre with limited processed meat, get recommended vaccinations (HPV, hepatitis B), and complete recommended screening for your age and risk profile. These claims are not controversial in the scientific literature — they reflect decades of converging evidence from epidemiology, mechanistic biology, and randomised trials. The 37.8% population-attributable fraction in the 2026 IARC study represents a genuine upper bound on what behavioural and institutional interventions could achieve if applied universally. The gap between what science knows and what populations practise remains the primary lever for reducing the cancer burden globally.

Related Reading

• Countries with the Most Cancer Rates
• Countries with the Most Diabetes
• Countries with the Most Mental Health Disorders
• Countries with the Highest Life Expectancy
• Countries with the Most Nurses Per Capita