Extra weight is linked to cancer. Nineteen types of it.

Meta-analysis of cohort studies. 1,520,512 cancer cases. 226 cohort articles published up to 23 April 2025; most studies followed people for more than 5 years.

Many studies, added up. People were watched, not told what to do. A clue, not proof.

Who: Adults (18+) in prospective cohorts and administrative health databases from 23 countries (cases 59.4% Europe, 29.9% East Asia, 10.4% North America); n = 1,520,512 incident cancer cases across 25 cancer types (n is cancer cases, not people)

• Womb cancer+58%per 5 BMI points
• Kidney cancer+30%per 5 BMI points
• Breast, pre-menopause-8%lower, not higher

More to know, from our read of the paper:

• Observational: pooled cohort associations, not proof that weight causes the cancers; MR (genetic) analyses agreed in direction for 9 of 22 cancers, others imprecise
• All risks are per 5 kg/m2 higher BMI (RR per 5 BMI points), converted from each study's HR/OR/RR; never phrase as 'obese people have X% more'
• The '19 cancers' include low-certainty links (head and neck and bladder in never-smokers, aggressive prostate with CI 1.00-1.09, glioma 1.03); authors rate high certainty for 11 types only
• The 3 inverse links (premenopausal breast, lung and oesophageal SCC in never-smokers) are associations; do not suggest weight protects. MR even found an inverse link for postmenopausal breast cancer, opposite to the observational result
• High heterogeneity for endometrial, breast and melanoma (I2 >= 90%); regional and sex differences come from tests not corrected for multiple comparisons

A new meta-analysis pooled 226 cohort studies covering 1.5 million cancer cases across 25 cancer types. Higher BMI was linked to higher risk of 19 of them: per 5 extra BMI points, 58% higher for womb (endometrial) cancer and 30% higher for kidney cancer, down to 3% for glioma. It was linked to lower risk of 3, including breast cancer before menopause (8% lower). Leukaemia, non-Hodgkin lymphoma, bladder cancer and glioma were new compared with earlier major reviews. These are observational studies: linked, not proven, though genetic studies mostly pointed the same way.

We read the abstract and the full text of this paper and checked every number on this page against it (17 facts, checked on 2026-10-29). The study design is named above. Cohort and other observational studies are described as linked, not proven. Found an error? Email theis@onestudyaday.com.