A sweeping analysis of 1.5 million cancer cases shows that excess body weight may shape cancer risk more broadly than previously recognized, with risks varying by cancer type, gender and region.
Study: Fat and cancer: a systematic review and meta-analysis. Image credit: Piyawat Nandeenopparit / Shutterstock
In a recent systematic review and meta-analysis published in the journal Metabolism of Natureresearchers combed through decades of peer-reviewed literature to reassess the global relationship between body mass index (BMI) and cancer risk.
The analyzes pooled data from more than 1.5 million documented cancer cases and found that elevated BMI was positively associated with 19 different types of cancer, significantly more than the 13 previously identified by consensus reports. The review further identified notable regional and gender variations in these risks and found that genetic evidence generally supported many of the observed associations, although not uniformly across all cancer types.
aSum of 25 types of cancer incidents. siIndividual cancers. Estimates from pooled studies covering multiple regions for which country-specific case numbers were not available were excluded from this figure (head and neck). Numbers may not add up due to rounding.
Background
The association between excess body weight and cancer risk is by no means a new concept. For years, major health organizations such as the World Cancer Research Fund (WCRF) and the International Agency for Research on Cancer (IARC) have warned that carrying excess weight increases the risk of at least 13 types of cancer.
However, as global obesity rates continue to show unprecedented growth, particularly in low- and middle-income countries, researchers highlight important gaps in our understanding of how these factors work biologically.
For example, it remains unclear whether obesity-related cancer risks apply equally to different global populations or whether alternative measurements, such as waist circumference, provide a clearer picture of the association between obesity and subsequent cancer risk.
Although previous reviews aimed to address these knowledge gaps, they lacked data from different geographic regions (most focused on American and European populations) and did not include data from next-generation genetic cohorts, thus necessitating a reevaluation of the variables that best explain these observed relationships.
About the review
The present review aimed to meet these requirements and inform future weight management and oncology policy by synthesizing comprehensive prospective cohort studies from online scientific repositories (PubMed, EMBASE and Scopus) from database inception to April 2025.
The final set of publications in the review included 226 separate peer-reviewed articles (n = 1,520,512 cancer cases) covering data from 23 countries (6 major geographic locations) and documenting 557 separate BMI-cancer risk associations in 25 common cancer types.
For the meta-analyses, all hazard ratios from the included publications were standardized to a scale measuring a 5 kg/m² increase in BMI, thus maintaining statistical uniformity and allowing direct comparisons between previously non-overlapping datasets.
Since most of the datasets were observational (identifying correlations), Mendelian randomization (MR) analyzes were used to strengthen causal inference. MRI analyzes use inherited genetic variations as receptors for lifetime exposure to the variable under consideration (here, increased body weight).
Finally, to minimize the effects of tobacco use (as a residual confounder), smoking-related cancers were assessed using data from lifetime nonsmokers.
Study Findings
Meta-analyses revealed statistically significant evidence linking participants’ higher BMI with increased risk of 19 different types of cancer, with risk estimates varying nearly 20-fold in magnitude between cancer types. For example, at the higher end, analyzes showed that each 5-point increase in BMI was associated with a 58% increase in endometrial cancer risk (relative risk [RR] = 1.58, 95% confidence interval [CI]: 1.51–1.67) and a 47% increase in the risk of esophageal adenocarcinoma (RR = 1.47).
Most importantly, the data revealed positive associations for leukemia (RR = 1.09), non-Hodgkin’s lymphoma (RR = 1.05), bladder cancer (RR = 1.04), and glioma (RR = 1.03), none of which had previously been identified as malignancies associated with BMI overstatement.
The authors also reported inverse associations for premenopausal breast cancer, lung cancer among nonsmokers, and esophageal squamous cell carcinoma among nonsmokers.
The study further identified significant regional disparities in the observed associations between BMI and cancer risk. For example, postmenopausal breast cancer risks associated with a 5-unit increase in BMI were found to present approximately twice the risk in East Asian cohorts (RR = 1.25) compared with corresponding European cohorts (RR = 1.11, p-heterogeneity = 7.6 × 10−6), emphasizing the non-generalizability of results from the second cohort to the first.
Similarly, gender differences were identified, as seen in the associations with colorectal cancer, which were significantly stronger in men (RR = 1.17) than in women (RR = 1.06, p-heterogeneity = 8.9 × 10−¹0). In contrast, the BMI-Gallbladder Cancer association was stronger in women (RR = 1.33) than in men (RR = 1.13, p-heterogeneity = 9.5 × 10−5).
Finally, when comparing BMI versus waist circumference as predictors of subsequent cancer risk, the review found that both variables yielded broadly similar risk estimates, although modest differences were observed for some cancer types.
conclusions
This review validates previous research indicating the substantial impact of obesity on cancer risk and the global burden of cancer, while highlighting that previous frameworks have greatly underrepresented regional risks, particularly in East Asian populations, where differences in hormone therapy use, estrogen exposure, gallbladder etiology, tumor subtype patterns, partial surveillance arrest or re-representation.
Furthermore, the review highlights that significant regional limitations remain, with Africa, South Asia and Central America (among other regions) remaining underrepresented by long-term cancer incidence cohorts even in the present study.
Future research should prioritize diverse understudied populations in order to elucidate a truly fair understanding of modifiable cancer risk factors.
