Global hotspots where breast cancer and atrial fibrillation overlap reveal smoking and alcohol as shared modifiable risk factors
核心洞察
A global analysis of women aged 55 and older across 204 countries found significant geographic overlap in breast cancer (搜索) and atrial fibrillation (搜索) incidence, with highest rates in North America, Europe, and Oceania.
Smoking and alcohol use emerged as shared modifiable risk factors for both conditions, with alcohol being the strongest contributor—linked to up to 12% higher BC incidence and 10% higher AF incidence.
Population-attributable fraction modeling estimated that reducing alcohol and smoking to theoretical minimum-risk levels could cut breast cancer (搜索) by 29% and atrial fibrillation (搜索) by nearly 12% globally.
A sweeping global analysis published in the Journal of the American Heart Association has mapped, for the first time, the geographic hotspots where breast cancer (搜索) (BC) and atrial fibrillation (搜索)/flutter (AF) most closely overlap in women aged 55 and older—and identified smoking and alcohol use as the key shared, modifiable risk factors driving both conditions.
Drawing on Global Burden of Disease (GBD) 2021 data spanning 204 countries and territories, researchers developed a unified analytical framework integrating spatial epidemiology with machine learning techniques, including random forest modeling and SHAP (Shapley Additive Explanations), to identify and rank candidate risk factors. Double Machine Learning was then applied to estimate potential causal associations at the population level.
"This is the first study combining global data with machine learning to show the relationship between the conditions, their location across the world and the shared risk factors of these two conditions," said Shu Wang, M.D., Ph.D., study co-author and director of the Breast Disease Center at Peking University People's Hospital in Beijing.
Three distinct global incidence patterns emerge
The analysis classified 202 countries (Tuvalu and Tokelau were excluded due to limited data) into three incidence patterns. Approximately 80 countries (39%) fell into "consistent" regions, where BC and AF occupied the same incidence quartile. About 65 countries (32%) were BC-dominant, concentrated mainly in Asia and northern South America, while 57 countries (28%) were AF-dominant, found in parts of Africa and southern South America.
The highest incidence rates for both diseases clustered in North America, Europe, and Oceania, while the lowest rates were observed in Northern and Central Africa. Positive correlations between BC and AF incidence held across all regional classifications, though the relationship was strongest in consistent regions and weakest in AF-dominant regions.
"One of the most surprising aspects of our findings was how common both breast cancer (搜索) and atrial fibrillation (搜索)/flutter diagnoses were among women ages 55 and older in high-income regions, which highlights the influence of lifestyle," Wang noted.
Alcohol and smoking: the dual drivers
From an initial pool of 58 possible risk factors, machine-learning screening narrowed the field to eight key candidate risks for each disease. For BC, these included sugar-sweetened beverages, smoking, alcohol, high BMI, low physical activity, red meat, high plasma glucose, and low whole grain intake. For AF, the main risks were high sodium, smoking, alcohol, processed meat, radon, high LDL cholesterol, sugar-sweetened beverages, and red meat.
Subsequent Double Machine Learning analysis confirmed four factors with significant associations: smoking and alcohol (common to both diseases), and high BMI and low physical activity (specific to BC). Alcohol emerged as the strongest risk factor overall—a one-standard-deviation increase in standardized exposure was associated with up to a 12% higher BC incidence and a 10% higher AF incidence.
Alcohol use was found to be a significant contributor to breast cancer (搜索) (9.27% of cases) and atrial fibrillation (搜索)/flutter (7.57% of cases). Population-attributable fraction modeling estimated that shifting these exposures to their theoretical minimum-risk levels could reduce BC by 29% and AF by nearly 12%, particularly in regions where both diseases overlap.
High-exposure regions and composite risk profiles
Western, Northern, and Southern regions—particularly North America and Oceania—recorded the highest smoking and alcohol exposure levels. The United States, Canada, and Australia were among the countries with high exposure to both smoking and alcohol, while Australia and Luxembourg recorded particularly high alcohol exposure. High BMI was most prevalent in the Pacific Islands and the Middle East, including Kuwait, Jordan, and Qatar, whereas low physical activity was highest in locations such as the Marshall Islands, American Samoa, and Qatar.
A composite exposure index ranging from 0 to 11 indicated the highest cumulative risk-factor exposure in Europe, Oceania, South America, and the United States, and the lowest in Africa and parts of Asia—a pattern that closely aligned with BC and AF incidence rates.
Clinical implications and integrated prevention
"Breast cancer (搜索) and atrial fibrillation (搜索)/flutter rise together across many regions of the world and share the same modifiable risk factors," said study co-authors Zeye Liu, M.D., Ph.D., and Yi Shi, M.D., Ph.D., of Peking University People's Hospital. "From a cardiovascular perspective, this means that reducing smoking and alcohol use could help lower the risk of both conditions at the same time."
Laxmi Mehta, M.D., FAHA, chair of the American Heart Association's Council on Clinical Cardiology, who was not involved in the study, underscored the significance: "Many of the same modifiable factors, including smoking, alcohol use, poor diet, physical inactivity and obesity, contribute to both breast cancer (搜索) and cardiovascular disease including atrial fibrillation (搜索)/flutter, as confirmed by this study's findings. This overlap underscores the importance of integrated lifestyle strategies to reduce risk of cardiovascular disease and cancer."
The newly developed spatial risk maps can help guide region-specific prevention strategies and provide precise targets for future research. The authors note that the next step will be to incorporate long-term research, genetic and metabolic data into the analysis, and to expand the study to include socioeconomic factors for more individualized and region-specific prevention strategies.
Study limitations
The researchers acknowledge that the ecological design cannot establish individual-level causality or confirm that women with one condition developed the other. Differences in disease reporting and diagnosis across countries may have influenced incidence estimates, particularly in lower-resource settings. Additionally, because the country-level data could not identify women with both conditions, separate models were used for BC and AF in the Double Machine Learning analysis.