Global patterns and risk factors of breast cancer in women under 35: a population-based study.

1
Introduction
Young breast cancer (YBC) is a distinct and clinically significant subset of breast cancer, characterized by unique biological and epidemiological features [1,2]. Despite representing a smaller proportion of overall cases, YBC is the most prevalent and lethal malignancy among young adults globally, with significant geographical variations in incidence and characteristics [3]. For instance, epidemiological data from 2019 indicate that only 4% of newly diagnosed breast cancer cases in the United States occurred in patients under 40 years of age [4]. In contrast, a 2017 study in China revealed a significantly higher prevalence, with 16.4% of new cases diagnosed in this younger population cohort [5].
This study defines YBC patients as those under 35 years of age, a definition that differs from the European Society for Medical Oncology (ESMO) guideline's 40-year cutoff [6]. This distinction is supported by two key observations. First, a large-scale studie [7] has shown significantly higher mortality rates among patients under 35 compared to those aged 35–39 and older, with the 35–39 age group exhibiting survival rates similar to older patients. Second, contemporary clinical guidelines, such as the National Comprehensive Cancer Network (NCCN) Adolescent and Young Adult (AYA) Oncology Guidelines and the ESO-ESMO fifth international consensus (BCY5), recognize age under 40 as a broad category for specialized management, while also noting that treatment intensification (e.g., ovarian suppression in hormone receptor-positive disease) is specifically indicated for women under 35 based on trial stratification (e.g., TEXT/SOFT) [8,9]. Recent prognostic models and regional consensus further underscore the distinct clinicopathological and outcome profiles of breast cancer diagnosed before age 35 [10,11].Thus, although age 40 remains a commonly used cutoff in practice, our use of <35 years aims to define a more homogeneous, highest-risk subgroup for assessing its unique global burden.
YBC presents unique clinical challenges characterized by distinct biological and prognostic features. It is associated with a higher prevalence of aggressive tumor phenotypes [12], including triple-negative breast cancer, and poorer prognoses regardless of disease stage [13,14]. These clinical challenges are compounded by the underrepresentation of young women in studies evaluating risk-stratification models and molecular tools, underscoring the need for tailored treatment strategies [15,16]. Furthermore, genetic predisposition plays a significant role in YBC development, with young patients showing higher rates of familial breast cancer and pathogenic variants in genes such as BRCA1, BRCA2, and CHEK2 [17]. Clinical guidelines recommend enhanced surveillance, including annual MRI, for high-risk carriers to improve early detection rates [9]. As genetic testing becomes increasingly accessible, especially in higher SDI regions, more young women are identified as high-risk and enter surveillance programs—a factor that likely contributes to rising YBC incidence rates and must be considered when interpreting global burden trends. However, significant geographical variations in YBC epidemiology and outcomes persist, highlighting the necessity of a comprehensive global analysis of trends and influencing factors to refine current strategies and clinical guidelines.
Given the rising burden of YBC and limited comprehensive global data, our study utilizes the 2021 Global Burden of Diseases data to evaluate incidence, mortality, and modifiable risk factors in YBC under 35. The objective is to inform effective interventions and policies that reduce the impact of YBC on young populations worldwide.

2
Methods
2.1
Study design and data sources
This population-based study analyzed the global burden of YBC, defined as breast cancer diagnosed in individuals under 35 years of age. The study adhered to STROBE and GATHER reporting guidelines and employed data from the 2021 Global Burden of Disease (GBD) Study, released in 2024.
Data were accessed via the Global Health Data Exchange (GHDx) online query tool (http://ghdx.healthdata.org/gbd-results-tool). The GBD 2021 breast cancer estimates, which form the basis of our analysis, are defined by specific International Classification of Diseases (ICD) codes and are modeled to represent invasive breast cancer only. The GBD 2021 dataset includes comprehensive epidemiological data (incidence, prevalence, mortality, YLDs, YLLs, and DALYs) for 204 countries and territories from 1990 to 2021, encompassing 371 diseases and injuries. Data were collected from multiple sources, including epidemiological surveys, hospital records, vital registration systems, and disease monitoring systems (https://ghdx.healthdata.org/gbd-2021/sources). Data standardization utilized International Classification of Diseases (ICD) codes, and complex modeling (DisMod-MR and spatiotemporal Gaussian process regression) generated estimates, incorporating uncertainty analyses via Monte Carlo simulations (https://www.healthdata.org/gbd/methods-appendices-2021/cancers).
Ethical considerations: The University of Washington Institutional Review Board approved a waiver of informed consent for GBD data; this study adheres to STROCSS reporting guidelines.

2.2
Measures and definitions
Disease burden was assessed using age-standardized incidence (ASIR), prevalence (ASPR), mortality (ASMR), and disability-adjusted life-year (DALY) (ASDR) rates. These rates, calculated per 100,000 population using the GBD 2021 standard population (Supplementary Table S1), facilitated comparisons across populations. The metrics represent, respectively: annual new cases, the proportion of individuals with the disease at a given time, annual deaths, and the combined burden of premature mortality and years lived with disability. The average annual percent change (AAPC) is used to assess temporal trends in age-standardized rates (ASRs) of diseases and the changing proportion of YBC among all breast cancer cases.

2.3
Geographic and socioeconomic context
Geographic analysis utilized the GBD 2021 hierarchical regional classification, which groups 204 countries and territories into seven super-regions and 21 regions based on geographic proximity and epidemiological similarity: (1) Central Europe, Eastern Europe, and Central Asia; (2) high-income regions including Australasia, high-income Asia Pacific, high-income North America, Southern Latin America, and Western Europe; (3) Latin America and the Caribbean regions including Andean Latin America, Caribbean, Central Latin America, and Tropical Latin America; (4) North Africa and the Middle East; (5) South Asia; (6) Southeast Asia, East Asia, and Oceania; and (7) Sub-Saharan Africa, including Central Sub-Saharan Africa, Eastern Sub-Saharan Africa, Southern Sub-Saharan Africa, and Western Sub-Saharan Africa. Socioeconomic status was assessed using the Socio-demographic Index (SDI), a composite index incorporating fertility rate, education level, and per capita income. Using four cutoff values of 0.47, 0.62, 0.71, and 0.81, countries and regions were categorized into five SDI quintiles (low, low-middle, middle, high-middle, high) to analyze the association between socioeconomic status and burden of YBC (https://ghdx.healthdata.org/record/global-burden-disease-study-2021-gbd-2021-socio-demographic-index-sdi-1950%E2%80%932021) (Supplementary Table S2).

2.4
Risk factor analysis
The GBD 2021 identifies 88 risk factors categorized as environmental/occupational, behavioral, and metabolic. Among these, tobacco, low physical activity, high fasting plasma glucose, dietary risks, alcohol use, and high body-mass index (BMI) are the only modifiable risk factors for YBC. This study reports YBC mortality and DALYs associated with these factors, utilizing the GBD comparative risk assessment framework. We employed advanced methods such as DisMod-MR 2.1 and ST-GPR to model exposure distributions across populations and regions. The theoretical minimum risk exposure level (TMREL) was defined for each factor to determine optimal exposure for minimizing YBC risk. By integrating exposure data, relative risk estimates, and TMRELs, we calculated population attributable fractions (PAFs) for each risk factor, stratified by region. This analysis underscores the significance of modifiable risk factors and provides insights for targeted interventions. Detailed methodologies for risk attribution are discussed elsewhere.

2.5
Statistical analysis
Age-standardized rates (ASIR, ASPR, ASMR, ASDR) were analyzed at global, regional, and national levels. Truncated ASR per 100,000 was calculated using the formula:
 × 100, 000 (where is age-specific rate for the -th age group, is the weight of the -th age group in the standard population; and A is the total number of age groups).
The AAPC is used to assess temporal changes in ASRs and proportions of YBC. The coefficient is derived from the natural logarithm of the ASRs, with representing and corresponding to calendar years. The AAPC, along with its 95% confidence interval (CI), was estimated using the following linear regression model: ; . A positive AAPC indicates an upward trend, a negative AAPC signals downward trend, and an AAPC with a 95% CI encompassing 0 indicates a stable trend. Bayesian Age-Period-Cohort (BAPC) models, implemented within the Integrated Nested Laplace Approximations (INLA) framework were used to project age-standardized rates of YBC from 2022 to 2050. These projections utilized Global Population Forecasts (2017–2100) data and the INLA and BAPC packages available at https://r-forge.r-project.org/and https://www.r-inla.org/,respectively. Additionally, Spearman's correlation analysis assessed the relationship between ASRs and SDI index. All analyses were performed using R software (version 4.3.2) and Python (version 3.12.7), with statistical significance set at P < 0.05.

3
Results
3.1
Global estimates of YBC burden in 2021
In 2021, the global burden of YBC across 204 countries and territories (Supplementary Table S2) was substantial. An estimated 81,856 new cases occurred worldwide (95% UI, 76,458–87,253), accompanied by 687,819 prevalent cases (95% UI, 642,400–733,238), 18,597 deaths (95% UI, 17,191–20,003), and 1,178,721 DALYs (95% UI, 1,090,902–1,266,541). The corresponding age-standardized rates per 100,000 population were as follows: ASIR, 1.7 (95% UI, 1.5–1.9); ASPR, 14.3 (95% UI, 13.0–15.9); ASMR, 0.4 (95% UI, 0.3–0.4); and ASDR, 24.6 (95% UI, 21.7–27.9).
Fig. 1 and Table 1 demonstrate the regional distribution of ASIR, ASPR, ASMR, and ASDR in 2021. Among the five SDI quintiles, the high SDI quintile had the highest ASIR (2.3; 95% UI, 2.2-2.5) and ASPR (20.3; 95% UI, 19.3–21.4), while the low-middle SDI region exhibited the highest ASMR (0.5; 95% UI, 0.4–0.6) and ASDR (32.9; 95% UI, 0.4–0.6). In the 21 GBD regions, High-Income North America reported the highest ASIR (2.7; 95% UI, 2.5–2.9) and ASPR (23.6; 95% UI, 22.0-25.2). The Oceania region recorded the highest ASMR (0.8; 95% UI, 0.5-1.2) and ASDR (49.4; 95% UI, 32.0-74.0) (Fig. 1A). The association between these rates and SDI is illustrated in Fig. 1B. A positive linear relationship exists between SDI and both ASIR (R = 0.47, P < 0.001) and ASPR (R = 0.51, P < 0.001), suggesting higher SDI correlates with increased ASIR and ASPR of YBC. However, a weak but significant negative correlation was observed between SDI and ASMR (R = −0.25, P < 0.001) and ASDR (R = −0.22, P < 0.001).
At the national level, Monaco had the highest ASIR (7.3; 95% UI, 4.2-12.1) and ASPR (64.0; 95% UI, 37.5-104.8), while Tokelau recorded the highest ASMR (1.6; 95% UI, 1.0-2.4) and ASDR (103.5; 95% UI, 65.0-154.1) among 204 countries and territories analyzed (Fig. 2, Supplementary Table S3).
The male proportion of YBC cases was 1.44% for incidence and 1.36% for Mortality. For both genders, YBC incidence, prevalence, mortality, and DALYs increased with age, showing consistent trends across sexes (Supplementary Table S4).

3.2
Global trends of YBC from 1990 to 2021
Between 1990 and 2021, significant increases were observed in ASIR, ASPR, ASMR, and ASDR. Specifically, the AAPC for ASIR was 1.3% (95% CI, 1.2-1.4), rising from 1.2 to 1.7. Similarly, ASPR increased with an AAPC of 1.3% (95% CI, 1.2-1.4), from 9.7 to 14.3. The ASMR showed a modest increase, with an AAPC of 0.3% (95% CI, 0.2-0.4), rising from 0.35 to 0.39, while ASDR exhibited an AAPC of 0.4% (95% CI, 0.3-0.5), increasing from 21.6 to 24.6. Notably, ASIR and ASPR demonstrated consistent upward trends during the periods of 1990-2000, 2001-2010, and 2011-2021. In contrast, ASMR declined from 1990 to 2000, stabilized from 2001 to 2010, and subsequently increased from 2011 to 2021. Likewise, ASDR remained stable from 1990 to 2010 before undergoing a significant rise during the period from 2011 to 2021 (Fig. 3A and B).
Low-middle SDI quintiles experienced the most pronounced increases in ASIR (AAPC, 2.5%; 95% CI, 2.4-2.5), ASPR (AAPC, 2.6%; 95% CI, 2.6-2.7), ASMR (AAPC, 1.3%; 95% CI, 1.2-1.4), and ASDR (AAPC, 1.3%; 95% CI, 1.3-1.4). Conversely, high SDI quintiles exhibited the smallest AAPCs for ASIR (0.1%; 95% CI, −0.02 to 0.2), ASPR (0.2%; 95% CI, 0.03-0.3), ASMR (−1.5%; 95% CI, −1.7 to −1.3), and ASDR (−1.4%; 95% CI, −1.6 to −1.2) (Fig. 3A and Supplementary Table S5). Among 21 geographic regions, North Africa and the Middle East reported the most substantial increases in ASIR (AAPC, 3.5%; 95% CI, 3.4-3.7), ASPR (AAPC, 3.6%; 95% CI, 3.4-3.7), ASMR (AAPC, 1.6%; 95% CI, 1.4-1.7), and ASDR (AAPC, 1.7%; 95% CI, 1.6-1.9). In contrast, Central Asia witnessed notable declines in ASIR (AAPC, −0.7%; 95% CI, −0.8 to −0.5) and ASPR (AAPC, −0.6%; 95% CI, −0.7 to −0.4), while Australasia experienced the most significant reductions in ASMR (AAPC, −2.3%; 95% CI, −2.5 to −2.1) and ASDR (AAPC, −2.1%; 95% CI, −2.3 to −1.9) (Fig. 3C and Supplementary Table S5).
Among the 204 countries and territories, changes in ASIR, ASPR, ASMR, and ASDR are presented in Fig. 4 and Supplementary Table S6.

3.3
Global trends of YBC proportion from 1990 to 2021
Between 1990 and 2021, YBC proportions within overall breast cancer significantly decreased, with AAPCs of −0.4% (95% CI: 0.5 to −0.4) for incidence, −0.5% (95% CI: 0.5 to −0.4) for prevalence, −0.51% (95% CI: 0.54 to −0.47) for mortality, and −0.40% (95% CI: 0.43 to −0.37) for DALYs. However, these proportion changes varied across different SDI quintiles and geographical regions. The middle SDI quintile saw the most significant declines, with AAPCs of −1.6% (95% CI: 1.7 to −1.5) for incidence, −1.7% (95% CI: 1.8 to −1.6) for prevalence, −2.0% (95% CI: 2.1 to −1.9) for mortality, and −1.75% (95% CI: 1.8 to −1.7) for DALYs. In contrast, the low SDI quintile experienced an increase in YBC proportions. Among the 21 GBD regions, Southern Sub-Saharan Africa had the largest reduction in incidence proportion, with an AAPC of −3.1% (95% CI: 3.9 to −2.4). The High-income Asia Pacific region recorded notable proportion declines with AAPCs of −3.0% (95% CI: 3.3 to −2.7) for prevalence, −4.3% (95% CI: 4.5 to −4.0) for mortality, and −3.3% (95% CI: 3.5 to −3.1) for DALYs (Supplementary Fig. S1 and Supplementary Table S7). Across 204 countries and territories, the Northern Mariana Islands reported the greatest proportion reductions in incidence, prevalence, and mortality, with AAPCs of −6.0% (95% CI: 7.3 to −4.8), −6.3% (95% CI: 7.6 to −5.1), and −6.6% (95% CI: 7.8 to −5.3), respectively. Saint Kitts and Nevis showed the largest decline in DALYs proportion, with an AAPC of −5.9% (95% CI: 7.0to −4.8) (Supplementary Fig. S2 and Supplementary Table S8).

3.4
Future forecasts of global burden of YBC
Globally, the absolute change in ASIR of YBC is projected to increase by 0.69 per 100,000, rising from 1.7 (95% UI: 1.5-1.9) in 2021 to 2.40 (95% UI: 1.6-3.1) by 2050 (Fig. 5A). Meanwhile, the ASPR is anticipated to increase by 2.2, from 14.3 (95% UI: 13.0–15.9) to 16.6 (95% UI: 9.1-23.7) per 100,000 globally (Fig. 5B). The ASMR and ASDR are expected to rise by 0.10, from 0.39 (95% UI: 0.3–0.4) to 0.5 (95% UI: 0.3–0.7) (Fig. 5C), and by 7.3, from 24.6 (95% UI: 21.7–27.9) to 31.9 (95% CI: 1.0–15.6) (Fig. 5D), respectively.

3.5
Risk factors [18] associated with mortality and DALYs of YBC
Fig. 6 illustrates the proportion of YBC mortality and DALYs attributable to risk factors in 2021. Globally, 17.89% of mortality and 17.99% of DALYs were linked to these risk factors. Dietary risks were the leading contributors to YBC deaths (11.0%, ranging from 6.9% to 13.7% across the 21 GBD regions) and DALYs (11.1%, ranging from 6.9% to 13.7%) in 2021. Other significant contributors included alcohol use, tobacco, high fasting plasma glucose, and low physical activity. Notably, a high body-mass index exhibited an inverse relationship with both mortality (−3.2%, ranging from −6.9% to −3.1%) and DALYs (−3.2%, ranging from −6.9% to −3.1%). The contribution of risk factors to YBC mortality and DALYs was highest in the high SDI quintile, accounting for 29.6% and 29.7% respectively. Among the 21 GBD regions, Australasia recorded the highest proportions of mortality and DALYs due to risk factors, both at 34.0% (Supplementary Table S9).

4
Discussion
To the best of our knowledge, this study presents the first comprehensive global analysis of YBC in patients under 35, revealing a growing global burden with significant disparities across SDI quintiles and geographic regions. From 1990 to 2021, ASIR increased universally, but ASMR trends diverged, rising in low-SDI but falling in high-SDI settings. Globally, YBC's proportion of all breast cancers decreased, except in the lowest SDI quintile. Dietary factors were a major driver of YBC mortality, while high BMI unexpectedly showed a protective association. These findings demonstrate a complex interplay of socioeconomic and established risk factors, necessitating context-specific interventions.
In 2021, a strong positive correlation existed between YBC ASIR and SDI. This correlation reflected significant regional and national variations in ASIR, mirroring global cancer burden trends [[19], [20], [21], [22], [23]]. High-SDI quintiles showed disproportionately higher ASIRs, consistent with global patterns in other young-onset cancers [3]. This disparity likely reflects several interacting factors: differences in genetic predisposition and risk factor exposures; variations in cancer detection and diagnostic capabilities; implementation of preventive measures; and challenges with data collection in low-SDI settings [24,25]. Our analysis also revealed significantly higher mortality rates for YBC in low-SDI quintiles compared to high-SDI quintiles. This unexpected finding, given lower incidence rates in low-SDI quintiles, highlights the urgent need for improved healthcare infrastructure and access to quality care in these regions. Effective prevention, early diagnosis, and treatment strategies are crucial for reducing YBC mortality in low-SDI settings [26].
Globally, breast cancer incidence has been rising overall, with an ASIR increasing at an AAPC of 1.57% from 1990 to 2021 [27]. In parallel, the incidence of young breast cancer (YBC) also showed an upward trend, though at a slower pace (AAPC = 1.27% annually). As a result, the proportion of YBC cases among all breast cancers declined worldwide, except in countries within the lowest SDI quintile. The global rise in YBC incidence may partly reflect evolving genetic testing practices. In high-SDI regions, expanding genetic testing identifies more young women with BRCA1/2 mutations who subsequently receive enhanced surveillance [28]. However, population-based screening typically begins at age 50 in most countries [29,30], meaning general screening doesn't affect YBC rates. Thus, reported incidence trends mainly reflect symptomatic detection and high-risk surveillance rather than population screening. These findings highlight how healthcare access and clinical practices differently influence YBC epidemiology across regions. Several factors likely contributed to the increased YBC incidence: improvements in screening and diagnostics (mammography in high-income countries, gradual screening implementation in middle-income regions) [[31], [32], [33], [34]]; changes in reproductive behaviors-such as delayed childbearing and reduced parity-which may extend the window of susceptibility to hormonal and environmental exposures over the life course [35,36]; as well as shorter breastfeeding duration in settings where early childbearing is less common, though its population-level impact may vary by region and socioeconomic context [37]. Furthermore, genetic susceptibility-particularly in carriers of high-penetrance genes such as BRCA1/2 and CHEK2-contributes to earlier onset and may amplify the detectability of YBC through intensified surveillance in identified high-risk individuals, thereby influencing incidence patterns especially in regions with greater genetic testing access [38]. These factors illustrate the multifaceted impact of healthcare improvements, lifestyle changes, and demographic transitions on global YBC incidence rates [39]. The observed decrease in the proportion of YBC globally is likely linked to population aging, particularly in high-SDI regions [40].
While high-SDI quintiles currently bear the greatest burden of YBC incidence and prevalence, low-SDI quintiles are experiencing faster growth rates. Furthermore, a striking difference emerged in the trends of ASMR and ASDR between high-SDI and low-SDI quintiles. In high-SDI quintiles, declining ASMR and ASDR for YBC likely reflect improved early detection, novel therapeutics, and advancements in treatment strategies, leading to better survival and reduced disability. In contrast, low- and lower-middle SDI regions experienced increases in ASMR and ASDR for YBC, underscoring the persistent lack of access to effective treatment and care [41].
Our projections suggest continued increases in ASIR, ASPR, ASMR, and ASDR for YBC, emphasizing the urgent and ongoing need for sustained efforts in prevention, early detection, and effective treatment.
Risk factor analysis identified dietary risks as the leading global contributor to YBC mortality and DALYs, followed by alcohol use, tobacco use, high fasting plasma glucose, and low physical activity. Dietary risks accounted for 11.04% of global YBC mortality in 2021 (13.68% in high-SDI quintiles), a finding in contrast to the established role of high BMI as a leading mortality risk factor in the overall breast cancer population [27]. These results highlight the importance of targeting modifiable risk factors in prevention strategies, with the substantial variation across SDI quintiles emphasizing the need for context-specific interventions. This study also reveals, for the first time globally in YBC, an unexpected inverse association between high BMI and YBC mortality, consistent with a previous large retrospective study [42]. This inverse association was consistent across all SDI quintiles. This association may be partially explained by the “obesity paradox” observed in some cancer populations. In the context of YBC, patients with higher BMI might exhibit better tolerance to systemic therapies such as chemotherapy. Emerging evidence suggests that body composition and nutritional status could influence pharmacokinetics, treatment completion rates, and thus overall survival in young patients receiving intensive regimens [43,44]. The complex relationship may also involve interactions between BMI, age, hormonal factors, and mammographic density [42]. However, this finding does not support weight gain as a preventive measure given the well-established health risks of obesity.
This study, while offering valuable insights, has several limitations. Data quality and availability vary considerably across regions, especially in low- and middle-income countries lacking robust cancer registries, potentially leading to underestimation of the YBC burden. The analysis is not exhaustive, omitting factors such as genetic predisposition and environmental exposures. The future projections presented are global-level estimates, and their predictive stability may not extend to sub-national or country-specific contexts due to uncertainties in local data and future trends. These projections are based on current trends and do not account for potential future changes in risk factors, technological advancements, or policy shifts. Finally, the study's methodology, although robust, is subject to inherent GBD limitations, including potential biases in data collection and modeling, and the use of disability weights derived from survey data [45,46]. Additionally, it should be noted that GBD estimates are modeled and may differ from directly observed national registry counts due to methodological harmonization across countries. Despite these limitations, the study provides valuable insights into the global YBC burden and informs future research and policy development. Cautious interpretation and acknowledgment of these limitations are essential.
In conclusion, this global study confirms a rising global burden of YBC, with substantial disparities across SDI quintiles and geographic locations. The observed trends underscore the urgent need for sustained and targeted interventions focused on modifiable risk factors, improvements in healthcare access and quality, particularly in low- and low middle-income countries, and enhanced efforts in prevention, early detection, and effective treatment for YBC. Further research is essential to fully elucidate the complex interplay of risk factors, including the unexpected inverse relationship between high BMI and mortality, and to develop effective, culturally sensitive, and resource-appropriate strategies for reducing the global YBC burden.

Contributors
DJ, JZ, DW, and JF are co-first authors and contributed equally to the work. DJ, HW, and ZL designed the study protocol and provided overall guidance. DJ, JZ, DW, and JF conducted data analysis and verified the underlying data. DJ and JF prepared the first draft and finalised the manuscript based on comments from all other authors. All other authors contributed to the review & editing of the current manuscript. All authors had full access to the data in the study and accepted responsibility for submitting this study for publication.

Data sharing statement
All data used in this study can be freely accessed at the GBD 2021 portal (http://ghdx.healthdata.org/gbd2021).

Declaration of competing interests
The author(s) declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.