GRAPHICAL ABSTRACT
INTRODUCTION
Cancer has become one of the most significant public health challenges of this century, threatening to surpass cardiovascular disease as the leading cause of premature death from noncommunicable diseases worldwide [1]. It is a multifactorial disease that predominantly affects individuals over the age of 50. However, the incidence of several types of cancer among adults younger than 50 years has been rising in many regions across the globe [2].
The term ‘late-onset’ is typically applied to cancers diagnosed at or after the age of 50, whereas ‘early-onset’ refers to cancers occurring in adults under 50 years of age [3]. Previous studies have reported that early-onset cancers often exhibit distinct characteristics, including later stage at diagnosis and greater aggressiveness, leading to a sharper increase in the cancer burden among younger populations [4]. This pattern has been linked to modern dietary and lifestyle factors that contribute to obesity and elevated blood glucose levels [5].
Metabolic risk factors, including high fasting plasma glucose (FPG) and high body mass index (BMI), are important modifiable determinants that substantially contribute to morbidity and mortality in early-onset cancers [6]. Adolescence and young adulthood represent critical developmental periods that present both unique challenges and opportunities for improving metabolic health [2]. High FPG and high BMI are associated with metabolic abnormalities, insulin resistance, and chronic inflammation, all of which may promote tumor initiation, progression, and metastasis [7].
Understanding the epidemiological patterns of these metabolic risk factors is crucial for elucidating the metabolic underpinnings of early-onset cancers. Such insights can guide the development of early screening, prevention, and intervention strategies, as well as the optimization of personalized treatment approaches to enhance therapeutic efficacy [8].
Previous research on early-onset cancers has primarily focused on overall cancer burden across all age groups and risk factors [5,9] or has briefly discussed associations between certain risk factors and specific early-onset cancers [10,11]. However, the global burden of early-onset cancers attributable specifically to metabolic risk factors remains insufficiently characterized. To address this gap, the present study utilized the latest Global Burden of Disease (GBD) 2021 data, stratified by age subgroup, sex, and sociodemographic index (SDI), to examine global trends in mortality and disability-adjusted life years (DALYs) attributable to high BMI and high FPG among persons aged 20–49 years.
Additionally, we projected future trends through 2040.
METHODS
Data source, data collection, and definitions
We obtained GBD 2021 estimates of cancer-related mortality and DALY burden among persons aged 20–49 years from the Global Health Data Exchange query tool (https://vizhub.healthdata.org/gbd-results/). Data were extracted from the GBD 2021 study for four key metrics attributable to metabolic risk factors: the number of deaths, the number of DALYs, the age-specific mortality rate, and the age-specific DALY rate. All metrics were obtained in accordance with the definitions provided in the GBD Data & Tools Guide, Appendix 3 [12].
Early-onset cancers were defined as those diagnosed before the age of 50 years. However, the GBD 2021 data used in the present analysis were stratified according to the age at which the health outcome occurred rather than the age at cancer diagnosis. Accordingly, to avoid conflating age at diagnosis with attained age at outcome, all study-specific analyses and results are described as cancer-related deaths and DALYs occurring among persons aged 20–49 years, corresponding to the available GBD age-specific estimates.
The nine cancer types (listed in Supplemental Table S1) were selected according to two a priori criteria: (1) sufficient data density within the GBD 2021 database for persons aged 20–49 years and (2) the presence of well-characterized etiologic links between metabolic risk factors and cancer incidence, as specified in the GBD 2021 comparative risk assessment framework. Cancer cases were mapped to the International Classification of Diseases, 10th Revision codes provided in Supplemental Table S1. We initially queried GBD 2021 estimates for persons aged 15–49 years.
However, no estimates were available for the 15–19-year age group for the selected risk–cancer outcome combinations. Consequently, the aggregate estimates returned for the 15–49-year age range were numerically equivalent to those for persons aged 20–49 years. The 15–19-year age group therefore did not contribute to the reported aggregate burden estimates.
To accurately reflect the age coverage of the available data, we describe the analyzed population as persons aged 20–49 years throughout the manuscript.
The SDI is a composite indicator of development that integrates per capita income, mean years of education, and fertility rates. Countries and territories were categorized annually into time-varying SDI quintiles—low, low-middle, middle, highmiddle, and high—according to the GBD 2021 classification. High FPG and high BMI were the two metabolic risk factors examined in this study.
Although the GBD 2021 comparative risk assessment framework quantifies six metabolic risk factors (high FPG, high BMI, high low-density lipoprotein cholesterol, high systolic blood pressure, kidney dysfunction, and low bone mineral density) [12], complete age-specific estimates for most of these risk factors were unavailable or insufficiently stable to support trend modeling for persons aged 20–49 years. In contrast, high FPG and high BMI have fully disaggregated, age-specific risk–outcome pairs for all incident cancers within this age group, providing sufficient statistical power for robust longitudinal analyses. Therefore, this investigation focused specifically on high FPG and high BMI.
In addition to the risk-specific estimates for high FPG and high BMI, the overall burden attributable to metabolic risks was obtained directly from the GBD 2021 parent-level “Metabolic risks” estimate. For the cancer outcomes included in this analysis, the GBD risk hierarchy confirmed that this aggregate category comprised high FPG and high BMI. The aggregate estimate was obtained directly from GBD 2021 and was not calculated by summing the two risk-specific estimates.
Risk-attributable burden estimates were obtained directly from the GBD 2021 comparative risk assessment framework. This framework estimates population-attributable fractions using population exposure distributions, risk–outcome relative risks, and theoretical minimum risk exposure levels. The applicable GBD methodology did not allow us to determine whether a cancer-specific latency period between high BMI or high FPG exposure and subsequent cancer outcomes was explicitly incorporated into the estimates used in this analysis.
We therefore did not independently assign or model an exposure lag.
This study does not contain personal or medical information about identifiable living individuals, and animal subjects were not involved. The Institutional Review Board of the First Hospital of China Medical University determined that the study did not need approval because it used publicly available data.
Statistical analysis
Counts of deaths and DALYs, corresponding rates, and their 95% uncertainty intervals (UIs) were obtained from the GBD 2021 study. Estimated annual percentage changes (EAPCs) and their 95% confidence intervals (CIs) were calculated by the authors based on annual rates from 1990 to 2021. The 95% UIs of GBD-derived estimates corresponded to the 2.5th and 97.5th percentiles of the ordered draw-level estimates.
Consistent with GBD analytical conventions, the lower bounds of 95% UIs for attributable burden estimates were truncated at zero when necessary because negative attributable fractions are not biologically meaningful [12]. Mortality rates, DALY rates, and EAPCs were used to characterize burden levels and temporal trends of cancers among persons aged 20–49 years. The EAPC was calculated as follows:
where t represents the calendar year and β represents the estimated annual change in the logarithm of the rate. The EAPC was calculated as 100[exp(β)-1]. The 95% CI of the EAPC was obtained by transforming the 95% CI of β, as follows:
Future mortality and DALY trends for 2020–2040 were projected using Bayesian age-period-cohort (BAPC) models [13] based on historical estimates from 1990–2019. The models were implemented using the BAPC package (version 0.0.36) and Integrated Nested Laplace Approximations (INLA) package (version 24.06.27) in R, with a Poisson likelihood, second-order random-walk priors for age, period, and cohort effects, and an iid overdispersion term. Population estimates for 2020–2040 were obtained from the Global Burden of Disease 2021 database through the Global Health Data Exchange.
Detailed model specifications are provided in the Supplementary Methods.
Rates reported for individual 5-year age groups were age-specific rates, whereas aggregate GBD estimates across persons aged 20–49 years were reported as crude rates. Age-standardized rates were used only for the BAPC projections presented in Fig. 1.
The EAPC was estimated through log-linear regression, which assumes a constant exponential trend over time. Although this provides a concise summary measure, it may not capture non-linear changes or abrupt inflections due to interventions, policy reforms, or diagnostic improvements. These projections represent statistical extrapolations of historical trends and should be interpreted as scenarios assuming continuation of current trajectories rather than as deterministic forecasts.
All statistical analyses were performed using R software version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria). It should also be noted that GBD-derived risk-attributable burden estimates reflect statistical associations based on population-attributable fractions and do not imply direct causation. Further methodological details are provided in the Supplemental Methods.
RESULTS
Global patterns
Both the numbers and rates of deaths and DALYs attributable to the two assessed metabolic risk factors (high FPG and high BMI) increased over the past 30 years. In 2021, an estimated 34,112 deaths (95% UI, 15,037 to 54,888) and 1,691,418 DALYs (95% UI, 762,923 to 2,698,657) were attributable to these two risks worldwide. The global crude mortality rate was 0.9 per 100,000 persons aged 20–49 years (95% UI, 0.4 to 1.4), and the corresponding crude DALY rate was 42.8 per 100,000 persons aged 20–49 years (95% UI, 19.3 to 68.3).
From 1990 to 2021, both mortality and DALY rates attributable to the combined metabolic risks increased, with EAPCs of 1.35% (95% CI, 1.26 to 1.44) and 1.31% (95% CI, 1.23 to 1.40), respectively (Table 1).
In 2021, among the two metabolic risk factors examined, high BMI was associated with higher mortality and DALY rates than high FPG (mortality: 0.6 [95% UI, 0.4 to 0.8] vs. 0.3 [95% UI, 0.0 to 0.6] per 100,000; DALY rate: 29.4 [95% UI, 18.1 to 41.9] vs. 13.7 [95% UI, 0.9 to 26.6] per 100,000).
Over the past three decades, mortality and DALY rates attributable to high FPG (EAPCs: 1.26% [95% CI, 1.18 to 1.35] and 1.25% [95% CI, 1.17 to 1.33], respectively) and high BMI (EAPCs: 1.42% [95% CI, 1.33 to 1.52] and 1.37% [95% CI, 1.29 to 1.45], respectively) showed consistent increases (Supplemental Tables S2-S5).
Among cancers occurring in persons aged 20–49 years in 2021, colon and rectal cancers exhibited the highest mortality and DALY rates attributable to high FPG and high BMI (0.3 per 100,000 [95% UI, 0.1 to 0.4] and 12.3 per 100,000 [95% UI, 5.6 to 18.7], respectively). Throughout the study period, all mortality and DALY rates attributable to these metabolic risks increased, with liver cancer showing the most pronounced growth (EAPCs: 2.67% [95% CI, 2.47 to 2.87] and 2.61% [95% CI, 2.42 to 2.80], respectively) (Tables 2, 3 and Supplemental Tables S6, S7).
Global patterns by sex
In 2021, males exhibited higher mortality and DALY rates attributable to high FPG and high BMI than females (mortality: 1.0 [95% UI, 0.4 to 1.6] vs. 0.8 [95% UI, 0.3 to 1.2] per 100,000; DALY: 48.2 [95% UI, 21.0 to 77.9] vs. 37.4 [95% UI, 17.5 to 57.2] per 100,000).
The mortality EAPC was approximately 9% higher in males than in females (1.40% [95% CI, 1.31 to 1.50] vs. 1.28% [95% CI, 1.19 to 1.38]), whereas the DALY-rate EAPC was approximately 8% higher in males (1.36% [95% CI, 1.28 to 1.44] vs. 1.26% [95% CI, 1.17 to 1.34]) (Table 1).
The cancer burden among persons aged 20–49 years attributable to high BMI was greater than that due to high FPG in both sexes, with corresponding EAPCs reflecting faster growth for high BMI. However, for cancer burden attributable to high FPG, the number of deaths among females surpassed that among males by 2021—a reversal of the 1990 pattern—driven by a more rapid increase in female mortality (EAPCs: 1.38% [95% CI, 1.31 to 1.46] vs. 1.14% [95% CI, 1.02 to 1.26]) (Supplemental Table S8).
A similar sex-specific trend was observed for DALY rates (Supplemental Tables S2-S5).
Global patterns by age
Between 1990 and 2021, mortality and DALY rates increased across all age groups. In 2021, the absolute numbers of deaths and DALYs and the corresponding mortality and DALY rates increased with age and were highest among persons aged 45–49 years. In contrast, the fastest temporal increases were observed among those aged 30–34 years, with EAPCs of 0.90% (95% CI, 0.80 to 0.99) for mortality and 0.93% (95% CI, 0.84 to 1.03) for the DALY rate.
These EAPCs were approximately 2.6 times those observed among persons aged 20–24 years (0.34% [95% CI, 0.31 to 0.37] for mortality and 0.36% [95% CI, 0.33 to 0.39] for the DALY rate) (Table 1, Supplemental Fig. S1). Age-specific mortality and DALY trends from 1990 to 2021 are presented in Supplemental Figs.
S1, S2.
Specifically, a high BMI was the primary factor contributing to early-onset cancers across various age groups. For a high BMI, the mortality growth trend was the highest in the 30–34- and 45–49-year-old age groups and was approximately three times as high as that in the 20–24-year age group. With respect to high FPG, the mortality growth trend was most pronounced in the 30–34-year-old age group and was approximately two times that observed in the 40–44-year age group.
The pattern for DALYs mirrored this trend (Supplemental Tables S2-S5).
Global patterns by the SDI
In 2021, the numbers of deaths and DALYs were greatest in middle-SDI regions, the mortality rate in high-middle-SDI regions was approximately five times as high as that in low-SDI regions and the DALY rate showed a similar pattern (70.5 [95% UI, 30.4 to 113.3] vs. 14.3 [95% UI, 7.4 to 21.8] per 100,000). Temporal trends varied across SDI regions rather than showing a monotonic association with SDI.
The largest increases were observed in low-middle-SDI regions, with EAPCs of 2.47% (95% CI, 2.42 to 2.51) for mortality and 2.42% (95% CI, 2.37 to 2.46) for DALY rates, which were markedly higher than those observed in high-SDI regions (Table 1).
High BMI remained the predominant risk factor contributing to the burden of cancers among persons aged 20–49 years across all regions. However, when examining temporal growth trends, in low-middle-SDI regions, the increase in mortality attributable to high FPG (EAPC=2.60% [95% CI, 2.54 to 2.65]) exceeded that attributable to high BMI (EAPC=2.39% [95% CI, 2.23 to 2.41]). A similar reversal was observed for DALY rate trends (high FPG: EAPC=2.60% [95% CI, 2.54 to 2.65]; high BMI: EAPC=2.32% [95% CI, 2.23 to 2.41]).
For high BMI, the most pronounced increases in mortality and DALY rates were observed in middle-SDI countries. In contrast, for high FPG, mortality and DALY rates rose most rapidly in low-middle-SDI regions (Supplemental Tables S2-S5).
Projection of mortality and DALY rates to 2040
Fig. 1 illustrates global trends in age-standardized mortality and DALY rates for cancers among persons aged 20–49 years attributable to high BMI and high FPG, based on historical estimates from 1990 to 2019 and BAPC projections from 2020 to 2040. Mortality rates attributable to both high BMI and high FPG are projected to continue increasing.
Similarly, DALY rates attributable to both metabolic risks are projected to increase over the projection period. By 2040, the age-standardized DALY rates for cancers among persons aged 20–49 years attributable to high BMI and high FPG are projected to reach 46.0 and 28.8 per 100,000 population, respectively.
DISCUSSION
This study provides a comprehensive analysis of global cancer-related mortality and DALY burden attributable to high FPG and high BMI among persons aged 20–49 years. Over the past three decades, the burden of cancers among persons aged 20–49 years linked to these metabolic risks has increased not only in terms of deaths and DALYs but also in mortality and DALY rates. The differential patterns by sex, age, and socioeconomic level highlight the critical importance of integrating strategies for glycemic control and weight management into targeted preventive interventions for at-risk populations [5,14].
It is imperative to emphasize that these GBD estimates, derived from population-attributable fractions, reflect statistical associations and do not establish direct causation. While the biological plausibility is strong, the observed burdens signify the potential reduction in cancer cases that might be achieved if exposure to these metabolic risks was minimized.
Over recent decades, despite improved understanding of the mechanisms linking obesity and metabolic dysfunction to cancer, high BMI has continued to rise globally. Currently, approximately 40% of adults and 16% of children are classified as obese or overweight, representing a steady increase over the past 40 years [15]. The growing prevalence of childhood obesity, coupled with the tendency for obese children to remain obese as adults, suggests that more individuals will experience prolonged exposure to obesity-related cancer-promoting conditions.
Cumulative evidence indicates that longer exposure duration is associated with higher cancer incidence [16]. Within the analyzed population, older age groups exhibited higher mortality from cancers attributable to high BMI, and with global aging and increasing life expectancy, absolute numbers of deaths and DALYs related to BMI-linked cancers have risen substantially. This trend correlates with increased sedentary behavior, disruptions in healthcare, and elevated stress levels [17].
The observed historical shift from a male-predominant to a female-predominant burden of high-FPG-related cancers is particularly concerning and suggests a rapidly evolving, sex-specific metabolic risk landscape. Within the GBD framework, sex-specific relative risks are incorporated where available [12]. The observed sex-specific pattern may partly reflect differences in exposure distributions and metabolic risk profiles, as previous studies have reported sex differences in glucose metabolism and diabetes-related risk factors [18,19].
However, the relative contributions of exposure prevalence, biological susceptibility, and other contextual factors cannot be determined from the present population-level analysis. Potential contributors reported in previous studies include socioeconomic circumstances, behavioral and reproductive factors, and lifestyle-related factors [19]. Nonetheless, given the overlapping UIs for sex-specific rates, these interpretations should be made cautiously and underscore the need for future research into the evolving determinants of sex-specific metabolic risk exposure.
Supporting this interpretation, a systematic review and meta-analysis found that cancer incidence among patients with diabetes was 70% higher than among those without diabetes [20]. Similarly, a study of the global burden of early-onset cancers attributed to high FPG reported that crude death rates increased with age [21]. Higher mortality and DALY rates in older age groups may be explained by the greater prevalence of diabetes and its comorbidities, which elevate both cancer incidence and mortality risk [22].
Marked disparities in metabolic risk-attributable cancer burden were observed across SDI quintiles. Importantly, the pattern differed according to the burden metric considered: middle-SDI regions accounted for the largest absolute numbers of deaths and DALYs, high-middle-SDI regions had the highest mortality and DALY rates, whereas low-middle-SDI regions experienced the fastest temporal increases. These contrasting patterns highlight the importance of distinguishing current burden levels from the pace of change when interpreting SDI-related differences.
Factors such as nutritional transition, urbanization, and access to preventive healthcare may provide context for these patterns, although their contributions were not directly evaluated in the present study [23,24]. These regional differences may help inform future research on context-specific prevention strategies, particularly in settings experiencing rapid increases in metabolic risk [25].
These findings indicate that the modeled cancer burden attributable to high BMI and high FPG among persons aged 20–49 years has increased over time. However, because the analysis was based on population-level GBD estimates, it cannot identify individual patients at high clinical risk or independently establish that these exposures directly caused cancer outcomes. These links may be mediated through mechanisms such as gut microbiome dysbiosis [26] and metabolic dysfunction-associated steatotic liver disease-related hepatocarcinogenesis [27].
The results also identify young adults with obesity or hyperglycemia as a distinct, high-risk clinical population warranting earlier risk assessment and preventive intervention.
Ultimately, the increasing burden of cancers among persons aged 20–49 years attributable to high BMI and high FPG highlights the potential importance of incorporating metabolic health considerations into cancer prevention research. A life-course perspective addressing metabolic risk from adolescence through adulthood may provide a useful framework for future prevention strategies [28,29]. However, the present study did not evaluate the effectiveness, feasibility, or cost-effectiveness of specific interventions, including metabolic risk screening, early cancer detection strategies, or targeted prevention programs.
Therefore, our findings should not be interpreted as evidence supporting any specific screening or intervention approach. Future studies are needed to determine whether targeted metabolic risk management and appropriate prevention strategies can reduce cancer burden among persons aged 20–49 years, particularly in populations experiencing rapid increases in metabolic risk.
This study has several limitations inherent to the data sources and methodological framework of the GBD study [5]. First, our findings depend on the accuracy and completeness of the GBD 2021 input data. Although the GBD framework strives for global comprehensiveness, data quality and availability remain uneven across regions, particularly in low-SDI countries.
These discrepancies may lead to underestimation or overestimation of the true burden in certain regions and affect the overall generalizability of our estimates. In addition, estimates for persons aged 15–19 years were unavailable for the selected risk–cancer outcome combinations in GBD 2021. Therefore, the quantitative findings do not represent the entire population younger than 50 years and should be interpreted as applying to persons aged 20–49 years.
Moreover, the age stratification in GBD reflects the age at which deaths or health loss occurred rather than the individual age at cancer diagnosis. Second, our analysis was necessarily restricted to cancer types with robust data and well-established metabolic risk–outcome relationships within the GBD framework. Consequently, other significant metabolic risk factors prevalent in younger populations—such as hypertension, dyslipidemia, and renal dysfunction—were not included.
Likewise, the analysis was limited to nine cancer types and did not encompass other high-burden cancer types relevant to younger adults which may also be influenced by metabolic factors. Third, the analytical models used in this study have inherent methodological constraints. The EAPC model, which assumes log-linearity, may smooth over short-term fluctuations or inflection points in temporal trends.
More importantly, the BAPC projections are based on historical associations and do not incorporate potential future paradigm shifts in metabolic health management, cancer prevention, or treatment. Thus, the projections represent a ‘status quo’ scenario and likely underestimate the potential impact of coordinated global interventions addressing metabolic risk factors. As with all GBD comparative risk assessments, our findings represent population-level attributable-burden estimates and do not establish individual-level causality.
In addition, we could not determine from the applicable GBD methodology whether a cancer-specific latency period between exposure to high BMI or high FPG and subsequent cancer development or outcomes was explicitly incorporated into the estimates used in this study. We did not independently model such latency. Consequently, the attributable burden reported here should not be interpreted as indicating that contemporaneous BMI or FPG exposure directly caused cancer deaths or DALYs occurring in the same period.
The timing of exposure relative to cancer development and subsequent outcomes could not be assessed using the aggregated GBD data. Despite these limitations, the GBD study remains the most comprehensive and standardized global effort to quantify health loss. Our analysis contributes important evidence on the substantial and rising burden of cancers among persons aged 20–49 years attributable to two key, modifiable metabolic risk factors.
In conclusion, cancer-related mortality and DALY burden attributable to high BMI and high FPG increased among persons aged 20–49 years from 1990 to 2021. Burden levels and temporal changes varied by sex, age, and SDI. These findings underscore the urgent need for integrated metabolic and oncologic prevention strategies.
Policymakers and healthcare systems should prioritize interventions targeting obesity and hyperglycemia, particularly in young populations, to mitigate the future burden of cancers among persons aged 20–49 years and enhance early cancer detection.