Global burden of brain and central nervous system cancer among people aged 20–64 years, 1992–2021 and projections to 2050: a population-based study

Abstract

Objectives

To estimate the burden, trends, and inequalities of brain and central nervous system cancer (CNS cancer) among adults at global, regional and national level from 1992 to 2021.

Design

Population-based study.

Population

Adults aged 20–64 years from 21 regions and 204 countries and territories (Global Burden of Disease and Risk Factors Study 2021) from 1992 to 2021.

Main outcome measures

Our primary outcomes comprised age-standardised prevalence, incidence, mortality and disability-adjusted life-years (DALYs) for CNS cancers. The analytical framework incorporated temporal trend analysis through annual percentage change (APC) and average APC (AAPC) metrics, complemented by Bayesian age-period-cohort modelling to assess demographic influences. We employed predictive modelling with decomposition techniques to evaluate contributions from age structure shifts, population dynamics and risk factor modifications, while spatiotemporal Gaussian process regression enabled robust smoothing and trend estimation across continuous time-space dimensions. The study specifically applied frontier analysis methodologies to examine epidemiological patterns of prevalence, incidence, mortality and DALYs within the 20–64 years adult population.

Results

From 1992 to 2021, the global age-standardised prevalence (AAPC 1.04 (95% CI 0.91 to 1.18); p<0.001) and incidence (AAPC 0.23 (95% CI 0.14 to 0.31); p<0.001) of CNS cancer among adults aged 20–64 years increased, while age-standardised DALYs (AAPC −0.28 (95% CI −0.34 to −0.22); p<0.001) and mortality (AAPC −0.27 (95% CI −0.34 to −0.2); p<0.001) decreased. The most significant increase in prevalence (AAPC 1.67 (95% CI 1.47 to 1.87); p<0.001) and incidence (AAPC 0.7 (95% CI 0.61 to 0.79); p<0.001) was observed among those aged 20–24 years. The most significant decrease in DALYs (AAPC −0.45 (95% CI −0.55 to −0.35); p<0.001) and mortality (AAPC −0.46 (95% CI −0.56 to −0.36); p<0.001) was observed among those aged 40–44 years. The rate of increase in prevalence and incidence was lower in high sociodemographic index (SDI) (prevalence: AAPC 0.95 (95% CI 0.8 to 1.11); p<0.001), (incidence: AAPC 0.95 (95% CI 0.8 to 1.11); p<0.001) countries compared with low-SDI countries (prevalence: AAPC 0.5 (95% CI 0.43 to 0.57); p<0.001), (incidence: AAPC 0.12 (95% CI −0.07 to 0.31); p<0.001). DALYs and mortality began to decline in high-SDI countries (DALYs: AAPC −0.52 (95% CI −0.73 to −0.31); p<0.001), (mortality: AAPC −0.5 (95% CI −0.7 to −0.3); p<0.001), but these indicators continue to rise in low-SDI nations (DALYs: AAPC 0.45 (95% CI 0.39 to 0.51); p<0.001), (mortality: AAPC 0.38 (95% CI 0.31 to 0.46); p<0.001). Our predictive analysis found that from 2021 to 2050, the number of CNS cancer cases among people aged 20–64 years will be on the rise globally, which is expected to increase from 186 891 to 245 942, an increase of 31.6%.

From 1992 to 2021, the global age-standardised prevalence and incidence of CNS cancer among adults aged 20–64 years increased, while age-standardised DALYs and mortality decreased. The most significant increase in prevalence and incidence was observed among those aged 20–24 years. The most significant decrease in DALYs and mortality was observed among those aged 40–44 years. The rate of increase in prevalence and incidence was lower in high SDI countries compared with low-SDI countries. DALYs and mortality began to decline in high-SDI countries, but these indicators continue to rise in low-SDI nations. Our predictive analysis found that from 2021 to 2050, the number of CNS cancer cases among people aged 20–64 years will be on the rise globally, which is expected to increase from 186 891 to 245 942, an increase of 31.6%.

Conclusions

Significant inequalities exist in age-standardised prevalence, incidence, DALYs and mortality of CNS cancer among countries with varying sociodemographic indices. These disparities highlight the urgent need for targeted clinical guidelines and equitable distribution of global health resources.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study provides a comprehensive 30-year analysis of central nervous system (CNS) cancer burden across 204 countries using advanced statistical methods.

• It highlights age-specific and sociodemographic disparities in prevalence, incidence, mortality and disability-adjusted life-years.

• Predictive modelling forecasts a 31.6% rise in CNS cancer cases among adults aged 20–64 by 2050.

• Data quality may be compromised in low-resource and conflict-affected regions.

• A time lag in data availability may not reflect recent epidemiological changes.

Introduction

Primary central nervous system (CNS) tumours, comprising neoplasms originating in the brain and spinal cord, represent a particularly devastating category of malignancies. The clinical and societal burdens of CNS cancers are magnified by their tendency to affect younger populations, with peak incidence occurring during prime working years (20–64 years), resulting in substantial losses in productive life years and significant caregiver burdens. CNS tumours are relatively rare compared with other cancer sites, accounting for less than 2% of the global cancer cases reported in 2020.¹ Nonetheless, these tumours exhibit one of the poorest 5-year survival rates among all cancers, at only 12.8%.² The aetiology of CNS cancer remains unclear, but several potential risk factors have been identified, including exposure to ionising radiation (such as atomic bomb radiation and childhood nasopharyngeal radium exposure),³,⁵ infections, exposure to viruses and allergens,⁶ as well as environmental factors like contaminated drinking water⁷ and lead.⁸ Additionally, a weakened immune system may also contribute to the risk.

Recent Global Burden of Disease (GBD) studies have significantly advanced our understanding of primary CNS cancers by providing comprehensive epidemiological estimates while simultaneously identifying critical research gaps.^{9 10} The latest GBD iterations have systematically documented the global burden distribution across demographic groups and geographic regions, revealing pronounced disparities in incidence and mortality that correlate strongly with socioeconomic development levels.⁹ Methodological improvements in recent studies include enhanced incorporation of cancer registry data and sophisticated modelling approaches to compensate for incomplete vital registration systems in certain regions. Of particular significance is the framework’s refined age stratification, which specifically highlights the substantial productivity losses experienced by the 20–64 years working-age population affected by these malignancies.¹¹ These findings underscore the necessity for detailed analyses of prevalence, incidence, mortality and disability-adjusted life- years (DALYs) across different age groups and geographic regions to inform targeted interventions.

Recent studies have indicated that advancements in medical technology have led to an improvement in the life expectancy of many CNS cancer patients.¹²,¹⁵ However, there is a notable lack of comprehensive data regarding the burden of CNS cancer in most countries and regions worldwide, highlighting an urgent need to fill this data gap. Understanding the mortality and changes in DALY for CNS cancer patients is crucial for implementing effective preventive interventions. Currently, there exists significant inequality in the global distribution of health resources,¹⁶ underscoring the need for targeted, reliable clinical data to support policymakers in developing CNS cancer prevention strategies in various regions.

This study investigates the prevalence, incidence, mortality and DALYs associated with CNS cancer among adults aged 20–64 years at the global, regional and national levels from 1992 to 2021, stratified by age, sex and socioeconomic development. Additionally, we forecast the incidence of CNS cancer cases through to 2050.

Methods

Data sources and study population

The data for this study were derived from publicly available sources: the Global Burden of Disease Database, accessible at https://ghdx.healthdata.org/gbd-resultstool.¹⁷ The GBD 2021 study employs the most recent epidemiological data and robust standardisation methods to comprehensively assess health losses related to 369 diseases, injuries and risk factors across 204 countries and regions. The GBD database uses complex methodologies to address missing data and adjust for confounding factors. Existing GBD literature has detailed the study’s design and methods.¹⁸ This study adhered to the Guidelines for Accurate and Transparent Health Estimates Reporting. The study population included male and female patients aged 20–64 years.

Patient and public involvement

Patients and/or the public were not involved in this study.

Measures of burden

Burden metrics at global, regional and country levels include incidence, prevalence, mortality and DALYs due to CNS cancer. Using cause aggregation models, the GBD estimates disease-specific mortality.¹⁹ DALYs are calculated by summing years lived with disability and years of life lost. Years lived with disability serve as a metric for disease burden, representing the years of life spent with a disease, accounting for the time spent ill and the disability weights associated with those conditions, thereby reflecting the potential severity of the disease. Years of life lost represent the number of years lost due to premature death from a disease. Both years lived with disability and years of life lost are estimated for each age, sex and location for specified years. All metrics are reported as raw values, rates per 100 000 population and age-standardised rates per 100 000 population, with age standardisation based on the WHO’s World Population Standard Age Structure.

Data processing and modelling

In GBD 2021, disease estimates were derived using spatiotemporal Gaussian process regression and DisMod-MR modelling.²⁰ Briefly, spatiotemporal Gaussian process regression is a set of regression techniques that smooths across age, time and location. DisMod-MR is a Bayesian meta-regression method used to synthesise all available data. The Gaussian, log-Gaussian, Laplace or log-Laplace likelihood functions were employed in DisMod-MR 2.1. Age and sex standardisation were performed according to the GBD reference population. More detailed methodologies can be found in previous studies.²¹

Sociodemographic index

The Sociodemographic Index (SDI) quantifies the level of development of a country or region using fertility rates, education levels and per capita income data. SDI values range from 0 to 1, with higher values indicating higher levels of socioeconomic development.²² It is well established that SDI is associated with disease incidence and mortality. In this study, countries and regions were categorised into five SDI levels (low <0.47, lower-middle 0.47–0.62, middle 0.62–0.71, upper-middle 0.71–0.81 and high >0.81) to explore the relationship between disease burden and socioeconomic development.

Joinpoint regression model

Joinpoint regression analysis was conducted using Joinpoint software (V.5.0.2, https://surveillance.cancer.gov/joinpoint/) to evaluate trends in CNS cancer from 1992 to 2021. The average annual percentage change (AAPC) represents the weighted average of the annual percentage changes (APC) from the joinpoint regression analysis. A Monte Carlo permutation test using 4499 random rearrangement datasets was employed to compute the AAPC of the disease and its 95% CI. An AAPC that significantly differs from 0 at α=0.05 indicates a statistically significant trend. If 0 falls within the 95% CI, the AAPC is considered a stable trend. If both limits of the 95% CI are positive or negative, the AAPC shows an increasing or decreasing trend.

Frontier analysis

To rigorously examine the association between CNS cancer burden and sociodemographic development across diverse populations, we implemented frontier analysis—an advanced econometric approach that establishes the optimal performance boundary by modelling the theoretically minimum achievable age-standardised incidence rates relative to SDI levels, thereby identifying potential efficiency gaps in disease burden management across varying developmental contexts; this methodological framework builds on well-established analytical paradigms as previously described,²³ using stochastic frontier regression techniques to account for both systematic variations and random noise in the observed data while controlling for potential confounding factors, with all statistical computations and model validations performed using SAS Enterprise Guide V.7.1 (SAS Institute) to ensure robust parameter estimation and geospatial visualisations of global disease patterns generated through ArcGIS V.10.5.1 (Environmental Systems Research Institute, Redlands, California, USA) to facilitate comprehensive interpretation of geographical disparities in CNS cancer burden relative to sociodemographic development gradients.

Decomposition analysis

To quantify the driving factors behind changes in the number of cases of CNS cancer, we estimated the relative contributions of changes due to population growth, age structure and variations in the prevalence of GBD risk factors. For this decomposition analysis, we employed a method developed by Das Gupta,²⁴ which summarises the contributions of various factors to observed changes by algebraically separating the standardised impacts of each contributory multiplicative factor.

Predictive analytics

Using estimated population projections and regression analyses,^{25 26} we predicted the incidence of CNS cancer globally and by region up to 2050, using the SDI as a predictive indicator, given the strong relationship between health outcomes and SDI. The predictive model did not account for specific risk factors or changes in case identification over time. Age-specific, location-specific and sex-specific incidence from GBD 2021 for the years 1992–2021 were logit transformed and subjected to a Bayesian age-period-cohort model, using a random effects model with a link function to forecast disease incidence for 2050, run in RStudio software.

Statistical analysis

We employed linear regression to calculate age-specific incidence and their AAPCs, where the logarithmic scale of the rates served as the dependent variable and year as the independent variable. AAPC is a summary measure of the trend over a prespecified fixed interval, calculated as a weighted average of the APC, allowing us to describe AAPC over multiple years with a single figure. APC is calculated as the geometric weighted average of various APC values obtained from the regression analysis. The value of AAPC represents the percentage change per year (increase, decrease or stability). For instance, an AAPC of 0.1 indicates an annual growth rate of 0.1%. Trend rates are reflected in the AAPC value and its 95% CI. Age standardisation was performed using the direct method, applying the global age structure from 1992 to 2021. Age-standardised rates of diseases for each country were calculated using the Global Burden of Disease World Population Standard and reported per 100 000 population. The 95% uncertainty intervals (95% UI) for each metric were reported using the 2.5% and 97.5% quintiles from 1000 samples. The differences for each metric value from 1990 to 2021 were calculated to derive total changes and percentages. We reported and interpreted the results of statistical tests, including effect sizes and CIs, ratios, UIs and exact p values. All necessary data analyses, tables and visuals were executed using RStudio software (V.4.3.3) (http://www.r-project.org, RRID: SCR_001905). Age-period-cohort effect analysis was conducted using the ‘apc-ie,’ command in STATA software (V.16.0).

Results

Global trends

Globally, the prevalence of brain and CNS cancer among adults aged 20–64 years increased significantly between 1992 and 2001 (AAPC 1.74 (95% CI 1.42 to 2.06); p<0.001), and this trend continued from 2002 to 2011 (AAPC 1.03 (95% CI 0.93 to 1.13); p<0.001), although at a slower rate from 2012 to 2021 (AAPC 0.39 (95% CI 0.11 to 0.67); p=0.007) (figure 1A; online supplemental table 1). Overall, the number of patients diagnosed with CNS cancer increased by 311 379 cases from 1992 to 2021, reflecting a substantial rise of 225.11% (AAPC 1.04 (95% CI 0.91 to 1.18); p<0.001). The prevalence rose from 9.10 (95% CI 8.03 to 10.01) per 100 000 population in 1992 to 12.30 (95% CI 10.73 to 14.12) per 100 000 in 2021 (figure 1A; online supplemental table 1).

Figure 1. Joinpoint regression analysis of global HIV prevalence (A), incidence (B), DALYs (C) and mortality (D) in CNS cancer among 20–64 years from 1992 to 2021. AAPC, average APC; APC, annual percentage change; CNS, central nervous system; DALYs, disability-adjusted life-years. ^*P<0.05.

The incidence of CNS cancer among this age group increased from 1992 to 2012 (AAPC 0.23 (95% CI 0.14 to 0.31); p<0.01), before stabilising from 2012 to 2021 (AAPC −0.07 (95% CI −0.28 to 0.14); p=0.506) (figure 1B; online supplemental table 1). Notably, both the age-standardised DALYs and mortality increased from 1992 to 2021, but decreased from 2002 to 2021, and these two indicators in 21 years were significantly lower than those recorded in 1992 (figure 1C,D; online supplemental table 1). Joinpoint regression analysis identified critical changes in the prevalence of CNS cancer in 1998, 2001, 2009 and 2019, while significant changes in incidence were noted in 1996, 2008, 2013 and 2018. Similarly, substantial changes in DALYs were observed in 2002, 2013 and 2018, as well as in mortality in 2002, 2014 and 2018 (figure 1A–D; online supplemental table 2).

Global trends by sex

From 1992 to 2021, the total number of individuals affected by CNS cancer among people aged 20–64 years increased for both men and women worldwide (men: from 0.13 million to 0.28 million; women: from 0.12 million to 0.28 million). The age-standardised prevalence of CNS cancer in this age group rose for both men (from 9.5 to 12.5 per 100 000 population) and women (from 8.7 to 12.1 per 100 000 population), as did the incidence of CNS cancer (men: from 4.2 to 4.5 per 100 000 population; women: from 3.4 to 3.6 per 100 000 population). Furthermore, the age-standardised prevalence increased more rapidly among women (AAPC 1.18 (95% CI 0.99 to 1.37), p<0.001) compared with men (AAPC 0.92 (95% CI 0.77 to 1.07), p<0.001), while there was little difference in the trends for the incidence between men (AAPC 0.22 (95% CI 0.11 to 0.32), p<0.001) and women (AAPC 0.24 (95% CI 0.14 to 0.35), p<0.001) (online supplemental table 3).

During the same period, the DALYs attributable to CNS cancer decreased for both men (from 139.8 to 132 per 100 000 population) and women (from 102.9 to 93.4 per 100 000 population), alongside reductions in age-standardised mortality (men: from 3.4 to 3.2 per 100 000 population; women: from 2.5 to 2.2 per 100 000 population). Furthermore, the decline in both DALYs and mortality was more pronounced among women compared with men (DALYs AAPC: −0.32 vs −0.20; Mortality AAPC: −0.19 vs −0.35) (online supplemental table 4). In 2021, across all SDI regions, the disease burden for women was consistently lower than that for men: high-middle SDI (women: 122 vs men: 176); middle SDI (women: 97 vs men: 134); low-middle SDI (women: 73 vs men: 97); low SDI (women: 48 vs men: 56); high SDI (women: 103 vs men: 160), particularly evident in countries with higher sociodemographic indices (see online supplemental appendix 1).

Global trends by age subgroup

Globally, from 1992 to 2021, the number of individuals diagnosed with CNS cancer showed an upward trend across all age subgroups within the 20–64 years population (20–24 years: from 22 540 to 42 968; 25–29 years: from 30 722 to 56 210; 30–34 years: from 32 870 to 70 644; 35–39 years: from 31 896 to 65 614; 40–44 years: from 25 652 to 55 779; 45–49 years: from 23 527 to 61 500; 50–54 years: from 25 903 to 68 674; 55–59 years: from 28 742 to 74 525; 60–64 years: from 27 023 to 64 342) (online supplemental table 3). The prevalence and incidence of CNS cancer similarly increased across all age subgroups. Notably, the most significant increase in both prevalence (AAPC 1.67) and incidence (AAPC 0.70) was in people aged 20–24 years (online supplemental table 3). Several factors may contribute to this observed trend, including improved detection through advances in neuroimaging technology that facilitate identification of previously undiagnosed cases, as well as potential increased exposure to environmental risk factors such as electromagnetic fields, nitrosamine compounds and viral infections. These findings highlight the importance of enhanced public health attention to this age group, particularly through measures aimed at reducing ionising radiation exposure, improving food safety standards and implementing effective infection control strategies.

There were notable sex differences in the prevalence of CNS cancer across all age subgroups. The upward trend in prevalence among males aged 20–24 years (AAPC 1.67) exceeds that of females (AAPC 1.62), as does the trend for males aged 25–29 years (AAPC 1.38 vs 1.24). In contrast, for all other age groups, females show a higher prevalence increase compared with males (see online supplemental appendix 2).

In all global age subgroups, the mortality associated with CNS cancer declined from 1992 to 2021, particularly among middle-aged individuals aged 40–44 years (AAPC −0.46) (online supplemental table 4). In 2021, the mortality for CNS cancer within the 20–64 years population increased with age, rising from 0.64 cases in the 20–24 age group to 8.91 cases in those aged 60–64 years (see online supplemental appendix 3). The DALYs attributable to CNS cancer decreased at different rates across all age subgroups, with the most significant decline observed in people aged 40–44 years (AAPC −0.45) (online supplemental table 4). In 2021, the highest DALYs were in individuals aged 60–64 years (260.6 per 100 000 population) (online supplemental table 4). Besides, joinpoint regression analysis revealed that significant changes of DALYs in those aged 60–64oc curred in the years 2002, 2007, 2010 and 2018 (see online supplemental appendix 4). Regarding the significant decrease in DALYs and mortality in the 40–44 age group, we propose this may reflect both improvements in diagnostic and treatment technologies and this demographic’s better treatment adherence, stronger immune function and favourable surgical outcomes.

Global trends by SDI

From 1992 to 2021, the age-standardised prevalence of CNS cancer among the global population aged 20–64 years increased across all SDI subgroups, particularly in middle SDI countries (AAPC 1.99 (95% CI 1.90 to 2.08)). In 2021, the highest prevalence of CNS cancer among the population aged 20–64 years was observed in high SDI countries, with a rate of 26.1 cases per 100 000 population (online supplemental table 3).

During the same period, mortality associated with CNS cancer in the 20–64 age group decreased in all sociodemographic subgroups, except for low SDI and low-middle SDI countries. High SDI countries experienced the most significant decline (AAPC −0.5), occurring at all rates 50 times faster than that seen in middle SDI countries (AAPC −0.01) (online supplemental table 4). In 2021, the highest mortality was in high-middle SDI countries, at 3.56 deaths per 100 000 population, nearly 2.9 times the mortality in low SDI countries (1.23 deaths per 100 000 population) (see online supplemental appendix 1). The age-standardised DALYs from CNS cancer exhibited varying trends across different SDI levels in the 20–64 years population. Countries classified as high SDI (AAPC −0.52, p<0.001) and high-middle SDI (AAPC −0.37, p<0.001) showed a decrease in DALYs (online supplemental table 3), while low SDI (AAPC 0.45, p<0.001) and low-middle SDI (AAPC 0.83, p<0.001) experienced an increase. The changes in DALYs for middle SDI countries (AAPC 0.00, p=0.95) were minimal over the same period. In 1992, DALYs were highest in high-middle SDI countries (165.53 per 100 000 population) and lowest in low SDI countries (45.42 per 100 000 population). By 2021, the highest DALYs were again in high-middle SDI countries (148.91 per 100 000 population), while the lowest DALYs were in low SDI countries (51.70 per 100 000 population) (see online supplemental appendix 1).

Regional trends

From 1992 to 2021, an increase in the prevalence of CNS cancer was observed across all 21 regions, with the rapid growth noted in East Asia (AAPC 3.39 (95% CI 3.15 to 3.64)), high-income Asia Pacific (AAPC 3.14 (95% CI 2.93 to 3.36)) and Andean Latin America (AAPC 2.85 (95% CI 2.17 to 3.53)). In contrast, the prevalence in Australasia (AAPC 0.12, (95% CI −0.23 to 0.48)), high-income North America (AAPC 0.23 (95% CI −0.04 to 0.51)) and Oceania (AAPC 0.32 (95% CI 0.07 to 0.56)) increased at a slower rate during the same period (see online supplemental appendix 5). In 2021, among the 21 regions, the highest age-standardised prevalence of CNS cancer among individuals aged 20–64 years was in high-income Asia Pacific (33.4 per 100 000 population), Western Europe (31.6 per 100 000 population), and high-income North America (25.7 per 100 000 population). After stratifying by sex, the prevalence was as follows: high-income Asia Pacific (men: 37.3 per 100 000; women: 29.3 per 100 000), Western Europe (men: 33.7 per 100 000; women: 29.6 per 100 000) and high-income North America (men: 27.9 per 100 000; women: 23.6 per 100 000) (see online supplemental appendix 6).

During the same period from 1992 to 2021, DALYs from CNS cancer among individuals aged 20–64 years showed different changes across different regions. The most significant increase in DALYs was reported in Western Sub-Saharan Africa (AAPC 1.44 (95% CI 1.32 to 1.56)), while the largest reduction occurred in high-income North America (AAPC −0.7219 (95% CI −0.91 to −0.54)) (see online supplemental appendix 5). In 2021, the three regions with the highest age-standardised DALYs among individuals aged 20–64 years were Central Europe (213 per 100 000 population), Central Asia (203.9 per 100 000 population) and Eastern Europe (185.3 per 100 000 population). The regions with the lowest DALYs were Western Sub-Saharan Africa (23.2 per 100 000 population), Oceania (28.9 per 100 000 population) and Central Sub-Saharan Africa (39.6 per 100 000 population) (online supplemental table 4). Stratifying by sex revealed notable disparities, with the lowest DALYs for men in Western Sub-Saharan Africa (22 per 100 000 population) and for women in Oceania (21.1 per 100 000 population). In contrast, the highest DALYs were observed for men in Central Europe (256.1 per 100 000 population) and for women in Central Europe (170.3 per 100 000 population) (see online supplemental appendix 6).

National trends

At the national level, from 1992 to 2021, Ecuador exhibited the largest increase in age-standardised prevalence of CNS cancer among individuals aged 20–64 years, with an AAPC (8.52 (95% CI 6 to 11.1)). This was followed by Turkmenistan (AAPC 7.19 (95% CI 5.68 to 8.72)) and Kyrgyzstan (AAPC 5.29 (95% CI 4.16 to 6.43)) (figure 2; see online supplemental appendix 7). In 2021, Norway exhibited the highest prevalence of CNS cancer (109.5 per 100 000 population), while Gambia exhibited the lowest prevalence (0.2 per 100 000 population) (online supplemental figure 1; see online supplemental appendix 8).

Figure 2. Age-standardised prevalence of CNS cancer among individuals aged 20–64 years across 204 countries worldwide from 1992 to 2021. ASR, age-standardized rate; CNS, central nervous system.

During the same period, the country with the largest decline in age-standardised DALYs was Luxembourg (AAPC −1.58 (95% CI −1.82 to −1.33)), followed closely by Greenland (AAPC −1.57 (95% CI −1.71 to −1.43)) and Belgium (AAPC −1.31 (95% CI −2.34 to −0.27)) (see online supplemental figure 2; online supplemental appendix 7). In 2021, Montenegro reported the highest DALYs for CNS cancer among individuals aged 20−64 years (326 per 100 000 population), while Gambia recorded the lowest DALYs for CNS cancer in this age group (4.5 per 100 000 population) (see online supplemental figure 3; online supplemental appendix 8).

Frontier analysis

To better understand the potential improvements in the incidence, prevention and management of CNS cancer across different countries and regions, we conducted a frontier analysis based on age-standardised incidence and SDI using CNS cancer data from 1992 to 2021. Overall, our findings indicate that as SDI increases, the incidence of CNS cancer decreases (figure 3A). In 2021, among high SDI countries (SDI>0.85), the five countries and regions with the largest frontier disparities are Norway, Iceland, Andorra, San Marino and Monaco. Conversely, in low SDI countries (SDI<0.5), the five countries with the smallest frontier disparities are Gambia, Mali, Somalia, Côte d'Ivoire and Guinea (figure 3B).

Figure 3. Frontier analysis of age-standardised incidence of CNS cancer among individuals aged 20–64 years across 204 countries worldwide from 1992 to 2021 (A) and in 2021 (B). CNS, central nervous system; SDI, sociodemographic index.

Across all countries, the 15 countries and regions with the largest frontier disparities are Norway, Iceland, Andorra, San Marino, Monaco, Greece, Denmark, Sweden, Finland, Montenegro, Ireland, Croatia, Belgium, Bulgaria and Latvia. These nations have the potential to significantly reduce incidence and thus decrease CNS cancer under similar SDI conditions through targeted efforts. In contrast, the 15 countries with the smallest distance from the frontier line include Benin, Burkina Faso, Sierra Leone, Chad, Kiribati, Malawi, Niger, Guinea, Côte d'Ivoire, Nigeria, Cook Islands, São Tomé and Príncipe, Somalia, Mali and Gambia, which have limited room for improvement under similar SDI conditions. Building on this analysis, we further explored the potential for improving CNS cancer incidence in countries across different SDI regions. Our research identified that, in low SDI regions (SDI<0.47), the countries with the greatest potential for improvement are Afghanistan, Yemen and Haiti. In low-middle SDI regions (SDI 0.47–0.62), the countries with the most potential for improvement are Tajikistan, Egypt and Kyrgyzstan. In middle SDI regions (SDI 0.62–0.71), the countries with the greatest potential for improvement are Turkmenistan, Armenia and Palestine. In high-middle SDI regions (SDI 0.71–0.81), the countries with the most potential for improvement are Greece, Montenegro and Croatia. Finally, in high SDI regions (SDI>0.81), the countries with the greatest potential for improvement are Norway, Iceland and Andorra (figure 3B; see online supplemental appendix 9 and 10).

Decomposition analysis by cause of CNS cancer

To better understand the key factors influencing the incidence of CNS cancer from 1992 to 2021, we conducted a decomposition analysis of the relative contributions to case increases across 21 global regions, focusing on population ageing, population growth and epidemiological changes. Our findings indicate the largest contributing factor to the increase in CNS cancer cases is globally population, followed by ageing and epidemiological change. The relative contributions of these three factors to the rising incidence vary by region. Specifically, in Central Europe, population contributes 5.03% to the increase in CNS cancer incidence, whereas in Western Sub-Saharan Africa, the contribution of population ageing reaches 283.65% (figure 4A,B; see online supplemental appendix 11).

Figure 4. Decomposition analysis of the incidence of CNS cancer among individuals aged 20–64 years globally and across 21 regions from 1992 to 2021 (A) and in 2021 (B). CNS, central nervous system.

Global incidence forecast

Predictive analysis indicates that from 2021 to 2050, the incidence of CNS cancers among the global population aged 20–64 years is projected to rise, increasing from 186 891 cases to 245 942 cases, representing a growth of 31.6%. By 2050, the number of new cases among females is expected to be lower than that of males (112 688 vs 133 254) (figure 5A; see online supplemental appendix 12). The age-standardised incidence for CNS cancers is anticipated to reach 4.02 (95% CI 0.57 to 7.53) per 100 000 population, with females exhibiting a lower incidence (AAPC 3.41 (95% CI 0.73 to 6.09)) than that of males (AAPC 3.96 (95% CI 0.84 to 7.08)) (figure 5A; see online supplemental appendix 13).

Figure 5. Predictive analysis of the number of new cases and incidence (A), the number of deaths and mortality (B) of CNS cancer among individuals aged 20–64 years globally from 1992 to 2050. ASR, age-standardized rate; CNS, central nervous system.

During the same period, the number of deaths from CNS cancers among the global population aged 20–64 years is expected to increase from 126 712 to 1 52 804, marking an increase of 20.59%. By 2050, the number of deaths among females is projected to be lower than that of males (66 314 vs 86 490) (figure 5B; see online supplemental appendix 12). The mortality for CNS cancers is projected to be 2.41 (95% CI 0.43 to 4.39) per 100 000 population, reflecting a decrease of 11.07%. In this context, the age-standardised mortality for females is estimated at 1.94 per 100 000, which is lower than that for males at 2.51 per 100 000. Furthermore, the decline in mortality is expected to be less pronounced for females (13.39%) compared with males (21.32%) (figure 5B; see online supplemental appendix 13).

Discussion

Current and comprehensive data on the burden of CNS cancers are lacking in many countries and territories worldwide. In this study, we conduct a thorough analysis of the prevalence, incidence, DALYs and mortality associated with CNS cancers globally from 1992 to 2021. Over the past 30 years, there has been a significant increase in the age-standardised prevalence and incidence of CNS cancers among the global population aged 20–64 years. Meanwhile, both mortality and DALYs have declined markedly. These findings indicate that, due to advancements in healthcare, CNS cancers are no longer a major contributor to reduced life expectancy. Notably, there are significant inequalities in age-standardised prevalence and mortality of CNS cancers between countries with different SDI and across different regions. Additionally, the determinants of incidence for CNS cancers differ among various countries and regions. This study expands our understanding of the increasing global burden of CNS cancers by focusing on trends among patients aged 20–64 years. Furthermore, it calls for urgent attention to the strategies for responding to CNS cancer patients worldwide, advocating for the equitable allocation of health resources and the development of targeted guidelines. These findings are not only significant for health practices but also provide valuable insights for future research, offering actionable evidence for the clinical prevention and management of all CNS cancer patients.

Age differences in the burden of CNS cancer

CNS cancer predominantly affects individuals between the ages of 20 and 50 years globally. To achieve a more comprehensive and objective analysis, we selected a study population aged 20–64 years. Our research reveals that from 1992 to 2021, the prevalence of CNS cancers increased across all age subgroups within this demographic. Technological advancements, such as improved resolution in CT and MRI, have facilitated the detection of CNS cancers. Additionally, the development of advanced treatment modalities, including gamma knife, radiosurgery, proton beam therapy and photon beam therapy, along with the emergence of targeted therapies and immunotherapies for CNS cancers, has contributed to increased 5-year survival rates and reduced mortality. Our findings corroborate a decline in mortality associated with CNS cancers from 1992 to 2021. Notably, the most significant increase in prevalence and incidence was observed in the 20–24 age group, indicating a trend towards younger patients being diagnosed with CNS tumours, which is consistent with previous reports.³ Furthermore, we found that mortality from CNS cancers rises with age, with individuals aged 60–64 experiencing mortality approximately 14 times higher than those aged 20–24. This disparity may be attributed to the more robust immune systems of younger individuals compared with their older counterparts, enhancing their ability to combat CNS cancers.

Sex differences in the burden of CNS cancer

Between 1992 and 2021, the age-standardised prevalence of CNS cancers among women aged 20–64 increased significantly compared with men, while the age-standardised mortality for female patients declined more sharply than those for males. The incidence of CNS cancers in women also rose notably when compared with men. The reduction in age-standardised DALYs associated with CNS cancers was more pronounced among women, particularly in countries with higher SDI, which is in line with previous studies.^{27 28} These findings suggest that, compared with females, greater attention should be paid to the burden of CNS cancers in males, which may be linked to sex hormones and genetic features, unhealthy lifestyle choices commonly associated with men, such as smoking and alcohol consumption.²⁹,³¹ Understanding the underlying causes of these differences is crucial for effective management and research on CNS cancers.

Sociodemographic differences in the burden of CNS cancer among older adults

From the analysis of the SDI, we found significant inequalities in age-standardised prevalence, incidence, DALYs and mortality of CNS cancer among countries with different SDI. In 2021, the highest prevalence of CNS cancer was observed in high SDI countries among individuals aged 20–64 years. However, between 1992 and 2021, the growth of incidence in these countries has been lower than that of low SDI countries. This may be attributed to the effective preventive measures adopted by high SDI countries facing a major burden of CNS cancers in recent years, including environmental improvements and increased awareness and advocacy regarding the disease.³²,³⁵ The USA exemplifies this approach through its coordinated efforts involving environmental protection regulations, radiation safety protocols and nationwide awareness programmes, which have collectively contributed to favourable epidemiological trends in CNS cancer outcomes as documented in national health statistics.^{34 35} Moreover, the DALYs and mortality associated with CNS cancer have started to decline significantly in high SDI and high-middle SDI countries. In contrast, these metrics continue to rise in low SDI and low-middle SDI countries, likely due to being reflective of a lack of access to the highly specialised services needed to treat these complex diseases.^{27 36} The objective reason is the lack of the relatively advanced medical equipment, medications and technologies in low SDI countries.^{37 38} These findings indicate a heavier burden of CNS cancer in low SDI countries. It is crucial to focus on effective strategies to strengthen the prevention and control of CNS cancer in low-SDI countries through targeted health policies, thereby reducing the overall burden of this disease. Currently, the prevention and treatment policies and advanced medical technologies adopted by high-SDI countries serve as valuable models for low-SDI countries.

Global trends in the development of CNS cancer

Overall, from 1992 to 2021, there has been a notable upward trend in the age-standardised prevalence and incidence of CNS cancers among adults aged 20–64 years, while age-standardised DALYs and mortality have exhibited a downward trend. This suggests significant advancements in the global management of CNS cancers. However, the absolute number of cases and DALYs remains alarmingly high. As research into CNS cancers progresses, we anticipate a deeper understanding of the disease, which will facilitate the development of new strategies for prevention and treatment, such as targeted immunotherapy drugs³⁴,³⁷ and fluorescence-guided tumour resection.³⁹,⁴¹

Strengths and limitations of this study

This study systematically summarises the global burden of CNS cancers over the past 30 years, highlighting differences across sex, age and region. It is encouraging to note a decline in mortality and DALYs related to CNS cancers among patients and their families. Furthermore, this study predicts changes in global incidence over the next 30 years, laying a foundational framework for future research.

However, several limitations of this study should be acknowledged. While the GBD study provides valuable standardised estimates across 204 countries and territories, several limitations regarding data representativeness merit consideration. The completeness and accuracy of source data vary substantially across regions, particularly in resource-limited settings where underdeveloped vital registration systems and cancer surveillance programmes may lead to incomplete case ascertainment. The standardised modelling approaches, though enabling cross-national comparisons, may obscure important local variations in diagnostic capabilities, clinical practices and reporting standards for CNS cancers. Reliance on modelling to compensate for data gaps in regions with limited primary sources could introduce estimation uncertainties, particularly for rare malignancies and age-stratified analyses. Additionally, inherent time lags in data availability and differences in healthcare systems across countries may further influence burden estimates. While these limitations suggest the need for cautious interpretation of country-specific findings, the GBD remains the most comprehensive effort to quantify global disease patterns, and its estimates provide important benchmarks when supplemented with local validation studies and considered alongside regional epidemiological evidence.

Conclusions

From 1992 to 2021, the prevalence and incidence of CNS cancers among adults aged 20–64 years globally increased, while mortality and DALYs significantly declined. Significant inequalities exist in age-standardised prevalence, incidence, DALYs and mortality of CNS cancer among countries with varying sociodemographic indices. These disparities highlight the urgent need for targeted clinical guidelines and equitable distribution of global health resources.

Data availability statement

Data are available in a public, open access repository. Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.