The GBD 2021 delivers through evaluation of disease, injury, and risk factor burden across 204 countries and territories, encompassing 88 risk factors (24, 25). This research offers findings on incidence, prevalence, mortality, DALYs, YLDs, and YLL for 371 diseases in 204 countries and regions. It uses data on 88 risk factors from 1990–2021, along with associated uncertainty intervals (UIs). Study data were sourced from the GBD 2021, accessible through https://ghdx.healthdata.org/gbd-2021. The risks are organized into a four-tier hierarchy: Level 1 encompasses environmental, occupational, behavioral, and metabolic risks; Level 2 details include 20 broader categories such as air pollution and high BMI; Level 3 encompasses more nuanced risks, including particulate matter pollution and child growth failure, representing some of the most detailed categorizations. Further refinement occurs at Level 4, which breaks down risks from Level 3 into even more specific classification, such as ambient particulate matter pollution and child stunning. The four-level risk hierarchy is based on the well-established comparative risk assessment (CRA) framework developed by the Global Burden of Disease (GBD) Study, which has been widely adopted as the international standard for risk factor quantification (26, 27). Data processing followed standardized GBD protocols (24, 25). Input data underwent: Quality grading (0–5 stars based on completeness, diagnostic specificity, and representativeness), Bias adjustment via spatial–temporal meta-regression, Ensemble modeling integrating 43 cause-of-death models. This framework replaces conventional systematic review tools by directly quantifying uncertainty from source heterogeneity.

The Socio-demographic Index (SDI) measures development by amalgamating income, education, and fertility data, classifying regions into five development stages from Low to High SDI, reflecting population wealth and education levels. Specifically, the SDI ranges are as follows: Low SDI from 0 to 0.4658, low-middle SDI from 0.4658 to 0.6188, Middle SDI from 0.6188 to 0.7120, High-middle SDI from 0.7120 to 0.8103, and High SDI from 0.8103 to 1.

In the 10th edition of the International Classification of Disease (ICD-10), subarachnoid hemorrhage is classified under codes I60-I60.9, I62.0, I67.0-I67.1, and I69.0. The ICD-10 assigns the code 430–430.9 to subarachnoid hemorrhage.

DALY (Disability-Adjusted Life Years): A summary measure of population health that quantifies the burden of disease by combining years of life lost due to premature mortality and years lived with disability, weighted by the severity of the disability (28). ASMR (Age-Standardized Mortality Rate): A mortality rate that has been adjusted to account for differences in age structure between populations, allowing for valid comparisons across different populations and time periods (29). ASDR (Age-Specific Death Rate): The number of deaths in a specific age group per 100,000 population in that same age group during a given time period (30).

Ambient particulate matter pollution refers to the annual average concentration of PM_2.5, particles smaller than 2.5 micrometers in diameter, in the air, weighted by population exposure. This estimation integrates data from various sources, such as satellite aerosol observations, ground-level air quality monitors, chemical transport models, demographic data, and land-use information. For the GBD 2021, the dataset was expanded with newer ground monitor readings from both pre-existing and newly added sites. Additional contributions to the database came from sources such as the European Environment Agency, the United States Environmental Protection Agency, and the OpenAQ initiative (31).

The Estimated Annual Percentage Change (EAPC) serves as a pivotal indicator for monitoring the progression of Age-Standardized Rate (ASR) across time. It is determined within a regression framework defined as y = α + βx + ε, where y corresponds to the annual rate change to per 100,000 individuals, α represents the intercept, β represents the slope, x is the calendar year, and ε is the error term. The EAPC calculation is based on the formula:

With the 95% confidence interval (CI) extracted directly from the linear regression model parameters. Statistical significance is established for two-sided p-values below 0.05.

In parallel, the Pearson correlation coefficient was applied to evaluate the relationship between ASR and the SDI, with significance indicated by p-values less than 0.001. These analyses were conducted using R software, version 4.4.1.

The Bayesian Age-Period-Cohort (BAPC) R package is a statistical tool designed for projecting future disease burdens using a Bayesian framework (29). We employed age-specific population data spanning from 1990 to 2021, along with projections for 2022 to 2050, to assess trends in mortality and DALYs among populations. For our analysis, we adhered to the standard parameters provided with the BAPC packages, leveraging its default settings to effectively model.

The trends in age-standardized mortality rates (ASMR) and age-standardized DALY rates (ASDR) for SAH attributed to ambient PM_2.5 have shown a downward trend from 1990 to 2021, except for regions with low-middle SDI (Figure 1 and Supplementary Figure S1). Despite a decline in ASMR from 0.99 (95% UI, 0.60–1.50) in 1990 to 0.63 (95% UI, 0.43–0.82) in 2021, the death toll rose from 38,129.84 (95% UI, 23,179.25–57695.62) in 1990 to 53,561.65 (95% UI, 36,516.80-69,717.63) in 2021. Within the spectrum of SDI regions, middle SDI exhibited the highest ASMR and ASDR at 1.07 and 27.42, respectively. In addition, high SDI regions experienced the most pronounced decline in ASMR and ASDR, from 0.66 and 20.9 in 1990 to 0.22 and 7.14 in 2021. Turning to gender differences, ASMR for SAH due to PM_2.5 was higher in male at 0.72 compared to female at 0.55 in 2021 (Tables 1, 2).

Among the 21 GBD regions, South Asia and Southeast Asia experienced the most notable increase in SAH deaths attributed to PM_2.5 by 2021 (Figure 2). Death numbers in South and Southeast Asia increased from 3,201.81 and 2,149.93 in 1990 to 11,091.32 and 5,309.19 in 2021. Nevertheless, East Asia bore the greatest burden, with the highest death and DALY figures peaking at 229,552.99 and 563,441.41 (Figure 2, Supplementary Figure S1, and Tables 1, 2). In contrast, high SDI regions such as Australasia and Oceania exhibited lower PM_2.5-attributable SAH burdens, with death numbers recorded at 66 (95% UI, 39.03–100.57) and 31.82 (95% UI, 9.85–71.45), and DALY numbers at 1,630.4 (95% UI, 964.32–2,422.15) and 1,204.59 (95% UI, 378.66–2,724.32) (Tables 1, 2). Furthermore, in South Asia, the proportion of female deaths increased significantly, from 31.0% in 1990 to 40.9% in 2021 (Figure 2).

Nationally, Mongolia had the highest burden of subarachnoid hemorrhage attributable to ambient PM_2.5, with an ASMR of 2.49 (95% UI, 1.23–3.82) and ASDR of 61.92 (95% UI, 30.60–93.24) in 2021. Moreover, the EAPC in ASMR showed the most significant increasing trend at 5.26 (95% CI, 4.69–5.84) (Supplementary Tables S1, S2). Similarly, increasing trends in ASMR and ASDR were noted across South Asia, Central Asia and Southern Sub-Saharan Africa (Figure 3).

Conversely, on a global scale, the EAPC for both ASMR and ASDR showed a decline in most countries, particularly in North America, Europe, Oceania, and South America. Notably, exhibited the most pronounced decreases in EAPC for ASMR and ASDR, at −7.01 (95% CI, −7.48 to −6.53) and −7.25 (95% CI, −7.66 to −6.84), respectively.

Between 1990 and 2021, the PM_2.5-attributable death and DALYs of SAH saw a consistent annual decline across all age groups. By 2021, the disease burden had lessened in comparison to 1990 for every age demographic (Figure 4 and Supplementary Figure S2). Notably, a downward trend in the burden for those over 50 years old was observed starting around 2000. For those under 50, although the death burden was comparatively minor, a decline has been noted since the early 2000s (Figure 4). The death among those aged 85 and above showed a brief increase between 1999 and 2003, followed by a decline. Between 1990 and 2021, there was a notable decrease in global age-stratified DALYs from SAH across all age groups. However, from 2012 to 2017, there was a slight increase in DALYs for individuals aged 70–74 (Supplementary Figure S2).

In the comparative analysis of PM_2.5-attributable SAH burden across countries and territories with varying SDI levels. It was found that the correlation between ASMR and SDI mirrored that between ASDR and SDI. However, a significant trend was not observed, with the p-value exceeding 0.05. The disease burden due to PM_2.5 in SAH was predominantly greater in countries or regions with an SDI ranging from 0.6 to 0.7 and lower in those with an SDI below 0.4 or above 0.8 (Figure 5).

Between 1990 and 2021, the ASMR and ASDR for SAH attributable to ambient PM_2.5 exhibited slightly fluctuations. This period witnessed a complex interplay of factors influencing the health outcomes related to ambient PM_2.5 exposure. Looking ahead, from 2022 to 2050, there is a projected increase in both ASMR and ASDR for SAH due to ambient PM_2.5, indicating a potential escalation in the disease burden (Figure 6 and Supplementary Figure S3).

This study provides a comprehensive analysis of the global disease burden of SAH attributable to ambient PM_2.5 pollution across regions, genders, age groups, and SDI levels from 1990 to 2021, with projections extending to 2050. While previous studies have investigated PM_2.5-associated stroke burden, this work uniquely quantifies the sex-specific temporal trends and aging-related vulnerability patterns in SAH burden, while integrating future projections through robust modeling approaches (32–34).

The paradoxical 40.4% increase in absolute SAH deaths from 38,129.84 to 53,561.65 despite a 36.4% decline in age-standardized mortality rates (from 0.99 to 0.63 per 100,000) reflects the complex interplay of demographic and epidemiological factors over the 32-year study period. This apparent contradiction can be attributed to several key demographic transitions: (1) population aging dynamics-the global population aged ≥60 years increased by approximately 180% during this period, creating a substantially larger at-risk population despite improved per-capita risk profiles (35). (2) overall population growth - the world population expanded from 5.3 billion in 1990 to 7.9 billion in 2021, representing a 49% increase in the denominator for absolute case calculations (36). and (3) persistent environmental health disparities - while high-income regions achieved substantial PM_2.5 reductions, rapid industrialization in South Asia and Sub-Saharan Africa maintained or increased exposure levels for large populations.

This demographic-epidemiologic paradox illustrates a critical limitation in public health assessment: age-standardized rates, while essential for comparing risk across populations and time periods, may underestimate the true societal burden when applied to aging populations with age-sensitive health outcomes. The divergence between these metrics emphasizes that successful risk reduction strategies at the individual level can be overwhelmed by demographic shifts, necessitating integrated approaches that address both exposure reduction and healthcare system capacity for aging populations. Furthermore, the absolute increase in deaths predominantly occurred in low-and middle-income countries (contributing 78% of the excess deaths), highlighting persistent global health inequities in environmental protection and healthcare access (37, 38).

The pronounced regional disparities underscore differential progress in environmental health. High SDI regions achieved the most substantial improvements (ASMR: 0.66 to 0.22), likely reflecting stringent air quality standards and advanced healthcare systems (39). The 65.8% ASDR reduction in high SDI regions (20.9 to 7.14) demonstrates the effectiveness of integrated environmental-health interventions, providing a roadmap for middle SDI countries (40).

Conversely, middle SDI regions bear the highest SAH burden (ASMR 1.07), trapped in a developmental phase combining industrializing economies with insufficient pollution controls. These regions experience rapid industrialization, environmental pollution, and limited medical resources, leading to a higher disease burden of PM_2.5-related SAH, particularly in South and East Asia (41). Low-middle SDI regions’ unfavorable trends may indicate persistent barriers in pollution control and healthcare access (42).

The striking geographical disparities in subarachnoid hemorrhage burden attributable to ambient PM_2.5 exposure reflect the complex interplay between environmental pollution levels, healthcare infrastructure, and socioeconomic development across different regions. Mongolia’s position as having the highest national burden, with an age-standardized mortality rate of 2.49 per 100,000 and a concerning annual increase of 5.26%, underscores the urgent need for targeted air quality interventions in this region.

South and Southeast Asia exhibited the steepest increases in PM_2.5-attributable SAH deaths (South Asia: +246.5%, Southeast Asia: +147.0%), aligning with satellite-derived PM_2.5 concentration trends showing population-weighted annual averages exceeding 75 μg/m³ in Bangladesh and India (43). These regions face compounded challenges including inadequate neurosurgical infrastructure (44) and limited implementation of WHO air quality guidelines (45, 46) creating a “double burden” of environmental and healthcare system deficiencies. The pronounced increasing trends observed in South Asia, Central Asia, and Southern Sub-Saharan Africa align with the rapid industrialization and urbanization occurring (Figure 3) (47, 48). Air pollution in South Asia results from a complex interplay of emission sources beyond industrial activities, including the combustion of solid fuels for cooking and heating, emissions from small industries such as brick kilns, the burning of municipal and agricultural waste, and cremation practices (49, 50). In contrast, East Asia’s substantial absolute burden (229,552.99 deaths in 2021) reflects legacy pollution effects from rapid industrialization, though recent policy interventions show promising declines in PM_2.5 levels (51). The declining trends in North America, Europe, Oceania, and South America reflect successful implementation of air quality improvement policies and stricter environmental regulations over the past decades, demonstrating the potential for effective public health interventions to reduce PM_2.5-related health burdens (36, 40). East Asia region includes China, North Korea, and Taiwan, with China contributing the vast majority of the 229,553 deaths in 2021 due to its large population size. Despite China’s middle SDI classification, this region showed significant improvements in age-standardized rates (ASMR decline), reflecting substantial investments in healthcare infrastructure and air quality management over the study period (52). The apparent contradiction between high absolute burden and improving rates reflects China’s demographic transition and successful implementation of pollution control policies since 2013 (53). High-income Asia-Pacific region encompasses Australia, Brunei, Japan, New Zealand, Singapore, and South Korea. These countries consistently demonstrate the lowest PM_2.5-attributable SAH burden, with marked improvements in both absolute and age-standardized metrics, directly correlating with their high SDI status and advanced healthcare systems (48). The steep declining trends in this region exemplify how combined high socioeconomic development and stringent environmental regulations effectively reduce pollution-related health burdens.

The male predominance in PM_2.5-related SAH mortality (ASMR 0.72 vs. 0.55 in females) likely reflects differential exposure patterns and biological susceptibility. Men experience higher occupational exposure through industrial work, while sex-linked differences in inflammatory responses and hormonal status may influence cardiovascular vulnerability to PM_2.5 (54, 55). However, evidence on gender differences in air pollution health effects remains inconsistent across studies (56).

The global increase in female SAH mortality proportion (particularly +9.9% in South Asia) may be explained through two complementary mechanisms: biological susceptibility via enhanced oxidative stress responses (20, 21), and socioeconomic factors limiting women’s access to preventive healthcare in LMICs (57). The age-dependent burden escalation (Figure 4 and Supplementary Figure S2) demonstrates a 3.2-fold higher mortality risk in populations >70 years compared to <50 years, likely mediated through PM_2.5-induced exacerbation of hypertension (58) and cumulative blood–brain barrier damage (16)^. This aging-related vulnerability is projected to intensify as the global population over 60 years grows by 56% by 2050 (23).

The increased burden of SAH projected by the model over the next 29 years (Figure 6 and Supplementary Figure S1), potentially due to a significant association between PM_2.5 exposure and SAH, with a particularly sharp increase in risk over short periods. As the global population ages, the older adults, who are more sensitive to the health effects of PM_2.5, become a larger proportion of society (59). Additionally, industrialization and urbanization have exacerbated PM_2.5 pollution, and regions lacking effective health protection measures and resources experience more significant adverse health effects from PM_2.5 (41, 60). Therefore, it is imperative that the world invests more time and effort into controlling PM_2.5 pollution and reducing the medical burden on SAH patients (61).

Previous research has examined the global burden of PM_2.5 in relation to SAH, yet this study extends these findings, revealing that between 1990 and 2021, the proportion of SAH attributed to PM_2.5 increased among women across all GBD regions. This underscores a concerning trend, particularly as the disease burden of SAH due to PM_2.5 escalating with advancing age. Our projections for future trends indicate a global increase in both SAH deaths and DALYs attributed to PM_2.5.

Since the database used in this study covers different countries and regions, cultural, geographical, climatic, and even genetic differences have influenced the study’s results. Although these factors are unlikely to fundamentally alter the relationship between ambient air pollution and the risk of SAH (62, 63). Genetic polymorphisms in oxidative stress pathways and inflammatory responses may modulate individual susceptibility to PM_2.5-induced cerebrovascular damage. This could potentially explain some of the striking regional differences, such as East Asia having a disproportionately high absolute burden despite relatively lower PM_2.5 concentrations (64, 65). Climatic factors, including temperature extremes and seasonal variations, may interact with PM_2.5 toxicity through altered particulate composition and enhanced inflammatory responses. This could partially account for Mongolia’s exceptionally high burden and the seasonal patterns observed in temperate regions (66). However, the consistent dose–response relationship between PM_2.5 exposure and SAH risk documented across diverse populations in epidemiological studies suggests that ambient particulate matter remains an independent risk factor regardless of these population-specific modifiers (67).

While the GBD 2021 methodology employs comparative risk assessment that inherently adjusts for major confounders through integrated exposure-response functions derived from epidemiological meta-analyses, residual confounding from traditional SAH risk factors may influence our estimates (68, 69). Chronic diseases (hypertension, diabetes), lifestyle factors (smoking, alcohol consumption), and healthcare access disparities may cluster geographically with PM_2.5 exposure patterns, potentially inflating the pollution-attributable burden in regions with limited diagnostic capacity and neurosurgical infrastructure (70). The stark regional disparities observed—particularly Mongolia’s exceptionally high burden and South Asia’s 246% mortality growth—likely reflect complex interactions between air pollution exposure and unmeasured socioeconomic, healthcare, and behavioral confounders that extend beyond PM_2.5 effects alone (71). Despite these limitations, the consistent global patterns and projected 2050 increases suggest that PM_2.5 reduction efforts would yield substantial health benefits, warranting urgent air quality interventions in high-burden regions regardless of residual confounding concerns.

Our study had some limitations including: (1) Data heterogeneity in cause-of-death certification and PM_2.5 exposure modeling across regions may affect burden estimates; (2) Residual confounding (e.g., unmeasured comorbidities, socioeconomic factors) could bias risk associations; (3) Limited granularity in occupational/lifestyle exposure data hinders precise gender-risk stratification; (4) Model uncertainties (e.g., counterfactual exposure thresholds) warrant sensitivity analyses in future work. Future there are urgent actions remain imperative: Prioritize air quality interventions in high-burden regions (Mongolia/South/Southeast Asia); Strengthen stroke care systems in low-resource settings; Implement gender-responsive strategies addressing occupational (male) and household (female) exposures; Target hypertension control in older populations. Integrating SAH burden metrics into global health frameworks is essential to mitigate this preventable crisis, despite current methodological constraints.