Global Asthma Burden due to Attributable Risk Factors, 1990-2021, and Forecast to 2050: A Systematic Analysis of the Global Burden of Disease Study 2021

Abstract

Background

Asthma, a common chronic respiratory disease worldwide, is strongly influenced by several attributable risk factors.

Objective

To systematically examine global asthma burden trends attributable to major risk factors from 1990 to 2021 and forecast changes to 2050.

Methods

Using 2021 GBD data, we estimated global asthma DALYs and age-standardized DALY rates (ASDRs) attributable to high BMI, smoking, occupational exposure, and NO₂ pollution, and explored their relationships with the Socio-demographic Index (SDI). Bayesian age-period-cohort (BAPC) modeling was used for forecasting, with 1990-2010 data serving as the training set for backtesting trends from 2011 to 2021, and with prediction accuracy evaluated with the mean absolute percentage error (MAPE).

Results

During 1990-2021, asthma burden related to most risk factors declined globally and across SDI quintiles. Backtesting confirmed BAPC model reliability (MAPE: high BMI 5.44%, smoking 1.16%, occupational asthmagens 3.35%). Worldwide, smoking-related burden declined most sharply: ASDR fell from 41.95 in 1990 to 16.19 in 2021 per 100,000 (-61.41%, estimated annual percentage change [EAPC] = -3.13). NO₂ pollution and occupational asthma burdens dropped 48.45% and 45.57%, respectively. High BMI showed the smallest reduction (-21.99%, EAPC = -0.85) and showed an increase in low-middle SDI regions (EAPC = 0.25). ASDRs for high BMI (ρ = -0.27), occupational risks, and smoking correlated negatively with SDI; NO₂ ASDR showed a significant positive correlation with SDI (ρ = 0.56, P < 0.001). By 2050, high BMI is projected to be the only risk factor with a continuously rising burden, with its ASDR rising from 39.42 in 2021 to 53.79 per 100,000 ( + 36.45%).

Conclusion

Despite declines in asthma burden from all four factors, high BMI poses a growing threat with significant regional disparities. Future strategies should be tailored by SDI region and high-risk population, prioritizing surveillance and intervention for high BMI to counter its rising disease risk.

Introduction

Asthma is a chronic, heterogeneous disorder characterized by airway inflammation and hyper-responsiveness, clinically manifesting as recurrent wheeze, cough, chest tightness, and breathlessness¹. According to the Global Burden of Disease (GBD) Study 2021, asthma affected approximately 260 million people worldwide, caused over 430,000 deaths, and led to 21.42 million disability-adjusted life-years (DALYs)². Although advances in diagnosis and management have markedly improved individual outcomes in recent decades³, substantial geographical disparities persist. Age-standardized DALY rates (ASDRs) in low and low-middle Socio-demographic Index (SDI) regions remain far higher than elsewhere⁴.

A wide spectrum of modifiable environmental, occupational, and lifestyle factors is consistently linked to both the onset and exacerbation of asthma. Ambient NO₂, occupational asthmagens, smoking, and high body-mass index (BMI) are among the most frequently confirmed risk factors⁵. Epidemiological evidence indicates that chronic NO₂ exposure significantly increases childhood asthma incidence and hospital admission rates^6,7. As important risk factors for asthma, it has long been proven by research that smoking and high BMI can enhance the body’s inflammatory response through various pathways, thereby increasing the risk of asthma occurrence and exacerbation^8,9. Retrospective analyses attribute approximately 16.2% of adult-onset asthma cases to workplace exposures, with exceptionally high risks documented in manufacturing, bakery, and health-care workers^10,11.

While these environmental drivers are known to shape the spatiotemporal evolution of asthma burden, their relative contributions across societies with divergent development levels have not been systematically quantified. Accurate tracking of risk-factor trends is fundamental for projecting future disease burden and designing targeted interventions. Using GBD 1990-2021 data, we quantified the DALYs attributable to high BMI, smoking, occupational asthmagens, and ambient NO₂ pollution, mapped their global and regional dynamics, aiming to provide evidence to inform tailored asthma prevention strategies.

Materials and Methods

Data Source

Using data from the GBD 2021, this study systematically analyzed asthma DALYs and the ASDRs attributable to four risk factors from 1990 to 2021 across 204 countries/territories and 21 regions globally. GBD 2021 categorized regions into five SDI regions (from low to high) based on fertility rate, per capita income, and education level. Data were sourced from the official GBD database (https://vizhub.healthdata.org/gbd-results/, accessed August 11, 2025), covering all-age, sex-, country-, and region-level asthma burden. Cases were defined according to ICD-10 codes (J45-J46.0). The study was approved by the Institutional Review Board of Washington University School of Medicine. As a secondary analysis of existing publicly available data compliant with the Guidelines for Accurate and Transparent Health Estimates Reporting, no additional ethical review or informed consent was required.

GBD 2021 ensured global data consistency through complex modeling. For regions lacking primary surveillance data, estimates were imputed by spatiotemporal Gaussian process regression and DisMod-MR 2.1 frameworks^12,13, leveraging statistical relationships across time, space, and relevant covariates. All reported data points include 95% uncertainty intervals (UIs) (based on the 2.5th and 97.5th percentiles) to quantify uncertainty arising from data sparsity, measurement bias, and model specification¹³. Details on data source transparency, missing data handling, and regional data accuracy are provided in Text S1.

Risk Factors

The GBD 2021 study adopted the comparative risk assessment (CRA) framework to estimate disease burdens attributable to various risk factors. This framework comprehensively integrates multi-source data, including global censuses, disease registries, healthcare service utilization records, vital statistics, and epidemiological surveys. It commonly employs advanced Bayesian statistical models such as spatiotemporal Gaussian process regression (ST-GPR) and DisMod-MR 2.1 to systematically assess the exposure levels and spatiotemporal distribution trends of diverse risk factors. On this basis, combined with population exposure distribution, relative risk (RR) between risk factors and asthma outcomes, as well as the theoretical minimum risk exposure level (TMREL), the framework calculates population attributable fraction (PAF), and further estimates DALYs and ASDR attributed to specific risk factors¹³. Through the aforementioned unified and standardized statistical modeling procedures, GBD effectively addresses the sparsity and heterogeneity of data from different sources, and greatly improves the comparability of estimation results across countries and regions.

The four key risk factors in this study were high BMI, smoking, occupational asthmagens, and NO₂ pollution.

High BMI is defined as BMI ≥ 25 kg/m² for adults ( ≥ 20 years) or BMI exceeding the age- and sex-specific thresholds set by the International Obesity Task Force for children and adolescents (2-19 years)¹⁴. Smoking includes both active and passive smoking. Occupational asthmagens included exposure to occupational asthmagens such as industrial dust, fumes, gases, and chemicals. NO₂ pollution refers to air pollutants primarily generated from motor vehicle emissions Exposure quantification was performed by mapping population employment patterns to job-exposure matrices (JEMs) in accordance with the International Standard Classification of Occupations (ISCO) and International Standard Industrial Classification (ISIC), to evaluate the exposure probability and intensity of specific substances¹⁵. The theoretical minimum risk exposure level (TMREL) was set to zero, adopting a no-threshold safety assumption¹³. NO₂ pollution refers to air pollutants primarily generated from motor vehicle emissions¹³.

Statistical Analysis

Descriptive Analysis

This study extracted asthma DALYs, ASDRs, the crude DALY rates (CDRs), and the estimated annual percentage changes (EAPCs) attributable to the four risk factors from 1990 to 2021 at the global level, across the five SDI regions, and within the 21 GBD regions from the GBD database. The proportion of DALYs and temporal trends was analyzed for different sexes and age groups globally and within the five SDI regions. The geographical distribution patterns of the asthma ASDRs were visually presented through global spatial distribution maps.

We presented the ASDR per 100,000 population using direct standardization and the World Health Organization (WHO) standard population. ASDRs and its 95% UI were obtained directly from the GBD database. Temporal trends were plotted as line graphs by calendar year. All analyses and graphing were performed using R software (v4.4.0). The study also used EAPC to reflect the trend of asthma burden. EAPC is a method using a regression model to describe the trend of the age-standardized rate (ASR), quantitatively calculating the average annual change rate of the ASR over specific intervals. This study fitted a log-linear model to the asthma ASDRs attributable to the four risk factors from 1990 to 2021: $\mathrm{In}\left({\mathrm{ASDR}}_t\right)=\alpha +\beta t+\epsilon$, where α represents the intercept; β represents the annual change; t represents the calendar year; and ε refers to the error term.

The EAPC value was calculated with the formula: $\mathrm{EAPC}=100\times (\exp \left(\mathrm{\beta }\right)-1)$. Its 95% confidence interval (CI) was derived from the standard error of β. If the EAPC value and its 95% CI were greater than 0, the ASDRs during the corresponding period showed an increasing trend; when the EAPC value and its 95% CI were less than 0, it indicated a decreasing trend. When the EAPC value and its 95% CI included 0, it suggested a stable trend.

SDI Correlation Analysis

To assess the relationship between the level of societal development and disease burden, the SDI was used as the explanatory variable. First, within the 21 GBD regions, the ASDR for each year from 1990 to 2021 was paired with the corresponding year’s SDI value to create “ASDR-SDI” temporal trajectory plots. Locally Estimated Scatterplot Smoothing (LOESS) was used to fit and display the overall trend line, illustrating the dynamic relationship over time. Second, based on cross-sectional data from 204 countries/territories in 2021, Spearman’s rank correlation coefficient (ρ) and two-sided P-value between ASDR and SDI were calculated. LOESS was also used to fit a curve showing the linear relationship between them¹⁶. The statistical significance threshold was set at P < 0.05.

Bayesian Age-Period-Cohort (BAPC) Prediction

To predict the global asthma DALY rates attributable to the four risk factors from 2022 to 2050, this study employed the BAPC model. This model builds upon the classic Age-Period-Cohort framework, using second-order random walks to smooth the priors for age, period, and cohort effects, balancing trend capture with avoiding overfitting, while also controlling for overdispersion to ensure robust fitting. The model constructed a Lexis array using 5-year age groups × calendar year data from 1990-2021. Count data were fitted using a Poisson regression with a log link function. Parameters were inferred via Integrated Nested Laplace Approximation (INLA), outputting the posterior mean and its 95% posterior CI. ASR was directly standardized using the GBD-recommended world standard population. All estimations and predictions were performed in the R environment (v4.3.0) using packages for BAPC and INLA. Historical data and forecast population data for 2022-2050 were sourced from GBD 2021 and the Institute for Health Metrics and Evaluation website (https://vizhub.healthdata.org/population-forecast/)¹⁷.

To validate the predictive performance of the BAPC model, a backtesting analysis was conducted. Historical data from 1990-2010 were used as the training set to forecast ASDRs for 2011-2021. The mean absolute percentage error (MAPE) was calculated to assess the agreement between predicted values and observed values in GBD 2021. The MAPE formula is as follows:

$$\mathrm{MAPE}=\frac{100\%}{n}\sum _{t=1}^n\left|\frac{{\mathrm{Observed}}_t-{\mathrm{Predicted}}_t}{{\mathrm{Observed}}_t}\right|$$

Here, n is the total number of backtesting years. Lower MAPE values indicate greater predictive robustness of the model. To comprehensively evaluate model stability across different risk factors and its ability to capture extreme fluctuations, the mean absolute error (MAE) and root mean square error (RMSE) were also introduced. MAE reflects the average absolute deviation of predicted values from observed values, while RMSE, by averaging squared errors and then taking the square root, can more sensitively capture and quantify predictive fluctuations induced by external policy interventions in specific years. Through this multi-metric assessment, the scientific validity and reliability of the BAPC model in extrapolating population structure-driven disease trends were systematically verified.

Ethics approval and consent to participate

Not applicable

Results

Temporal and Spatial Distribution of Asthma Burden Attributable to the Four Risk Factors Globally and Across the Five SDI Regions, 1990-2021

During 1990-2021, both asthma DALYs and ASDRs linked to the four risk factors declined in most of the world (Table 1; Supplementary Table 1). The steepest reduction was observed for smoking-related asthma (EAPC = -3.13), whereas the smallest decrease was seen for high BMI (EAPC = -0.85), leaving high BMI as the dominant risk source. Trends across the five SDI quintiles largely mirrored the global pattern. However, low-middle-SDI countries experienced a slight increase in high BMI-attributable ASDR (EAPC = 0.25).

Table 1.

ASDR and EAPC for asthma attributable to the four risk factors, 1990-2021, global and five SDI quintiles.

Location	Indicator	NO₂ pollution	Smoking	Occupational asthmagens	High BMI
Global	1990 (95% UI)	4.80 (-4.29,17.75)	41.95 (5.00,79.54)	38.05 (30.56,48.67)	50.53 (24.72,78.10)
	2021 (95% UI)	2.48 (-2.26,10.30)	16.19 (1.89,31.74)	20.75 (16.70,26.50)	39.42 (19.62,60.16)
	EAPC (95% CI)	-1.93 (-2.13,-1.73)	-3.13 (-3.20,-3.07)	-2.07 (-2.11,-2.02)	-0.85 (-0.96,-0.73)
Low SDI	1990 (95% UI)	1.32 (-0.99,7.02)	54.02 (6.16,112.79)	87.12 (66.33,122.41)	58.99 (26.95,93.07)
	2021 (95% UI)	0.96 (-0.77,4.73)	25.28 (2.64,52.19)	45.63 (34.30,61.72)	58.28 (26.47,96.83)
	EAPC (95% CI)	-0.59 (-1.11,-0.07)	-2.56 (-2.68,-2.43)	-2.29 (-2.41,-2.17)	-0.10 (-0.14,-0.06)
Low-middle SDI	1990 (95% UI)	2.16 (-1.69,10.42)	76.27 (9.16,164.36)	66.90 (49.38,102.09)	52.89 (24.28,82.79)
	2021 (95% UI)	1.42 (-1.15,6.23)	34.26 (3.77,71.52)	35.89 (28.00,54.15)	55.50 (24.53,88.23)
	EAPC (95% CI)	-0.97 (-1.47,-0.46)	-2.49 (-2.55,-2.43)	-1.99 (-2.16,-1.81)	0.25 (0.20,0.29)
Middle SDI	1990 (95% UI)	3.88 (-3.42,15.34)	34.27 (4.18,63.45)	29.60 (24.29,37.23)	36.88 (18.02,55.64)
	2021 (95% UI)	2.82 (-2.49,11.44)	12.57 (1.49,23.92)	15.32 (12.69,18.56)	32.06 (16.57,47.76)
	EAPC (95% CI)	-0.68 (-1.06,-0.30)	-3.44 (-3.54,-3.35)	-2.33 (-2.42,-2.23)	-0.65 (-0.73,-0.57)
High-middle SDI	1990 (95% UI)	5.04 (-4.36,18.86)	26.89 (3.34,50.36)	21.31 (16.49,28.12)	43.40 (20.60,67.87)
	2021 (95% UI)	3.47 (-3.24,13.89)	8.45 (1.00,16.96)	9.42 (6.60,12.98)	25.40 (12.75,39.05)
	EAPC (95% CI)	-0.84 (-1.13,-0.54)	-4.13 (-4.28,-3.97)	-3.05 (-3.23,-2.87)	-2.15 (-2.36,-1.94)
High SDI	1990 (95% UI)	17.48 (-19.16,57.81)	45.60 (5.47,91.14)	36.19 (25.64,49.92)	76.66 (36.67,121.16)
	2021 (95% UI)	7.35 (-7.18,30.90)	16.36 (1.80,35.16)	20.74 (14.09,29.47)	62.10 (30.61,95.22)
	EAPC (95% CI)	-2.77 (-3.09,-2.45)	-3.26 (-3.53,-2.98)	-1.72 (-1.96,-1.49)	-0.48 (-0.74,-0.22)

Note ASDR Age-Standardized DALYs Rate per 100,000 population, EAPC Estimated Annual Percentage Change, NO₂ Nitrogen Dioxide, BMI Body-mass Index, SDI Sociodemographic Index, UI Uncertainty Interval, CI Confidence Interval.

In percentage terms, global smoking-attributable asthma burden fell by 61.41%, while burdens from NO₂ pollution, occupational asthmagens, and high BMI dropped by 48.45%, 45.57%, and 21.99%, respectively (Fig. 1A). Regional trajectories differed: low- and low-middle-SDI regions showed fluctuating upward trends for the ASDR related to high BMI but continuous declines for occupational asthmagens (Fig. 1B-C). Middle- and high-middle-SDI regions displayed a three-phase pattern for high BMI (rise-fall-rise), and high-middle-SDI regions saw a rebound in occupational asthmagens (fall then rise) (Fig. 1D-E). High-SDI regions had a rising ASDR for both high BMI and occupational asthmagens after initial decreases (Fig. 1F). Smoking-related ASDR fell in every region, although the magnitude varied (Supplementary Table 2).

Fig. 1. Temporal trends in ASDRs for asthma attributable to the four risk factors, globally and across the five SDI regions, 1990-2021.

(A) Global; (B) Low SDI; (C) Low-middle SDI; (D) Middle SDI; (E) High-middle SDI; (F) High SDI. ASDR, age-standardized DALY rate; NO₂, nitrogen dioxide; BMI, body-mass index. Red = high BMI; green = NO₂ pollution; orange = occupational asthmagens; blue = smoking.

Notably, NO₂-attributable ASDR remained the lowest of the four factors globally and in every SDI stratum, whereas high BMI ranked first in all regions except low- and low-middle-SDI areas (Fig. 1).

Asthma Burden Attributable to the Four Risk Factors by Sex and Age group Across the Five SDI Quintiles, in 1990 and 2021

In 2021, clear age- and sex-specific patterns emerged in the global asthma burden. High BMI was the leading risk factor; its burden rose steadily from adolescence into old age and was consistently higher in females than in males. Smoking and occupational asthmagens predominantly affected middle-aged and older adults, with males carrying a heavier load than females. NO₂-related burden was the smallest and confined to children and adolescents (Fig. 2A; Supplementary Table 3).

Fig. 2. CDRs for asthma attributable to the four risk factors by sex and age group, globally and across five SDI regions, in 2021.

(A) Global; (B) Low SDI; (C) Low-middle SDI; (D) Middle SDI; (E) High-middle SDI; (F) High SDI. CDR, crude DALY rate; NO₂, nitrogen dioxide; BMI, body-mass index. Red = high BMI; green = NO₂ pollution; orange = occupational asthmagens; blue = smoking. Dashed lines = females; solid lines = males.

Across SDI quintiles, the global pattern was largely replicated, except in low-middle-SDI countries, where male smoking-related burden exceeded that of high BMI. Notably, the peak age for occupational asthmagens shifted from 60-64 years in low-SDI regions to 55-59 years for males and 45-49 years for females in high-SDI regions. NO₂-attributable CDR remained the lowest at every SDI level. Sex differences widened with rising SDI, with males gradually exhibiting higher rates (Fig. 2B-F).

Temporal and Spatial Patterns of Asthma Burden Attributable to the Four Risk Factors Across 21 GBD Regions, 1990-2021

During 1990-2021, the majority of the 21 GBD regions experienced downward trends in both absolute DALYs and ASDRs for asthma linked to the four risk factors. However, the relative importance of each factor varied markedly across regions (Table 2; Supplementary Table 4). DALYs attributable to high BMI rose in most regions, with South Asia recording the largest increase ( + 173.34%), whereas High-income Asia Pacific showed the sharpest decline (-55.02%). Occupational asthmagen burdens increased across many developing regions. Smoking-related DALYs fell in most developed regions but continued to climb in regions such as Southeast Asia. NO₂-attributable burdens, consistently the smallest, declined in the majority of the world, yet slight increases were observed in Southeast Asia and Western Sub-Saharan Africa.

Table 2.

ASDR and EAPC for asthma attributable to the four risk factors in 21 GBD regions, 1990-2021.

Location	Indicator	NO₂ pollution	Smoking	Occupational asthmagens	High BMI
Andean Latin America	1990 (95% UI)	13.44 (-10.81,48.37)	6.26 (0.67,12.68)	11.25 (8.83,14.40)	44.30 (26.75,66.86)
	2021 (95% UI)	7.80 (-6.68,30.07)	2.36 (0.23,5.15)	7.62 (5.25,10.62)	29.97 (17.16,46.10)
	EAPC (95% CI)	-1.73 (-2.45,-1.00)	-3.23 (-3.43,-3.04)	-1.14 (-1.27,-1.01)	-1.33 (-1.53,-1.12)
Australasia	1990 (95% UI)	8.06 (-7.07,30.09)	53.25 (5.58,105.42)	52.80 (37.95,72.13)	112.95 (54.91,176.91)
	2021 (95% UI)	2.84 (-2.56,14.49)	16.77 (1.68,36.36)	26.42 (17.93,37.50)	76.14 (37.92,119.92)
	EAPC (95% CI)	-3.92 (-4.44,-3.40)	-3.95 (-4.15,-3.75)	-2.49 (-2.72,-2.26)	-1.51 (-1.68,-1.35)
Caribbean	1990 (95% UI)	5.23 (-4.43,22.57)	26.59 (2.85,52.75)	27.76 (21.93,35.06)	66.91 (34.65,101.78)
	2021 (95% UI)	5.04 (-4.38,22.72)	12.62 (1.33,25.51)	25.26 (19.81,33.80)	63.90 (33.83,96.30)
	EAPC (95% CI)	-1.06 (-1.37,-0.75)	-2.80 (-2.97,-2.63)	-0.33 (-0.38,-0.29)	-0.33 (-0.41,-0.25)
Central Asia	1990 (95% UI)	3.17 (-2.70,11.73)	30.20 (3.48,58.18)	38.49 (32.24,46.16)	85.50 (39.26,131.67)
	2021 (95% UI)	2.24 (-2.03,9.12)	12.52 (1.48,24.34)	16.54 (13.14,20.69)	47.44 (24.10,71.81)
	EAPC (95% CI)	-0.94 (-1.79,-0.08)	-3.60 (-4.00,-3.19)	-3.65 (-4.08,-3.23)	-2.69 (-3.09,-2.29)
Central Europe	1990 (95% UI)	7.50 (-7.10,26.54)	47.17 (5.83,92.21)	32.56 (23.17,45.28)	76.72 (35.91,122.34)
	2021 (95% UI)	4.50 (-4.01,21.97)	14.89 (1.74,30.77)	14.66 (9.37,21.20)	43.43 (21.73,67.69)
	EAPC (95% CI)	-1.97 (-2.59,-1.34)	-3.81 (-4.10,-3.53)	-2.79 (-3.07,-2.52)	-2.00 (-2.28,-1.72)
Central Latin America	1990 (95% UI)	7.35 (-6.85,26.28)	17.08 (2.01,33.74)	19.61 (15.72,24.76)	53.63 (28.29,80.74)
	2021 (95% UI)	3.29 (-2.83,13.27)	3.42 (0.36,7.20)	8.94 (6.50,11.91)	29.05 (15.92,43.67)
	EAPC (95% CI)	-2.58 (-2.95,-2.20)	-5.57 (-5.77,-5.38)	-2.73 (-2.86,-2.60)	-2.23 (-2.39,-2.06)
Central Sub-Saharan Africa	1990 (95% UI)	1.13 (-0.88,4.46)	28.20 (2.90,58.71)	82.96 (56.75,141.13)	62.32 (26.79,132.40)
	2021 (95% UI)	0.69 (-0.59,3.04)	15.44 (1.71,31.98)	45.26 (30.30,80.74)	75.95 (30.95,167.77)
	EAPC (95% CI)	-0.47 (-1.27,0.33)	-1.93 (-2.03,-1.83)	-2.18 (-2.30,-2.06)	0.55 (0.49,0.61)
East Asia	1990 (95% UI)	2.92 (-2.70,12.43)	22.88 (2.86,42.40)	19.26 (14.80,25.62)	17.46 (8.22,27.24)
	2021 (95% UI)	2.26 (-2.07,10.05)	6.89 (0.83,13.13)	8.02 (5.67,11.01)	14.18 (7.25,21.82)
	EAPC (95% CI)	-0.04 (-0.65,0.58)	-4.07 (-4.20,-3.94)	-3.34 (-3.56,-3.11)	-0.94 (-1.15,-0.74)
Eastern Europe	1990 (95% UI)	7.74 (-8.54,25.13)	29.68 (3.76,56.09)	25.59 (19.54,33.73)	61.01 (28.63,94.95)
	2021 (95% UI)	3.53 (-3.30,13.42)	6.37 (0.79,12.98)	7.27 (4.83,10.29)	21.34 (10.59,33.50)
	EAPC (95% CI)	-2.70 (-3.05,-2.35)	-5.81 (-6.09,-5.53)	-4.61 (-4.78,-4.44)	-4.19 (-4.42,-3.96)
Eastern Sub-Saharan Africa	1990 (95% UI)	0.69 (-0.53,4.64)	32.21 (3.57,65.82)	88.82 (69.49,110.99)	48.37 (25.04,73.07)
	2021 (95% UI)	0.46 (-0.36,3.12)	15.66 (1.72,31.35)	49.84 (37.74,67.22)	48.14 (22.82,82.59)
	EAPC (95% CI)	-0.70 (-1.19,-0.21)	-2.53 (-2.60,-2.46)	-2.16 (-2.25,-2.07)	-0.19 (-0.24,-0.14)
High-income Asia Pacific	1990 (95% UI)	7.90 (-7.18,29.90)	43.96 (5.39,85.08)	35.32 (26.64,47.52)	45.15 (21.63,73.03)
	2021 (95% UI)	5.93 (-5.52,21.84)	6.04 (0.68,12.56)	9.24 (5.93,13.68)	15.92 (8.11,25.25)
	EAPC (95% CI)	-0.29 (-0.86,0.28)	-7.03 (-7.26,-6.80)	-4.75 (-4.92,-4.58)	-3.90 (-4.18,-3.62)
High-income North America	1990 (95% UI)	31.52 (-35.11,100.03)	33.66 (3.51,70.18)	29.04 (20.06,40.01)	81.44 (40.39,125.56)
	2021 (95% UI)	10.75 (-10.13,46.56)	22.80 (2.33,49.06)	28.29 (19.36,39.75)	98.56 (48.43,150.97)
	EAPC (95% CI)	-3.51 (-3.88,-3.15)	-0.65 (-1.05,-0.24)	0.32 (0.09,0.54)	1.23 (0.94,1.53)
North Africa and Middle East	1990 (95% UI)	5.45 (-4.19,20.03)	40.80 (4.76,78.70)	30.20 (24.10,39.01)	101.45 (47.21,154.49)
	2021 (95% UI)	3.67 (-2.88,13.73)	16.21 (1.77,32.15)	14.61 (11.79,18.21)	71.41 (37.10,107.87)
	EAPC (95% CI)	-0.94 (-1.33,-0.55)	-3.21 (-3.30,-3.12)	-2.63 (-2.75,-2.50)	-1.37 (-1.46,-1.28)
Oceania	1990 (95% UI)	0.06 (-0.04,0.44)	109.95 (11.39,221.76)	50.50 (37.42,69.11)	197.64 (90.12,328.94)
	2021 (95% UI)	0.11 (-0.08,0.66)	66.92 (6.84,133.47)	49.39 (34.91,73.28)	162.79 (73.99,285.55)
	EAPC (95% CI)	2.33 (1.60,3.07)	-1.77 (-1.86,-1.68)	0.06 (-0.03,0.14)	-0.75 (-0.82,-0.68)
South Asia	1990 (95% UI)	1.90 (-1.46,9.70)	83.05 (9.72,186.18)	72.08 (50.35,119.01)	41.85 (18.35,70.47)
	2021 (95% UI)	1.44 (-1.21,6.28)	33.74 (3.44,73.25)	32.76 (24.44,55.26)	49.83 (20.64,83.93)
	EAPC (95% CI)	-0.07 (-0.81,0.68)	-2.81 (-2.90,-2.73)	-2.54 (-2.75,-2.32)	0.76 (0.67,0.86)
Southeast Asia	1990 (95% UI)	1.45 (-1.17,8.46)	72.70 (8.82,135.78)	55.12 (45.37,65.35)	43.98 (21.26,68.59)
	2021 (95% UI)	2.00 (-1.76,8.55)	32.74 (4.01,61.46)	36.68 (31.36,43.11)	42.96 (20.05,66.78)
	EAPC (95% CI)	1.14 (0.83,1.45)	-2.67 (-2.77,-2.57)	-1.26 (-1.38,-1.13)	-0.12 (-0.25,0.01)
Southern Latin America	1990 (95% UI)	7.02 (-6.26,25.59)	36.97 (4.47,73.25)	26.77 (18.73,37.13)	70.26 (33.30,112.19)
	2021 (95% UI)	5.74 (-5.19,22.19)	18.07 (2.11,37.95)	18.45 (12.15,26.52)	61.93 (29.78,96.74)
	EAPC (95% CI)	-0.31 (-0.67,0.06)	-2.73 (-2.88,-2.58)	-1.32 (-1.41,-1.23)	-0.65 (-0.73,-0.57)
Southern Sub-Saharan Africa	1990 (95% UI)	0.73 (-0.65,4.28)	65.21 (7.64,123.46)	50.31 (43.57,58.03)	116.52 (52.92,187.94)
	2021 (95% UI)	0.56 (-0.49,3.11)	30.66 (3.43,59.48)	36.87 (31.72,42.76)	107.94 (50.65,163.83)
	EAPC (95% CI)	-0.37 (-0.91,0.17)	-2.62 (-3.04,-2.20)	-1.17 (-1.73,-0.61)	-0.33 (-0.83,0.18)
Tropical Latin America	1990 (95% UI)	11.30 (-9.17,43.61)	20.94 (2.52,41.49)	18.01 (12.88,23.95)	42.63 (22.09,65.57)
	2021 (95% UI)	6.36 (-5.45,24.79)	5.78 (0.55,12.66)	11.27 (7.84,15.61)	37.47 (20.79,56.40)
	EAPC (95% CI)	-1.85 (-2.07,-1.64)	-5.05 (-5.42,-4.67)	-2.10 (-2.51,-1.69)	-0.91 (-1.10,-0.72)
Western Europe	1990 (95% UI)	11.58 (-11.84,37.14)	53.39 (6.15,107.66)	40.93 (27.43,58.55)	81.52 (37.96,132.10)
	2021 (95% UI)	4.24 (-4.10,17.65)	17.50 (1.98,37.27)	20.41 (12.94,29.26)	49.71 (23.14,80.19)
	EAPC (95% CI)	-3.34 (-3.68,-3.00)	-3.56 (-3.76,-3.35)	-2.10 (-2.32,-1.89)	-1.58 (-1.78,-1.38)
Western Sub-Saharan Africa	1990 (95% UI)	1.23 (-1.03,5.95)	20.21 (2.11,43.06)	67.39 (53.67,83.07)	64.57 (30.62,102.52)
	2021 (95% UI)	1.03 (-0.85,4.88)	9.43 (0.88,19.50)	39.99 (32.28,49.18)	60.72 (28.99,93.97)
	EAPC (95% CI)	-0.24 (-0.63,0.16)	-2.43 (-2.48,-2.39)	-1.74 (-1.77,-1.70)	-0.21 (-0.27,-0.15)

Note: ASDR Age-Standardized DALYs Rate per 100,000 population, EAPC Estimated Annual Percentage Change, NO₂ Nitrogen Dioxide, BMI Body-mass Index, SDI Sociodemographic Index, UI Uncertainty Interval, CI Confidence Interval.

As for the trends of ASDRs, high BMI ASDR decreased in the majority of regions, but rose in Central Sub-Saharan Africa, High-income North America, and South Asia. Smoking-related ASDR recorded the steepest reductions: Central Latin America (EAPC = -5.57), Eastern Europe (EAPC = -5.81), and High-income Asia Pacific (EAPC = -7.03). Occupational asthmagen ASDR declined fastest in developed regions, notably Central Asia (EAPC = -3.65), Eastern Europe (EAPC = -4.61), and High-income Asia Pacific (EAPC = -4.75). NO₂-related ASDR fell overall, yet modest increases occurred in Southeast Asia and Oceania.

During 1990-2021, the proportional contribution of each risk factor to global asthma DALYs shifted substantially (Fig. 3). High BMI’s DALYs grew steadily from 36.9% to 49.6%, while that of smoking and occupational asthmagens declined from 29.1% to 21.2% and to 26.6%, respectively. NO₂ pollution remained the smallest contributor, falling from 4.8% to 2.7%. Across SDI quintiles, the contribution of high BMI rose with increasing SDI, whereas that of smoking and occupational asthmagens declined. Occupational asthmagens still dominated in the low-SDI quintile.

Fig. 3. Proportion of asthma DALYs attributable to the four risk factors, 1990-2021, shown for the world, five SDI quintiles and 21 GBD regions.

BMI, body-mass index; NO₂, nitrogen dioxide. Red = high BMI; orange = occupational asthmagens; blue = smoking; green = NO₂ pollution.

Regionally, High-income North America exhibited the largest increase in high BMI contribution ( + 13.9%) and the greatest decrease in NO₂ (-9.6%). Oceania consistently displayed the highest proportion of high BMI DALYs (56.2% in 2021). Southern Sub-Saharan Africa retained the world’s largest proportion of occupational asthmagens. Southeast Asia showed only a modest drop in smoking-attributable proportion (-9%), whereas High-income Asia Pacific recorded the steepest decline (-11.8%).

Country/Territory-Level Trends in Asthma Burden Attributable to the Four Risk Factors Across 21 GBD Regions, 1990-2021

In 2021, the spatial distribution of ASDRs attributable to each of the four risk factors among 204 countries/territories is shown in Fig. 4 and Supplementary Table 5. The highest BMI-related asthma ASDR was recorded in Fiji (285.03 per 100,000; 95% UI: 138.37-435.79), followed by Kiribati and Swaziland (Fig. 4A). Zimbabwe displayed the greatest occupational asthmagen burden with an ASDR of 115.46 per 100,000 (95% UI: 84.00-154.23), with Madagascar and Somalia ranking next (Fig. 4B). In 2021, smoking-attributable asthma ASDRs were low in most countries, and the highest values clustered in Pacific island nations, South Asia, and Africa. Kiribati again led the list (118.55 per 100,000), followed by Fiji (82.46 per 100,000) and Papua New Guinea (70.60 per 100,000) (Fig. 4C). Country/territory-level NO₂-attributable ASDRs were universally the lowest of the four factors. The top three rates were observed in Peru, Qatar, and the United States (Fig. 4D).

Fig. 4. Global spatial distribution of ASDRs attributable to the four risk factors, 2021.

(A) High BMI; (B) Occupational asthmagens; (C) Smoking; (D) NO₂ pollution. BMI, body-mass index; NO₂, nitrogen dioxide. Red = high BMI; orange = occupational asthmagens; blue = smoking; green = NO₂ pollution.

Association Between Asthma ASDRs Attributable to the Four Risk Factors and SDI, 1990-2021

The relationship between asthma ASDR and the SDI varied markedly by risk factor across the 21 GBD regions from 1990 to 2021 (Fig. 5). In the 21 GBD regions, the asthma ASDR burden related to high BMI increased with rising SDI levels, peaking in low-middle SDI regions before declining slowly (Fig. 5A). Cross-sectional analysis revealed a significant negative correlation between high BMI-related asthma ASDR and SDI (ρ = -0.27, P < 0.001) (Fig. 5B). SDI showed a significant negative non-linear relationship with asthma ASDR due to occupational asthmagens at both the regional (ρ = -0.73, P < 0.001) and country/territory levels (ρ = -0.55, P < 0.001) (Fig. 5C-D). Similarly, a significant negative correlation was observed between SDI and asthma ASDR due to smoking at both the regional (ρ = -0.51, P < 0.001) and country/territory levels (ρ = -0.20, P = 0.006) (Fig. 5E-F). In contrast, NO₂-attributable ASDR showed a significant positive non-linear relationship with SDI (region ρ = 0.65, country/territory ρ = 0.56; both P < 0.001), with burden rising as SDI increased (Fig. 5G-H).

Fig. 5. Correlation analyses between ASDR and SDI for the four risk factors.

Left panels (A, C, E, G): trajectories of ASDR changing along with SDI across 21 GBD regions, 1990-2021. Right panels (B, D, F, H): correlation between ASDR and SDI for 204 countries/territories in 2021. (A-B) High BMI; (C-D) Occupational asthmagens; (E-F) Smoking; (G-H) NO₂ pollution. BMI, body-mass index; NO₂, nitrogen dioxide. P < 0.05 was considered statistically significant.

Global Asthma Burden Projections (2022-2050) and Model Validation Using BAPC

Backtesting confirmed the BAPC model’s robustness in capturing asthma burden trends (Supplementary Table 6). Quantitative assessment showed that the model fitted well for high BMI, smoking, and occupational asthmagens, with all error metrics at low levels: high BMI (MAE = 2.13, RMSE = 2.29, MAPE = 5.44%), occupational asthmagens (MAE = 0.75, RMSE = 0.83, MAPE = 3.35%), and smoking (MAE = 0.22, RMSE = 0.25, MAPE = 1.16%), indicating strong predictive accuracy (Supplementary Table 7). Although NO₂ pollution showed moderately higher errors (MAE = 0.81, RMSE = 0.99, MAPE = 31.43%), reflecting disruptions from nonlinear policy interventions to historical statistical trends, the model still accurately captured its long-term declining macro-trajectory.

Global projections (Fig. 6; Supplementary Table 8) indicate that high BMI will remain the leading contributor to asthma burden, with ASDR expected to rise from 39.42 per 100,000 in 2021 to 53.79 in 2050 ( + 36.45%). In contrast, burdens attributable to occupational asthmagens, smoking, and NO₂ pollution are projected to decline by 24.10%, 35.70%, and 47.98%, respectively (Fig. 6A).

Fig. 6. BAPC-based projections of asthma DALY rates attributable to the four risk factors, 2022-2050.

(A) All-age ASDR. (B) CDR for ages 0-19 years. (C) CDR for ages 20-64 years. (D) CDR for ages 65+ years. ASDR, age-standardized DALY rate; CDR, crude DALY rate; BMI, body-mass index; NO₂, nitrogen dioxide. Red = high BMI; orange = occupational asthmagens; blue = smoking; green = NO₂ pollution.

In terms of the disparity across age groups, 0-19 years had increasing CDRs for both high BMI and NO₂, reaching an estimated 24.23 and 19.50 per 100,000, respectively, by 2050 (Fig. 6B). 20-64 years had a constantly rising trend in high BMI-attributable burden, and occupational asthmagens declined initially, then rebounded slightly, while NO₂ showed persistent decreases (Fig. 6C). For 65+ years, NO₂ burden keeps falling; high BMI and occupational asthmagens dropped modestly before rising again toward 2050 (Fig. 6D).

Discussion

This study provides a comprehensive assessment of the global asthma burden attributable to four major risk factors during 1990-2021, and quantifies projected trends up to 2050 using a backtest-validated BAPC model. The findings indicate that while asthma burdens from smoking, occupational asthmagens, and NO₂ pollution have generally declined worldwide, high BMI has emerged as the primary driver of increasing asthma burden both globally and in selected low SDI regions. This pattern exhibits pronounced heterogeneity across age groups and geographical settings, revealing a fundamental transition in global respiratory health governance from traditional environmental exposures to metabolic risk factors as emerging determinants of disease burden. Notably, a more rigorous epidemiological perspective is required when interpreting the overall declining trend of disease burden attributable to the aforementioned conventional risk factors. Such burden reduction cannot be simply attributed to decreased risk exposure alone. Instead, it is jointly driven by the remarkable improvement in global primary healthcare accessibility, standardized optimization of clinical diagnostic practices, and widespread implementation of global prevention and treatment guidelines over the past three decades. These systematic advances in healthcare systems serve as indispensable macro-level synergistic contributors to the overall decline in disease burden.

We identified high BMI as the principal risk factor of the globally increasing asthma burden, which aligns with existing evidence¹⁸. As one of the important modifiable risk factors for asthma¹⁹, obesity exacerbates asthma symptoms and increases the frequency of attacks through immunostimulation, increased oxidative stress, heightened airway hyper-responsiveness, and accelerated loss of lung function^20,21. The effect was markedly larger in females, consistent with a previous meta-analysis²². Women are more prone to subcutaneous fat accumulation. Their adipose tissue typically secretes more leptin and less adiponectin, creating a stronger pro-inflammatory state^23,24, thereby increasing the risk of asthma development. Furthermore, estrogen may exacerbate asthma symptoms by enhancing Th2 immune responses and airway smooth muscle contractility²⁵.

Notably, high BMI was the only one of the four risk factors whose attributable burden continued to rise. In high-income countries such as the United States, the unit calorie price of ultra-processed foods is significantly lower than that of healthy foods (0.55 USD vs 1.45 USD per 100 kcal), and their relative price has risen more slowly²⁶, making them the default choice for low-income households. The food industry spends billions marketing directly to adolescents²⁷, shaping lifelong preferences for energy-dense products. Chronic psychological stress further reinforces consumption of “comfort” foods high in sugar and fat²⁸, creating a self-perpetuating stress-eating cycle that promotes the obesity epidemic. Relatively, low- and middle-SDI countries face a more complex, structural predicament. Rapid nutrition transitions coincide with under-resourced public-health systems and scarce capacity for obesity and chronic disease control^29,30. Peru, for example, devotes only 3% of GDP to public health, far below the Latin-American average of 8.3%^31,32, severely constraining effective prevention of obesity-related asthma.

We therefore recommend differentiated strategies. High-SDI regions should use fiscal measures to regulate food prices, such as taxes on sugar-sweetened beverages³³, subsidies for healthy foods, and stricter food labeling regulations³⁴, so as to integrate weight management and asthma control into primary care. Low- and middle-income countries should prioritize improving the accessibility of healthy foods, incorporate weight screening and basic nutrition counselling into primary care, focus maternal and child nutrition interventions, and establish surveillance systems for high-risk groups to maximize the impact of limited resources.

Smoking is a well-established, modifiable risk factor for asthma. Over the past three decades, its ASDR has fallen markedly worldwide, largely reflecting benefiting from the cumulative impact of comprehensive tobacco-control policies implemented across most countries, such as smoke-free public places, advertising bans, and higher excise taxes³⁵. Meanwhile, the decreased ASDR related to smoking is also profoundly affected by improved global medical accessibility. With the long-term promotion of guidelines such as the Global Initiative for Asthma (GINA), primary healthcare has achieved remarkable standardization in chronic disease management¹⁹. Standardized therapy centered on inhaled corticosteroids (ICS) has been widely applied at the population level, which substantially reduces the risks of acute asthma exacerbations and mortality³⁶. Nevertheless, accumulating clinical evidence demonstrates that tobacco smoke exposure markedly diminishes airway responsiveness to ICS and induces relative drug insensitivity, thereby greatly impairing the clinical efficacy of ICS in suppressing airway inflammation and preventing severe acute attacks³⁷. Accordingly, the sustained decline in smoking-attributable disease burden worldwide is predominantly driven by the shrinking population exposed to tobacco. On the premise of reduced tobacco exposure, advances in healthcare systems exert strong synergistic protective effects alongside tobacco control policies. This pathological mechanism also explains why the benefits of standardized routine medical intervention are largely offset in regions with weak tobacco control and persistent high tobacco exposure. Consequently, the local asthma burden remains severe or even presents an upward trend against the general trend^38,39. For instance, although Vietnam has enacted tobacco control legislation, enforcement coverage is limited, and the rate of secondhand smoke exposure among adolescents remains as high as 71.2%⁴⁰. Based on this, it is recommended to establish a specialized monitoring system for tobacco use and asthma burden, and to strengthen technical and financial support for resource-limited countries through international cooperation. Only by simultaneously implementing policies, building healthcare capacity, and ensuring financial support can the smoking-related asthma burden be effectively reduced, and public health levels improved in resource-constrained environments.

Occupational exposure to asthmagens is a major cause of adult-onset asthma and displays pronounced geographical and sex-specific patterns. In low-income settings, such as Eastern Sub-Saharan Africa, the proportional burden of occupational asthmagens has declined, but absolute DALYs remain high, reflecting weak occupational-health infrastructure, limited regulation, and low worker awareness⁴¹. High-income countries have achieved substantial reductions in traditional industries through stringent oversight^42–44. Furthermore, the global reduction in the burden of work-related asthma is fundamentally attributable to the effective correction of case misclassification bias brought about by standardized diagnosis and treatment. Guided by internationally authoritative guidelines and multidisciplinary collaboration, more rigorous etiological differentiation, combined with objective physiological monitoring such as PEF detection, enables accurate distinction between occupational asthma and mimicking disorders, including chronic obstructive pulmonary disease and pneumoconiosis. In addition, timely cessation of hazardous exposure and standardized clinical management have substantially reduced the incidence and recurrence risk of this disease at the population level⁴⁵. Nevertheless, despite increasing standardization within traditional industries, global occupational health governance still faces emerging challenges posed by novel allergens in newly developed sectors, which has led to a rebound in ASDR in certain countries⁴⁶. Males carry a heavier burden than females, largely because manufacturing, construction, and other high-exposure sectors remain male-dominated⁴⁷. At a deeper level, gender norms restrict women to lower-paid, lower-exposure jobs, while their health concerns are frequently overlooked, and their access to diagnosis and compensation is poorer^48,49. Research paradigms have historically centered on male workers, leading to systematic underestimation of the female burden⁵⁰. Addressing regional disparities requires strengthening the protection systems and regulatory capacities in low-resource areas, while enhancing comprehensive risk control for both traditional and emerging industries in high-income countries. Regarding gender aspects, it is essential to promote occupational equity, improve women’s health rights, and rectify gender biases in research.

Elevated ambient NO₂ concentrations are consistently associated with reduced lung function and increased emergency visits for childhood asthma^7,51. Children are intrinsically more vulnerable than adults: their respiratory and immune systems are still developing, their minute ventilation per kilogram of body weight is higher, they breathe more through the mouth, and they spend greater time outdoors, together resulting in children’s disproportionately higher risks and levels of exposure to air pollution⁵². Co-exposure to indoor allergens and environmental tobacco smoke may act synergistically with NO₂ to amplify respiratory harm⁵³. Notably, many high-SDI regions retain elevated NO₂ levels despite stringent emission controls, simply because of high vehicle ownership and traffic density⁵⁴. Conversely, several countries with the highest SDI scores have preliminarily decreased the NO₂-related burden by deploying electric vehicles and building green-transport infrastructure⁵⁵. Furthermore, improved medical accessibility also serves as a crucial buffer driving this burden decline. Regions with an extremely high SDI have well-established early screening systems for childhood asthma and efficient primary care intervention networks. Timely and standardized chronic disease management effectively halts the progression from pollutant exposure to severe acute episodes. These findings indicate that the prior decline of disease burden in extremely high SDI regions stems essentially from the synergistic protective effects achieved through dual interventions: source-based risk control and end-stage medical security. This study, through backtesting validation, found that the BAPC model exhibited remarkable stability in projecting indicators with distinct biological characteristics and intergenerational continuity, such as high BMI, owing to its capacity to capture long-term statistical associations between population age structure dynamics and historical exposure risks. However, for the burden attributable to NO₂ pollution, a certain degree of discrepancy was observed between predicted and observed values, likely attributable to the nonlinear effects of exogenous environmental policy interventions. Specifically, NO₂-attributable burden has been substantially influenced by global public health governance measures. In 2021, the WHO released updated Global Air Quality Guidelines, tightening the recommended annual mean concentration limit for NO₂ from 40 μg/m³ to 10 μg/m³ ⁵⁶. This landmark revision prompted many countries to adopt emission reduction measures with unprecedented intensity, such as establishing Low Emission Zones and accelerating the electrification of public transport^57,58. Abrupt and robust policy interventions accelerated the decline in actual burden beyond the linear extrapolation of historical trends observed over the preceding three decades. Accordingly, the 2050 projections in this study should be interpreted as an important “business-as-usual scenario” warning illustrating the substantial public health challenges that may emerge, particularly from the continued rise in high BMI, in the absence of additional interventions. These findings provide a critical evidence base for policymakers to formulate forward-looking preventive strategies.

This study has several limitations. First, asthma cases in this study were defined according to ICD-10 codes (J45-J46). This approach, which relies heavily on original disease registration and surveillance systems, has inherent limitations. Globally, especially in low- and middle-income countries short of medical resources and objective diagnostic tools such as pulmonary function testing devices, a large number of asthma cases tend to be underdiagnosed or misdiagnosed⁵⁹. Such marked regional disparities in diagnostic capacity may lead to systematic underestimation of the actual prevalence and overall disease burden of asthma in low SDI regions. Moreover, although sophisticated algorithms were adopted in GBD 2021 for missing data imputation, the accuracy of relevant estimates remains highly dependent on raw surveillance data. Due to severe data sparsity, estimates in low SDI regions rely heavily on statistical models. While 95% UIs have quantified such uncertainty, additional baseline survey data are still required to optimize and revise global prediction models in future research. Second, exposure assessments for all risk factors in the present study were conducted within the unified standardized modeling framework of GBD. Nevertheless, objective discrepancies exist in the sources and quality of surveillance data across nations and regions, which may give rise to measurable uncertainty in exposure measurement and relevant classification bias, with varying severity among different risk factors¹³. Notably, occupational exposure carries a particularly high risk of misclassification. Substantial heterogeneity persists worldwide in occupational classification systems, occupational disease surveillance capacity and occupational health and safety criteria, making fully standardized data unification difficult to achieve⁴¹. Accordingly, the actual exposure levels to occupational asthmagens and their attributable disease burden may be misclassified or underestimated in regions with inadequate occupational supervision and scarce basic epidemiological data. In conclusion, the combined effects of limitations in diagnostic coding and deviations in occupational exposure classification collectively weaken the robustness of cross-national and cross-SDI regional comparisons in this study. Third, the GBD framework did not fully account for potential interactions among risk factors. For instance, synergistic effects between occupational dust exposure and smoking, or joint effects of NO₂ exposure and asthmagens, which are incompletely quantified, may lead to underestimation of the attributable burden for individual factors. Fourth, while the BAPC model excels in fitting chronic risk trends, it is subject to limitations in capturing nonlinear shifts strongly driven by external policy interventions, as demonstrated by the backtesting bias for NO₂. Finally, the projections are predicated on the baseline assumption that exposure trends for risk factors will maintain historical inertia, without explicitly modeling the impact of future public health interventions within the predictive framework. Should countries implement more targeted interventions (e.g., stricter energy transition policies or obesity control strategies), the actual asthma burden by 2050 may be lower than projected.

Conclusion

The study provides a comprehensive analysis of the spatiotemporal dynamics of the global asthma burden attributable to four major risk factors from 1990 to 2021, revealing its strong association with the SDI and projecting future risk patterns through 2050. While asthma burdens related to smoking, occupational exposures, and NO₂ pollution have declined substantially, high BMI stands out as the sole risk factor with a consistently rising trajectory, signaling an emerging and escalating public health challenge. Global asthma governance is thus undergoing a fundamental shift in its dominant drivers. Moving forward, prevention and control strategies should prioritize population-level weight management and occupational health protection, while maintaining progress in air quality control and tobacco cessation. It should be emphasized that as a macro observational study based on global public databases, the study’s findings are primarily intended to provide baseline evidence for population-level public health planning and resource allocation, rather than to establish definitive causal inferences or directly guide individualized clinical diagnosis and treatment. Overall, this study offers critical scientific evidence and early warning insights to support the formulation of precise, differentiated strategies tailored to countries and populations across the SDI spectrum.

Author contributions

Conception and design of study: Qiming Dai, Jin Xiang Acquisition of data: Jin Xiang Analysis and interpretation of data: Qiming Dai Drafting of the manuscript: Qiming Dai, Jin Xiang Revising the manuscript critically for important intellectual content: Jin Xiang Approval of the version of the manuscript to be published: Qiming Dai, Jin Xiang.

Funding

None

Data Availability

The data and materials in the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41533-026-00530-5.