1. Introduction
Asthma is a complex chronic inflammatory disorder of the airways, leading to recurrent episodes of wheezing and breathlessness.[1] The global burden of disease (GBD) 2021 study reported a global prevalence of around 262 million asthma cases, equating to an age-standardized rate (ASR) of approximately 3340/100,000 people. A critical and paradoxical trend has emerged: Although the age-standardized mortality rate (ASMR) has steadily declined since 1990, the absolute number of individuals living with asthma and dying from it has risen significantly. This escalating burden places a particularly heavy strain on vulnerable groups, including children and older adults.[2] These trends suggest that despite therapeutic advances, medical interventions alone are insufficient to curb the asthma crisis.
The pathogenesis of asthma involves a complex interplay of genetic predisposition,[3] environmental exposures,[4] and immune dysregulation.[5]
Body mass index (BMI) is a frequently used indicator of obesity and represents a key modifiable risk factor for asthma. According to GBD 2021, a high BMI is a significant risk factor for asthma-related disability-adjusted life years (DALYs) worldwide, accounting for 14.93% of the total disease burden.[6] A large number of observational studies have consistently linked an elevated BMI to an increased risk of asthma.[7,8] Potential biological mechanisms underlying this association include altered airway mechanics, systemic inflammation, metabolic dysfunction, and gut microbiome dysbiosis.[9]
Previous GBD-based studies on the asthma burden attributable to a high BMI have mostly focused on older adults,[10] and some relied on the outdated GBD 2019 dataset.[11] Consequently, these studies could not capture recent trends, including the potential impact of the COVID-19 pandemic. This research gap highlights the need for a comprehensive, up-to-date assessment spanning all ages, genders, and socio-demographic index (SDI) categories, and one that integrates genetic evidence from Mendelian randomization (MR). To address these gaps, this study presents the 1st comprehensive analysis of the spatiotemporal trends in high BMI-related asthma burden using the GBD 2021 database.
We examined variations by SDI, age, and sex, identified key drivers of burden changes, and projected trends to 2050. Furthermore, to minimize confounding, we performed a 2-sample MR analysis to evaluate the association between genetically predicted BMI and asthma risk. By integrating descriptive epidemiology with MR, this research aims to elucidate the role of a high BMI in driving the evolution of asthma burden and to provide evidence for shaping obesity prevention and global asthma control strategies.
2. Methods
2.1. Data collection
This analysis used the GBD 2021 dataset to quantify the worldwide burden of asthma resulting from high BMI. The GBD 2021 provides modeled estimates for 371 health conditions and 88 risk factors across geographical scales from national to global.[12,13] From the global health data exchange platform, we retrieved metrics including death numbers, DALYs, and ASRs for high BMI-related asthma. All estimates were obtained with their 95% uncertainty intervals (UIs) for the years 1990 through 2021.
The dataset has a global scope, encompassing 204 countries and territories categorized into 21 GBD regions and 5 SDI quintiles. Subgroup analyses were conducted by sex and age group (5-year intervals from 1 to 94 years, plus a category for ≥95 years). Corresponding SDI values were also obtained.
For the MR analysis, summary-level genetic data were sourced from the MRC integrative epidemiology unit (IEU) OpenGWAS database. Genetic instruments for high BMI were obtained from a genome-wide association study (GWAS) of 457,756 individuals of European ancestry[14] (accession: ebi-a-GCST90025994). Summary-level data for asthma were derived from an independent GWAS meta-analysis including 449,500 European-ancestry participants[15] (accession: ebi-a-GCST90018795).
Since the exposure (BMI) and outcome (asthma) data came from nonoverlapping studies, sample overlap was assumed to be minimal. The IEU OpenGWAS database was queried with the filters: trait = “body mass index” or “asthma,” population = “European,” and sample size >400,000. The GBD 2021 estimates used in this study are publicly available, de-identified, aggregate-level data that contain no individual-level information.
All original GWAS from which we obtained summary-level data had received ethical approval, and participants had provided informed consent. The use of these de-identified summary statistics for the present analysis is considered exempt from additional ethical review.
2.2. Definitions
We defined asthma as a persistent respiratory disorder marked by symptoms of bronchospasm and shortness of breath, most commonly triggered by allergic responses. A clinical diagnosis generally necessitates confirmation by a healthcare professional along with a history of wheezing within the previous 12 months. For this study, asthma cases were primarily identified using International Classification of Diseases, 10th revision codes J45 and J46.
Supplementary case definitions included self-reported physician diagnosis, self-reported wheezing episodes or asthma within the past year, or any self-reported lifetime asthma.[16] BMI was computed as the ratio of weight (in kilograms) to the square of height (in meters). Individuals aged 20 years or above with a BMI of 25 kg/m2 or greater were defined as having a high BMI, in accordance with established criteria.[17] For participants below 20 years of age, we adhered to the age-specific reference criteria set forth by the international obesity task force. The proportion of the asthma burden that could be attributed to high BMI, quantified as the population attributable fraction, was calculated following the comparative risk assessment methodology.[12] This estimation process incorporated 3 key elements: the exposure–response relationship derived from meta-analyses of epidemiological studies, the prevailing global distribution of BMI levels, and the theoretical minimum risk exposure level.
DALY is a comprehensive measure that combines years of life lost due to premature mortality with years lived with disability, providing a standardized quantification of overall disease burden. SDI serves as a summary indicator of regional development, synthesized from 3 fundamental components: income per capita, average educational achievement, and the total fertility rate. Within the GBD 2021 analytical framework, all geographical units were stratified into 5 distinct SDI tiers: high (>0.81), high-middle (0.70–0.81), middle (0.61–0.69), low-middle (0.46–0.60), and low (<0.46).
2.3. Statistical analysis
The analysis comprised 4 components: burden description, trend analysis, decomposition and prediction, and MR analysis.
2.3.1. Burden description
To facilitate comparisons of the asthma burden by different populations and temporal periods, we relied on ASMR and age-standardized DALY rate (ASDR). These metrics, computed for every 100,000 individuals based on the GBD World Standard Population, adjust for variations in age structure. The standardization follows this calculation:
ri is the age-specific rate for the ith age group, wi is the corresponding weight in the reference population, and A is the total number of age groups.
The estimated annual percentage change (EAPC) was derived to quantify the temporal trend of the ASR. This metric was obtained by 1st applying a linear regression model to the natural log-transformed rates, defined by the equation: ln(ASR)=α+β·year+ε. The EAPC value was then calculated based on the regression coefficient β using the formula:
A trend was considered significantly increasing if both the EAPC estimate and its lower 95% confidence limit were >0, and significantly decreasing if both were <0.
2.3.2. Joinpoint regression analysis
To characterize temporal patterns, we employed Joinpoint regression. This technique models trends by identifying points where the slope changes significantly, calculating the annual percent change (APC) for each segment and the average annual percent change (AAPC) for the entire 1990 to 2021 span, along with their 95% confidence interval (CI). The best-fitting model, allowing for a maximum of 5 joinpoints, was identified by minimizing the akaike information criterion.
A permutation test (10,000 replicates; significance threshold P < .05) was used to validate the final model selection. An overall trend was considered meaningful if the AAPC was statistically significant (P < .05). All Joinpoint analyses were conducted with the Joinpoint Regression Program (Version 5.0.1 from the National Cancer Institute).
2.3.3. Decomposition analysis and prediction analysis
We applied the Das Gupta decomposition method[18] to assess the contributions of population growth, population aging, and changes in epidemiological rates to the observed changes in deaths and DALYs due to high BMI-related asthma. This method is symmetric in that the factor contributions do not depend on the order in which the factors are considered, and no additional standard-population weighting is applied beyond the population counts and rates for the 2 endpoint years.
Projections of the future disease burden for 2022 to 2050 were generated using a Bayesian age–period–cohort (BAPC) model. The model was fitted utilizing the integrated nested Laplace approximation technique,[19] which offers computational advantages and avoids convergence problems common with conventional Markov Chain Monte Carlo sampling. Temporal smoothing was ensured by incorporating 2nd-order random walk (RW2) priors for the age, period, and cohort effects, following the default specification of the BAPC framework.
RW2 priors allow locally linear trends to be learned from the data while penalizing excessive volatility, and are therefore well-suited to capturing nonlinear temporal dynamics. The model’s predictive performance was evaluated by training it on data from 1990 to 2010 and forecasting the period 2011 to 2021. The strong concordance between the projected estimates and the actual observed values during this validation period provided confidence in the model’s forecasts up to 2050.
In sensitivity analyses, all 4 sex- and measure-specific models were refitted with 1st-order random walk (RW1) priors to assess the robustness of the projections to the prior specification.
2.3.4. MR analysis
A bidirectional 2-sample MR analysis was employed to evaluate the association between genetically predicted BMI and asthma risk. This analysis rests on 3 key instrumental variable assumptions[20]: relevance: genetic instruments must be strongly associated with BMI; independence: instruments are independent of confounders affecting the BMI–asthma relationship; and exclusion restriction: instruments influence asthma risk only via BMI, not through alternative pathways. Genetic instruments for BMI were selected as single-nucleotide polymorphisms (SNPs) reaching genome-wide significance (P < 5 × 10−8) from a large-scale GWAS of 457,756 individuals of European ancestry.[14] To ensure independence, linkage disequilibrium (LD) clumping was performed (window size = 10,000 kb, r2 < 0.001).
SNPs associated with asthma (P < 5 × 10−8) were excluded to minimize pleiotropy. This process yielded 368 independent SNPs, of which 361 were retained after harmonization as final instruments. All MR analyses were performed with the TwoSampleMR R package using summary-level data and the European LD reference panel from the OpenGWAS database.
The strength of these instruments was confirmed, with all SNP-specific F statistics exceeding 10 (mean F statistic = 68.4, minimum = 28.5; Table S1, Supplemental Digital Content 1), indicating minimal weak instrument bias. The primary estimate was derived using the inverse-variance weighted (IVW) method. To evaluate robustness, we conducted sensitivity analyses including MR-Egger regression, the weighted median method, the weighted mode and simple mode estimators, Cochran’s Q test for heterogeneity, the MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) test, and leave-one-out analysis.
The MR-PRESSO test was implemented using the MR-PRESSO R package, with 10,000 Monte Carlo simulations performed to ensure the precision of the empirical P values of the global, outlier, and distortion tests (for 361 instruments, at least 7220 simulations – 20 per instrument – are required to reach the nominal significance level of 0.05 for the outlier test). MR-Egger regression further relies on the InSIDE assumption (i.e., that instrument strength is independent of the direct pleiotropic effects), a condition that cannot be directly tested in summary-data MR and whose violation may bias the MR-Egger estimate.
In order to evaluate the possibility of reverse causality, we conducted a reverse MR analysis with asthma as the exposure and BMI as the outcome, using the same GWAS summary datasets.[14,15] We identified 67 independent SNPs associated with asthma at genome-wide significance (P < 5 × 10−8), applying the same LD clumping criteria and quality control procedures as in the forward analysis. After harmonization and proxy-based lookup, 48 instruments were retained. These instruments were robust (mean F statistic = 65.5, minimum = 29.9), substantially exceeding the recommended threshold of 10, thus mitigating concerns of weak instrument bias.
All MR analyses followed the Strengthening the Reporting of Observational Studies in Epidemiology for Mendelian Randomization (STROBE-MR) guidelines.[21]
2.4. Statistical statement
The reported values for all GBD-derived estimates correspond to the mean of 1000 draws from the posterior distribution. The 95% UIs are defined by the 2.5th and 97.5th percentiles of these posterior draws. All rates are presented per 100,000 people.
For the MR analysis, results are reported as odds ratios (ORs) accompanied by 95% CIs. A 2-sided P value below .05 served as the threshold for determining statistical significance. The statistical computing environment R (version 4.3.2; R Foundation for Statistical Computing, Vienna, Austria) was used for all analyses, leveraging key packages such as ggplot2, JoinpointR, BAPC, and TwoSampleMR (version 0.7.11).
This study is reported in accordance with the STROBE-MR guidelines for Mendelian randomization studies and the Guidelines for Accurate and Transparent Health Estimates Reporting statement for studies using GBD data and global health estimates; a completed STROBE-MR checklist accompanies this submission, and the completed Guidelines for Accurate and Transparent Health Estimates Reporting checklist is available from the corresponding author upon reasonable request.
3. Results
3.1. Global burden of high BMI-associated asthma
A steady decline was observed in the global ASMR and ASDR for high BMI-related asthma between 1990 and 2021. The ASMR decreased from 1.00 (95% UI: 0.43–1.58) to 0.74 (95% UI: 0.31–1.19)/100,000, with an EAPC of −1.13 (95% CI: −1.23 to −1.04). Similarly, the ASDR fell from 50.53 (95% UI: 24.72–78.10) to 39.42 (95% UI: 19.62–60.16)/100,000, corresponding to an EAPC of −0.85 (95% CI: −0.96 to–0.73) (Tables 1 and 2; Fig.
1).
Conversely, the absolute number of asthma deaths attributable to high BMI increased by 63.94%, from 38,040.53 (95% UI: 16,523.50–59,933.15) in 1990 to 62,363.20 (95% UI: 26,317.27–100,543.22) in 2021 (Table 1). Similarly, DALYs increased by 45.67%, from 2,253,937 (95% UI: 1,125,455–3,432,736) to 3,283,296 (95% UI: 1,620,702–5,034,164) over the same period (Table 2).
3.2. Regional burden of high BMI-associated asthma
The low-middle SDI region recorded the greatest absolute number of asthma-related deaths (24,557.84) and DALYs (874,703.70) associated with high BMI in 2021 (Tables 1 and 2). A distinctive pattern was observed in this region, which was the only SDI category to record concurrent increases in both ASMR and ASDR throughout the period (1990–2021; Fig. 1).
In stark contrast, all other SDI regions exhibited universal declining trends. The most pronounced decrease in ASMR occurred in the high SDI region, while the greatest reduction in ASDR was seen in the high-middle SDI region (Fig. 1).
Across the 21 GBD regions, South Asia accounted for the largest share of asthma deaths (25,159.69; 95% UI: 10,125.98–45,192.89) and DALYs (776,006.01; 95% UI: 327,108.28–1,289,399.31) caused by high BMI in 2021 (Tables 1 and 2; Fig. 2A and B). A different pattern emerged upon examining ASRs: Oceania presented the most severe burden globally, with an ASMR of 6.10 and an ASDR of 162.79/100,000 population (Tables 1 and 2; Fig.
2C and D). An analysis of temporal trends from 1990 to 2021 revealed that high-income Asia Pacific experienced the steepest mortality rate reduction (EAPC = −7.84), whereas Eastern Europe showed the most rapid decline in DALY rates (EAPC = −4.19). On the other end of the spectrum, South Asia displayed the most substantial upward trend in ASMR (EAPC = 1.02), and high-income North America led in ASDR increases (EAPC = 1.23) (Tables 1 and 2).
3.3. National burden of high BMI-associated asthma
In 2021, India recorded the highest number of asthma deaths attributable to high BMI (21,018.30; 95% UI: 8407.39–36,835.71), followed by China (4321.91; 95% UI: 1949.72–7146.99) and Indonesia (3231.94; 95% UI: 1439.41–5316.53) (Table S2, Supplemental Digital Content 2). A comparable trend was noted in relation to DALYs, with India again having the highest burden (632,587.57; 95% UI: 261,685.39–1,046,890.64), followed by the United States (389,253.62; 95% UI: 187,273.68–606,143.36) and China (218,662.32; 95% UI: 107,327.86–339,079.19) (Table S3, Supplemental Digital Content 3). Regarding ASRs, Fiji recorded the greatest ASMR (12.84/100,000; 95% UI: 5.96–20.04) and ASDR (285.03/100,000; 95% UI: 138.37–435.79) globally in 2021, with Kiribati ranking 2nd (Fig.
3A and B; Tables S2 and S3, Supplemental Digital Content 2). A longitudinal analysis from 1990 to 2021 identified 26 countries with rising ASMR trends. Additionally, 41 countries showed increasing ASDR.
Among the countries with rising trends, Zimbabwe stood out with the sharpest increases in both ASMR and ASDR, exhibiting EAPCs of 2.97 (95% CI: 2.56–3.38) and 3.01 (95% CI: 2.55–3.48) (Tables S2 and S3, Supplemental Digital Content 2).
3.4. High BMI-associated asthma burden stratified by age group and sex
The distribution of asthma deaths associated with high BMI in 2021 exhibited a clear age-dependent pattern (Fig. 4A). The number of deaths was relatively low during childhood, then gradually increased with advancing age.
Both men and women reached their peak in the 70 to 74 age group, and then it decreased again. A consistent sex disparity was apparent throughout the lifespan, with female death numbers persistently surpassing those of males in every age group. When examining mortality rates, a steady age-related increase was observed, culminating in the highest values among women aged 95 years and above.
For men, the mortality rate maximized in the 90 to 94 year age group before stabilizing.
An analogous profile was evident for DALYs (Fig. 4B). The number of DALYs decreased from ages 5 to 19 before gradually increasing to their maximum.
The peak occurred in the 60 to 64 age group for men and the 55 to 59 age group for women, followed by a subsequent decline. A notable sex-specific crossover was identified: while males below 15 years of age exhibited greater DALY numbers compared to females, this pattern reversed after age 15, from which point females shouldered a progressively larger proportion of the DALY burden. The ASDR demonstrated a near-linear ascending trajectory from ages 15 to 79 for both males and females.
3.5. Correlation of SDI with high BMI-associated asthma burden
A significant inverse relationship was identified between the SDI and ASMR of asthma attributable to high BMI, both globally and across the 21 GBD regions from 1990 to 2021 (ρ = −0.70, P < .001; Fig. 5A). Higher SDI was generally associated with a lower mortality burden.
A weaker, but still significant, inverse correlation was found between SDI and ASDR (ρ = −0.12, P < .001; Fig. S1A, Supplemental Digital Content 4). Notably, regions including Oceania and sub-Saharan Africa exhibited ASMR and ASDR values that markedly exceeded the expectations based on their SDI levels.
This inverse association persisted at the national level in 2021 for both ASMR (ρ = −0.64, P < .001; Fig. 5B) and ASDR (ρ = −0.25, P < .001; Fig. S1B, Supplemental Digital Content 4).
Fiji was a notable outlier, with observed rates far exceeding the predictions based on its SDI.
3.6. Joinpoint regression analysis of high BMI-associated asthma burden
Joinpoint analysis for the period 1990 to 2021 revealed an overall declining trend in the global ASMR (AAPC = −0.97) and ASDR (AAPC = −0.79) for asthma attributable to high BMI. The most substantial decline in ASMR was recorded in the 1996 to 2000 interval (APC = −2.29), after which the rate of decline decelerated (Fig. 6A).
A similar pattern was observed for ASDR, with the steepest decrease from 1996 to 2000 (APC = –1.90), followed by a slower rate of decline. Of note, the ASDR exhibited a slight but significant increase between 2015 and 2021 (APC = 0.2, P < .05; Fig. 6B).
3.7. Decomposition analysis of high BMI-associated asthma burden
Decomposition analysis indicated that the increases in deaths and DALYs from high BMI-related asthma between 1990 and 2021 were primarily driven by population growth and population aging, while the overall contribution of changes in epidemiological rates was relatively modest (Fig. 7; Table S4, Supplemental Digital Content 5).
Notable disparities were observed across SDI regions. High and high-middle SDI regions experienced a net decrease in deaths, largely due to favorable epidemiological changes that reduced death rates, especially in high-SDI regions. In middle SDI and low SDI regions, although epidemiological changes also contributed to risk reduction, this effect was more than offset by the strong positive contributions from population growth and aging, resulting in a net increase in deaths (Fig.
7A). Trends for DALYs broadly mirrored those for mortality (Fig. 7B), with the exception that high-SDI regions showed a small net increase in DALYs despite a net decrease in deaths, driven by the overwhelming contributions of population growth and aging.
It is important to highlight that unfavorable epidemiological changes, which increased the burden of asthma-related deaths and DALYs, were observed exclusively in the low-middle SDI regions (Fig. 7).
3.8. Prediction of high BMI-associated asthma burden
The BAPC model was employed to project the high BMI-attributable asthma burden through the year 2050. Globally, a steady decline in ASMR is forecast for both males and females throughout this period (Fig. 8).
The decline is expected to be slightly faster in females than in males. By 2050, the ASMR is predicted to be 0.60/100,000 individuals for males and 0.65/100,000 individuals for females (Fig. 8A and B).
In contrast, the ASDR is forecast to increase modestly. By 2050, the male ASDR is expected to reach 43.40/100,000, and the female ASDR is projected to be 54.68/100,000 (Fig. 8C and D).
These projections indicate that females will likely continue to experience a disproportionately higher burden in the future. In sensitivity analyses, refitting all 4 sex- and measure-specific models with RW1 priors yielded materially different 2022 to 2050 trajectories for some outcomes (Fig. S2, Supplemental Digital Content 6).
The RW2-based projections were therefore retained as the primary results.
3.9. Bidirectional MR of the association between high BMI and asthma
Results from the MR suggested a positive, genetically supported association between BMI and asthma risk. The primary IVW analysis showed that each 1-unit increase in genetically predicted BMI was associated with higher odds of asthma (OR = 1.389, 95% CI: 1.303–1.479, P = 3.3 × 10−24; Fig. 9; Table 3).
Using 361 genetic instruments, we observed substantial heterogeneity (Cochran’s Q = 669.07, P = 7.4 × 10−21; Table 3), which is consistent with widespread horizontal pleiotropy or other violations of the MR assumptions; the IVW estimate should therefore be interpreted with caution. The weighted median (OR = 1.282, 95% CI: 1.169–1.405, P = 1.2 × 10−7) and MR-Egger (OR = 1.317, 95% CI: 1.118–1.551, P = 1.1 × 10−3) estimates, which are more robust to horizontal pleiotropy than the IVW method, were directionally consistent with the IVW estimate (Table 3). The MR-PRESSO global test indicated the presence of pleiotropic outliers (P < 1 × 10−4).
After removal of 3 outlier SNPs (rs1800437, rs429358, and rs6559365), the corrected estimate remained essentially unchanged (OR = 1.388, 95% CI: 1.305–1.476, P = 1.6 × 10−22; Table 3; Fig. S3, Supplemental Digital Content 7), and the distortion test showed no significant difference between the estimates before and after correction (P = .989; Table 3). Leave-one-out analysis confirmed that the association was not driven by any single SNP (Fig.
S4, Supplemental Digital Content 8), and the funnel plot did not suggest pronounced asymmetry (Fig. S5, Supplemental Digital Content 9). The forest plot of individual SNP effects showed predominantly positive effect directions across variants (252/361, 69.8%), although most individual estimates had confidence intervals crossing the null, consistent with weak instrument strength (Fig.
S6, Supplemental Digital Content 10). The MR-Egger intercept test did not indicate directional pleiotropy (P = .492; Table 3); however, this test has limited statistical power, particularly in the presence of substantial heterogeneity, and residual pleiotropy cannot be excluded. Therefore, the consistency of estimates across methods should be interpreted as supportive of, but not definitive for, a robust association.
To evaluate reverse causation, we performed a reverse MR analysis using 48 independent genetic instruments for asthma. The IVW analysis found no significant effect of genetically predicted asthma on BMI (OR = 0.988, 95% CI: 0.965–1.011, P = .296; Fig. S7, Supplemental Digital Content 11; Table 3).
This null association was broadly consistent across all sensitivity analyses, with no evidence of directional pleiotropy (MR-Egger intercept P = .997; Table 3).
Collectively, these findings support a unidirectional, genetically supported association between higher BMI and asthma risk, with no evidence of a reverse association.
4. Discussion
This study integrates data from the GBD and GWAS to provide a comprehensive assessment of the burden of high BMI-related asthma and of the genetically supported association between high BMI and asthma risk. Key findings reveal a substantial growth in the absolute asthma burden attributable to high BMI over the 3 decades between 1990 and 2021. In contrast, ASMR and ASDR exhibited a downward trend.
This pattern, indicative of “falling rates but rising numbers” underscores that population growth and aging, rather than increasing individual-level risk, are the main drivers of the growing absolute burden. These findings highlight an escalating public health challenge, necessitating a deeper exploration of the biological mechanisms and historical context of asthma management.
Epidemiological and clinical research has demonstrated a significant association between elevated BMI and the risk of asthma.[7,22] Our study strengthens this evidence by integrating population-level burden estimates with genetic association evidence from MR. This approach not only corroborates the association but also quantifies its population-wide impact over 3 decades, extending previous research.[23] These robust findings underscore the importance of weight management as a cornerstone of asthma clinical care and prevention, particularly for high-risk individuals.[24] Beyond this modifiable risk factor, advancements in asthma treatment have also played a crucial role. The pronounced decline in ASMR and ASDR between 1996 and 2000 coincides with a pivotal shift in global asthma management, following the dissemination and implementation of the 1995 global initiative for asthma guidelines, which established inhaled corticosteroids as 1st-line anti-inflammatory therapy.[25] This temporal association suggests that the standardization of care globally contributed to reducing asthma-related mortality and disability, a benefit supported by extensive evidence for inhaled corticosteroids in preventing exacerbations, hospitalizations, and death.[26,27]
Our analysis revealed substantial geographic heterogeneity in the asthma burden attributable to high BMI. India accounted for the greatest total count of deaths and DALYs, likely reflecting its large population size and the rapidly rising prevalence of obesity.[28] In contrast, Pacific Island nations, notably Fiji and Kiribati, bore the heaviest burden in terms of ASRs, despite their smaller populations. This disparity may be influenced by genetic susceptibility,[29] dietary transitions towards Western patterns, and declining physical activity associated with urbanization.[30] Meanwhile, Zimbabwe experienced the most rapid increase in disease burden, a trend characteristic of the “dual burden of disease” often seen during socioeconomic transition.
A significant inverse association was found between the SDI and ASRs of high BMI-related asthma. Consistent with prior studies,[9] high-SDI regions observed a consistent decline in both ASMR and ASDR. Conversely, the low-middle SDI quintile was the only group to experience an increase in ASRs from 1990 to 2021.
Decomposition analysis revealed that in these regions, population growth and aging, compounded by unfavorable epidemiological changes, jointly drove the rising burden. This disparity likely reflects challenges unique to low-middle SDI settings, such as rapidly escalating obesity rates,[31] deteriorating air quality,[32] and constrained access to essential medicines, including a critical shortage of inhaled corticosteroids for asthma management.[33] Addressing these inequities requires multisectoral strategies aimed at reducing dietary salt and sugar consumption, enforcing stricter air quality standards, and strengthening primary healthcare systems to enhance the coverage and quality of asthma diagnosis and treatment.
The asthma burden associated with high BMI also demonstrated distinct variations by age and sex, with the most pronounced impact observed among older adult females – a finding consistent with previous reports.[9] This disparity may be explained by the interplay of hormonal fluctuations, body fat distribution, and inflammatory pathways. For example, the postmenopausal decline in estrogen promotes central adiposity,[34] which can elevate levels of pro-inflammatory cytokines and leptin[35] and impose mechanical loads on lung function.[36] Furthermore, comorbidities such as cardiovascular disease and depression are highly prevalent in older women with obesity[37,38] and can exacerbate asthma-related disability. These findings underscore the value of a life-course perspective.
Implementing comprehensive strategies that integrate weight management, physical rehabilitation, and psychosocial support is essential to alleviate the disease burden in this vulnerable demographic.
The COVID-19 pandemic provided a unique context in which to assess the impact of high BMI on asthma burden. Although the GBD 2021 data encompass the initial years of the pandemic (2019–2021), our trend analysis did not reveal a significant deviation in the high BMI-related asthma burden trajectory during this period. This observation might be explained by countervailing mechanisms.
On 1 hand, lockdowns and restrictions likely reduced physical activity and promoted unhealthy dietary shifts, contributing to rising obesity rates globally,[39] a trend that could potentially increase asthma incidence. On the other hand, public health measures such as vaccination, mask-wearing, and remote work may have reduced exposure to respiratory viruses and other environmental triggers, thereby lowering the rate of acute asthma exacerbations.
Our projections suggest divergent future pathways: a continuing decline in ASMR but a slight rise in ASDR. This pattern of decreasing mortality alongside increasing disability likely reflects improvements in acute care and survival, yet underscores persistent chronic morbidity and health loss among people with high BMI.[40] Consequently, global asthma management strategies should evolve beyond a primary focus on survival to also prioritize the improvement of long-term functional status and quality of life through sustained behavioral and clinical support.
Several limitations of this study warrant consideration. First, as our analysis relies on the GBD database, potential under-ascertainment of cases in low-income regions may lead to an underestimation of the true disease burden in these settings. Second, the MR analysis is subject to substantial heterogeneity across genetic variants (Cochran’s Q = 669.07, P = 7.4 × 10−21) and to pleiotropic outliers identified by MR-PRESSO, both of which violate the exclusion restriction assumption.
Although multiple sensitivity analyses were performed, the IVW estimate may be biased by horizontal pleiotropy, and the InSIDE assumption underlying MR-Egger regression cannot be tested. The MR findings should therefore be interpreted as genetically supported associations rather than causal effects. Third, all GWAS summary statistics used in the MR analysis were derived from populations of European ancestry, whereas the settings that bear the highest absolute burden of high-BMI-attributable asthma, such as India and China, are predominantly non-European.
Differences in LD structure, allele frequencies, and gene–environment interactions may therefore limit the transportability of our estimates. The genetically supported association reported here should be extrapolated to non-European populations with caution, and replication using ancestry-diverse GWAS is warranted. Fourth, whereas the forward MR analysis employed strong genetic instruments for BMI, the reverse MR analysis was constrained by a relatively limited number of independent SNPs for asthma.
Although these instruments were robust (high F statistics), the smaller set may affect the power and robustness of sensitivity analyses. Fifth, despite the use of MR to minimize confounding, residual bias from unmeasured environmental factors (e.g., air pollution and smoking[41,42]) cannot be fully ruled out. Sixth, the BAPC projections extrapolate historical mortality and DALY trends, do not incorporate BMI as a time-varying covariate, and should therefore be interpreted as trend-based scenarios conditional on the continuation of past patterns.
In regions undergoing rapid nutritional and obesity transitions – particularly low- and middle-SDI settings where obesity prevalence is rising fastest – the projections may underestimate future asthma burden attributable to high BMI. Finally, BMI is an imperfect proxy for adiposity, as it does not differentiate between fat and lean mass, potentially obscuring variations in obesity-related risk.
5. Conclusions
By integrating the GBD framework with MR, this study identifies high BMI as a risk factor genetically associated with the global asthma burden. The increasing absolute burden, driven largely by demographic forces, masks improvements in ASRs and exhibits substantial inequality across regions, SDI levels, age groups, and sexes. Moving forward, public health efforts must prioritize population-level obesity prevention, mitigate socioeconomic barriers to healthcare, and deliver tailored interventions that account for demographic and sex-specific disparities.
Acknowledgments
We acknowledge the immense efforts of the Global Burden of Disease Study 2021 collaborators and the broader GBD network in producing and maintaining this vital public resource. We also thank the participants and investigators of the genome-wide association studies from which we obtained summary statistics.
Author contributions
Conceptualization: Xiaokun Yao, Chuanli Zhang.
Data curation: Xiaokun Yao.
Formal analysis: Xiaokun Yao.
Investigation: Chuanli Zhang.
Methodology: Xiaokun Yao, Chuanli Zhang.
Project administration: Chuanli Zhang.
Resources: Chuanli Zhang.
Supervision: Chuanli Zhang.
Visualization: Xiaokun Yao.
Writing – original draft: Xiaokun Yao.
Writing – review & editing: Chuanli Zhang.