Outdoor air pollutants and asthma risk in adolescents: evidence from a systematic review and meta-analysis

With an estimated 260 million cases globally in 2021 and a projection to reach 275 million by 2050, asthma poses a substantial and growing public health challenge worldwide (1). This chronic, complex, and heterogeneous respiratory disorder remains the most common chronic condition among adolescents globally (2). A recent meta-analysis conducted in 2025, synthesizing data from 164 population-based studies, estimated the global prevalence of childhood asthma at 10.2% [95% confidence interval (95% CI): 9.5–11.0] (3). Notable geographic disparities in adolescent asthma prevalence were identified, ranging from as low as 0.3% in India to as high as 29.9% in Spain (4). Asthma imposes a considerable and multifaceted burden, as evidenced by frequent emergency department visits, hospitalizations, school absenteeism, activity limitations, and reliance on long-term pharmacological therapy (1–4). These adverse outcomes can interfere with children’s development, lower their academic performance, affect their mental health, and reduce their quality of life in the long term (1–4). To tackle this global health problem, it is important to identify modifiable risk factors, especially common environmental exposures, in order to develop effective, evidence-based prevention strategies (5, 6).

Air pollution, a complex mixture of particulate matter, gaseous pollutants, and biological or chemical agents, has been causally linked to both the development and worsening of asthma, particularly due to variations in exposure across time and location (7). These variations arise from differences in pollutant sources, as outdoor air pollution mainly originates from fossil fuel combustion (e.g., traffic, industrial activities, power generation), industrial emissions, and natural events such as wildfires. Among these, particulate matter with diameter ≤ 2.5 micrometers (PM_2.5), nitrogen dioxide (NO₂), and ozone (O₃) are considered the most significant in terms of health impact (8). Long-term exposure to outdoor air pollution can lead to respiratory irritation and increase the risk of chronic respiratory diseases, including asthma, particularly among vulnerable populations such as children. According to the World Health Organization (WHO) (9), over 99% of the global population lives in areas where air pollution levels exceed WHO air quality guidelines, and an estimated 4.2 million deaths each year are attributed to outdoor air pollution. These statistics highlight outdoor air pollution as a critical global public health crisis.

Adolescence represents a critical and vulnerable period in the human life course. During this stage, the respiratory system continues to develop, with lung function maturing into early adulthood (10), and the immune system undergoes significant changes, including the maturation of T helper type 1 (Th1) and type 2 (Th2) cell balance (11). Compared to adults, adolescents have higher minute ventilation relative to body weight, and their typical behaviors, such as increased outdoor activity and proximity to ground-level air, may result in greater exposure to outdoor air pollutants (12). In addition to causing acute respiratory symptoms and impaired lung function, both early and prolonged exposure to air pollution can significantly affect asthma onset, progression, treatment outcomes, and long-term respiratory health (13).

Although numerous epidemiological studies have investigated the association between outdoor air pollution and asthma, several important knowledge gaps remain, particularly in relation to adolescents. One key limitation is the lack of systematic evaluation of both independent and combined effects of multiple air pollutants. In real-world settings, adolescents are simultaneously exposed to various pollutants, which may exhibit collinearity and interact in synergistic, additive, or antagonistic ways. However, most prior studies have examined single pollutants or isolated environments, limiting their ability to capture the complexity of actual exposure scenarios (14–16). Another major gap is the inconsistency in identifying critical exposure windows. Existing studies vary widely in the timing of exposure assessed (e.g., intrauterine, infancy, preschool), making it unclear which developmental period is most sensitive to air pollution. In addition, substantial heterogeneity in study findings for the same pollutant complicates efforts to draw firm conclusions regarding its effect on asthma outcomes. A further challenge lies in the variability of exposure units and increments used across studies, with measurements reported in microgram per cubic meter (μg/m³), parts per million (ppm), or parts per billion (ppb), hindering comparability. These limitations highlight the need for a rigorous, comprehensive, and adolescent-focused systematic review and meta-analysis, which holds significant scientific and public health relevance.

This study systematically synthesizes global observational evidence, including cohort, case–control, and cross-sectional studies, to evaluate the association between exposure to specific outdoor air pollutants and asthma outcomes among adolescents aged 10–19 years. A meta-analysis was conducted to quantify the effect sizes of outdoor air pollutant exposure on asthma risk, and a dose–response relationship model was developed. Subgroup analyses were performed to explore potential effect modification by study design, exposure window, exposure duration, and asthma subtype, aiming to identify sources of heterogeneity. Risk of bias was assessed using standardized evaluation tools to determine the overall certainty of the evidence. The findings from this review provide important scientific support for revising policy standards such as the WHO global air quality guidelines, advancing precision environmental risk management in clinical settings, and informing the design of intervention trials targeting mixed exposure effects. Ultimately, this research contributes to an evidence-based foundation for reducing the burden of preventable respiratory diseases in adolescents and promoting health equity.

This systematic review was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (17). The primary objectives were: (1) to systematically synthesize epidemiological evidence on the association between outdoor air pollution exposure and asthma, and (2) to conduct a quantitative meta-analysis of exposure–response relationships. The methodology followed the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines (18, 19). The study protocol was prospectively registered in the PROSPERO database (CRD42024622246; 15 December 2023; https://www.crd.york.ac.uk/prospero/↗) and received ethical approval from the Research Ethics Committee, Faculty of Medicine, Chiang Mai University, Thailand (Approval No. Exemption 0391/2025, approved on 29 May 2025).

To our knowledge, this is the systematic review and meta-analysis to comprehensively synthesize epidemiological evidence on the association between outdoor air pollution and asthma in adolescents. Quantitative analyses of seven major combustion-related pollutants revealed significant positive associations between exposure to CO, NO₂, O₃, and TRAP and increased asthma risk. In contrast, PM_2.5, PM₁₀ and SO₂ were not significantly associated with asthma in the pooled estimates.

For PM_2.5, a primary meta-analysis restricted to cohort studies (n = 7) was conducted to strengthen causal inference. The pooled estimate was not statistically significant (aOR = 1.11; 95% CI: 0.89–1.34; I² = 32.3%). Crucially, this null association remained non-significant, and the effect estimate was often attenuated, following adjustment for substantial publication bias. Accordingly, the most methodologically robust evidence within our review does not support a significant association between long-term PM_2.5 exposure and adolescent asthma. Kamarehei et al. (79), using the AirQ⁺ model in Poldokhtar, Iran, provided a mechanistic and methodological explanation for this finding, showing that the health effects of long-term PM_2.5 exposure vary substantially by concentration, emission source, and health outcome. Regional heterogeneity in PM_2.5 composition contributes to divergent toxicological profiles, with rural or natural-source PM_2.5 often differing markedly from urban anthropogenic particles (80–83), potentially reducing its ability to induce allergic Th2 immune responses (84, 85). The characteristics and outcomes of the seven included cohort studies further support this interpretation. Moreover, the attributable burden of PM_2.5 is predominantly observed in adult chronic diseases such as chronic obstructive pulmonary disease, ischemic heart disease, and lung cancer, while adolescent asthma has not shown a statistically significant association (86). As a multifactorial disease, its weaker signal may be masked by the dominant burden of adult morbidity. Additionally, short-term PM_2.5 exposure appears to exert a stronger effect on respiratory morbidity than long-term exposure (86). In summary, this meta-analysis found no significant association between long-term PM_2.5 exposure and adolescent asthma, and this null finding persisted after adjustment for publication bias. Clarifying this relationship will require large, preregistered prospective cohorts with result-independent publication.

Interestingly, in the secondary analysis, PM_2.5 exposure was significantly associated with adolescent asthma (aOR = 1.27; 95% CI: 1.04–1.51) with significant heterogeneity (I² = 77.3%) arose from multiple methodological sources with significant heterogeneity (I² = 77.3%) arose from multiple methodological sources. First, study design contributed to bias, while one cross-sectional study reported strong associations (aOR = 1.59, 95% CI: 1.40–1.77), cohort studies (n = 7) showed null effects (aOR = 1.11, 95% CI: 0.89–1.34), likely due to reverse causation and residual confounding. Adjustments varied across studies. Some studies accounted for indoor pollutants (55), others adjusted only for smoking (74), affecting effect size estimates. Second, inconsistent asthma definitions introduced misclassification: studies using current asthma showed stronger associations (aOR = 1.54), possibly due to symptom-based overreporting, whereas ever asthma definitions may underestimate risk by including remitted cases and relying on incomplete records. Third, exposure misclassification likely resulted from variation in PM_2.5 assessment methods, e.g., gridded models (75) vs. LUR models (67), with differing spatial and temporal resolutions, potentially biasing associations toward the null, as seen in Dockery et al. (36) and To et al. (68). Fourth, only three studies explicitly assessed early-life exposure across multiple time points (58, 68, 75); others lacked clear exposure windows. Fifth, limited eligible studies constrained analysis. Of 12 identified, three were excluded due to unspecified increments (27, 28, 31, 74), leaving eight. The absence of short-term exposure data and unadjusted co-pollutants (e.g., NO₂, O₃) further obscured dose–response patterns. Additional variability likely stems from unmeasured modifiers such as age (11 years in Zanobetti et al. (75) vs. 17 years in To et al. (68)) and regional differences in PM_2.5 composition.

A critical finding from our meta-analysis of PM₁₀ warrants cautious interpretation. Although the pooled estimate was not statistically significant, the high heterogeneity (I² = 76.5%) indicates substantial inconsistency across studies rather than uniform evidence of no effect. Thus, it would be premature to conclude the absence of an association between PM₁₀ and asthma; instead, the evidence remains inconclusive and context-dependent. The non-significant pooled estimate for long-term PM₁₀ exposure and adolescent asthma likely reflects multiple, overlapping factors. First, PM₁₀ composition varies widely across regions (e.g., crustal dust vs. anthropogenic emissions), resulting in differential toxicity (87). Second, adolescents may exhibit distinct exposure patterns or physiological resilience compared with adults or younger children (88). Third, long-term exposure tends to produce weaker health effects than short-term exposure (89). Finally, methodological differences, such as exposure assessment precision and adjustment for confounders like indoor allergens or socioeconomic status, may have attenuated the observed associations (90, 91).

Of the 13 eligible studies, five did not report exposure increments and were excluded to ensure comparability and data integrity (37, 40, 45, 66, 74). The remaining eight studies were included in the analysis. Notably, substantial heterogeneity was observed, particularly among cross-sectional designs and studies examining ever asthma or long-term exposure. This variability may reflect methodological inconsistencies and differences in population characteristics. Substantial heterogeneity in the PM₁₀–asthma association largely reflects differences in study design. Cohort studies (I² = 0.0%) reported no significant association, contrasts sharply with the significant positive association in the cross-sectional study (I² = 91.0% despite very high heterogeneity), likely influenced by reverse causation and recall bias. Although cohort designs better infer temporality, their small sample sizes (n = 4) and exposure misclassification from residential mobility may obscure associations. Outcome definitions also contributed: current asthma analyses showed more consistent associations (I² = 22.8%) compared to ever asthma (I² = 84.7%), reflecting diagnostic variability and recall differences. Exposure assessment methods varied widely—e.g., personal monitors (60) vs. fixed-site data (67)—along with differences in temporal windows and co-pollutant adjustment. Gruzieva et al. (53) found early-life exposure associated with higher asthma risk (aOR = 1.96), unlike later exposure (aOR = 1.02), highlighting critical developmental windows. Gehring et al. (58) reported location-based differences (birth vs. current address), illustrating exposure misclassification from migration. Population-level modifiers also shaped associations: Sahsuvaroglu et al. (49) found null effects in girls without hay fever, whereas Kuiper et al. (70) observed elevated risk in other subgroups (aOR = 1.90), pointing to gene–environment interactions and host susceptibility. Thus, the non-significant pooled estimate for PM₁₀ should be viewed not as evidence of absence, but as evidence of inconsistency across studies. Future research employing rigorous longitudinal designs with precise exposure assessment is needed to resolve these discrepancies and clarify the true association.

Meta-analysis revealed that exposure to CO was significantly associated with adolescent asthma (aOR = 1.31, 95% CI: 1.08–1.53). CO contributes to the onset and exacerbation of adolescent asthma through three interrelated mechanisms. CO’s high binding affinity to hemoglobin induces systemic hypoxia (92), which promotes bronchial smooth muscle contraction and activates hypoxia-inducible factors (93), intensifying airway inflammation characterized by increased mucus secretion, mucosal edema, and immune cell infiltration (94). CO also disrupts redox homeostasis by elevating reactive oxygen species (ROS) and impairing antioxidant defenses (95, 96), thereby initiating oxidative stress and damaging airway epithelial integrity. This, in turn, activates proinflammatory pathways such as nuclear factor-kappa B (NF-κB), amplifying asthma-related inflammation (97). Lastly, the combined effect of hypoxia, oxidative stress, and inflammation lowers airway reactivity thresholds, enhancing susceptibility to environmental triggers and worsening symptom severity in adolescents (98). Substantial heterogeneity in the CO meta-analysis (I² = 68.9%) arose from both methodological and biological variability. Study design played a major role. Cross-sectional studies (n = 4; I² = 76.6%) reported higher effect estimates (aOR = 1.32), likely due to temporal ambiguity and incomplete confounder control, such as partial adjustment for smoking and unaccounted co-pollutants (37, 74). In contrast, the null finding in underpowered single cohort study (aOR = 1.10, 95% CI: 0.02–2.17) likely reflects insufficient power. Exposure assessment varied widely: annual means, 8-h means, geographic information system-based estimates, and hourly maxima with inconsistent increments (37, 39, 41, 45, 74). Despite unit harmonization efforts, some metrics remained non-convertible, though excluding Delfino et al. (41) had minimal impact on pooled estimates (aOR = 1.32 vs. 1.31). Long-term exposures were vulnerable to spatial misclassification and seasonal variation. Short-term exposure showed stronger associations (aOR = 1.80, I² = 0.0%) than long-term (aOR = 1.17, I² = 52.8%), likely due to acute hypoxic effects and latency bias. Clinical heterogeneity persisted due to inconsistent asthma definitions (self-report vs. diagnosis), limited stratification by modifiers (e.g., gender in Ho et al. (45)), and lack of endotype-specific analyses. Data on ever asthma was limited to a single recall-based study.

The meta-analysis identified NO₂ as a significant risk factor for adolescent asthma (aOR = 1.18; 95% CI: 1.08–1.29). The persistence of a statistically significant association after this adjustment reinforces the robustness of the finding and reduces the likelihood that it reflects a publication bias effect. NO₂ exacerbates asthma through multiple synergistic pathways. Upon inhalation, it dissolves in airway mucosa, generating reactive nitrogen species and ROS (99), depleting antioxidants like glutathione, inducing oxidative stress, disrupting epithelial barriers, and enhancing allergen penetration (100). It also activates NF-κB and upregulates interleukin-33 (IL-33) and thymic stromal lymphopoietin (TSLP) (101), promoting dendritic cell-driven Th2 polarization (increased interleukin-4 (IL-4), interleukin-5 (IL-5), interleukin-13 (IL-13)) (102), leading to eosinophilic inflammation, immunoglobulin E (IgE) production, and mucus hypersecretion (103). Concurrently, oxidative stress upregulates transient receptor potential vanilloid 1 (TRPV1) and transient receptor potential ankyrin 1 (TRPA1) channels (104), triggering neurogenic inflammation and airway hyperresponsiveness (105). Long-term exposure alters immune gene expressions (e.g., forkhead box protein P3, IL-4) via DNA methylation, impairing regulatory T cells (Treg) function and contributing to airway remodeling and asthma persistence (106). However, the findings should be interpreted with caution due to substantial heterogeneity (I² = 67.9%). Substantial heterogeneity in the NO₂ meta-analysis stems from multiple methodological sources. First, exposure measurement inconsistencies: of 23 eligible studies, 8 were excluded for lacking standardized effect estimates or using incompatible units (e.g., μg/m³, ppm, ppb). The remaining 15 studies were standardized to 10 μg/m³, but conversion assumptions (1 atm, 25 °C) may not reflect regional conditions, introducing systematic error. Second, population variability contributed, as studies spanned ages 10–18 with wide age ranges and differing sex distributions (e.g., Kuiper et al. (70); Liu et al. (67)). Third, study design influenced heterogeneity: while both cohort and cross-sectional studies showed positive associations, the latter may inflate estimates due to temporal ambiguity, while the former may retain residual confounding. Fourth, sample size varied widely (n = 19–22,574), with smaller studies often reporting larger, less precise effects (e.g., Delfino et al. (41)). Although trim-and-fill analysis showed minimal publication bias impact, imbalance in study size remained influential. Fifth, exposure assessment methods differed: studies used LUR models, satellite-integrated predictions, and ground monitoring with varying spatial resolutions, leading to classification inconsistencies. Sixth, asthma definitions varied, with higher heterogeneity in ever asthma likely due to recall bias. Seventh, exposure duration influenced findings—long-term exposure was generally positively associated with asthma (e.g., Yang et al. (61)), whereas short-term exposure findings were mixed (e.g., Zhao et al. (46)). Lastly, confounder adjustment strategies differed, ranging from socioeconomic status (62) to indoor pollution control (68), further contributing to inter-study variability.

O₃ exposure is associated with increased adolescent asthma risk (aOR = 1.01, 95% CI: 1.00–1.03; I² = 43.9%). Its pathogenesis centers on potent oxidative activity that compromises airway epithelial integrity. O₃ reacts with alveolar surface lipids to generate ROS (107), depletes antioxidants like glutathione, and disrupts tight junction proteins (e.g., e-cadherin, zonula occluden-1), impairing epithelial barrier function (108, 109). This epithelial injury triggers the release of alarmins IL-33 and TSLP (110), activates innate lymphoid cell types 2 (ILC2) and T cells, and promotes IL-5/IL-13-driven eosinophilic and interleukin-17A-mediated neutrophilic inflammation—resulting in mixed airway inflammation (111, 112). O₃ also disrupts adaptive immunity and reduces treatment efficacy by enhancing Th2/ T helper type 17 cell (Th17) polarization, impairing Treg function, and increasing glucocorticoid receptor β expression while inactivating surfactant protein D through oxidative stress (113–116). These mechanisms may explain observed associations between long-term O₃ exposure and impaired small airway function (117). However, the primary source of heterogeneity in the O₃ meta-analysis was study design. Cohort studies (I² = 34.7%, aOR = 1.01) showed a weak positive association, while cross-sectional studies (I² = 0.0%, aOR = 0.46) indicated an inverse relationship, likely reflecting methodological differences. Exposure duration also contributed: long-term analyses (I² = 38.0%, aOR = 1.02) supported a chronic risk, possibly linked to oxidative stress, whereas short-term studies (I² = 0.0%, aOR = 0.72) may reflect seasonal confounding or insufficient acute exposure data. Outcome definitions influenced estimates, with current asthma (I² = 48.4%, aOR = 0.85) showing weaker associations than ever asthma (I² = 19.1%, aOR = 1.01), likely due to recall bias and case misclassification. Most studies relied on self-reported diagnoses, limiting validity. Exposure assessment varied—ranging from ultraviolet photometry and passive diffusion to LUR models (41, 43, 55), introducing spatial and temporal misclassification, particularly relevant for O₃ with smaller effect sizes. Only 8 of 13 studies were included in meta-analysis due to inconsistent or unconvertible exposure units. Despite efforts to harmonize increments (e.g., ppb to μg/m³ using standard conditions), regional variability likely introduced additional error. Overall, the observed heterogeneity stems largely from methodological variation rather than true biological differences.

Of the 13 studies on SO₂ and adolescent asthma, only 4 were included in the meta-analysis due to unclear exposure increments in the remaining 9 (e.g., Chiang et al. (15); Faraji et al. (74); Radhakrishnan et al. (27)). The pooled estimate for SO₂ was not statistically significant, but this finding is characterized by high heterogeneity, sensitivity analysis confirmed the robustness of the SO₂ association after adjusting for publication bias. Substantial heterogeneity was observed (I² = 53.6%), with greater variability in cohort (I² = 73.4%), current asthma (I² = 67.0%), and short-term exposure (I² = 67.0%) subgroups. Contributing factors include: (1) small subgroup sizes (n = 2), amplifying random error; (2) inconsistencies in unit conversion, most studies used ppb, converted using a factor of 2.619 under standard conditions, though regional climate and pollution differences (e.g., US, Canada, China) likely introduced error; (3) contradictory effect directions (e.g., Dockery et al. (36): protective vs. Delfino et al. (41): harmful); (4) methodological heterogeneity, cohort studies (aOR = 1.27) used dynamic monitoring, while cross-sectional studies (aOR = 1.00) relied on retrospective data prone to recall bias; (5) variation in exposure duration, Delfino et al. (41) found 1-h peaks associated with increased asthma risk (aOR = 2.36), while longer exposures showed attenuated effects (aOR = 1.91 for 8-h max); (6) effect modification, Sahsuvaroglu et al. (49) reported stronger associations in girls without hay fever, while Zhao et al. (46) found null results in unstratified analyses. These findings suggest that observed heterogeneity reflects an interplay between exposure timing and host immune status. Acute SO₂ effects in sensitized individuals may be genuine, while inverse associations may reflect survivor bias (e.g., severe asthmatics relocating from polluted areas).

The meta-analysis indicated that TRAP significantly increases adolescent asthma risk (aOR = 1.15, 95% CI: 1.10–1.21). Mechanistically, inhaled TRAP components—such as diesel particles, NO₂, and O₃, disrupt the respiratory epithelial barrier by degrading tight junction proteins (e.g., occludin, claudin-5), enhancing allergen penetration (118, 119). Concurrently, ROS depleted antioxidant defenses, a vulnerability heightened in adolescents due to an underdeveloped nuclear factor erythroid 2-related factor 2 pathway (120, 121). Epithelial injury triggers release of alarmins (TSLP, IL-33), activating ILC2 and initiating type 2 innate inflammation (122, 123). Polycyclic aromatic hydrocarbons (PAHs) further impair Treg function via the aryl hydrocarbon receptor, promoting Th2/Th17 polarization (124). Neuro-immune crosstalk amplifies damage through TRAP-induced activation of TRPV1/TRPA1 channels, releasing neuropeptides and triggering neurogenic inflammation that interacts with ILC2 to drive bronchial hyperresponsiveness and airway remodeling (125, 126).

This systematic review and meta-analysis synthesize robust epidemiological evidence linking outdoor air pollution to increased adolescent asthma risk, though key limitations remain. Most studies focused on NO₂ and TRAP, with limited data on PM_2.5, PM₁₀, CO, O₃, and SO₂, warranting cautious interpretation for these pollutants. A key limitation of this meta-analysis is the widespread reliance on fixed-site ambient monitoring data for individual exposure assessment. This method does not capture individual mobility patterns or time spent indoors, leading to non-differential exposure misclassification. Because such errors are typically unrelated to health outcomes, it tends to bias effect estimates toward the null (127). Therefore, the pooled risk estimates in this analysis likely underestimate the true association between ambient air pollution and adolescent asthma. If exposure were measured with greater individual precision, the true effect sizes would likely be stronger. This limitation has important policy implications, suggesting that the actual public health burden of air pollution may exceed current estimates and that reducing exposure could yield even greater respiratory health benefits for adolescents. Asthma definitions also varied (self-report vs. clinical diagnosis), contributing to endpoint heterogeneity. Despite a broad search strategy, non-English or unpublished studies may have been missed, introducing potential language and publication bias (128, 129). However, funnel plots and Egger tests indicated minimal bias for most pollutants, and trim-and-fill analysis confirmed the robustness of PM_2.5, NO₂ and SO₂ effect estimates. Many studies lacked adjustment for key confounders (e.g., allergy history, socioeconomic status) and varied in exposure metrics and time windows, affecting estimate validity. Predominantly cross-sectional designs further limit causal inference (130). High heterogeneity was evident in several subgroups (e.g., PM_2.5, PM₁₀, CO, NO₂), likely driven by differences in study design, population levels, and analytical methods. Additional gaps include limited analysis of developmental stage–specific effects, underrepresentation of low-income or genetically vulnerable populations, and lack of data on secondary pollutants (e.g., O₃-terpene–derived formaldehyde). Moreover, subgroup analyses by finer age, region, or socioeconomic factors were not feasible due to data limitations.

To address existing methodological limitations and refine causal inference, future research should adopt synergistic strategies. Emphasis should be placed on life-course exposure modeling within well-designed cohort studies to capture temporal variations and developmental susceptibility. Standardization of asthma diagnosis is essential and should involve a composite reference incorporating spirometry, validated biomarkers (e.g., fractional exhaled nitric oxide), and harmonized questionnaires, particularly adapted for use in resource-constrained settings. Moreover, accurate differentiation between long- and short-term exposure windows requires the application of exposure assessment models with high spatiotemporal resolution to minimize misclassification bias. Finally, future analyses should incorporate stratification by co-pollutant exposures, individual susceptibility factors, and epigenetic profiles, to better elucidate effect modification and improve the precision of risk estimates.

The interpretation of our findings should consider the methodological quality of the included studies. Common risks of bias in study design, exposure assessment, and confounder control may limit generalizability, though the overall low risk in analytical domains supports the internal validity of the pooled estimates. Our results, highlighting a significant association between outdoor air pollutants and adolescent asthma, must be viewed within global and national public health contexts. The WHO’s 2020 physical activity guidelines recommend that adolescents engage in at least 60 min of moderate-to-vigorous activity daily (131), a key strategy in combating non-communicable diseases (NCD), also prioritized in Iran’s national NCD prevention plan. However, in areas with poor air quality, encouraging outdoor activity may inadvertently increase exposure to asthma-inducing pollutants, posing a dilemma for adolescents, especially those with pre-existing asthma. Achieving global health targets thus requires integrated approaches that address both physical inactivity and air pollution. Effective interventions may include air quality alert systems to guide outdoor activity scheduling, development of urban “clean air zones” and green spaces, and provision of indoor exercise facilities during high-pollution episodes. Aligning environmental health and NCD prevention policies can foster synergistic strategies that protect respiratory health while promoting overall adolescent wellbeing.

This systematic review and meta-analysis provide evidence that exposure to specific outdoor air pollutants, including CO, NO₂, O₃, and TRAP, is significantly associated with increased risk of adolescent asthma. Among these, NO₂ and CO showed the strongest associations, underscoring their critical role in adolescent respiratory pathogenesis. However, the interpretation of pooled estimates must be approached with caution due to notable heterogeneity across studies, arising from methodological disparities in exposure assessment, asthma definition, study design, and population characteristics. While publication bias and residual confounding may influence effect estimates, sensitivity and trim-and-fill analyses affirmed the robustness of the core findings.

These results underscore the urgent need for refined epidemiological approaches and international collaboration to advance exposure science, harmonize diagnostic criteria, and identify vulnerable subpopulations. Without such improvements, current risk estimates likely underestimate the true pulmonary burden of ambient air pollution on adolescents. Public health policies aimed at reducing air pollution exposure remain critical to mitigating the rising burden of asthma and promoting respiratory health across the course of life.