Air pollution is a global health threat increasingly linked to mental disorders. Exposure to particulate matter (PM10), nitrogen dioxide (NO2), and ozone (O3) has been linked to higher risk of these conditions. Evidence is strongest for NO2, while findings for PM10 and O3 remain inconsistent, particularly in Southern Europe.
MethodsWe conducted a cross-sectional study geolinking 2019 mental health diagnoses from the PADRIS-PRESTO cohort (n=1,415,944) with air pollution data from 30 health districts in Catalonia, Spain. Annual mean concentrations of NO2, O3, and PM10 were estimated. Logistic regression models, adjusted for age, sex, socioeconomic status, and health district, examined associations between pollutant exposure tertiles and the prevalence of major depressive disorder (MDD), anxiety disorder (AD), bipolar disorder, schizophrenia (SCZ), attention deficit hyperactivity disorder (ADHD), and autism spectrum disorder (ASD).
ResultsNO2 exposure showed the strongest associations, including a moderate effect for ADHD (OR=2.04, 95%CI: 1.94–2.15) and small positive associations with ASD, SCZ, and AD (OR range: 1.27–1.31). O3 demonstrated a small positive association with ADHD (OR=1.41, 95%CI: 1.36–1.46) but showed no substantive associations with other disorders. PM10 showed mostly null or inverse associations, including a small inverse association with ADHD (OR=0.76, 95%CI: 0.73–0.79).
ConclusionsThis large population-based study found consistent associations between NO2 exposure and neurodevelopmental and psychotic disorders, whereas O3 and PM10 showed weaker or inconsistent relationships. These findings highlight the need for longitudinal research to clarify causality.
Air pollution represents one of the major environmental threats, contributing to over eight million deaths annually.1 Beyond respiratory and cardiovascular disease, air pollution also affects the central nervous system through mechanisms such as neuroinflammation, oxidative stress and neuronal damage, potentially disrupting neurotransmission and cognitive–emotional regulation.1–3
Epidemiological studies link air pollution to anxiety, depression, and neurodevelopmental and psychotic disorders.4–7 The most consistent associations involve NO2 and PM2.5, whereas evidence for PM10 and O3 is weaker. Long-term exposure to PM2.5 and NO2 has been associated with depression, while short-term pollution peaks have been linked to symptom exacerbation and suicidal behaviour. Evidence also connects these pollutants with adverse neurodevelopment, particularly autism, whereas findings for ADHD remain mixed.8,9
Most studies have been conducted in North America, Asia, or Northern Europe, leaving Southern Europe underrepresented. Catalonia (north-eastern Spain) provides an appropriate context to examine these associations due to its comprehensive health databases and systematic air-quality monitoring. The PADRIS-PRESTO cohort, a population-based dataset covering around 1.4 million residents, enables large-scale analyses of mental health outcomes.10
This study examined cross-sectional associations between major air pollutants (PM10, NO2, O3) and mental health disorders in the PADRIS-PRESTO cohort in Catalonia. We hypothesized that higher exposure would be associated with greater disorder prevalence. To our knowledge, this is the largest population-based study of air pollution and mental health in Southern Europe.
Material and methodsStudy designWe conducted a cross-sectional study geolinking 2019 mental health diagnoses in the PADRIS-PRESTO cohort with publicly available air pollution data from Catalonia. The analysis was restricted to 2019 because PADRIS-PRESTO currently provides validated data only up to that year and to avoid COVID-19-related diagnostic and mobility distortions. The cohort includes sociodemographic and clinical information for 1,421,510 residents (473,812 patients receiving specialized mental health care and 947,698 controls); details published elsewhere.10 Air pollution data were obtained from the Xarxa de Vigilància i Previsió de la Contaminació Atmosfèrica (XVPCA), providing monthly concentrations for 30 health districts.11 Approved by the Hospital Clínic de Barcelona Ethics Committee (HCB/2020/0735).
VariablesMental health disorders commonly investigated in air-pollution research were included.1 Diagnoses were identified using ICD-10 codes, retaining the latest diagnosis recorded in 2019. Disorders of interest included major depressive disorder (MDD; F32–F33), anxiety (AD; F40–F41), bipolar disorder (BD; F31), schizophrenia (SCZ; F20), attention deficit hyperactivity disorder (ADHD; F90.0), and autism spectrum disorders (ASD; F84). As PADRIS–PRESTO provides ICD-10 codes only at the block level, diagnoses could not be further disaggregated. Covariates included age, sex, socioeconomic status, and health district.
Pollutants (NO2, O3, PM10) were selected based on literature relevance and the availability of systematic monitoring in Catalonia. Monthly concentrations for the 30 health districts were obtained from the XVPCA monitoring system and averaged to calculate annual means. Missing values were imputed using spatial interpolation (>99% recovery; r>0.95). Exposure levels were categorized into tertiles.
Statistical analysisPearson correlations examined associations between disorder prevalence and pollutant concentrations (NO2, O3, PM10). Multivariable logistic regression models were fitted separately for each disorder to evaluate associations between pollutant tertiles (highest vs lowest), adjusting for age, sex, socioeconomic status, and health district. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs) and interpreted using established thresholds: ORs<1.2 negligible, 1.2–1.5 small, 1.6–3.0 moderate, and >3.0 large (or reciprocal values when <1).12 Analyses were conducted in R4.4.0.
ResultsStudy populationThe final analytical dataset comprised 1,415,944 individuals after exclusion of those with incomplete exposure records. Mean age was 41.6 years (SD=23.4) and 53.7% were female (Table 1). Mean 2019 concentrations were 29.0μg/m3 for NO2 (11.6–49.5), 47.8μg/m3 for O3 (36.7–81.6), and 21.8μg/m3 for PM10 (15.2–27.3).
Demographic and clinical characteristics of the sample.
| Characteristic | Participants (N=1,415,944) |
|---|---|
| Age, mean (SD) | 41.6 (23.4) |
| Female, % | 53.7 |
| Socioeconomic status, % | |
| Exempt status | 15.9 |
| Low income (<€18.000) | 50.9 |
| Middle income (€18.000 to €100.000) | 32.3 |
| High income (>€100.000) | 0.9 |
| Mental health diagnosis, n (%) | |
| Major Depressive Disorder | 99,075 (6.99) |
| Anxiety Disorders | 79,733 (5.63) |
| Bipolar Disorder | 15,617 (1.10) |
| Schizophrenia | 22,542 (1.59) |
| Attention Deficit Hyperactivity Disorder | 37,029 (2.61) |
| Autism Spectrum Disorders | 7,710 (0.54) |
At the ecological level, NO2 was positively correlated with SCZ (r=0.34, p<0.001) and ASD (r=0.25, p=0.002), while negatively correlated with ADHD (r=−0.12, p=0.045). O3 was inversely correlated with ASD (r=−0.45, p<0.001) and ADHD (r=−0.12, p=0.044). PM10 showed positive correlations with ASD (r=0.36, p<0.001) and SCZ (r=0.22, p=0.004).
Multivariable logistic regressionORs for each pollutant are shown in Fig. 1 and reflect comparisons between the highest and lowest exposure tertiles.
Associations between air pollutant exposure and mental health disorders. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for exposure to NO2, O3, and PM10 in relation to major mental health disorders in Catalonia, 2019. Estimates are adjusted for age, sex, socioeconomic status, and health area; the dashed line indicates the null value (OR=1). Abbreviations: AD, anxiety disorders; ADHD, attention deficit hyperactivity disorder; ASD, autism spectrum disorders; BD, bipolar disorder; MDD, major depressive disorder; SCZ, schizophrenia; PM10, particulate matter ≤10μm; NO2, nitrogen dioxide; O3, ozone.
NO2 exposure was positively associated with several disorders. The strongest association was for ADHD (OR=2.04, 95%CI=1.94–2.15), followed by smaller associations for ASD (OR=1.31, 95%CI=1.18–1.45), SCZ (OR=1.29, 95% CI=1.21–1.38), and AD (OR=1.27, 95%CI=1.22–1.32). MDD and BD showed negligible inverse associations.
For O3, ADHD showed a small positive association (OR=1.41, 95%CI=1.36–1.46), whereas SCZ (OR=1.13, 95%CI=1.08–1.19) and AD (OR=1.08, 95%CI=1.04–1.12) showed no meaningful associations. MDD and ASD showed negligible inverse associations, whereas BD showed no significant association.
For PM10, ADHD showed a small inverse association (OR=0.76, 95%CI=0.73–0.79). AD and MDD showed negligible inverse associations, whereas BD showed no significant association.
DiscussionThis study examined cross-sectional associations between three air pollutants and mental health disorders in a population-based cohort in Catalonia, Spain. Three patterns appeared: (1) NO2 showed moderate associations with ADHD and small associations with ASD, SCZ, and AD; (2) O3 showed a small positive association with ADHD; and (3) PM10 showed a small inverse association with ADHD. Overall, NO2 emerged as the pollutant most consistently linked to mental health outcomes, whereas associations with O3 and PM10 were weaker and heterogeneous across disorders.
Differences in effect sizes across disorders may reflect heterogeneous environmental influences. Results for NO2 align with prior evidence linking traffic-related pollutants to neurodevelopmental, anxiety and psychotic disorders.9 These associations are biologically plausible, as experimental studies show that NO2 induces neuroinflammation, oxidative stress, and blood–brain barrier disruption.3 Such mechanisms are particularly relevant during brain development and may explain the stronger associations observed for ADHD and ASD. Psychotic disorders may also show short-term symptom exacerbations following pollution peaks, further supporting their susceptibility to traffic-related pollutants.8,9
In contrast, we did not observe a positive association with MDD despite prior evidence suggesting such a link. Several factors may explain this discrepancy. First, most studies linking NO2 to depression rely on long-term exposures, whereas our one-year cross-sectional window may not capture the chronic processes underlying depression onset. Second, depression is strongly influenced by psychosocial factors not captured in administrative records. Third, NO2 often correlates with PM2.5—the pollutant most strongly linked to depression—and its absence in our models may limit the ability of NO2 to capture traffic-related exposure. Finally, depression diagnoses are clinically heterogeneous and less precisely captured in routine data.
Evidence for O3 and PM10 was weaker. Prior reviews report little or no association between O3 and mental health, consistent with our findings.9,13 Only a small positive association with ADHD emerged in regression models, whereas ecological correlations were negative for ASD and ADHD. This likely reflects the spatial distribution of ozone, which reaches higher concentrations in rural or coastal areas with lower population density and fewer diagnosed psychiatric conditions.
For PM10, a small inverse association was observed for ADHD. However, this finding is unlikely to reflect a true protective effect. Toxicological evidence indicates that coarse particles deposit mainly in the upper airways, with lower systemic penetration and consequently limited neurodevelopmental effects.14 The inverse association is more likely an artefact of residual confounding or exposure misclassification, further amplified by the limited spatial variability of PM10 following extensive imputation.
Several limitations should be considered. First, the cross-sectional design and single-year exposure limit causal inference and cumulative exposure effects.15 Second, area-level exposure assignment and imputation may introduce exposure misclassification. Third, single-pollutant models were used; GVIF values (1.51–1.88; threshold<2.24) indicated acceptable collinearity, although residual confounding by co-exposures remains possible. Fourth, adjustment for health districts may not fully capture urbanicity, and unmeasured urban factors (e.g., noise, light pollution, or access to mental health services) may have confounded associations. Finally, PADRIS–PRESTO provides ICD-10 codes only at the block level, preventing analysis of diagnostic subtypes.
Nonetheless, the large population coverage and robust exposure estimates strengthen the validity of these findings. Future longitudinal studies should examine cumulative exposures and more detailed diagnostic coding. Spatial analyses combining pollutant levels and disorder prevalence could help identify high-risk areas for public health interventions.
ConclusionsIn this population-based study in Catalonia, NO2 exposure was associated with neurodevelopmental and psychotic disorders, whereas O3 and PM10 showed weaker or inconsistent associations. These findings highlight NO2 as a key pollutant for mental health and support the need for longitudinal studies.
Ethical considerationsNone.
FundingSupported by the Fundació Clínic per a la Recerca Biomèdica through the Pons Bartran 2020 grant (PI046549); the Spanish Foundation for Psychiatry and Mental Health, the Spanish Psychiatric Society, and the Spanish Society of Biological Psychiatry (PI046813); the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR) – PANDÈMIES 2020 grant (2020PANDE00081), Generalitat de Catalunya; the Ministerio de Sanidad through the Pla Director de Salut Mental i Addiccions and the Direcció General de Planificació i Recerca en Salut, Departament de Salut, Generalitat de Catalunya; and by the Ministerio de Ciencia, Innovación y Universidades (MCIN) and the Agencia Estatal de Investigación (AEI, project TED2021-131999BI00, Strategic Projects Oriented to the Ecological Transition and the Digital Transition 2021), with funding from the European Union NextGenerationEU/PRTR.
Conflict of interestVLB has received travel and congress registration support from Adamed, Angelini, Casen-Recordati, and Lundbeck, with no financial or other relationship relevant to the subject of this article.
DHM has received CME-related honoraria and served as a consultant for Abbott, Angelini, Neuraxpharm, Ethypharm Digital Therapy, Lundbeck and Viatris.
DI has received continuing medical education support from Rubió, Esteve and Takeda.
MB has been a consultant for, received grant/research support and honoraria from, and been on the speakers/advisory board of has received honoraria from talks and/or consultancy of Adamed, Angelini, Casen-Recordati, Exeltis, Ferrer, Janssen, Lundbeck, Neuraxpharm, Otsuka, Pfizer, Rovi and Sanofi.
EV has received grants and served as a consultant, advisor, or CME speaker for the following entities: AB-Biotics, AbbVie, Angelini, Biogen, Biohaven, Boehringer-Ingelheim, Celon Pharma, Dainippon Sumitomo Pharma, Ferrer, Gedeon Richter, GH Research, Glaxo-Smith Kline, Idorsia, Janssen, Lundbeck, Novartis, Orion Corporation, Organon, Otsuka, Sage, Sanofi-Aventis, Sunovion, and Takeda, outside the submitted work.
None declared.
CVP is supported by the FPU grant for University Teaching Training (FPU23/01555) from the Spanish Ministry of Science, Innovation, and Universities.
MDP is supported by the Translational Research Programme for Brain Disorders, IDIBAPS.
VLB thanks the support of the Spanish Ministry of Health financed by the Instituto de Salud Carlos III (ISCIII) and co-financed by the Fondo Social Europeo Plus (FSE+) (CM24/00074); the Fundación Española de Psiquiatría y Salud Mental (FEPSM); and the Fundació Vila Saborit.
VO is supported by a Rio Hortega 2024 grant (CM24/00143) from the Spanish Ministry of Science, Innovation and Universities financed by the Instituto de Salud Carlos III (ISCIII) and co-financed by the Fondo Social Europeo Plus (FSE+).
DHM gratefully acknowledges the support of the Spanish Ministry of Health, Instituto de Salud Carlos III (PI049759) and the Pons-Bartran Legacy grant 2024 (FCRB_IPB1_2024).
DI gratefully acknowledges the support of the Pons-Bartran Legacy grant 2025.
MB has received grants from Spanish Ministry of Health, Instituto de Salud Carlos III (PI20/01066), Fundació La Marató de TV3 (202206-30-31) and Pons-Bartran legacy (FCRB_IPB1_2023).
