Introduction

Depression is a prevalent mental health issue that is characterized by persistent feelings of sadness, reduced interest in activities, fatigue, and low energy levels1. According to research, approximately 264 million individuals globally experience depression2. The health consequences of depression encompass diminished quality of life, impaired social functioning3, and heightened susceptibility to physical ailments, disability, and suicide4. Depression is highly prevalent among middle-aged and elderly adults, making them one of the most vulnerable age groups for this mental health condition5. With the acceleration of the aging process in China, the health challenges caused by depression are becoming increasingly serious. The risk factors of depression include genetic susceptibility, social demographic factors, lifestyle behaviors, and environmental factors6. The prevalence of depression is on the rise, however, obtaining effective treatment is often challenging7. Therefore, identifying risk factors for depression to develop preventive interventions holds significant importance in public health.

Experimental studies have demonstrated that the inhalation of air pollutants could lead to depression through the potential biological mechanism of increased inflammation and oxidative stress in the nervous system7. Similarly, there is a growing body of epidemiological evidence indicating that both short-term and long-term exposure to ambient air pollutants could potentially affect neurobehavioral and psychological outcomes. Specifically, there is an increasing risk of developing depression associated with such exposure5. However, the results are contradictory and inconsistent. According to a cross-sectional study conducted in Spain, it was found that higher levels of outdoor air pollutants are associated with an increased risk of depressive symptoms among adults8. However, in a study conducted on European populations, no conclusive evidence was found to support the association between outdoor air pollutants and depression9. According to estimates, approximately 2.4 billion people globally, particularly in low- and middle-income countries like China, continue to rely on solid fuels for cooking10. The incomplete combustion of indoor solid fuels leads to the production of sulfur dioxide, black carbon, carbon monoxide, and particulate matter11. This process results in high levels of indoor air pollution, which has been identified as a major contributor to the burden of disease12. Different from outdoor air pollutants, indoor air pollutants are more closely related to the health of middle-aged and elderly people because they spend more time indoors and have a longer exposure to indoor air pollutants13. Limited evidence exists regarding the relationship between indoor air pollution caused by the use of solid fuels indoors and depression, in comparison to outdoor air pollutants11. To gain a comprehensive understanding, it is essential to investigate the correlation between outdoor and indoor air pollution and its potential impact on depression.

To the best of our knowledge, there are still few studies in China examining the relationship between exposure to outdoor and indoor air pollution and depressive symptoms. Most existing research primarily focuses on either outdoor or indoor air pollution exposure, with limited studies evaluating both simultaneously about depressive symptoms. In addition, considering the similar underlying mechanisms of depression caused by outdoor and indoor air pollutants14,15, we hypothesize that there could be an interaction between indoor and outdoor air pollutants. Therefore, to fill this knowledge gap, this study aimed to investigate the association and interaction between outdoor air pollutants (fine particulate matter [PM2.5], inhalable particulate matter [PM10], sulfur dioxide [SO2], nitrogen dioxide [NO2], ozone [O3], and carbon monoxide [CO]), indoor air pollutants from solid fuel use, and depressive symptoms in middle-aged and elderly adults based on the China Health and Retirement Longitudinal Study (CHARLS).

Materials and methods

Study population

The CHARLS is a national longitudinal survey that employs a multistage probability sampling method to gather data from residents aged 45 years and above in China. The survey covered 150 counties in 28 provinces and collected information on participants’ basic information, health status and function, family characteristics, social and economic status, and retirement information. The National baseline survey was conducted in 2011, and participants were followed up every 2–3 years. Subsequently, the Wave 2 study was conducted in 2013, followed by the Wave 3 study in 2015, and finally the Wave 4 study in 2018. This study utilized data from the 2018 (Wave 4 study) and included a total of 18,152 participants. Then, we excluded (1) 2802 participants who had a change of residential address, (2) 188 participants who were younger than 45 years of age, (3) 412 participants who had an intellectual disability and brain damage, and (4) 208 participants who had Alzheimer’s disease, brain atrophy, Parkinson’s disease, and other memory-related disorders. A total of 14,542 participants were included in the study for analysis. The detailed selection process of participants is illustrated in Fig. S1. This study utilized publicly available data from the CHARLS (http://charls.pku.edu.cn). The study protocol of CHARLS received approval from the Ethics Committee of Peking University (approval number: IRB00001052-11015), and all participants provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki.

Assessment of exposure to outdoor air pollutants

The monitoring data of outdoor air pollutants PM2.5, PM10, SO2, NO2, O3, and CO were obtained from the China National Environmental Monitoring Centre (CNEMC, http://www.cnemc.cn/). All monitoring was performed at the state-controlled air sampling (SCAS) site in each city, and multiple SCAS sites were set up in each city16. The layout of sampling points in each city followed technical regulations for the selection of ambient air quality monitoring stations (on trial). Before the data release, data quality assurance and control were carried out according to the technical guidelines on environmental monitoring quality management. These monitoring stations adhered to uniform construction standards, enabling them to provide a more accurate representation of the national air pollution concentration. The daily mean concentrations of air pollutants PM2.5, PM10, SO2, NO2, and CO were obtained based on the city where the participants were located. Additionally, the 8-hour sliding mean concentrations of O3 were also collected. The exposure windows were 30-day, 60-day, 90-day, 180-day, and 365-day before the interview date.

Assessment of indoor air pollutants from solid fuel use

A primary source of indoor air pollution is from combustion of gas or other fuel used to cook or heat the home. In this study, the participants’ use of indoor heating and cooking fuels was assessed through questionnaires. Questions regarding information on household fuel types included such as: “Does your residence have coal gas or natural gas supply?”, “What is the main source of cooking fuel?”, “Does your residence have heating? (does not include AC with heating or self-made heating)” and “What is the main heating energy source?”. Based on the above standardization questions, heating fuel was divided into (1) solar; (2) coal; (3) natural gas; (4) liquefied petroleum gas; (5) electric; (6) crop residue/wood burning; and (7) others. Cooking fuel was divided into (1) coal; (2) natural gas; (3) marsh gas; (4) liquefied petroleum gas; (5) electric; (6) crop residue/wood burning; and (7) others. Based on previous research, we classified heating and cooking fuels into two categories. The first category includes solar, natural gas, liquefied petroleum gas, electric, marsh gas, and other similar fuels, which are considered clean fuels (non-solid fuels). The second category includes coal and burning of crop residue/wood, which are classified as solid fuels17.

Assessment and definition of depressive symptoms

The CHARLS employed the 10-item Center for Epidemiologic Studies Depression Scale (CES-D 10) to evaluate the depressive symptoms of participants over the previous week. Due to the established reliability and validity found in previous studies, the CES-D 10 scale is widely utilized for measuring depression in middle-aged and elderly adults18. The CES-D 10 scale is comprised of 10 questions that assess feelings or behaviors. In the CES-D 10 answers, participants rate eight negative feelings or behaviors using a four-point scale: 3 = 5–7 days (Most or all of the time); 2 = 3–4 days (Occasionally or a moderate amount of the time); 1 = 1–2 days (Some or a little of the time); 0 = Less than 1 day (Rarely or none of the time). Additionally, two positive feelings or behaviors are inversely rated using the same scale. The total score on the CES-D 10 scale in this study ranged from 0 to 30, with higher scores indicating more severe depressive symptoms. Based on earlier published research results19,20, a cutoff value of 10 was used to define the dichotomous depression variable. Participants who scored 10 or more were considered to have depressive symptoms.

Assessment and definition of covariates

Drawing on previous research5,21, to control for potential bias arising from covariates associated with depressive symptoms or indoor and outdoor air pollutants, these covariates were categorized into four groups: demographic characteristics, lifestyle, health status, and household economic level. Demographic characteristics included age (“< 60 years” and “≥ 60 years”), gender (“Male” and “Female”), race (“Han” and “Others”), education level (“Illiterate”, “Primary school or below”, “Middle school”, and “High school or above”), marital status (“Married and cohabiting” and “Divorced, separated, widowed, and never married”), and residence (“Rural” and “Urban”). Lifestyle included smoking status (“Non-smoker” and “Smoker”) and drinking status (“Non-drinker” and “Drinker”). Health status included chronic disease status (“Yes” and “No”). Participants with hypertension, heart disease, stroke, diabetes, cancer, and dyslipidemia were classified as having chronic diseases. The household economic level was determined based on annual household income (“Low”, “Middle” and “High”).

The selection of covariates in this study warrants further explanation. Prior research has indicated that age and gender are significant risk factors for depressive symptoms22. The influence of age on depressive symptoms among the elderly could be attributed to alterations in their physical, psychological, and social roles, as they often face increased health issues and feelings of loneliness, both of which may exacerbate depressive symptoms23. Additionally, gender contributes to the prevalence of depressive symptoms, with women generally exhibiting a higher likelihood than men of experiencing such symptoms. This disparity may be linked to biological differences, hormonal fluctuations, and various social and psychological factors24. Research indicates that certain ethnic groups may be at an elevated risk for depression, potentially due to factors such as cultural background, social status, and experiences of discrimination25. Regarding educational attainment, lower levels of education may serve as a risk factor for depressive symptoms, likely due to limited employment opportunities, reduced income, and diminished social support23. Previous research indicated that individuals who were divorced, separated, widowed, or never married may be more susceptible to experiencing depressive symptoms compared to their married or cohabiting counterparts. This disparity may be linked to differences in social support and life stability22. Additionally, residing in rural areas may pose a risk factor for depressive symptoms due to limited resources and insufficient medical services26. The effects of smoking and drinking on depressive symptoms remain contentious, as varying cutoffs can yield conflicting results. The presence of chronic diseases, the number of chronic diseases, and the occurrence of new chronic diseases all increase the risk of depression27. Individuals suffering from chronic conditions may be more vulnerable to depressive symptoms due to the stress inherent in their illnesses and the anxiety surrounding their health28. Additionally, lower family economic status may contribute to heightened stress, further increasing the risk of depression29.

Statistical analysis

The baseline characteristics of participants were described based on their depressive status. Continuous variables were presented as mean ± standard deviation and tested using a t-test. Categorical variables were presented as percentages and tested using a chi-square test. Daily missing data on outdoor air pollutants were imputed using the daily mean concentrations of each pollutant to reduce bias caused by missing data on air pollutants30. To investigate the effects of outdoor air pollutant exposure windows, we calculated average levels of outdoor air pollutants for 30-day, 60-day, 90-day, 180-day, and 365-day. We employed multivariate logistic regression models to estimate the odds ratio (OR) (95%CI) for the association between outdoor air pollutant exposure (measured as an increment per standard deviation [SD] of outdoor air pollutants) and depressive symptoms. We employed multivariate logistic regression to investigate the association between indoor air pollutants from solid fuel use and depressive symptoms. Four models were used for the analysis. Model 1 represented the unadjusted model. Model 2 adjusted for age and gender. Model 3 further adjusted for smoking status, drinking status, and chronic disease. Model 4 included additional controls for race, marriage status, education level, residence, and annual household income.

In addition, we performed a multiplicative and additive interaction analysis to examine the effects of outdoor and indoor air pollutants on the occurrence of depressive symptoms. The interaction analysis used a 365-day exposure window for outdoor air pollutants as the exposure measure. To examine multiplicative interaction, the interaction term of outdoor air pollutants and the type of indoor fuel use was included in the multivariate logistic model, and stratified analysis was performed based on the type of indoor fuel use. In the case of additive interactions, the participants were stratified based on the median concentration of outdoor air pollutants. The following categories were defined: participants with low exposure to outdoor air pollutants and indoor use of clean fuels (low clean), participants with low exposure to outdoor air pollutants and indoor use of solid fuels (low solid), participants with high exposure to outdoor air pollutants and indoor use of clean fuels (high clean), and participants with high exposure to outdoor air pollutants and indoor use of solid fuels (high solid). We utilized three measures to assess the magnitude of the additive interaction: relative excess risk due to interaction (RERI), attributable proportion due to interaction (AP), and synergy index (S). In addition, we performed two sensitivity analyses. By utilizing a cutoff value of 12 for CES-D 10 scores, previous studies demonstrated the effectiveness of this threshold in identifying clinically significant depression. We reanalyzed the relationship between outdoor air pollutants, indoor air pollutants from solid fuel use, and depressive symptoms31. In addition, linear models were employed to analyze the association between outdoor air pollutants, indoor air pollutants from solid fuel use, and CES-D 10 scores. All statistical analysis was conducted using R software (version 4.1.1). In this study, significance levels were considered p values < 0.05 (two-tailed).

Results

Descriptive statistics

A total of 14,542 participants were selected from the wave 4 study of CHARLS. The distribution of these participants across 28 provinces in China is illustrated in Fig. 1. Table 1 presents an overview of baseline characteristics and summary statistics of the participants. The average age of the 14,542 participants was 62.1 ± 10.2 years. Among the total number of participants, 56.4% were 60 years old or older, and 47.8% were male. The study observed that participants who developed depressive symptoms were more likely to exhibit certain characteristics. These characteristics included being non-smokers and non-drinkers, having chronic diseases, residing in rural areas, being illiterate, being divorced, separated, widowed, or never married, having a low annual household income, and using solid fuels for heating and cooking (p < 0.05). Table S1 presents the average levels of outdoor air pollutant concentrations in the five exposure windows. The average concentrations of PM2.5, PM10, SO2, NO2, O3 and CO in the past year were 45.55 (SD = 14.71), 82.33 (SD = 28.45), 16.43 (SD = 8.23), 32.18 (SD = 9.88), 62.94 (SD = 9.46), 928.93 (SD = 249.81), respectively.

Fig. 1
figure 1

Sample distribution.

Full size image
Table 1 Baseline characteristics of participants were described based on depressive symptoms.
Full size table

Relationship between exposure to outdoor air pollutants and depressive symptoms

Figure 2 illustrates the correlation between each SD increase in outdoor air pollutant concentration and the occurrence of depressive symptoms. It was observed that each SD increase in PM2.5 over both the 180-day and 365-day exposure windows was significantly associated with the occurrence of depressive symptoms. This association remained significant even after adjusting for covariates in the logistic regression model. Notably, the effect was found to be strongest in the 365-day exposure window (OR = 1.215, 95% CI: 1.073, 1.375). In Model 4, the occurrence of depressive symptoms was significantly associated with SO2 exposure in 30-day, 60-day, and 365-day exposure windows. The effect was found to be greatest in the 365-day exposure window (OR = 1.068, 95% CI: 1.006, 1.134). However, in the model adjusted for all covariates, each SD increment of NO2 and O3 in the 60-day, 90-day, 180-day, and 365-day exposure windows was associated with a lower occurrence of depressive symptoms. The greatest effect was observed for NO2 in the 365-day exposure window (OR = 0.793, 95% CI: 0.741, 0.848), followed by O3 in the 60-day exposure window (OR = 0.846, 95% CI: 0.781, 0.917). In addition, PM10 and CO were not found to be statistically associated with the occurrence of depressive symptoms in this study (p > 0.05) (Table S2).

Fig. 2
figure 2

Odds ratio and 95% CI of each SD increment of outdoor air pollutants for depression symptoms. Notes: *p < 0.05, **p < 0.001. Abbreviations: CI, confidence interval. Model 1: the unadjusted model; Model 2: model 1 adjusted for age and gender; Model 3: model 2 with additional adjustment for smoking status, drinking status, and chronic disease (hypertension, heart disease, stroke, diabetes, cancer, and dyslipidemia); Model 4: model 3 with additional controlling for race, marriage status, education level, residence, and annual household income.

Full size image

Relationship between indoor air pollutants from solid fuel use and depressive symptoms

Table 2 describes the association between indoor air pollution from using solid fuels for heating and cooking and the occurrence of depressive symptoms. In Model 1, without covariate adjustment, indoor air pollution from using solid fuels for heating was found to be associated with the occurrence of depressive symptoms compared to clean energy (OR = 1.88, 95% CI: 1.52, 2.32). In Model 2, Model 3, and Model 4, with further adjustment of covariates, the relationship between indoor air pollution from solid fuel use for heating and the occurrence of depressive symptoms exhibited a consistent trend, albeit with a reduced magnitude. In Model 4, indoor air pollution from the use of solid fuels for heating was found to be associated with the occurrence of depressive symptoms, with an OR of 1.44 (95% CI: 1.05, 1.97). Comparable findings were observed in the context of cooking, as indoor air pollution from using solid fuels for cooking was linked to the occurrence of depressive symptoms when contrasted with the use of clean fuels. In logistic regression analyses adjusting for all covariates, indoor air pollution from using solid fuels for cooking was associated with the occurrence of depressive symptoms with an OR of 1.43 (95% CI: 1.29, 1.58).

Table 2 Odds ratio and 95% CI of indoor fuel use types for depression symptoms.
Full size table

Interaction of outdoor and indoor air pollutants on depressive symptoms

Figure 3 illustrates the multiplicative interaction between outdoor air pollutants and indoor air pollution from the use of solid fuels on depressive symptoms. Indoor air pollution from using solid fuels for cooking and exposure to outdoor PM2.5 (interaction p = 0.049) and PM10 (interaction p = 0.019) had an antagonistic effect on the occurrence of depressive symptoms, as observed on a multiplicative scale. The stratified analysis results indicated that indoor air pollution from cooking with solid fuels might modify the effect of PM2.5 (OR = 1.227, 95% CI: 0.990, 1.521 vs. OR = 1.246, 95% CI: 1.063, 1.461) and PM10 (OR = 0.684, 95% CI: 0.526, 0.896 vs. OR = 1.106, 95% CI: 0.847, 1.219) on the occurrence of depressive symptoms when compared to the use of clean fuel. In this study, we did not find any evidence of a multiplicative interaction between indoor air pollution from solid fuel use for heating and exposure to outdoor air pollutants. Table S3 presents the additive interaction between outdoor air pollutants and indoor air pollution from solid fuel use on depressive symptoms. The results indicated that there was no additive interaction between indoor air pollution from using solid fuels for heating and cooking and outdoor air pollutants on the occurrence of depressive symptoms.

Fig. 3
figure 3

The relationship between each SD increment of 365-day outdoor air pollutants exposure and depressive symptoms in multivariate models: modification by indoor fuel use types. Notes: *p < 0.05, **p < 0.001. The p-value is the multiplicative interaction. Abbreviations: OR, odds ratio; CI, confidence interval. Multivariate logistic regression adjusted for age, gender, smoking status, drinking status, chronic disease (hypertension, heart disease, stroke, diabetes, cancer, and dyslipidemia), race, marriage status, education level, residence, and annual household income.

Full size image

Sensitive analysis

The results of the sensitivity analysis were generally consistent with those of our primary models. Table S4 and Table S5 present the effects of outdoor air pollutants and type of indoor fuel use on the risk of occurrence of depressive symptoms (cutoff = 12), respectively. This result is consistent with the main result. Figure 4 and Table S6 present the interaction of outdoor air pollutants and indoor fuel use on the occurrence of depressive symptoms (cutoff = 12). The findings from the interaction analysis aligned with the main results. In addition, solid fuel use for cooking and outdoor O3 (interaction p = 0.020) exposure were found to have antagonistic effects on depressive symptoms on a multiplicative scale. Solid fuel use for heating and outdoor NO2 exposure had antagonistic effects on depressive symptoms on an additive scale. The relationship between outdoor air pollutants, indoor fuel types, and CES-D 10 scores is further detailed in the supplementary material (Tables S7-S9). The relationship (including interactions) between outdoor air pollutants, indoor air pollutants from solid fuel use, and depression scores was similar to the results for depressive symptoms.

Fig. 4
figure 4

The relationship between each SD increment of 365-day outdoor air pollutants exposure and depressive symptoms (cutoff 12) in multivariate models: modification by indoor fuel use types. Notes: *p < 0.05, **p < 0.001. The p-value is the multiplicative interaction. Abbreviations: OR, odds ratio; CI, confidence interval. Multivariate logistic regression adjusted for age, gender, smoking status, drinking status, chronic disease (hypertension, heart disease, stroke, diabetes, cancer, and dyslipidemia), race, marriage status, education level, residence, and annual household income.

Full size image

Discussion

In this study, we found that exposure to outdoor air pollutants such as PM2.5 and SO2, as well as indoor air pollutants from the use of solid fuels for heating and cooking, were significantly associated with the occurrence of depressive symptoms. In addition, we observed a significant interaction between outdoor and indoor air pollutants, suggesting that indoor air pollutants caused by cooking with solid fuels may weaken the association between outdoor PM2.5 and PM10 exposure and depressive symptoms. We conducted sensitivity analyses to ensure the reliability and consistency of our findings.

Outdoor air pollutants and depressive symptoms

Several studies have reported a significant association between short-term exposure to PM2.5 and SO2 and an increased risk of hospital admission for depression7,32. Long-term exposure to PM2.5 and SO2 was also found to be associated with a higher prevalence of depression when compared to short-term exposure33,34. In this study, we discovered a significant association between outdoor exposure to PM2.5 and SO2 and the occurrence of depressive symptoms, which is consistent with previous studies. The relationship between O3 exposure and depressive symptoms has yielded inconsistent results in past studies35. The findings from both this study and the study conducted in Northeast China do not provide support for the hypothesis that exposure to O3 will lead to an increase in the occurrence of depressive symptoms36. In contrast to previous studies36,37, this study found no association between PM10 and CO exposure and depressive symptoms. Possible reasons for these differences may include variations in the exposure level of the study population, composition of pollutants, meteorological conditions, demographic characteristics, study design, and adjustment for confounding factors. Notably, the association of NO2 with depressive symptoms was found to be abnormal in our study. Contrary to most previous studies, we observed that NO2 exposure was associated with a lower incidence of depressive symptoms34,38. Several potential explanations may account for this unexpected finding. First, our exposure assessment relied on ambient NO₂ concentrations from fixed-site monitoring stations as a surrogate for individual-level exposure. This may introduce non-differential exposure misclassification, potentially biasing estimates toward the null. Second, regions with higher NO₂ concentrations are often urban cores that benefit from better access to healthcare, social services, and higher socioeconomic status. Inadequate control for these contextual factors may have introduced residual confounding, attenuating or even reversing the observed associations. Third, unlike fine particulate matter (PM2.5), which has well-established neuroinflammatory and neurotoxic effects, the biological mechanisms linking NO₂ to depression remain unclear. It is also plausible that the observed effect may be attributed to co-pollutants highly correlated with NO₂, such as polycyclic aromatic hydrocarbons (PAHs), which have been shown to exert neurotoxic effects39. Moreover, the hormesis hypothesis in environmental toxicology suggests that low levels of certain pollutants may trigger transient adaptive responses, resulting in non-linear or even inverse associations in specific subpopulations or exposure ranges. Taken together, these factors suggest that the observed inverse or null association should be interpreted with caution. Therefore, further research is warranted to clarify the mental health effects of NO₂, and to systematically examine the associations between exposure to outdoor O3, PM10, CO, and depressive symptoms using individualized exposure assessments and multi-pollutant modeling approaches.

Indoor air pollutants from solid fuel use and depressive symptoms

Limited evidence exists regarding the correlation between mental health and indoor air pollutants resulting from the utilization of solid fuels in comparison to outdoor air pollutants17. A cross-sectional study conducted in six low- and middle-income countries revealed a significant association between the use of unclean cooking fuels in older adults and higher odds of depression40. According to a longitudinal study conducted in China, the prolonged use of solid fuels for heating and cooking among older adults was found to be significantly associated with a higher risk of depression17. In line with previous studies, our findings indicate that the use of solid fuels for indoor heating and cooking, leading to indoor air pollutants, is linked to the presence of depressive symptoms among middle-aged and elderly adults in China. The underlying mechanism between indoor air pollutants caused by solid fuel use and the occurrence of depressive symptoms is not well understood. One possible explanation is that the use of solid fuel indoors is linked to elevated levels of indoor air pollutants41. Increased levels of indoor air pollutants may potentially raise the risk of experiencing depressive symptoms through various pathways, including neuroinflammation, oxidative stress, vascular damage, and neurodegeneration42. Given the prevalent use of solid fuels for household heating and cooking in China, coupled with the increasing aging population and the significant burden of depression, policymakers must develop appropriate health policies that promote the adoption of clean energy sources and the reduction of indoor air pollution exposure.

The interaction of outdoor air pollutants and indoor air pollutants from fuel use on depressive symptoms

The results of the interaction analysis indicated a significant interaction between outdoor PM2.5 and PM10 exposure and indoor air pollutants caused by solid fuel use with depressive symptoms when considering the multiplicative scale. However, no significant interaction was observed when considering the additive scale. The concept of an “interaction continuum” is used to explain the occurrence of inconsistent results between multiplicative and additive interactions43. Scale dependencies can influence the way interactions occur. It is important to note that a negative multiplicative interaction can still be significant even if there is no additive interaction43,44. This study discovered that indoor air pollution from cooking with solid fuels had a mitigating effect on the link between outdoor PM2.5 and PM10 exposure and depressive symptoms. A cross-sectional study conducted in northeast China revealed that the utilization of solid fuel for indoor heating had a diminishing effect on the association between outdoor PM2.5 exposure and depressive symptoms36. This study suggests a potential interaction between outdoor and indoor air pollutants and their impact on depressive symptoms. The use of indoor solid fuel, which releases harmful air pollutants such as particulate matter, carbon monoxide, sulfur oxides, and polycyclic aromatic hydrocarbons45, could potentially contribute to alterations in overall pollution exposure. These changes in exposure could potentially amplify or mitigate the harmful effects of specific air pollutants36. Further research is required to investigate the impact of outdoor and indoor air pollutants on depression, as there is limited existing literature on this topic. Additional population cohorts and experimental studies are necessary to validate and establish conclusive findings.

Underlying mechanisms of outdoor air pollutants and indoor fuel use on depressive symptoms

The biological mechanisms underlying the association between exposure to air pollution and depressive symptoms remain incompletely understood. However, accumulating evidence suggests several plausible pathophysiological pathways. First, exposure to air pollutants—particularly PM2.5, NO₂, and indoor pollutants such as cooking smoke and volatile organic compounds—can induce systemic oxidative stress and inflammation46. These stressors have the capacity to cross the blood–brain barrier, initiating neuroinflammatory responses that can disrupt neuronal function and integrity. Oxidative stress, in particular, has been implicated in dopaminergic neuronal damage, with dopamine deficiency recognized as a neurochemical hallmark of depression47. In parallel, exposure to air pollutants may elicit a systemic inflammatory response48, resulting in cytokine production that promotes neurotoxicity and neuroinflammation, which may ultimately lead to cerebrovascular injury and neurodegeneration49. Furthermore, the entry of air pollutants into the body can trigger the activation of the hypothalamus-pituitary-adrenal (HPA) axis50. This could elevate cortisol secretion, a hormone closely linked to the pathogenesis of depression51. Another proposed mechanism involves the depletion of tryptophan, an essential precursor for serotonin synthesis. Exposure to air pollutants has been shown to reduce circulating tryptophan levels, thereby impairing serotonin production and contributing to depressive symptomatology52. Additionally, chronic exposure to air pollutants has been associated with hippocampal atrophy—a structural brain change frequently observed in individuals with depression53,54. Beyond the biological pathways, air pollution may also affect mental health through psychosocial mechanisms. Deterioration in air quality can restrict outdoor physical activity and limit opportunities for social interaction, both of which are important for psychological well-being55,56. This may further exacerbate depressive symptoms by promoting sedentary behavior, loneliness, and social isolation. Taken together, these multifaceted mechanisms—encompassing oxidative stress, inflammation, neuroendocrine dysregulation, neurotransmitter imbalance, and psychosocial deprivation—partially elucidate the adverse impact of air pollution on mental health. Nevertheless, further longitudinal and mechanistic studies are warranted to fully characterize the complex interplay between air pollutant exposure and the development of depressive disorders.

Implications of air pollution–related depression for promoting healthy ageing

Depression is not only a clinical syndrome but also a critical determinant of overall quality of life, particularly among middle-aged and older adults. Elevated depressive symptoms in this population have been associated with reduced physical functioning, limited social engagement, and lower life satisfaction. Our findings that both outdoor and indoor air pollutant exposures are linked to higher levels of depressive symptoms suggest that poor air quality may adversely affect multiple dimensions of well-being beyond emotional health alone. Long-term exposure to air pollutants may contribute to systemic inflammation and neuroendocrine dysregulation, which can, in turn, disrupt daily functioning, increase the risk of social isolation, and negatively influence perceived health status. These results highlight the potential of air quality interventions not only to support mental health but also to promote overall quality of life in ageing populations.

Study strengths and limitations

The study is the first to examine the effects of indoor and outdoor air pollutant exposure on depressive symptoms in middle-aged and older Chinese adults. This study encompasses data from 150 counties across 28 provinces in China, providing a high level of representativeness and reliability in its results. Although this study demonstrates innovation and has several advantages, it is important to acknowledge its limitations. First, due to the unavailability of detailed residential addresses, exposure assessment was conducted at the city level. This coarse spatial resolution may lead to exposure misclassification by failing to capture intra-city variability, potentially biasing effect estimates toward the null. Additionally, reliance on interpolated average outdoor pollutant concentrations may not fully address missing data, further limiting exposure accuracy. Second, direct measurement of indoor air pollution was not feasible given the constraints of large population-based surveys. Instead, proxy variables such as the use of solid versus clean cooking fuels were employed. While this comparative approach intuitively demonstrates the impact of solid fuel use on indoor air quality, it overlooks other important determinants, such as volatile organic compounds, formaldehyde from indoor materials, indoor vs. outdoor cooking, ventilation conditions (e.g., chimneys, windows, range hoods), building characteristics, and resident behaviors. These factors interact complexly in real environments, and their omission may oversimplify indoor pollution exposure and limit interpretability. In particular, the CHARLS dataset lacks detailed information on cooking location and ventilation; although the presence of an independent kitchen was considered as a proxy indicator, this remains insufficient to capture exposure heterogeneity fully. Future studies should incorporate these variables to enable more comprehensive exposure assessments. Third, seasonality is an important potential confounder, especially in China with distinct heating periods differing widely by region. The timing and intensity of heating activities can differ substantially from cooking fuel use due to climatic and geographic heterogeneity. Although seasonality was considered in the study design, limited availability of detailed seasonal heating data in CHARLS precluded direct adjustment. This residual confounding may dilute or mask associations between cooking-related exposures and depressive symptoms. Heating-related exposures may also interact with or obscure the health impacts attributed to cooking fuels. Our current work represents an initial exploration of these relationships; ongoing studies will further investigate the interplay and combined effects of heating and cooking exposures on depressive symptoms. Fourth, while the CES-D 10 scale is widely used, potential biases in assessing depressive symptoms cannot be ruled out18. Moreover, the cross-sectional study design limits causal inference, as exposure and outcome were assessed simultaneously. Fifth, due to the restrictive inclusion and exclusion criteria applied to the study population, our decision to exclude participants with intellectual disabilities, brain injuries, Alzheimer’s disease, brain atrophy, Parkinson’s disease, and other memory-related conditions may have led to an underestimation of the prevalence of depressive symptoms. This limitation potentially reduces the statistical power to detect the effects of indoor and outdoor air pollutants. Consequently, the limitation of this study may not stem from the sample size, but rather from the relatively small number of cases for the outcome variable assessed8. Changes in residential address may increase the likelihood of selection bias57, as exposure data for outdoor air pollutants are only available for participants who have not changed their residential address. In contrast, exposure data for participants who have changed their residential address are unavailable. Finally, it is important to note that the findings of this study may not be directly applicable to populations in other countries, as the research was specifically conducted on a sample of middle-aged and elderly adults in China. Our study population comprises middle-aged and older adults in China, where average life expectancy, pollution sources, and socioeconomic conditions differ from those in other regions. These factors may modify both exposure intensity and individual vulnerability, potentially affecting the magnitude of observed risks. Therefore, we caution against directly applying our effect estimates to populations with substantially different demographic or environmental profiles. Nevertheless, cohort studies conducted in Europe⁴ and North America⁵ have similarly linked long-term air pollution exposure to an increased risk of depression, suggesting that the underlying biological mechanisms are broadly relevant across diverse settings.

Conclusion

In this study, we found a significant association between exposure to outdoor air pollutants, specifically PM2.5 and SO2, and the occurrence of depressive symptoms among middle-aged and older adults in China. Additionally, indoor air pollution resulting from the use of solid fuels for heating and cooking was also significantly associated with elevated depressive symptoms. Notably, our findings suggest that indoor air pollutants from solid fuel use may attenuate the relationship between outdoor air pollution and depressive symptoms, highlighting the complex interplay between indoor and outdoor environmental exposures. Given that depressive symptoms represent a key indicator of quality of life, these findings underscore the importance of addressing both indoor and outdoor air pollution as part of broader public health strategies. Integrated approaches—such as enforcing stricter air quality regulations, promoting cleaner household energy sources, and implementing comprehensive pollution control policies—are essential to improving mental health and overall well-being in ageing populations.