Systematic Review and Meta-Analysis of the Association between Environmental Noise Exposure and Depression and Anxiety Symptoms in Community-Dwelling Adults

Abstract

Background:

Environmental noise pollution has emerged as an increasingly significant public health concern, particularly in relation to its potential impact on the mental health of adults aged 35 years or older. This study aims to systematically examine the association between environmental noise exposure and the incidence of depression and anxiety symptoms within this demographic.

Methods:

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, a comprehensive search of multiple electronic databases was conducted. A total of 11 eligible studies were identified, comprising 15,525 adults aged 35 years or older, among whom 2205 (14.2%) individuals reported experiencing symptoms of depression or anxiety. A random-effects model was used for meta-analysis with heterogeneity assessment, sensitivity analysis and exploratory subgroup analyses to evaluate the robustness and variability of the findings.

Results:

The meta-analysis demonstrated a significant association between long-term environmental noise exposure and an increased risk of depression and anxiety symptoms in adults aged 35 years or older (risk ratio [RR] = 1.29, 95% confidence interval: 1.16–1.51, P < 0.05). Substantial heterogeneity was observed across the included studies (I² = 68.1%), suggesting moderate variability in study design, geographic location and exposure assessment methodologies. Sensitivity analyses confirmed the robustness of the findings, and recalculated RRs ranged from 1.27 to 1.35. Exploratory subgroup analyses revealed that road traffic noise was the most commonly investigated source and consistently exhibited positive associations. Notably, both aircraft noise and community-level or multi-source noise were linked to adverse mental health outcomes.

Conclusion:

Long-term exposure to environmental noise increases the risk of depression and anxiety in adults aged 35 or older. Key factors such as socioeconomic status, residential environment and individual sensitivity to noise moderate this association. Future research should further explore the role of socioeconomic factors.

KEY MESSAGES

• (1)Long-term environmental noise exposure is significantly associated with increased depression and anxiety symptoms in adults aged 35 years and older.
• (2)Socioeconomic factors, noise sensitivity, and residential environment may moderate the relationship between environmental noise and mental health outcomes.
• (3)Further research, particularly using longitudinal designs and controlling for confounders like socioeconomic status (SES), is necessary to establish causality.

INTRODUCTION

In recent years, increasing attention has been directed towards the influence of environmental noise on psychological well-being. A growing body of evidence suggests that noise exposure may elevate the risk of depression, anxiety, suicidal ideation and behavioural disorders. The proposed underlying mechanisms include neuroinflammation, oxidative stress and disruptions in neurochemical homeostasis, all of which may contribute to the onset and progression of psychological disorders.[1] However, the precise relationship between environmental noise exposure and mental health outcomes remains a subject of debate, particularly in adults aged 35 years or older and in ageing populations. Individuals with pre-existing physiological or cognitive vulnerabilities may exhibit heightened sensitivity to environmental stressors such as noise pollution.[2,3,4] While long-term exposure to air pollutants has been consistently associated with an increased risk of depression across diverse age groups, findings regarding the mental health effects of environmental noise exposure in adults aged 35 years or older have been inconclusive, underscoring the need for further investigation.

In recent years, environmental noise has emerged as a pervasive stressor, particularly affecting vulnerable populations, including adults aged 35 years or older. Despite growing concern, studies investigating the association between environmental noise exposure and mental health outcomes, such as depression and anxiety, have yielded inconsistent findings.[5] These inconsistencies are partly attributable to substantial heterogeneity across studies in terms of population characteristics; study design (e.g., cross-sectional versus cohort); noise exposure assessment methods (e.g., modelled Lden versus self-reported annoyance) and mental health measurement instruments (e.g., General Health Questionnaire, Short Form-36 Health Survey [SF-36] and Patient Health Questionnaire-9 [PHQ-9].

Given these methodological discrepancies and the growing global burden of age-related mental health concerns, there is a critical need for a comprehensive synthesis of the available evidence.[6,7] Accordingly, this study conducts a systematic review and meta-analysis to evaluate the relationship between environmental noise exposure and the prevalence of depressive and anxiety symptoms amongst adults aged 35 years or older. By integrating findings from 11 high-quality studies spanning diverse geographic regions; noise sources (including road traffic, aircraft and community-level noise) and measurement tools, this review seeks to clarify both the magnitude and consistency of the observed associations. Ultimately, the findings aim to provide a robust scientific foundation to inform public health strategies and guide targeted interventions aimed at mitigating the psychological impacts of environmental noise on this population.

Although prior studies and international guidelines [e.g., those of the World Health Organization (WHO)] have commonly defined ‘middle age’ as beginning at 45 years, this review adopts a threshold of 35 years for two principal reasons as follows: (1) several high-quality studies included in our synthesis used 35 years as the lower age boundary for reporting mental health outcomes; and (2) evidence suggests that vulnerability to environmental noise, which is reflected in cumulative exposure burden and the onset of stress-related symptoms, may begin as early as the mid-thirties. Thus, the inclusion of adults aged 35 years or older ensures both the inclusivity of relevant data and the comparability of findings across studies.

MATERIALS AND METHODS

Search Strategy

To ensure a comprehensive and systematic analysis, an extensive literature search was conducted across both English- and Chinese-language databases. The databases included CNKI, Wanfang Data, VIP Chinese Science and Technology Journals Database, CBM, PubMed, Embase, the Cochrane Library and Web of Science. The search spanned from the inception of each database through March 2024, with the objective of capturing all potentially relevant studies. A combination of controlled vocabulary (e.g., MeSH, Emtree) and free-text keywords was employed to maximise sensitivity. Key search terms included environmental noise, traffic noise, aircraft noise, depression, anxiety and age-related descriptors, such as adults aged 35 years or older, middle-aged and elderly people.

However, the actual inclusion criterion was restricted to studies that reported data specifically for adults aged 35 years or older, and studies without stratified data for this age group were excluded during screening. We employed Boolean operators (AND, OR) to enhance search specificity and breadth. We systematically searched PubMed using the following search strategy: (‘Environmental Noise’ OR ‘Traffic Noise’ OR ‘Aircraft Noise’ OR ‘Community Noise’ OR ‘Noise Pollution’ OR ‘road traffic noise’ OR ‘aircraft noise’ OR ‘community noise’) AND (‘Depression’ OR ‘Anxiety’ OR ‘Mental Health’ OR ‘Psychological Distress’ OR ‘depression symptoms’ OR ‘anxiety symptoms’) AND (‘Adult’ OR ‘Middle Aged’ OR ‘Older Adults’ OR ‘Aged’) AND (‘35 years’ OR ‘aged 35 and older’) NOT (‘Children’ OR ‘Adolescents’).

In addition to database searches, manual screening of reference lists from eligible articles and targeted reviews of relevant professional journals were conducted for the identification of additional studies. Literature screening was independently performed by two researchers using predefined inclusion and exclusion criteria to ensure methodological rigour and reproducibility.

INCLUSION AND EXCLUSION CRITERIA

Inclusion Criteria

The inclusion criteria for this review were defined as follows:

Studies were deemed eligible when they met all of the following criteria: first, the study population consisted of adults aged 35 years or older according to the definition used in each study. While sample size was generally considered in the selection process, studies with slightly smaller sample sizes (e.g., Hardoy et al.,[13] with n = 71 and Nitschke et al.,[14] with n = 63) were retained when they demonstrated high methodological rigour and contributed unique population characteristics relevant to the review’s objectives. For instance, the study by Hardoy et al.[13] (n = 71), which focused on a cohort of adults aged ≥45 years, and the study by Nitschke et al.[14] (n = 63), which investigated a highly noise-annoyed subpopulation in South Australia, were included because of their high-quality study designs and relevance to the research question.

Additionally, the study of Wright et al.,[10] which provided census-based data to examine self-reported mental health outcomes, was included despite not employing validated diagnostic tools. This decision was justified by the study’s large sample size (n = 2939), which allowed for its inclusion in the pooled effect estimates. The limitations associated with self-reported, non-validated mental health outcomes are acknowledged and critically examined in the Discussion section.

Eligible studies were required to employ either a cohort or cross-sectional study design, with the primary objective of investigating the association between environmental noise exposure, symptoms of depression and/or anxiety in adults aged 35 years or older. Environmental noise exposure, including road traffic noise, aircraft noise, railway noise, neighbour noise, community-level noise, urban ambient noise and multi-source combinations, had to be clearly defined and assessed using objective monitoring instruments, predictive modelling or standardised questionnaires. A minimum sample size of 50 participants was required for inclusion. No restrictions were imposed on the duration of follow-up or observational period, allowing for the inclusion of both cross-sectional and longitudinal designs.

Additionally, at least one mental health outcome was reported, such as depression scores (e.g., PHQ-9, Center for Epidemiologic Studies Depression Scale [CES-D]; anxiety scores (e.g., Hospital Anxiety and Depression Scale–Anxiety subscale, Generalized Anxiety Disorder 7-item scale and mental health subscales [e.g., SF-36] or data on the use of antidepressant or anxiolytic medications.

Exclusion Criteria

Studies were excluded when they met one or more of the following criteria:

Population mismatch: Articles involving mixed-age samples that did not provide specific statistical results for adults aged 35 years or older were excluded to maintain population consistency and reduce potential confounding arising from age heterogeneity.

Insufficient follow-up duration: Longitudinal cohort studies were included. Cross-sectional studies were also considered eligible, provided they reported valid and extractable data on chronic or habitual environmental noise exposure and corresponding mental health outcomes. Therefore, no exclusions were applied based on follow-up duration, as long as the study addressed chronic or habitual noise exposure rather than acute events.

Inappropriate study design: Non-original research, including reviews, case reports, editorials and animal studies, was excluded. Studies were also excluded when they lacked a clear environmental noise exposure evaluation or did not use validated mental health outcome measures, such as those focusing exclusively on hearing impairment or cardiovascular parameters.

Data incompleteness: Studies without sufficient statistical data, such as missing risk estimates, standard errors or confidence intervals (CIs), were excluded because of their limited utility in quantitative synthesis.

Duplicate publications: In instances where multiple articles were derived from the same cohort or included overlapping participant data, only the most comprehensive or up-to-date version was retained to avoid data duplication.

The application of these exclusion criteria ensured scientific rigour, population specificity and methodological comparability across the included studies, thereby enhancing the validity of the meta-analytic findings and supporting more robust and accurate conclusions regarding the impact of environmental noise exposure on the mental health of adults aged 35 years or older.

Study Inclusion and Subgroup Extraction

For the purposes of meta-analysis, we extracted specific subgroups from each eligible study population to ensure comparability across studies in terms of age, noise exposure levels and the definitions used for mental health outcomes. Many of the included studies involved large general population cohorts or administrative datasets. However, only participants who met the following criteria were included in the pooled effect estimates:

• (1)Adults aged 35 years or older, consistent with the focus of this review;
• (2)Exposure to high levels of environmental or occupational noise, as defined by study-specific thresholds (e.g., Lden ≥65 dB, ≥70 dB) or inclusion in ‘high annoyance’ groups where quantitative thresholds were not specified;
• (3)Assessment of mental health outcomes using validated and standardised measures, such as PHQ-9 scores ≥10, international statistical classification of diseases and related health problems 10th revision (ICD-10) physician diagnoses, SF-36 Mental Health subscale scores <30, center for epidemiological studies depression scale (CES-D) scores ≥17 or equivalent clinical proxies.

For example, in Leijssen et al.,[9] although the full sample included 23,293 adults, only the subgroup exposed to Lden ≥70 dB (n = 279) was included in the analysis. Likewise, Seidler et al.[15] reported data on more than 578,000 insured individuals, but only those with confirmed depression diagnoses and exposure to road traffic noise ≥70 dB (n = 8865) were analysed. In studies such as Sygna et al.[16] and Nitschke et al.,[14] the included subsamples consisted of individuals reporting poor sleep and high noise annoyance, respectively.

Study Quality Evaluation

The methodological quality of the studies included in this meta-analysis was assessed using the Newcastle–Ottawa Scale (NOS), a tool specifically developed for evaluating observational studies, including both cross-sectional and cohort designs. The NOS assesses study quality across three key domains: selection, comparability and outcome. Each domain comprises a set of criteria, and studies receive points based on the extent to which they meet these criteria. The total NOS score ranges from 0 to 9, with higher scores reflecting higher methodological quality.

Selection evaluates the adequacy of sample selection, including the representativeness of the study population and the appropriateness of control groups.

Comparability examines whether the study controlled for potential confounding factors.

Outcome assesses the validity and reliability of the outcome measurements used in the study.

The studies were rated as follows based on the NOS criteria:

high quality: NOS score of ≥7

moderate quality: NOS score of 5–6

low quality: NOS score of <5

Statistical Analysis

To account for heterogeneity across studies and improve the comparability of results, a random-effects model was employed in the meta-analysis. Differences in study design, such as the use of cohort studies with temporal data versus cross-sectional studies with snapshot data, may have influenced the pooled estimates. Although the random-effects model adjusts for some degree of between-study variability, it does not fully eliminate the effects of methodological differences, and residual confounding may persist, particularly in studies that did not sufficiently control for key variables. Given the variability in mental health assessment tools across the included studies (e.g. PHQ-9, CES-D, ICD-10 diagnoses), subgroup analyses were conducted to evaluate whether the association between environmental noise exposure and mental health outcomes remained consistent when limited to studies using validated clinical diagnoses (e.g., ICD-10) or standardised instruments with established thresholds (e.g., PHQ-9 ≥10). These analyses aimed to address the potential influence of measurement heterogeneity on the overall findings. Despite these methodological adjustments, the results of this meta-analysis should be interpreted with caution, particularly in light of unmeasured or inadequately adjusted factors such as socioeconomic status, pre-existing mental health conditions and differential access to healthcare. For studies that reported odds ratios (ORs) rather than risk ratios (RRs), the Zhang and Yu[6] method was applied to convert ORs to RRs, provided that baseline incidence (P₀) data were available. This conversion was necessary to ensure that the effect estimates accurately reflected relative risk.

where OR is the odds ratio reported by the study. P₀ is the baseline incidence (prevalence) of the outcome in the unexposed group (control group). If the baseline incidence (P₀) was not reported, a reasonable estimate (e.g., 0.1 or 0.2) was assumed for the conversion. This formula is appropriate when the P₀ (baseline incidence) is available. When P₀ was not provided, we either used an assumed value for P₀ or retained the OR in the analysis, depending on the data available. The method is primarily applicable to studies with cohort or cross-sectional designs, where the baseline incidence is known or can be estimated.

Specifically, six studies required conversion from ORs to RRs using this method. For example, Leijssen et al.[9] used available baseline prevalence data. In cases where conversion was not feasible (e.g., Leijssen et al.[9], the reported ORs were retained. For these cases, we discussed the potential for bias due to the use of ORs rather than RRs in the meta-analysis. This approach improves interpretability, especially when the outcome is not rare. When conversion was infeasible, the original OR was retained with the corresponding sensitivity analysis.[8]

For dichotomous outcomes, such as the prevalence of depression or anxiety, pooled effect sizes were expressed as RRs with corresponding 95% CIs. For continuous outcomes, including symptom severity scores for depression or anxiety, the standardised mean difference was used for the estimation of effect size. Between-study heterogeneity was assessed using the I² statistic, and values exceeding 50% were considered indicative of moderate to substantial heterogeneity.

Potential sources of heterogeneity, including noise type (e.g., road traffic noise, aircraft noise), exposure assessment methods (e.g., modelled decibel levels vs. self-reported annoyance) and study design (e.g., cohort vs. cross-sectional), were explored using subgroup analysis. Furthermore, we conducted sensitivity analyses by sequentially excluding each study to examine its influence on the overall pooled estimate and to ensure the stability of the results. All statistical tests were two-sided, and a P-value of less than 0.05 was considered statistically significant.

In the forest plot [Figure 2], each study’s contribution to the overall pooled effect was weighted according to the inverse variance method under the random-effects model. Specifically, study weights reflected the inverse of the sum of the within-study variance and the between-study variance (tau²), as implemented in standard meta-analysis software such as Stata and R. For instance, Seidler et al.,[15] due to its large sample size and narrow CI, contributed approximately 60.0% to the overall pooled estimate. These weights were automatically generated during the meta-analysis and verified manually for accuracy.

Figure 2.

Forest map of total efficiency.

Given the observational nature of the included studies, causal inferences should be made with caution. Although several studies adjusted for major confounding variables (e.g. socioeconomic status [SES], comorbid conditions and access to green space), the extent and rigour of adjustment varied. For studies that reported adjusted effect estimates (e.g., adjusted OR or RR), those values were prioritised for pooling. In the absence of adjusted estimates, unadjusted values were used, which may introduce residual confounding due to unaccounted variables. We assessed whether adjustments for important confounders were applied by reviewing each study’s methodology and reporting. We also evaluated the quality of evidence using the grading of recommendations, assessment, development and evaluation (GRADE) approach. This assessment considers factors such as risk of bias, consistency of results, directness of evidence, precision of estimates and potential for publication bias. Based on GRADE, the overall quality of evidence for the association between environmental noise exposure and mental health outcomes was rated as moderate, reflecting the generally strong methodological quality of included studies as well as the inherent limitations associated with observational research designs.

RESULTS

Literature Search and Selection

The literature search and study selection process adhered strictly to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A comprehensive search was conducted across multiple databases, including PubMed, the Cochrane Library, Medline, CNKI and Embase, encompassing publications up to February 2025. Although search terms such as ‘elderly’ and ‘middle-aged’ were included to maximise sensitivity, the inclusion criterion was restricted to studies reporting data for adults aged 35 years or older. The initial search yielded 3216 records. After removing 1034 duplicates, 2182 unique articles remained for title and abstract screening. Two independent reviewers conducted the screening process, excluding 1678 articles that clearly did not meet the predefined eligibility criteria.

Subsequently, 504 full-text articles were retrieved and assessed for eligibility. Of these, 493 were excluded for the following reasons: the study population did not meet the inclusion criteria (e.g., participants under 35 years of age; n = 196); absence of ambient noise exposure assessment (n = 154); and inadequate characterization of noise exposure (e.g., vague or missing noise metrics; n = 143). Ultimately, 11 studies fulfilled all predefined inclusion criteria and were included in both the qualitative and quantitative syntheses [Figure 1]. All screening and review procedures were conducted independently by two researchers, with discrepancies resolved by a third reviewer to ensure objectivity and methodological rigour.

Figure 1.

Flow diagram of study inclusion procedure.

Study Characteristics

A total of 11 studies met the predefined inclusion criteria and were incorporated into the meta-analysis. Most employed observational designs, including eight cross-sectional and three cohort studies. No randomised controlled trials or case–control studies were included, as these designs are generally less suited for evaluating chronic environmental exposures. The studies spanned diverse regions, including Europe (Netherlands, Germany, Norway and Italy); Asia (South Korea and Japan) and Oceania (Australia), drawing participants from population-based cohorts, health insurance records and municipal survey systems. Sample sizes ranged from 59 (Sygna et al.[16] to 8865 (Seidler et al.[15], yielding a total pooled sample of 15,525 adults aged 35 years or older, of whom 2205 (14.2%) exhibited depression or anxiety symptoms or both.

Several studies specifically focused on adults aged ≥40 years (e.g., Orban et al.,[11] Seidler et al.[15] and Nitschke et al.[14], whereas others, such as Leijssen et al.[9] and Wright et al.,[10] included general adult populations (≥18 years) but allowed for age- and exposure-based subgroup extraction. Similarly, Hammersen et al.[12] and Yoon et al.[20] included participants across a wide age range but reported stratified results, enabling consistent inclusion of the target age group (≥35 years) across all studies.

Environmental noise exposure was assessed through various methods, including predictive modelling based on residential address data (e.g., Lden, Lnight; Orban et al.,[11] Nitschke et al.[14] and Seidler et al.[15], self-reported noise annoyance via structured questionnaires (Hammersen et al.,[12] Wright et al.[10] and Yoon et al.[20] or hybrid approaches that combined objective and subjective measures (e.g., Zock et al.[19]. Road traffic noise was the most frequently studied source, examined in 9 of the 11 studies, followed by aircraft noise (e.g., Hardoy et al.,[13] Wright et al.[10] and Seidler et al.[15] and mixed-source or community-level noise (e.g., Hammersen et al.[12] and Zock et al.[19].

Mental health outcomes were measured using a variety of validated instruments, including the PHQ-9 (Leijssen et al.[9], CES-D (Orban et al.[11], SF-36 or Mental Health Inventory-5 [MHI-5] (Nitschke et al.,[14] Yamazaki et al.[17] and Hammersen et al.[12] and clinical interview schedule [CIDIS] interviews based on diagnostic and statistical manual of mental disorders fourth edition (DSM-IV) criteria (Hardoy et al.[13]. Seidler et al.[15] used ICD-10 diagnostic codes, while other studies reported broader psychological distress measures, including the hopkins symptoms check list - 25 (HSCL-25) (Sygna et al.[16], self-reported chronic mental illness (Wright et al.[10] and suicidal ideation (Yoon et al.[20].

Although Seidler et al.[15] included over 578,000 insured individuals, only the subgroup with confirmed depression diagnoses and exposure to road traffic noise ≥70 dB (n = 8865) was included in the meta-analysis. Likewise, Wright et al.[10] contributed self-reported mental health data from a large census-based sample, which, despite the absence of validated tools, was retained because of the study’s scale and relevance. However, the lack of validation in the study of Wright et al.[10] should be considered in the interpretation of results. Despite some heterogeneity in design and measurement tools, the included studies collectively offer a robust and diverse empirical foundation for examining how long-term environmental noise exposure affects the mental health of adults aged 35 years or older.

(Notes: ‘Sample Size [Total]’ refers to the total population reported in each original study, while ‘Sample Size [Meta]’ reflects the effective subset used in the current meta-analysis based on inclusion criteria (e.g., age ≥ 40, high noise exposure or symptom-defined subgroups). For example, studies of Wright et al.[10] and Seidler et al.[15] provided large administrative or census-based datasets, but only a fraction matched the eligibility for meta-analysis. Subgroup selection details are provided in the Methods section. ‘High noise exposure’ thresholds vary by study, typically defined as ≥55 dB or ≥57 dB, based on modelled annual average or participant-reported annoyance levels. Some studies began at age 35, which was treated as consistent with the predefined inclusion threshold.). ‘Population’ reflects the age range or subgroup reported in the original study. Studies marked with ‘*’ indicate that only a stratified subset of adults aged 35 years or older was extracted for this meta-analysis (e.g., high noise exposure, poor sleep quality or defined age strata). CES-D indicates Center for Epidemiologic Studies Depression Scale; CIDIS, Clinical Interview Schedule; MHI-5, Mental Health Inventory-5; PHQ-9, Patient Health Questionnaire-9;SF-36,Short Form-36 [Table 1].

Table 1.

Basic features included in the study

Study	Country	Noise type	Population	Sample Size (adults aged 35 years or older)	Mental Health Measure	Study year	Study design	GRADE quality rating	Comments
Leijssen et al.[9]	Netherlands	Road traffic noise	18–70 years (subgroup: ≥70 dB)*	279	PHQ-9 (cutoff ≥10)	2011–2014	Cross-sectional	Moderate	High sample size but cross-sectional design limits causal inference.
Wright et al.[10]	United Kingdom	Aircraft noise	≥18 years (stratified subgroup)*	2939	Self-reported chronic mental health (Census)	2011	Cross-sectional (Census-based)	Moderate	Self-reported data, potential bias due to non-validated measures.
Orban et al.[11]	Germany	Road traffic noise	45–75 years	1179	CES-D (cutoff ≥17) + antidepressant use	2008–2011	Cohort	High	Cohort design with follow-up, minimal bias risk.
Hammersen et al.[12]	Germany	Road, neighbour, aircraft noise	18–99 years (high annoyance group ≥35 years)*	1054	MHI-5 (cutoff ≤52)	2012	Cross-sectional	Moderate	Self-reported annoyance; high noise exposure, but cross-sectional.
Hardoy et al. [13]	Italy	Aircraft noise	≥45 years	71	CIDIS (DSM-IV structured interview)	2003	Cross-sectional	Low	Small sample size, non-validated clinical interview.
Nitschke et al. [14]	Australia	Road traffic noise	≥45 years (high annoyance group)*	63	SF-36 Mental Health (QOL dimension)	2000–2003	Cross-sectional	Moderate	Small sample, cross-sectional, but robust methodology.
Seidler et al.[15]	Germany	Aircraft, road, railway noise	≥40 years (stratified ≥35 years)*	8865	ICD-10 physician-diagnosed depression (insurance record)	2005–2010	Retrospective cohort	High	Large sample, validated diagnosis, minimal bias risk.
Sygna et al. [16]	Norway	Road traffic noise	18–77 years (poor sleepers ≥45 years)*	59	HSCL-25 (cutoff ≥1.55)	2011–2012	Cross-sectional	Moderate	Small sample, limited adjustment for confounders.
Yamazaki et al.[17]	Japan	Road traffic noise (proximity proxy)	≥20 years (proximity subgroup ≥35 years)*	193	SF-36 Mental Health domain (score <30)	2002	Cross-sectional	Low	Small sample, wide confidence intervals, imprecise exposure.
Zock et al. [19]	Netherlands	Road and railway noise	≥18 years (subgroup: ≥65 dB)*	534	GP-recorded depression (ICPC-code, EHR based)	2012–2014	Cross-sectional	High	Large sample, validated diagnosis, robust methodology.
Yoon et al.[20]	South Korea	Occupational noise annoyance	≥19 years (male subgroup ≥40 years)*	289	Self-reported depressive symptoms or suicidal ideation	2007–2009	Cross-sectional (KNHANES)	Moderate	Self-reported depression and suicide ideation, potential bias.

Quality Evaluation of Included Studies

The assessment of evidence quality using the NOS revealed a range of methodological rigour across the included studies. Most studies in this meta-analysis were rated as moderate quality, reflecting the inherent limitations of observational research, including susceptibility to confounding and potential bias. Nevertheless, several studies received high-quality ratings, particularly those employing cohort designs with comprehensive confounder adjustment and validated outcome measures. Table 2 summarises the NOS ratings for each study, detailing scores across the three core domains, namely, selection, comparability and outcome, and the overall quality classification. These evaluations are critical for contextualising the strength and reliability of the meta-analytic findings. The predominance of moderate-quality studies underscores common limitations in environmental health research, such as residual confounding due to unmeasured variables. Despite these challenges, the meta-analysis identified a significant association between environmental noise exposure and adverse mental health outcomes, reinforcing the need for future longitudinal research utilising more robust exposure assessments and standardised clinical evaluations.

Table 2.

Newcastle–Ottawa Scale (NOS) quality assessment of the included studies

Study	Selection criteria (Max 4 points)	Comparability (Max 2 points)	Outcome (Max 3 points)	NOS quality rating
Leijssen et al.[9]	3	2	2	7/9
Wright et al.[10]	3	1	2	6/9
Orban et al.[11]	4	2	3	9/9
Hammersen et al. [12]	3	2	1	6/9
Hardoy et al.[13]	2	1	1	4/9
Nitschke et al. [14]	3	1	2	6/9
Seidler et al. [15]	4	2	3	9/9
Sygna et al. [16]	3	1	1	5/9
Yamazaki et al. [17]	2	1	1	4/9
Zock et al. [19]	4	2	3	9/9
Yoon et al.[20]	3	1	2	6/9

Comprehensive Assessment of the Impact of Environmental Noise Exposure on Mental Health in Adults Aged 35 Years or Older

Meta-analysis Results

The meta-analysis, employing a random-effects model, synthesised data from 11 eligible studies to estimate the pooled RR for depression and anxiety symptoms associated with environmental noise exposure among adults aged 35 years or older. In six studies, ORs were converted to RRs using the method of Zhang and Zhou[6] because baseline incidence data (P₀) were available. When conversion was infeasible, the original ORs were retained. The pooled analysis demonstrated a significant association between long-term environmental noise exposure and an increased risk of depression and anxiety symptoms (pooled RR = 1.29, 95% CI: 1.16–1.51; P < 0.05). Substantial heterogeneity was observed (I² = 68.1%), likely reflecting differences in geographic context, noise exposure assessment methods and mental health outcome measures. Nonetheless, the application of a random-effects model was appropriate, and heterogeneity remained within acceptable limits. Sensitivity analyses affirmed the robustness of the results as follows: sequential exclusion of individual studies yielded recalculated RRs ranging from 1.27 to 1.35, indicating that no single study disproportionately influenced the overall estimate. Importantly, even after excluding the most heavily weighted study (Seidler et al., which contributed 60.0% of the weight), the pooled RR remained statistically significant and within the original CI.

The meta-analysis was conducted using a random-effects model to account for inter-study variability, synthesising data from 11 eligible studies to estimate the pooled RR for depression and anxiety symptoms associated with environmental noise exposure in adults aged 35 years or older. In six studies, ORs were converted to RRs using the Zhang[6] method because baseline incidence (P_₀) data were available. For studies where conversion was not feasible, such as the study of Leijssen et al.,[9] the original ORs were retained. Adjusted estimates were used wherever available to control for potential confounding factors, including socioeconomic status and comorbidities. The pooled analysis indicated a significant association between long-term environmental noise exposure and increased risk of depression and anxiety symptoms (pooled RR = 1.29, 95% CI: 1.16–1.51, P < 0.05). This finding suggests that adults aged 35 and older who are chronically exposed to environmental noise are at elevated risk of psychological distress.

Substantial heterogeneity was observed (I² = 68.1%), reflecting variation across studies in geographic context, exposure modelling and mental health outcome measures. Nevertheless, the use of a random-effects model was appropriate, and heterogeneity remained within acceptable bounds. Sensitivity analyses confirmed the robustness of the findings as follows: exclusion of any single study did not significantly alter the overall effect estimate, with recalculated RRs ranging from 1.27 to 1.35. Notably, the exclusion of the most heavily weighted study (Seidler et al.[15] contributing 60.0% of the weight) did not affect the statistical significance of the pooled result, which remained within the original CI.

Subgroup analyses further revealed that studies employing objective noise exposure assessments (e.g., modelled decibel levels) demonstrated a slightly stronger association (RR = 1.35) than those relying on subjective assessments (RR = 1.24). Likewise, cross-sectional studies reported a marginally higher pooled RR (1.34) compared to cohort studies (1.25), suggesting that study design and exposure assessment methods may influence the magnitude of the observed association. Meta-regression analyses identified noise type (road traffic vs. aircraft) and study design (cross-sectional vs. cohort) as significant moderators (P = 0.025 and 0.015, respectively), and stronger associations were observed in studies examining road traffic noise and those utilising a cohort design.

These findings support the conclusion that environmental noise exposure represents a meaningful and modifiable risk factor for mental health outcomes in adults aged 35 years or older. The consistent positive associations across diverse populations and exposure definitions underscore the need for public health efforts for mitigating chronic noise exposure in ageing societies. To evaluate publication bias, we performed Egger’s test for asymmetry. The results showed significant asymmetry (P = 2.83 ×  10 ⁻⁵), indicating potential for publication bias. Additionally, Begg’s test showed a Kendall’s Tau of 0.5636 (P = 0.0165), which also suggests the presence of publication bias. This adjustment indicates that publication bias, particularly the omission of small studies with null results, inflated the original effect size. The adjusted RR suggests a more accurate estimate of the true effect, accounting for the missing studies. These findings support the conclusion that environmental noise exposure represents a meaningful and modifiable risk factor for mental health outcomes in adults aged 35 years or older. The consistent positive associations across diverse populations and exposure definitions underscore the need for public health efforts for mitigating chronic noise exposure in ageing societies.

Exploratory Subgroup Synthesis

To further investigate potential sources of heterogeneity, we conducted a descriptive synthesis of study characteristics based on noise type (e.g., road traffic and aircraft), exposure assessment method (e.g., modelled decibel levels vs. self-reported annoyance) and mental health outcome measures (e.g., PHQ-9 and CES-D). These variables were used to structure the subgroup analyses and examine their influence on the observed associations. Among the 11 studies included, road traffic noise was the most frequently examined exposure source (n = 8; e.g., Orban et al.,[11] Nitschke et al.[14] and Zock et al.[19], followed by aircraft noise (n = 2; Hardoy et al.[13] and Seidler et al.[15] and mixed-source, community-level, or occupational noise (n = 3; Yoon et al.,[20] Hammersen et al.[12] and Zock et al.[19].

Studies employing modelled noise metrics (e.g. Lden and Lnight) generally reported higher effect sizes than those relying on self-reported annoyance or subjective noise perception. Regarding outcome measurement, depression-related symptoms were evaluated in 10 of the 11 studies using instruments such as the PHQ-9, CES-D, SF-36, MHI-5 and CIDIS. Anxiety-specific outcomes were directly assessed in three studies (Hardoy et al.,[13] Yoon et al.[20] and Sygna et al.[16]. Among the included studies, the study of Yamazaki et al.[17] demonstrated the most substantial deviation from the pooled risk estimate, with a reported RR of 2.05 (95% CI: 0.41–8.02). This wide CI and elevated point estimate likely reflect its small sample size (n = 193) and the use of a proximity-based exposure proxy (i.e., residential distance to arterial roads), which may have introduced exposure misclassification. However, sensitivity analyses indicated that excluding this outlier did not materially affect the overall pooled estimate, confirming the robustness of the meta-analytic findings.

Most studies employed validated mental health instruments and demonstrated consistent positive associations across different symptom domains. While the limited number of studies per subgroup (generally ≤3) precluded formal statistical comparisons, the observed associations remained directionally consistent across all subgroups. These findings suggest that the relationship between long-term environmental noise exposure and mental health symptoms in adults aged 35 and older is robust across noise sources, exposure assessment methods and outcome measures. This descriptive synthesis complements the quantitative meta-analysis by contextualising methodological variability and providing a foundation for more targeted subgroup analyses in future studies as the evidence base continues to grow.

DISCUSSION

This study provides a comprehensive evaluation of the association between environmental noise exposure and the risk of depression and anxiety symptoms among adults aged 35 years or older. The quality of the evidence was assessed using the GRADE framework, with the overall rating deemed moderate primarily because of the observational nature of the included studies. The meta-analysis, which synthesised findings from 11 eligible studies, demonstrated a significant association between chronic environmental noise exposure and increased psychological distress in this population, with a pooled RR of 1.29 (95% CI: 1.16–1.51). The level of heterogeneity across studies was moderate (I² = 68.1%), yet the results were consistent, suggesting a robust and stable relationship irrespective of variations in geographic region, exposure assessment method, mental health instruments or study design.

The inclusion of both cross-sectional and cohort studies introduced methodological heterogeneity, potentially affecting the precision of the pooled estimates. Cohort studies offer temporal insights, allowing for a clearer understanding of causality between noise exposure and mental health outcomes. In contrast, cross-sectional studies provide only a snapshot in time, limiting causal inference and increasing susceptibility to confounding. This distinction is crucial in interpreting the strength and directionality of observed associations.

A variety of instruments were used to assess mental health outcomes, including the PHQ-9, CES-D and ICD-10 diagnoses. While all are widely recognised and validated, they differ in sensitivity and specificity, which may have contributed to some variability in effect estimates. To address this, subgroup analyses were conducted, restricting the analysis to studies using clinical diagnoses (e.g., ICD-10) or validated cutoffs (e.g., PHQ-9 ≥ 10). The results remained consistent across these subgroups, supporting the robustness of the association regardless of diagnostic method. Nonetheless, variability introduced by differing measurement tools should be acknowledged in interpreting the overall findings.

Although the random-effects model used in this meta-analysis accounts for between-study variation, it cannot fully adjust for differences in study design. The lack of temporality in cross-sectional studies, for example, increases susceptibility to unmeasured confounding. Factors such as SES, pre-existing mental health conditions and access to healthcare are known to influence mental health but were not consistently controlled for across the included studies. This inconsistency raises the possibility of residual confounding, which may have biased the pooled estimates, either inflating or attenuating the observed association.

The observed pooled RR of 1.29 indicates a moderate association between environmental noise exposure and adverse mental health outcomes; however, this estimate should be interpreted cautiously in light of potential unmeasured confounders. Future research should prioritise longitudinal designs that can better isolate the temporal dynamics of exposure and outcome, while rigorously controlling for confounding factors such as SES and healthcare access. Such studies will be instrumental in clarifying whether environmental noise exposure represents an independent and modifiable risk factor for mental health symptoms in ageing populations.

Environmental noise exposure affects mental health through a complex interplay of SES, residential environment and individual susceptibility.[20,21] Populations with lower income levels are disproportionately exposed to high noise levels, often residing in densely populated urban areas near major traffic corridors and in housing with inadequate sound insulation. This chronic exposure, combined with financial stress and limited access to mental healthcare, amplifies psychological distress. Poor housing conditions further exacerbate the effects of noise by increasing the likelihood of sleep disturbances and physiological stress responses.

Policy interventions are essential for mitigating these adverse outcomes. Measures, such as improving housing quality, expanding access to urban green spaces and enforcing stricter noise control regulations have demonstrated effectiveness.[18,22,23] The EU Noise Directive offers a comprehensive framework for assessing and managing environmental noise across member states. It sets exposure limits and encourages national implementation of measures to reduce harmful noise levels. Evaluations of the directive have shown that countries adopting active noise reduction strategies, such as enhanced urban planning, road traffic noise control and stringent enforcement, have achieved measurable reductions in noise-related health risks.

The residential environment plays a critical role in shaping individuals’ experience of noise pollution. The presence of green spaces and noise-buffering infrastructure is associated with lower levels of anxiety and depression, whereas high-noise urban zones are linked to increased psychological distress. Urban development strategies that incorporate green buffers, sound barriers and traffic-calming measures may significantly reduce the adverse mental health effects of environmental noise.[24]

Individual susceptibility also influences the degree of psychological impact. People with heightened noise sensitivity, particularly those with pre-existing mental health conditions, such as depression, anxiety or sleep disorders, are more vulnerable to noise-induced distress. Adults aged 35 years or older, as well as those with cognitive impairments, may experience greater difficulty adapting to chronic noise exposure.[25] Interventions targeting these high-risk groups, such as enhanced indoor noise insulation, stress reduction programs and access to mental health support, can help buffer the psychological burden of environmental noise.

This study highlights the urgent need for integrated noise mitigation strategies that combine urban planning, public health policy and targeted individual interventions. Future research should prioritise longitudinal studies and investigate the biological mechanisms underlying noise-induced psychological distress to inform more effective and equitable policy responses.

The underlying mechanisms through which environmental noise influences mental health remain an area of active investigation. Several biological pathways have been proposed, most notably the chronic activation of the hypothalamic–pituitary–adrenal (HPA) axis, increased cortisol secretion and heightened sympathetic nervous system activity, all of which may lead to sustained stress responses and impaired emotional regulation.[26,27,28] While dysregulation of the HPA axis has been frequently cited as a key mechanism linking environmental noise exposure to adverse mental health outcomes, none of the studies included in this meta-analysis directly measured physiological biomarkers such as cortisol levels or markers of neuroinflammation.

These proposed biological mechanisms remain hypothetical and require further empirical validation. Future research incorporating biomarker assessments and neuroimaging is needed to establish a causal relationship between noise exposure and the observed psychological effects. In addition to HPA axis involvement, prolonged exposure to environmental noise has also been associated with neuroinflammatory processes, oxidative stress and disruptions in neurotransmitter function, all of which may contribute to the onset or exacerbation of mood disorders such as depression and anxiety.

Such physiological responses are particularly concerning in adults aged 35 years or older, who may be more vulnerable because of age-related declines in cognitive resilience and stress adaptability.

Several studies included in this meta-analysis, such as those by Hardoy et al.[13] and Zock et al.,[19] utilised physician-diagnosed depression or electronic health record data as proxies for mental health outcomes rather than psychometric scale scores. While these registry-based indicators provide valuable population-level data, they are potentially influenced by extraneous factors such as healthcare accessibility, diagnostic variability and patients’ help-seeking behaviours. Consequently, such indicators may not capture subclinical symptoms or the full spectrum of psychological distress severity, potentially leading to misclassification or underestimation of mental health outcomes.

One study of particular note, Yamazaki et al.,[17] reported an unusually high RR (2.05, 95% CI: 0.41–8.02), accompanied by a wide CI. This outlier result likely reflects the study’s small sample size (n = 193) and imprecise exposure assessment, which relied on residential proximity to arterial roads rather than modelled noise levels. Although this level of variability raised concerns regarding potential bias or effect overestimation, sensitivity analyses revealed that the inclusion or exclusion of this study did not materially affect the overall pooled estimate. Therefore, we retained it in the final meta-analysis to maintain data comprehensiveness. Nonetheless, studies with small samples and methodological limitations should be interpreted cautiously, and future meta-analyses should systematically assess the influence of such outliers on pooled outcomes.

While our findings confirm a significant association between environmental noise exposure and increased symptoms of depression and anxiety in adults aged 35 years or older, it is important to note that none of the included studies directly evaluated noise mitigation strategies, such as urban design modifications or home-based noise insulation interventions. This represents a major research gap. Future studies should prioritise interventional research designs to assess the mental health benefits of policy-driven or environmental interventions aimed at reducing noise exposure.

A fundamental limitation of this meta-analysis is the observational nature of the included studies, which inherently limits the ability to draw causal inferences. Although many of the studies adjusted for known confounders, such as SES, pre-existing comorbidities and access to green spaces, residual confounding remains a concern. SES, in particular, is a well-established and complex confounder in environmental health research. While most of the included studies identified SES as a potential confounder, adjustment for SES across studies was inconsistent. This inconsistency may lead to residual confounding, potentially affecting the pooled risk ratio. For example, studies with unadjusted or poorly adjusted SES data may have overestimated the effect of noise exposure on mental health outcomes. In future research, more rigorous and consistent adjustment for SES, along with other potential confounders, is needed to clarify the actual association between environmental noise and mental health.

Future studies employing longitudinal designs with rigorous confounder control are critical to strengthening the evidence base. A key limitation of the studies included in this meta-analysis is that most were cross-sectional, making the establishment of clear causal relationships between environmental noise exposure and psychological symptoms difficult. Cross-sectional studies capture only a snapshot of a relationship at a specific time and cannot assess temporal changes or causality. Although these studies offer valuable insights, the absence of temporal data hampers causal inference. To address this gap, future research should prioritise longitudinal designs that monitor individuals over time, allowing researchers to evaluate whether sustained exposure to environmental noise leads to the development or worsening of depression and anxiety symptoms. Additionally, natural experiments, such as studies conducted before and after the implementation of noise-reduction policies, can provide more robust evidence for causality by assessing the direct impact of environmental interventions on mental health outcomes.

Such approaches would considerably enhance our understanding of whether reducing noise exposure can lead to measurable improvements in psychological well-being, particularly among vulnerable populations. In conclusion, this meta-analysis highlights the substantial impact of environmental noise on depression and anxiety in adults aged 35 years or older. These findings underscore the urgent need for comprehensive noise reduction strategies and mental health interventions. Addressing this public health issue will require a multidisciplinary approach that integrates environmental policy, urban planning and clinical mental health support to effectively protect the well-being of ageing populations.

CONCLUSION

Long-term exposure to environmental noise is significantly associated with an increased risk of depression and anxiety symptoms among adults aged 35 years or older. These findings highlight the critical need for effective noise-reduction strategies, particularly those aimed at protecting vulnerable populations, including older adults and individuals residing in high-exposure areas. Future research should prioritise evaluating the effectiveness of targeted interventions and assessing their impact on mental health outcomes across diverse demographic and geographic populations.

Availability of Data and Materials

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author Contributions

Xinling Hu conceptualized and designed the study, conducted data analysis, and wrote the manuscript.

Ethics Approval and Consent to Participate

This study does not involve ethical approval. Informed consent was obtained from all individual participants included in the study.

Conflict of Interests

The authors declare no conflict of interest.

Funding Statement

None.