Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • View all journals
  • My Account Login
  • Explore content
  • About the journal
  • Publish with us
  • Sign up for alerts
  • Open access
  • Published: 04 July 2022

A systematic review and meta-analysis on the association between ambient air pollution and pulmonary tuberculosis

  • Christian Akem Dimala 1 , 2 &
  • Benjamin Momo Kadia 3 , 4  

Scientific Reports volume  12 , Article number:  11282 ( 2022 ) Cite this article

2578 Accesses

9 Citations

5 Altmetric

Metrics details

  • Environmental sciences
  • Health care
  • Medical research
  • Risk factors

There is inconclusive evidence on the association between ambient air pollution and pulmonary tuberculosis (PTB) incidence, tuberculosis-related hospital admission and mortality. This review aimed to assess the extent to which selected air pollutants are associated to PTB incidence, hospital admissions and mortality. This was a systematic review of studies published in English from January 1st, 1946, through May 31st, 2022, that quantitatively assessed the association between PM 2.5 , PM 10 , NO 2 , SO 2 , CO, O 3 and the incidence of, hospital admission or death from PTB. Medline, Embase, Scopus and The Cochrane Library were searched. Extracted data from eligible studies were analysed using STATA software. Random-effect meta-analysis was used to derive pooled adjusted risk and odds ratios. A total of 24 studies (10 time-series, 5 ecologic, 5 cohort, 2 case–control, 1 case cross-over, 1 cross-sectional) mainly from Asian countries were eligible and involved a total of 437,255 tuberculosis cases. For every 10 μg/m 3 increment in air pollutant concentration, there was a significant association between exposure to PM 2.5 (pooled aRR = 1.12, 95% CI: 1.06–1.19, p < 0.001, N = 6); PM 10 (pooled aRR = 1.06, 95% CI: 1.01–1.12, p = 0.022, N = 8); SO 2 (pooled aRR = 1.08, 95% CI: 1.04–1.12, p < 0.001, N = 9); and the incidence of PTB. There was no association between exposure to CO (pooled aRR = 1.04, 95% CI: 0.98–1.11, p = 0.211, N = 4); NO 2 (pooled aRR = 1.08, 95% CI: 0.99–1.17, p = 0.057, N = 7); O 3 (pooled aRR = 1.00, 95% CI: 0.99–1.02, p = 0.910, N = 6) and the incidence of PTB. There was no association between the investigated air pollutants and mortality or hospital admissions due to PTB. Overall quality of evidence was graded as low (GRADE approach). Exposure to PM 2.5 , PM 10 and SO 2 air pollutants was found to be associated with an increased incidence of PTB, while exposure to CO, NO 2 and O 3 was not. There was no observed association between exposure to these air pollutants and hospital admission or mortality due to PTB. The quality of the evidence generated, however, remains low. Addressing the tuberculosis epidemic by 2030 as per the 4th Sustainable Development Goal may require a more rigorous exploration of this association.

Introduction

Pulmonary tuberculosis (PTB), a bacterial infection of the lungs caused by Mycobacterium tuberculosis is one of the top 10 causes of death worldwide and the leading cause of death from a single infectious agent 1 . PTB remains a global health emergency despite the significant progress that has been made worldwide in its control over the past two and a half decades 2 . Much still needs to be done to end the tuberculosis epidemic by 2030 3 as per the World Health Organisation’s (WHO) 4th sustainable development goal (SDG). This includes addressing important predisposing factors to tuberculosis infection such as smoking, diabetes, human immunodeficiency virus (HIV) and social determinants of health such as poverty, malnutrition, poor ventilation and over-crowding among others 1 , 4 . A multi-faceted and multi-sectorial approach to tuberculosis prevention, case identification, management and control of its health and social determinants is therefore required 4 , 5 .

Air pollution, currently on several global health agendas, has rapidly become a global problem with the increasing global urbanisation, transportation-related emissions, and increased energy consumption. Air pollution could therefore be an important factor to address on the journey to ending tuberculosis as there are growing concerns of its association to increased tuberculosis-related hospital admissions and deaths 6 , 7 .

There is a well-known association between different air pollutants and cardio-respiratory diseases 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 . However, there is still no conclusive evidence of an association between PTB and outdoor air pollution despite its well-known association to indoor pollution from activities such as smoking and biomass fuel burning 16 , 17 , 18 The review by Popovic et al. indicated a possible association between PM 2.5 and PTB outcomes (incidence, hospital admissions and mortality) and reported the contrasting findings from several earlier studies on the association between PM 10 , NO 2 , and SO 2 and PTB 19 , but did not synthesise these findings to determine to what extent these air pollutants are associated to PTB. Also, several studies have been published on this subject after the review by Popovic et al. This review therefore had as objectives to determine if there is an association between the selected air pollutants (PM 2.5 , PM 10 , NO 2 , SO 2 , CO, O 3 ) and PTB incidence, hospital admissions and mortality, and to what extent, by systematically reviewing and quantitatively combining published evidence on this topic.

This was a systematic literature review and meta-analysis of articles published in English from January 1st, 1946, through May 31st, 2022, that quantitatively assessed the association between ambient air pollution and PTB. The study protocol for this review was registered with the international prospective register of systematic reviews (PROSPERO) with trial registration number CRD42020165888 and has been published 20 . This review was reported according to The RepOrting standards for Systematic Evidence Syntheses (ROSES) for systematic review 21 as presented in Additional file 1 .

Deviations from the protocol

There were no deviations from the published study protocol.

Search for articles

A comprehensive search strategy (Additional file 2 ) combining medical subject headings (MeSH) and free-text searches for the appropriate keywords was developed by the authors and used to search the databases: Medline, Embase, Scopus and The Cochrane Library. The keywords ‘air pollution’, ‘carbon monoxide’, ‘nitrogen dioxide’, ‘sulphur dioxide’, ‘ozone’, and ‘particulate matter’ were combined with the keywords ‘tuberculosis’, ‘incidence’, ‘mortality’, ‘hospital admission’ and their respective synonyms, using the Boolean operator ‘AND’ in the search strategy. The search was run by the principal investigator (CAD), all searches were limited to the language English and grey literature search was not conducted given the lack of relevant studies from preliminary searches of the grey literature. Search dates of interest were January 1st, 1946, through May 31st, 2022. The search language was in English, and all the database searches were done on the same day, June 5th, 2022. The search was run twice to ensure replicability of results and the same results were obtained with each search run.

Article screening and study eligibility criteria

Screening process.

Articles returned by the search were saved on Zotero Version 5.0 reference management software and duplicates of the studies were manually removed by the principal investigator (CAD) with the assistance of the reference management software. More articles were added to the search output by the principal investigator by reviewing the reference list of relevant articles. The titles and abstracts of all the remaining articles were then independently screened for eligibility according to the set eligibility criteria by each of the two independent reviewers (CAD and BMK). The full texts of all the articles retained after the title and abstract screen, were then independently reviewed by the same two independent reviewers (CAD and BMK) for eligibility and inclusion to the analysis. The two independent reviewers compared their findings at the end of both the title and abstract screening and the full text review stages of the article selection process to ensure concordance in their final selection. There were no reviewer disagreements at all stages of the study selection process and no third reviewer to settle discordances as had been planned in the study protocol, was therefore needed due to concordance in the findings of the two independent reviewers.

Eligibility criteria

The following criteria were used during the article selection process to determine the eligible studies.

The following studies were included:

Population: Studies focused on adults aged 18 and above with PTB

Exposure: Studies that reported direct measurements on any of the air pollutants; carbon monoxide (CO), nitrogen dioxide (NO 2 ), sulphur dioxide (SO 2 ), ozone (O 3 ), particulate matter ≤ 2.5 µm (PM 2.5 ) and/or particulate matter ≤ 10 µm (PM 10 ) in any country, region, city or locality;

Outcomes: Studies that reported measures of association on the risk of PTB incidence, hospital admission and/or mortality from PTB;

Study design/Other: Cross-sectional, case–control, cohorts, case-crossover, ecological and time-series studies that reported on the association between ambient air pollution and PTB.

The following studies were excluded:

Population: Studies that reported on respiratory diseases other than PTB

Exposure: Studies that reported on other forms of air pollution such as indoor air pollution

Outcomes: Studies that reported outcomes related to PTB in combination with other respiratory diseases. Studies that reported on measures of effect/association other than risk ratios and odds ratios or that provided data from which these measures could not be calculated.

Other: Conference abstracts, editorials, letters, opinion papers, unpublished studies, same studies published in different journals with the same or a different title.

Study validity assessment

Assessment of study quality of each included study was done by both independent reviewers (CAD and BMK) using the respective Study Quality Assessment Tools of the National Health Institute/National Heart, Lung and Blood Institute (NHI/NHLBI) 22 depending on their study designs. There was no discordance in the overall rating of the quality of the eligible studies. Study quality indicators were included in the meta-regression.

The overall quality of the evidence provided by the studies with regards to the primary outcome of interest was assessed and graded as very low, low, moderate or high, using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) 23 .

Data coding and extraction strategy

Data on the publication details, study methods and outcomes of interest were extracted from the eligible studies into a Microsoft excel office 365 data extraction sheet (Additional file 3 ) by the principal investigator (CAD) and independently rechecked by a second reviewer (BMK) for accuracy. The following data were extracted: First author, year of publication, study location, study design, socio-demographic and clinical characteristics of study participants, study duration, number of tuberculosis cases and new tuberculosis cases, annual incidences of tuberculosis, mean and median concentration data on air pollutants of interest (CO, NO 2 , SO 2 , O 3 , PM 2.5 and PM 10 ), data on incidence, hospital admission and mortality from tuberculosis, including measures of effect/association (risk ratios, odds ratios and percentage change in the incidence of PTB) and their respective confidence intervals, and confounders reported by the respective studies and if studies adjusted for confounders or not. PM 2.5 and PM 10 air pollutants were measured in µg/m 3 and NO 2 , SO 2 and O 3 in parts per billion (ppb) and CO in parts per million (ppm). For studies that reported air pollutant concentrations in units other than the above, the Air Pollution Information System 24 , was used to convert air pollutant concentrations to appropriate units, taking into consideration the average yearly temperatures reported for the various cities or countries. The average annual outdoor temperature obtained from public sources was used for studies that did not report them. In studies where several measures of effect were reported for different quintiles or levels of exposure to air pollutants, the largest numerical estimates of the measures of effect were considered, to quantify the maximum extent of the association of air pollutants to PTB. When protective effects were observed among the measures of effect, the lowest numerical measures of effect were used. Default measures of effect reported by the studies were considered. Adjusted measures of effect were chosen over crude measures, and both single-pollutant models and multi-pollutant models were reported as appropriate. All data was transferred to STATA version 14.0 statistical software for analysis.

Potential effect modifiers/reasons for heterogeneity

Between-study heterogeneity was anticipated given the differences in study designs, settings, duration, sample sizes, and population characteristics based on review of existing literature.

Data synthesis and presentation

Meta-analyses were done through random effects models to account for the possibility of between-study heterogeneity. Risk ratios and odds ratios on the incidence of PTB following exposure to the selected air pollutants, and their respective confidence intervals from the various studies, were log-transformed, and the corresponding standard errors derived. Pooled summary estimates for the respective log-transformed measures of association were computed and presented on forest plots. Studies were pooled according to their study designs with ecologic studies and studies that used time-series analysis pooled together, separate from cohort and case–control studies. Heterogeneity between studies was assessed using the Cochrane’s Q test, and the I 2 test statistic was reported as a measure of the extent of this heterogeneity. The Begg’s and Egger’s statistical tests were used for the statistical assessments of publication bias and small study effect 25 , 26 . All statistical tests and plots were done on STATA version 14.0 statistical software.

Ethics approval and consent to participate

This systematic review does not require ethical approval as it entails a synthesis of data collected from several primary studies. No primary data collection from patients will be done for this systematic review.

Review descriptive statistics

Figure  1 summarises the study selection process.

figure 1

PRISMA flow chart.

A total of 12,652 records were returned by the search. Following removal of duplicates, screening of titles and abstracts, addition of studies from the reference list of relevant studies, full-text reviews, 24 eligible studies were retained. Figure  1 summarises the PRISMA flow chart of the study selection process. The studies excluded following full-text review and the reasons for exclusion are presented in Additional file 4 .

Narrative synthesis including study validity assessment

Most studies were from Asian countries and a total of 437,255 tuberculosis cases were reported across the 22 studies that reported the number of tuberculosis cases over their study periods (1996–2019) 7 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 . Of the 24 studies included in the review, 10 were time series, 5 were cohort studies (3 retrospective, 2 prospective), 5 were ecologic, 2 were case–control studies (1 nested, 1 retrospective), 1 was a retrospective case cross-over and 1 was cross-sectional. Average male participation was at 64.9% (N = 13 studies) 7 , 27 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 37 , 40 , 43 , 46 , mean age of 46.3 years (N = 7 studies) 27 , 30 , 31 , 32 , 37 , 40 , 43 and average annual tuberculosis incidence was 45.3 per 100,000 population (N = 10 studies) 7 , 27 , 29 , 32 , 35 , 36 , 45 , 46 , 47 , 48 . Study and participant characteristics are summarised on Table 1 .

The average of the annual mean concentrations of the various air pollutants are presented on Table 2 .

Twelve studies were of good quality, eleven of fair quality and one of poor quality (Additional file 6 ). The overall quality of evidence for the association of all 6 air pollutants to the incidence of PTB was graded as low based on the study limitations affecting generalisability of the findings, and some inconsistency across the studies due to the significantly elevated between-study heterogeneity (Additional file 7 ).

Data synthesis

Association between air pollutants and pulmonary tuberculosis incidence.

There was a significant association between exposure to PM 2.5 and incidence of pulmonary tuberculosis (PTB), pooled adjusted RR = 1.12 (95% CI: 1.06–1.19), p < 0.001, N = 6, I 2  = 72.4% 7 , 29 , 38 , 39 , 43 , 49 . There was no evidence of publication bias (Begg’s test, p = 0.133 and Egger’s test, p = 0.203). Begg’s test, p = 1. Likewise, Xiong et al. 46 reported an association (RR = 3.10, 95% CI: 1.10–8.79) for a 50 µg/m 3 increase in PM 2.5 concentration. The study by Lai et al. 32 (RR = 1.39, 95% CI: 0.95–2.03) which was cohort in design did not find a significant association. Jassal et al. 28 reported an odds ratio of 25.3 (95% CI: 3.38–29.1).

There was a significant association between exposure to PM 10 and incidence of PTB, pooled adjusted RR = 1.06 (95% CI: 1.01–1.12), p = 0.022, N = 8, I 2  = 97.6% (Begg’s test, p = 0.536 and Egger’s test, p = 0.204) 7 , 29 , 35 , 39 , 40 , 43 , 44 , 49 . The studies by Lai et al. 32 (HR = 0.95, 95% CI: 0.78–1.17) and Hwang et al. 27 (male RR = 1.00, 95% CI: 0.96–1.05 and female RR = 1.01, 95% CI: 0.98–1.05) did not find a significant association. Likewise, the pooled adjusted OR was 1.03 (95% CI: 1.01–1.04), p = 0.001, N = 3, I 2  = 0% (Begg’s test, p = 1 and Egger’s test, p = 0.211) (Fig.  2 ) 31 , 34 , 37 .

figure 2

Forest plot showing the individual and pooled risk ratios and odds ratios for pulmonary tuberculosis incidence for PM 2.5 and PM 10 . The dashed line on the Forest plot represents the overall pooled estimate. The grey squares and horizontal lines represent the vaccine acceptance rate of each study and their 95% confidence intervals. The size of the grey square represents the weight contributed by each study in the meta-analysis. The diamond represents the pooled vaccine acceptance rate and its 95% confidence intervals.

There was no significant association between exposure to CO and the incidence of PTB, pooled adjusted RR = 1.04 (95% CI: 0.98–1.11), p = 0.211, N = 4, I 2  = 87.4% (Begg’s test, p = 0.734 and Egger’s test, p = 0.355) 39 , 43 , 47 , 49 . The studies by Lai et al. 32 (HR = 1.89, 95% CI: 0.78–4.58) and Hwang et al. 27 (male RR = 0.99, 95% CI: 0.95–1.03 and female RR = 1.01, 95% CI: 0.98–1.04) had similar findings. The pooled adjusted OR was 1.22 (95% CI: 0.84–1.76), p = 0.305, N = 3, I 2  = 78.5% (Begg’s test, p = 1 and Egger’s test, p = 0.364) (Fig.  3 ) 31 , 34 , 37 . However, Xiong et al. 46 (RR = 1.436, 95% CI: 1.004–2.053) reported a significant association for a 100 µg/m 3 increase in CO concentration.

figure 3

Forest plot showing the individual and pooled risk ratios and odds ratios for pulmonary tuberculosis incidence for CO and NO 2 . The dashed line on the Forest plot represents the overall pooled estimate. The grey squares and horizontal lines represent the odds ratios of each study and their 95% confidence intervals. The size of the grey square represents the weight contributed by each study in the meta-analysis. The diamond represents the pooled odds ratio and its 95% confidence intervals.

There was no association between exposure to NO 2 and the incidence of PTB, pooled adjusted RR = 1.08 (95% CI: 0.99–1.17), p = 0.057, N = 7, I 2  = 98% (Begg’s test, p = 1 and Egger’s test, p = 0.437) (Fig.  3 ) 7 , 35 , 39 , 40 , 43 , 48 , 49 . Lai et al. 32 (HR = 1.33, 95% CI: 1.04–1.70) found a significant association, while Hwang et al. 27 (male RR = 1.00, 95% CI: 0.96–1.05 and female RR = 1.01, 95% CI: 0.98–1.05) did not. Likewise, the pooled adjusted OR was 1.05 (95% CI: 0.95–1.17), p = 0.322, N = 3, I 2  = 72.4% (Begg’s test, p = 0.296 and Egger’s test, p = 0.145) (Fig.  3 ) 31 , 34 , 37 . However, Xiong et al. 46 (RR = 1.8, 95% CI: 1.11–2.91) reported a significant association for a 5 µg/m 3 increase in NO 2 concentration.

There was an association between exposure to SO 2 and incidence of PTB, pooled adjusted RR = 1.08 (95% CI: 1.04–1.12), p < 0.001, N = 9, I 2  = 94.4% (Begg’s test, p = 0.517 and Egger’s test, p = 0.356) (Fig.  4 ) 7 , 35 , 39 , 40 , 43 , 44 , 47 , 48 , 49 . Hwang et al. 27 (male RR = 1.07, 95% CI: 1.03–1.12 and female RR = 1.02, 95% CI: 0.98–1.07) reported similar findings in males. Likewise, Xiong et al. 46 reported an association (RR = 1.62, 95% CI: 1.12–2.33) for a 5 µg/m 3 increase in SO 2 concentration.

figure 4

Forest plot showing the individual and pooled risk ratios and odds ratios for pulmonary tuberculosis incidence for SO 2 and O 3 . The dashed line on the Forest plot represents the overall pooled estimate. The grey squares and horizontal lines represent the odds ratios of each study and their 95% confidence intervals. The size of the grey square represents the weight contributed by each study in the meta-analysis. The diamond represents the pooled odds ratio and its 95% confidence intervals.

There was no significant association between O 3 exposure and incidence of PTB, pooled adjusted RR = 1.01 (95% CI: 0.97–1.06), p = 0.560, N = 4, I 2  = 75.6% (Begg’s test, p = 0.734 and Egger’s test, p = 0.734) (Fig.  4 ) 39 , 43 , 47 , 49 . While Hwang et al. 27 had similar findings (male RR = 0.99, 95% CI: 0.94–1.03 and female RR = 1.01, 95% CI: 0.97–1.05), Lai et al. 32 rather found a protective effect (HR = 0.69, 95% CI: 0.49–0.98). Xiong et al. 46 reported an association (RR = 0.96, 95% CI: 0.93–1.0) for a 5 µg/m 3 increase in O 3 concentration.

Table 3 summarises the percentage change in the number of PTB cases for the respective changes in air pollutant concentrations.

Association between air pollutants and hospital admissions and mortality due to pulmonary tuberculosis

Two studies reported a significant association between PM 2.5 and PTB mortality; OR = 1.46 (95% CI: 1.15–1.85) 33 and percentage change in cases of 0.08% (95% CI: 0.06–0.09) 45 . There was no significant association between CO, SO 2 , and O 3 and PTB mortality 47 (Table 4 ). Likewise, there was no significant association between PM 10 , CO, SO 2 , O 3 and hospital admission 30 , 47 . NO 2 was associated with hospital admission due to PTB, OR: 1.21 (95% CI: 1.10–1.33) (Table 4 ).

Subgroup analysis and meta-regression

Studies were categorised according to their duration (less than 5 years and 5 years or more), location (Asia and others), number of PTB cases (less than 5000 and 5000 or more) and study quality (good and fair/poor). None of these study characteristics could explain the observed heterogeneity across studies, except for study location with regards to exposure to PM 2.5 air pollutant. There was a higher risk of PTB incidence with PM 2.5 exposure in studies conducted out of Asia (Additional file 5 ).

Exposure to PM 2.5 , PM 10 and SO 2 air pollutants was found to be associated with an increased incidence of PTB, while exposure to CO, NO 2 and O 3 was not. There was no observed association between exposure to these air pollutants and hospital admission or mortality due to PTB. The findings of this review are particularly relevant given the increasing global concentrations and exposure to some air pollutants such as SO 2 and PM 2.5 over the past decades 50 , 51 . Public health strategies aimed at ending the tuberculosis epidemic would therefore have to work alongside interventions aimed at improving overall air quality and addressing air pollution 51 .

Air pollutants including O 3 and NO 2 mainly originate from volatile organic compounds, combustion processes including heating, power generation, the engines of vehicles and ships and also from industry emissions 52 . SO 2 originates from the burning of fossil fuels for power generation and the smelting of sulfur-containing mineral ores 52 . PM 2.5 and PM 10 which consist of particles of organic and inorganic substances are typically suspended in the air 52 . Air pollutants have been previously associated with the development of cardio-respiratory diseases in both children and adults 8 , 9 , 11 . Traffic-related pollution and several air pollutants such as O 3 , NO 2 , PM 2.5 and PM 10 , have not only been associated with exacerbations of asthma and chronic obstructive pulmonary disease, but have also been implicated in the development of these conditions especially in childhood 11 , 53 , 54 . Air pollutants are known to increase the risk of infection when inhaled as they dampen the natural defence barriers of the respiratory tract, inhibit muco-ciliary clearance, inhibit macrophages and initiate a chronic inflammatory response with the release of pro-inflammatory mediators 55 , 56 . In a similar way, exposure to particulate matter for example has immunomodulatory effects on antimycobacterial activity through impaired expression of important cytokines and chemokines which are important in controlling mycobacterial infections 57 , 58 . This reduced antimycobacterial host immune response predisposes to tuberculosis infection.

Measures and policies in various sectors such as the transport, housing and industry sectors are known to reduce air pollutions, including; prioritising walking and cycling in cities, using low-emission vehicles; using clean technologies that reduce industrial emissions; improving access to clean household energy for heating, lighting and cooking; making cities more green; using low-emission fuels and combustion-free power sources, among others 52 . In 2015, the WHO member states adopted a resolution for enhanced global response to the adverse health effects of air pollution, and the WHO has been overseeing this response through; the production of air quality guidelines and exposure limits to these air pollutants 52 .

Even though the studies by Ge et al. 59 and Xu et al. 60 reported a possible association between short-term exposure to SO 2 , our review did not assess outpatient PTB visits as an outcome. This is therefore a subject amenable to further exploration.

The studies in this review were conducted over a 24-year period and we did not observe a particular change or variation in the trend of the reported associations between exposure to the air pollutants and PTB incidence over time across the older and newer studies. Close to four fifth of the studies in our review were conducted in Asia and up to half of the studies were in China, which could affect the generalisability of the findings of this review, however, China is still a high-burden country for tuberculosis 61 , 62 . The observed between-study heterogeneity highlights the need for more uniform study designs and methods in future studies aiming to assess this association.

Interpretation of the findings from this review should take into consideration some limitations. This review did not assess the contribution of indoor air pollution and other comorbidities to the increased risk of PTB incidence, hospital admission and mortality. The different study designs and methodologies affected the types of confounders that could be adjusted for in the different studies and therefore introducing inconsistency in the adjustment of confounders across studies. This review, therefore, focused on the strongest reported associations between air pollutant exposure and PTB incidence rather than on the duration of exposure to the air pollutants.

Exposure to PM 2.5 , PM 10 , NO 2 and SO 2 air pollutants was found to be associated with an increased incidence of PTB, while exposure to CO and O 3 was not. These findings of this study and the overall quality of the evidence highlight the need for more rigorous exploration of this association.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

World Health Organization. Global Tuberculosis Report 2019 (World Health Organization, 2019).

Google Scholar  

GBD Tuberculosis Collaborators. Global, regional, and national burden of tuberculosis, 1990–2016: Results from the global burden of diseases, injuries, and risk factors 2016 study. Lancet Infect Dis. 18 (12), 1329–1349 (2018).

Article   Google Scholar  

United Nations. Sustainable Development Goals. Goal 3: Ensure healthy lives and promote well-being for all at all ages [Internet]. https://www.un.org/sustainabledevelopment/health/ [cited 2020 Jan 17].

Hargreaves, J. R. et al. The social determinants of tuberculosis: From evidence to action. Am. J. Public Health 101 (4), 654–662 (2011).

Article   PubMed   PubMed Central   Google Scholar  

Grobusch, M. P. & Kapata, N. Global burden of tuberculosis: Where we are and what to do. Lancet Infect. Dis. 18 (12), 1291–1293 (2018).

Article   PubMed   Google Scholar  

Rajaei, E. et al. Outdoor air pollution affects tuberculosis development based on geographical information system modeling. Biomed. Biotechnol. Res. J. BBRJ 2 (1), 39 (2018).

Li, Z. et al. Long-term effect of exposure to ambient air pollution on the risk of active tuberculosis. Int. J. Infect. Dis. 1 (87), 177–184 (2019).

Article   CAS   Google Scholar  

Requia, W. J. et al. Global association of air pollution and cardiorespiratory diseases: A systematic review, meta-analysis, and investigation of modifier variables. Am. J. Public Health 108 (S2), S123–S130 (2017).

Atkinson, R. W., Kang, S., Anderson, H. R., Mills, I. C. & Walton, H. A. Epidemiological time series studies of PM2.5 and daily mortality and hospital admissions: A systematic review and meta-analysis. Thorax 69 (7), 660–665 (2014).

Article   CAS   PubMed   Google Scholar  

Ab Manan, N., Noor Aizuddin, A. & Hod, R. Effect of air pollution and hospital admission: A systematic review. Ann. Glob. Health 84 (4), 670–678 (2018).

Kelly, F. J. & Fussell, J. C. Air pollution and airway disease. Clin. Exp. Allergy J. Br. Soc. Allergy Clin. Immunol. 41 (8), 1059–1071 (2011).

Moore, E. et al. Global associations between air pollutants and chronic obstructive pulmonary disease hospitalizations. A systematic review. Ann. Am. Thorac. Soc. 13 (10), 1814–1827 (2016).

PubMed   PubMed Central   Google Scholar  

Song, Q., Christiani, D. C., Wang, X. & Ren, J. The global contribution of outdoor air pollution to the incidence, prevalence, mortality and hospital admission for chronic obstructive pulmonary disease: A systematic review and meta-analysis. Int. J. Environ. Res. Public Health 11 (11), 11822–11832 (2014).

Schwartz, J. PM10 ozone, and hospital admissions for the elderly in Minneapolis-St. Paul, Minnesota. Arch. Environ. Health Int. J. 49 (5), 366–374 (1994).

Dockery, D. W. et al. Effects of inhalable particles on respiratory health of children. Am. Rev. Respir. Dis. 139 (3), 587–594 (1989).

Sumpter, C. & Chandramohan, D. Systematic review and meta-analysis of the associations between indoor air pollution and tuberculosis. Trop. Med. Int. Health 18 (1), 101–108 (2013).

Kolappan, C. & Subramani, R. Association between biomass fuel and pulmonary tuberculosis: A nested case–control study. Thorax 64 (8), 705–708 (2009).

Kurmi, O. P., Sadhra, C. S., Ayres, J. G. & Sadhra, S. S. Tuberculosis risk from exposure to solid fuel smoke: A systematic review and meta-analysis. J. Epidemiol. Community Health 68 (12), 1112–1118 (2014).

Popovic, I. et al. A systematic literature review and critical appraisal of epidemiological studies on outdoor air pollution and tuberculosis outcomes. Environ. Res. 170 , 33–45 (2019).

Dimala, C. A., Kadia, B. M. & Hansell, A. The association between ambient air pollution and pulmonary tuberculosis: A systematic review protocol. Environ. Evid. 9 (1), 29 (2020).

Haddaway, N. R., Macura, B., Whaley, P. & Pullin, A. S. ROSES RepOrting standards for Systematic Evidence Syntheses: Pro forma, flow-diagram and descriptive summary of the plan and conduct of environmental systematic reviews and systematic maps. Environ. Evid. 7 (1), 7 (2018).

Study Quality Assessment Tools—NHLBI, NIH [Internet]. https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools [cited 2017 Nov 2].

What is GRADE?—BMJ Best Practice [Internet]. https://bestpractice.bmj.com/info/toolkit/learn-ebm/what-is-grade/ [cited 2020 Jan 17].

Unit Conversion | Air Pollution Information System [Internet]. http://www.apis.ac.uk/unit-conversion [cited 2020 Feb 10].

Begg, C. B. & Mazumdar, M. Operating characteristics of a rank correlation test for publication bias. Biometrics 50 (4), 1088–1101 (1994).

Article   CAS   PubMed   MATH   Google Scholar  

Egger, M., Smith, G. D., Schneider, M. & Minder, C. Bias in meta-analysis detected by a simple, graphical test. BMJ 315 (7109), 629–634 (1997).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Hwang, S. et al. Impact of outdoor air pollution on the incidence of tuberculosis in the Seoul metropolitan area, South Korea. Korean J. Intern. Med. 29 (2), 183–190 (2014).

Jassal, M. S., Bakman, I. & Jones, B. Correlation of ambient pollution levels and heavily-trafficked roadway proximity on the prevalence of smear-positive tuberculosis. Public Health 127 (3), 268–274 (2013).

Smith, G. S., Schoenbach, V. J., Richardson, D. B. & Gammon, M. D. Particulate air pollution and susceptibility to the development of pulmonary tuberculosis disease in North Carolina: An ecological study. Int. J. Environ. Health Res. 24 (2), 103–112 (2014).

Álvaro-Meca, A., Díaz, A., de Miguel, D. J., Resino, R. & Resino, S. Environmental factors related to pulmonary tuberculosis in HIV-infected patients in the combined antiretroviral therapy (cART) era. PLoS One 11 (11), e0165944 (2016).

Article   PubMed   PubMed Central   CAS   Google Scholar  

Chen, K. Y. et al. Particulate matter is associated with sputum culture conversion in patients with culture-positive tuberculosis. Ther. Clin. Risk Manag. 12 , 41 (2016).

Lai, T. C. et al. Ambient air pollution and risk of tuberculosis: A cohort study. Occup. Environ. Med. 73 (1), 56–61 (2016).

Peng, Z., Liu, C., Xu, B., Kan, H. & Wang, W. Long-term exposure to ambient air pollution and mortality in a Chinese tuberculosis cohort. Sci. Total Environ. 15 (580), 1483–1488 (2017).

Article   ADS   CAS   Google Scholar  

Smith Geneé, S. et al. Air pollution and pulmonary tuberculosis: A nested case-control study among members of a Northern California Health Plan. Environ. Health Perspect. 124 (6), 761–768 (2016).

Zhu, S. et al. Ambient air pollutants are associated with newly diagnosed tuberculosis: A time-series study in Chengdu, China. Sci Total Environ. 631–632 , 47–55 (2018).

Article   ADS   PubMed   CAS   Google Scholar  

Joob, B. & Wiwanitkit, V. Incidence of pulmonary tuberculosis and particulate matter 2.5 pollutant level: The association analysis for 2019 air pollution Crisis, Bangkok, Thailand. Biomed. Biotechnol. Res. J. 3 (2), 126–126 (2019).

Yao, L. et al. Ambient air pollution exposures and risk of drug-resistant tuberculosis. Environ. Int. 124 , 161–169 (2019).

Article   PubMed   CAS   Google Scholar  

Carrasco-Escobar, G., Schwalb, A., Tello-Lizarraga, K., Vega-Guerovich, P. & Ugarte-Gil, C. Spatio-temporal co-occurrence of hotspots of tuberculosis, poverty and air pollution in Lima, Peru. Infect. Dis. Poverty 9 (1), 32 (2020).

Huang, K. et al. Association between short-term exposure to ambient air pollutants and the risk of tuberculosis outpatient visits: A time-series study in Hefei, China. Environ. Res. 184 , 109343 (2020).

Wang, W. et al. Epidemiological characteristics of tuberculosis and effects of meteorological factors and air pollutants on tuberculosis in Shijiazhuang, China: A distribution lag non-linear analysis. Environ. Res. 22 , 110310 (2020).

Yang, J. et al. A study on the relationship between air pollution and pulmonary tuberculosis based on the general additive model in Wulumuqi, China. Int. J. Infect. Dis. 1 (96), 42–47 (2020).

You, S., Tong, Y. W., Neoh, K. G., Dai, Y. & Wang, C. H. On the association between outdoor PM2.5 concentration and the seasonality of tuberculosis for Beijing and Hong Kong. Environ. Pollut. Barking Essex 2016 (218), 1170–1179 (1987).

Liu, Y., Cui, L. L., Hou, L. J., Yu, C. B., Tao, N. N., Liu, J. Y. et al . Ambient air pollution exposures and newly diagnosed pulmonary tuberculosis in Jinan, China: A time series study. Sci. Rep. [ Internet ] 8 (1). https://www.scopus.com/inward/record.uri?eid=2-s2.0-85057222564&doi=10.1038%2fs41598-018-35411-6&partnerID=40&md5=688c7ce37da23c36a007e125ddddadbc (2018).

Kim, H., Yu, S. & Choi, H. Effects of particulate air pollution on tuberculosis development in seven major cities of Korea from 2010 to 2016: Methodological considerations involving long-term exposure and time lag. Epidemiol. Health 42 , e2020012 (2020).

Liu, Y. et al. Effect of ambient air pollution on tuberculosis risks and mortality in Shandong, China: A multi-city modeling study of the short- and long-term effects of pollutants. Environ. Sci. Pollut. Res. Int. 28 (22), 27757–27768 (2021).

Xiong, Y. et al. Association of daily exposure to air pollutants with the risk of tuberculosis in Xuhui District of Shanghai, China. Int. J. Environ. Res. Public Health 19 (10), 6085 (2022).

Sohn, M. et al. Association of social deprivation and outdoor air pollution with pulmonary tuberculosis in spatiotemporal analysis. Int. J. Environ. Health Res. 29 (6), 657–667 (2019).

Liu, F., Zhang, Z., Chen, H. & Nie, S. Associations of ambient air pollutants with regional pulmonary tuberculosis incidence in the central Chinese province of Hubei: A Bayesian spatial-temporal analysis. Environ. Health 19 (1), 51 (2020).

Wang, H., Tian, C., Wang, W. & Luo, X. Temporal cross-correlations between ambient air pollutants and seasonality of tuberculosis: A time-series analysis. Int. J. Environ. Res. Public Health [ Internet ] 16 (9). https://www.scopus.com/inward/record.uri?eid=2-s2.0-85065795216&doi=10.3390%2fijerph16091585&partnerID=40&md5=6b4fa1102eabb1f409e8770c89f3a35f (2019).

Butt, E. W. et al. Global and regional trends in particulate air pollution and attributable health burden over the past 50 years. Environ. Res. Lett. 12 (10), 104017 (2017).

Shaddick, G., Thomas, M. L., Mudu, P., Ruggeri, G. & Gumy, S. Half the world’s population are exposed to increasing air pollution. Npj Clim. Atmos. Sci. 3 (1), 1–5 (2020).

Ambient (outdoor) air pollution [Internet]. https://www.who.int/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health [cited 2022 Jun 11].

McConnell, R. et al. Asthma in exercising children exposed to ozone: A cohort study. Lancet Lond. Engl. 359 (9304), 386–391 (2002).

Gauderman, W. J. et al. The effect of air pollution on lung development from 10 to 18 years of age. N. Engl. J. Med. 351 (11), 1057–1067 (2004).

Gasser, M. et al. Toxic effects of brake wear particles on epithelial lung cells in vitro. Part Fibre Toxicol. 6 (1), 30 (2009).

Behndig, A. F. et al. Airway antioxidant and inflammatory responses to diesel exhaust exposure in healthy humans. Eur. Respir. J. 27 (2), 359–365 (2006).

Article   ADS   CAS   PubMed   Google Scholar  

Rivas, C. E., Cantarella, P., Sarkar, S., Rockafellow, M., Osornio Vargas, A. R., Torres, M. et al . Particulate air pollution matter (PM) modifies innate immunity of type II pneumocytes (A549) against mycobacterium tuberculosis. Am. J. Respir. Crit. Care Med . [ Internet ]. 189 , Meeting Abstracts. https://doi.org/10.1164/ajrccm-conference.2014.189.1_MeetingAbstracts.A2489 (2014).

Sarkar, S. et al. Suppression of the NF-κB pathway by diesel exhaust particles impairs human antimycobacterial immunity. J. Immunol. 188 (6), 2778–2793 (2012).

Ge, E. et al. Ambient sulfur dioxide levels associated with reduced risk of initial outpatient visits for tuberculosis: A population based time series analysis. Environ. Pollut. 228 , 408–415 (2017).

Xu, M. et al. Association of air pollution with the risk of initial outpatient visits for tuberculosis in Wuhan, China. Occup. Environ. Med. 76 (8), 560–566 (2019).

Tuberculosis (TB) [Internet]. https://www.who.int/news-room/fact-sheets/detail/tuberculosis [cited 2021 Apr 3].

Tuberculosis China [Internet]. https://www.who.int/westernpacific/health-topics/tuberculosis [cited 2021 Apr 3].

Download references

Author information

Authors and affiliations.

Health and Human Development (2HD) Research Network, Douala, Cameroon

Christian Akem Dimala

Department of Medicine, Reading Hospital, Tower Health System, West Reading, PA, USA

Health Education and Research Organisation (HERO) Cameroon, Buea, Cameroon

Benjamin Momo Kadia

Department of Clinical Sciences, Liverpool School of Tropical Medicine, Liverpool, UK

You can also search for this author in PubMed   Google Scholar

Contributions

C.A.D. conceived and designed the experiments. C.A.D. and B.M.K. conducted the experiments. C.A.D. produced the manuscript. B.M.K. reviewed the manuscript. All authors approved the final copy of the manuscript.

Corresponding author

Correspondence to Benjamin Momo Kadia .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary information 1., supplementary information 2., supplementary information 3., supplementary information 4., supplementary information 5., supplementary information 6., supplementary information 7., supplementary information 8., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Dimala, C.A., Kadia, B.M. A systematic review and meta-analysis on the association between ambient air pollution and pulmonary tuberculosis. Sci Rep 12 , 11282 (2022). https://doi.org/10.1038/s41598-022-15443-9

Download citation

Received : 21 December 2021

Accepted : 23 June 2022

Published : 04 July 2022

DOI : https://doi.org/10.1038/s41598-022-15443-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

  • Explore articles by subject
  • Guide to authors
  • Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

systematic literature review on air pollution

  • Research article
  • Open access
  • Published: 28 November 2017

A systematic review of data mining and machine learning for air pollution epidemiology

  • Colin Bellinger   ORCID: orcid.org/0000-0002-3567-7834 1 ,
  • Mohomed Shazan Mohomed Jabbar 1 ,
  • Osmar Zaïane 1 &
  • Alvaro Osornio-Vargas 2  

BMC Public Health volume  17 , Article number:  907 ( 2017 ) Cite this article

26k Accesses

145 Citations

12 Altmetric

Metrics details

Data measuring airborne pollutants, public health and environmental factors are increasingly being stored and merged. These big datasets offer great potential, but also challenge traditional epidemiological methods. This has motivated the exploration of alternative methods to make predictions, find patterns and extract information. To this end, data mining and machine learning algorithms are increasingly being applied to air pollution epidemiology.

We conducted a systematic literature review on the application of data mining and machine learning methods in air pollution epidemiology. We carried out our search process in PubMed, the MEDLINE database and Google Scholar. Research articles applying data mining and machine learning methods to air pollution epidemiology were queried and reviewed.

Our search queries resulted in 400 research articles. Our fine-grained analysis employed our inclusion/exclusion criteria to reduce the results to 47 articles, which we separate into three primary areas of interest: 1) source apportionment; 2) forecasting/prediction of air pollution/quality or exposure; and 3) generating hypotheses. Early applications had a preference for artificial neural networks. In more recent work, decision trees, support vector machines, k-means clustering and the APRIORI algorithm have been widely applied. Our survey shows that the majority of the research has been conducted in Europe, China and the USA, and that data mining is becoming an increasingly common tool in environmental health. For potential new directions, we have identified that deep learning and geo-spacial pattern mining are two burgeoning areas of data mining that have good potential for future applications in air pollution epidemiology.

Conclusions

We carried out a systematic review identifying the current trends, challenges and new directions to explore in the application of data mining methods to air pollution epidemiology. This work shows that data mining is increasingly being applied in air pollution epidemiology.

The potential to support air pollution epidemiology continues to grow with advancements in data mining related to temporal and geo-spacial mining, and deep learning. This is further supported by new sensors and storage mediums that enable larger, better quality data. This suggests that many more fruitful applications can be expected in the future.

Peer Review reports

The decreasing costs of remote sensors for measuring airborne agents, along with the increasing availability of environmental and clinical data, has led to an explosion in the number of pollution datasets available for analysis. These datasets often have a very large number of samples and tend to have a significant number of variables with mixed degrees of dependencies. These big datasets come with complexity that renders it difficult to rely on traditional epidemiological or environmental health models to analyze them. As a result, new methods of analysis are required in order to advance our understanding of the data. Data mining and machine learning methods from computing science present a wide array of scalable and reliable methods that have performed well on similar problems in other domains. This has inspired a burgeoning field of research within Environmental Health aimed at the adoption of data mining methods to analyze modern, big datasets in air pollution epidemiology inefficient and effective ways.

Data mining is the computational process that is often applied to analyze large datasets, discover patterns, extract actionable knowledge and predict outcomes of future or unknown events. Methods used in this process come from a combination of computational disciplines including Artificial Intelligence, Statistics, Mathematics, Machine Learning, and Database Systems. Apart from the core methods used to carry out the analysis, the process of data mining can involve various preprocessing steps prior to executing the mining algorithm. In addition, a post-processing stage is typically employed to visualize the results of the analysis (i.e. recognized patterns or retrieved information) in an intuitive and easy-to-communicate manner. In this review, we limit our scope to focus on core data analysis techniques as they have been applied to the field of air pollution epidemiology and reported within the air pollution epidemiology literature.

In a broad sense, there are two major paradigms of algorithms: prediction and knowledge discovery. Within these, there are four sub-categories: 1) Classification and regression, 2) Clustering, 3) Association Rule Mining, and 4) Outlier/Anomaly Detection. In addition, there are some relatively new and exciting areas of data analysis, such as spatial data mining and graph data mining, that have been made possible via the building blocks of data mining methods.

According to the best of our knowledge, there are no studies that investigate the depth and breadth of the application of data mining methods within air pollution epidemiology. With this in mind, we perform an investigation to identify which data mining methods have been applied, and to which areas of air pollution epidemiology they have been applied to. Our goal is to point domain researchers to preexisting data mining applications in their areas, and related areas, as well as advance their understanding of the potential of data mining and inspire them to explore further research avenues.

Methodology and paradigms of data mining algorithms

Data mining algorithms are particularly beneficial on complex datasets with a large number of variables and samples. With respect to knowledge discovery, they add insight into high-dimensional problems where traditional statistical methods often fail. Similarly, machine learning algorithm can induce accurate predictor functions from complex, high-dimensional datasets where statistical and mathematical methods, such as regression, can be prone to inaccuracies and be difficult to apply due to their underlying assumptions.

Considerations for applying data mining

In order to implement a successful data mining solution, the user must analyze and formalize their objective. The problem objective guides the user to the appropriate paradigm of learning algorithm. If the objective is to identify hidden groups in data or identify associations between key variables in the data, the users are interested in knowledge discovery and will want to select a clustering or association mining algorithm. Alternatively, the objective might be to induce a predictive model that can classify samples as belonging to a particular category, such as poor air quality, or a real-valued outcome, such as the air quality index.

A large and growing number of algorithms belong to the prediction paradigm and the knowledge discovery paradigm. How to choose between the methods within each paradigm is a topic in its own right. To assist practitioners that are new to the application of machine learning algorithms, Domingos discusses the some of the key considerations in [ 1 ].

When making this decision, the user should consider the complexity of the problem and the amount of data available. A simple, linear classifier, for example, will be ineffective on a complex non-linear classification problem. A large volume of data will facilitate the use of advanced learning algorithms, such as deep artificial neural networks [ 2 ], however, it also forces users to consider questions related to storage, memory and training time.

In general, it is widely understood that there is no silver bullet when it comes to learning algorithms. From an application perspective, a good practice is to select a small, diverse set of algorithms from the paradigm of relevant methods, test them individually and select the one that best meets the performance objectives. Alternatively, grouping a diverse set of models to form an ensemble of predictors has been demonstrated to be an effective solution in theory and practice [ 3 ]. Within the surveyed literature, for example, [ 4 ] applied an ensemble formed of neural networks, support vector machines, Gaussian processes, decision trees and random forests.

Once a set of potential algorithms has been selected, the models that are induced by each algorithm over the available dataset must be evaluated in order to select the one model, or ensemble, that is most likely to preform the best on the prediction task in the future. This is an area of research that is presented in Error Estimation for Pattern Recognition [ 5 ] and Evaluating Learning Algorithms [ 6 ].

The paradigms of learning that have most widely been applied in air pollution epidemiology can be categorized as prediction-based or knowledge discovery methods.

Value prediction

Value prediction is a common and widely applicable area of data mining in which the objective is to take in a set of variables related to an instance drawn from an underlying sample population and predict the corresponding value. Depending on the nature of the application, the user will choose either a data mining algorithm that makes categorical predictions (a classifier) or numeric predictions (similar to regression.) Typical classification algorithms include decision trees, Bayesian classifiers, support vector machines and multilayer perceptrons. Artificial neural networks, support vector regression and regression trees are typical data mining methods for performing numeric predictions. The standard approach to select the most appropriate method for a given problem, such as a classification problem, is to perform repeated trials with multiple classifier algorithms and select the approach that performs the best on the learning problem.

More formally, prediction algorithms are typically induced through a process of supervised learning. The objective is, thus, to make predictions y about instances x of the target problem. For this, a parametrized function \(\mathcal {F}: x \rightarrow y\) is induced. The prediction problem can be one of discrete value prediction, such as classifying breast cancer, or continuous value prediction, much like regression.

In order to perform supervised learning, a dataset X of examples, such as patient information, and corresponding values (or labels) Y , are compiled and used for model induction. Each row of X is a feature vector x =( x 1 , x 2 ,..., x n ). The features x i are equivalent to the data variables in a statistical context. The label set, Y ∈ { y 1 , y 2 ,... y n }, specifies the value that each corresponding instance x i takes. In discrete prediction tasks, the class labels are typically mutually exclusive but do not necessarily have to be [ 7 ]. For continuous values prediction, the value space typically involves real numbers, \(Y \in \mathcal {R}\) , but can also apply to integers, \(Y \in \mathcal {I}\) .

Decision trees, Bayesian methods, support vector machines and artificial neural networks are the most common supervised learning algorithms. We provide a brief overview and direct the reader to [ 8 ] for a detailed description of these algorithms.

Decision trees are simple, but an often effective form of learning classifiers, regressors, and rules. The induction process applies a divide-and-conquer strategy which partitions the data space based on the feature values. Decision trees are often preferred over the more sophisticated models that we discuss below in fields such as medicine because the decisions leading to their predictions can be understood by humans. A very simple example of the interoperability comes from a hypothetical flu classifier which makes predictions { FEVER = TRUE ∧ HEADACHE = TRUE ∧ COUGH = TRUE → FLU = TRUE }.

The standard tree induction algorithms are CART, ID3 and C4.5 [ 9 – 11 ]. Decision trees are induced in a top-down manner by recursively selecting a feature that best divides the training instances according to their labels. A notion of purity known as information that is measured in units of bits is commonly used to measure purity in the determination of the best feature f i at the current level l i . Branches from level l i to level l i −1 are then created; one branch is made for each potential value of f i . The training set is partitioned based on the branches from l i to l i −1 and the process is repeated for each node in level l i −1 . The recursive process stops when the leaves only contain instances from a single class. It should be noted, however, that a form of pruning must be applied to the tree to avoid overfitting.

Artificial neural networks are a powerful form of learning algorithm with a long tradition in pattern recognition and machine learning. Their foundation comes from mathematical attempts at replicating information processing in biological systems [ 12 ]. In modern applications, however, they deviate significantly from the roots of their biological inspiration.

With modern memory and processing power, there is a great potential for complex artificial neural network architectures such as convolutional networks and recurrent network that have seen recent success in deep learning [ 2 ]. The standard architecture, however, is a feedforward network known as a multilayer perceptron. The name refers to the fact that the network is a directed graph that is typically composed of three or more layers. The nodes in the first layer are connected to the nodes in the second layer and so on. The first layer is the input layer. This is where the feature vector x enters the network. It is passed successively through the layers of the network until it reaches the final layer, the output layer. The layers between the input and the output layers are known as hidden layers. Each hidden layer is composed of a user-specified number of hidden units (the nodes in the directed graph).

For each unit i of each hidden layer l , the value of the unit \(h^{(l)}_{i}\) is calculated as the values of the units connected to \(h^{(l)}_{i}\) from the layer below as:

where i is the number of units in the previous layer, j is the specifies the unit in the current layer, ω ji is the parametrized weights connecting layer l −1 to the current layer, l , for unit j , and b l is the bias applied to the current layer.

An activation function is applied to hidden value \(\mathbf {h}^{(l)}_{j}\) . The choice of a non-linear activation, such as sigmoid , enables the model to learn a non-linear representation of the data. However, regularized linear units have recently been found to be useful in the hidden layers [ 13 ].

Multilayer perceptrons are typically trained via back-propagation with gradient descent. This involves updating the weights of the network over multiple iterations of the training set. This is a non-convex optimization process, and thus, training may get stuck in local minima. In practice, however, the models have been found to be very effective.

Support vector machines (SVM) are a powerful method for solving classification and regression problems based on the calculation of the maximum margin hyperplane [ 14 , 15 ]. For non-linear SVM, the data is mapped to a higher dimensional space via a user-specified kernel, such as a polynomial kernel or a radial basis function. The maximal margin hyperplane is implicitly found in this higher dimensional space, the result of which can be a non-linear decision boundary in the original space. A key property of SVM is that model induction is a convex optimization problem. As a result, any local minima is also a global minima.

The maximum margin classifier is of the form y ( x )= w T θ ( x )+ b , where x is a query instance, w is the maximum margin hyperplane, θ is a kernel function, and b is an offset.

The maximum margin hyperplane is solved via:

where x n and y n are the training instances and labels. Directly solving this optimization problem is very complex, however, it can be converted to a simpler, but equivalent problem using the Lagrangian dual which is solvable via quadratic programming. Finally, for kernels satisfying the property k ( x i , x j )= θ ( x i )· θ ( x j ) the kernel trick is used to avoid performing the computations in the kernel-space.

Knowledge discovery

Clustering algorithms are a form of knowledge discovery performed via unsupervised learning. They group the instances of a dataset X into k clusters based on an algorithm specific notion of similarity. The process is termed unsupervised because the algorithms do not use a label set for learning. As a result, the process is one of knowledge discovery that infers the groupings from the data.

Similar to classification and regression, a wide variety of clustering algorithms have been developed. Selecting the right algorithm is domain dependent. Nonetheless, the k-means algorithm remains one of the most prominent clustering techniques. It is often preferred for its simplicity and theoretical foundation.

K-means employs an iterative process of updating the cluster centres that repeats until convergence. The k in k-means refers to the user-specified number of clusters. Initially, the k centres are set at random. Subsequently, each instance in X is assigned to the cluster of its nearest centre. The k centres are then updated to be at the centre of their assigned group. Convergence occurs when the centres stop moving.

In spite of its popularity, k-means has some well-known weaknesses, such as susceptibility to outliers. The Density-based clustering algorithm DBSCAN is an alternative method designed to account for noisy instances and outliers. In addition, one can manufacture scenarios in which k-means will fail to define good clusters under certain conditions.

Hierarchical clustering is a form of distance-based clustering that creates hierarchies of clusters. The clusters are either built agglomeratively or divisively. The former commences by assuming each instance of X belongs to its own cluster and builds up the hierarchy by successively merging clusters. Alternatively, the divisive approach starts with all instances in one big cluster and recursively splits the clusters into smaller clusters down the tree. This form of clustering is very effective for visualizing the groupings and different levels of granularity.

Association rules are similar to the rules extracted from decision trees and produced by rule-based classifiers. The key difference is that in association rule mining, the notion of class categories is not utilized in the rule induction process.

In association rule mining, a dataset X is given in which the rows are instances and the columns are the feature, F ∈ { f 1 , f 2 ,..., f n }, that quantify the instances. In medical domains, the features could be has _ cough ∈ { yes , no }, fever _ level ∈ { none, low, medium, high }, has _ headache ∈ { yes, no }, etc .

Through association rule mining we aim to generate a set of interesting rules from X of the form A → B , where A ⊂ F and B ⊆ F . In contrast, rule-based classifiers learn rules of the form A → B , where A ⊂ F and B ∈ Y ; here, Y is the set of possible class labels.

Given the definition of an association rule, any unique combination of the features, F , can appear on the left side and the right side of the implication. As a result, an enormous number of rules can be generated. Many of these, perhaps the majority, would be uninteresting according to any reasonable assessment. Thus, the rules must be filtered or pruned, as to only keep the valuable rules. Individually assessing each rule in a brute-force manner is prohibitive, and thus, more efficient methods of rule induction have been developed.

The APRIORI algorithm is the most common technique of association mining [ 16 ]. The key to their strategy is the employment of an iterative process that builds up frequent item sets and association rules from their simplest form (one-item sets) to the complex (two-item sets, three-item sets,..., n-item sets). An example of a one-item set and a two-item set from our medical domain is has _ cough = yes , and has _ cough = yes ∧ fever _ level = none . The items are deemed to be frequent if they have more than a user-specified number of necessary occurrences s in the dataset.

The algorithm gains its efficiency from the realization that if a one-item set, such as has _ cough = yes , is not frequent in the dataset, then no two-item set including the one-item set, such as has _ cough = yes ∧ fever _ level = none , can be frequent. Therefore, the algorithm can ignore all higher-order rules involving has _ cough = yes . In general, the algorithm commences by finding all frequent one-itemsets and then finds candidate two-items sets from the frequent one-item sets. The two-itemsets that are frequent are kept, and the process repeats until some point, k , is reached where no k -itemsets are frequent.

In the last stage, all of the frequent itemsets are used to form association rules. The frequent item set A 1 ∧ A 2 ∧ A 3 , for example, generates A 1 → A 2 ∧ A 3 , A 1 ∧ A 2 → A 3 , etc . A similar bottom-up methodology is applied here to efficiently generate rules that meet the minimum frequency requirement.

We have undertaken this survey in a systematic manner guided by the work of Kitchenham in [ 17 ] and the PRISMA standards [ 18 ]. Accordingly, the strategy for conducting this survey is detailed in the following sub-sections. In addition, we have taken motivation for the organization of this survey from a related survey on dengue disease surveillance [ 19 ].

Research questions

The primary research questions considered in this survey are:

R1 To what degree has data mining been applied in air pollution epidemiology?

R2 Are there any hotbeds of this research area?

R3 To which sub-fields of air pollution epidemiology has data mining been applied?

R4 Which data mining methods have been applied?

R5 What are the limitations of the current work?

R6 What potentially fruitful directions remain unexplored?

With respect to R1, we searched the relevant epidemiological literature for research employing data mining techniques. We did not place any bounds on the dates, however, it is clear that the active period is relatively small. Moreover, there is an upward trend in the frequency as the benefits of data mining become more widely known, and tools that lower the barriers to use are made available.

Following from R1, R2 considered if the existing research is uniformly spread around the countries and institutions of the world, or if particular countries and institutions have a more keen focus on researching this area.

To address R3, we filtered through the identified articles to find any reasonable sub-categorization of the epidemiological work in terms of the application areas. This process revealed three categories of epidemiological studies of air pollution in the literature involving data mining.

In R4, we looked to see which paradigms, and which algorithms, have been applied in the air pollution epidemiology literature. From this vantage point, we found that four classes of methods have been applied.

For research question R5, we considered if, given the objectives, the data and/or the mining algorithms applied had any limitations. Given our backgrounds in data mining, we were particularly focused on the data used, algorithms applied and the processes by which the methods were evaluated.

Finally, in R6 we considered the reasonable next steps. Once again, our consideration here took a data mining perspective. To this end, we were interested in identifying new ways of using the existing data and cutting edge data mining algorithms that should be tested within this research domain.

Search process

We performed a temporally unbounded search for articles listed in the PubMed database 1 , the Public Library of Science (PLOS) 2 and Google Scholar 3 . This includes articles published up to the time of writing in October 2017.

The articles reported herein result from a three-part search procedure. This involved: a ) a query-based search to produce a long list of potential articles designed and conducted by CB and MSMJ, b ) a fine-grained manual evaluation of the long-listed articles by one author performed by CB and MSMJ, and c ) identified articles were reviewed by the remaining authors (AOV and OZ). The queries applied to the database and with the number of articles returned are reported in Table  1 .

We excluded articles that did not go through a peer review process in recognized biomedical publication, and articles that did not apply one or more data mining algorithms. Many environmental health articles, for example, mention, and/or discuss, the potential for data mining but did not applying data mining methods. Articles that discuss data mining in the future work were returned by our queries, but are not appropriate for inclusion in our survey.

Data extraction and synthesis

The following information was extracted from each of the selected articles:

The source (journal or conference) and full reference.

A summary of the objective of the study.

The air pollutants of interest in the study.

The data mining method applied to achieve the objective.

A summary of the findings of the study.

This information was extracted by CB and MSMJ and validated by AOV and OZ. Any disagreements were handled via discussion and common consensus. After the raw details of the articles were tabulated, data synthesis was performed. In addition, AOV extracted information about each article related to the biomedical objectives.

Data synthesis involved analyzing the objectives, data mining methods, and the target pollutants in order to identify categories to effectively group the various studies. This exercise was performed by CB and reviewed by the remaining authors. Our goal in the categorization was to identify a hierarchy of categories that provided a sketch of the research landscape. In addition, the purpose was to facilitate quick and easy locating of the studies that are related to the reader’s area of interest. The identified categories are listed below:

Physical Area

Outdoor (Rural, Urban and General 4 )

Forecasting and Prediction

Source Apportionment

Hypothesis Generation

Data Mining Method

Classification

Association Mining

Aspects of data mining in air pollution epidemiology

Environmental setting: overview.

In this section we discuss the target areas of interest (the environmental setting). We have separated these into indoor, outdoor and general. Indoor refers the studies focused on indoor air pollution, such as air pollutants measured within the home or workplace. Outdoor refers to studies interested in outdoor air pollution, such as air pollution measured at a specific intersection or the dispersion of pollutants across an area of interest. It can be further separated into urban, metropolitan and rural. Given that the current breadth of research is still relatively sparse, we focus on the top level of abstraction in this article. We note, however, that a large portion of the research in the outdoor category has been applied to urban and/or metropolitan settings. This is, perhaps, not surprising given that the high population density in metropolitan areas can lead to high impact research. Nonetheless, it suggests rural environments as a potential direction for future work.

The general category covers research that applies data mining methods to study the health impacts of combinations of chemicals common in air pollution. These studies were typically conducted in laboratory settings rather than in the field (or relying on data collected from the field). Table  2 includes a categorized list of articles in relation to their environmental settings.

Categorized study objectives: overview

We grouped the selected articles into the following general study objectives: forecasting and prediction, source apportionment and hypothesis generation. A large percentage of the articles identified in our survey dealt with forecasting or predicting pollution levels based on various climatic and/or pollutant values. These studies considered: a ) forecasting future pollution levels at a specific location given some specific data for that location, b ) forecasting current pollution levels at a specific site given some regional data, and c ) forecasting the geo-spatial distribution of air quality or the spread of pollutants.

Closely related are the studies that were designed to predict increases in sickness or hospitalization from climatic and pollution measurements or to classify sickness in individuals given an air quality or pollution assessment.

Studies classified into the source apportionment category aimed to trace a given decrease in air quality or increase in a given pollutant back to its emission source given a set of pollutant and climatic variables.

Finally, a large number of articles performed hypothesis generation. These studies take in the wide variety of data available about the evolution of air pollution at a specific location, its spread across a region or the globe, health indicator variables, etc ., and use data mining algorithms to identify hidden associations between the variables. These associations are used to test existing assumptions and generate new ones. Exemplary associations might indicate that a certain chemical combination X,Y,Z is associated with increased volume in the emergency department at a hospital of interest, or that climatic conditions W,R combined with heavy seaport traffic, lead to a decrease in the air quality index. These associations can serve to motivate focused trials to study the discovered relationship in depth.

Table  3 includes the articles in a list sorted according to the objectives of the research. It is worth noting that a given article may have more than one objective, and thus, may appear multiple times in the table.

Summary statistics

Prisma results.

The summary statistics recording the numbers of articles returned from our search process, excluded, and included are presented in the PRISMA flow chart in Fig.  1 . Our initial search returned 400 articles. In addition to these, one article ([ 20 ]) was suggested during the review process. After the initial screening and eligibility assessment, 47 articles were included in this survey.

PRISMA flow diagram. Overview of the PRISMA results from our search process

Regional and temporal overview

We have found that eighteen of the studies were from Europe and the UK, sixteen were from the USA, ten were from China, and four were from other Asian countries. The detailed breakdown of this is provided in Fig.  2 . The papers were published between 2000 and October 20, 2017. Figure  3 illustrates a strong upward trend in recent years. We believe this to be owing to better access to data and computing power, along with a growing awareness and access to data mining tools that are accessible to users outside of the data mining community. These tools include the Weka data mining software, which enables users to directly apply data mining algorithms to their data through Java interfaces or a graphical user interface [ 21 ].

Publications Per Country. The number of publications per country identified a predominance in the filed by European countries, the USA and China

Publications Per Year. Number of articles per year between January 2000 and October 2017. We identified an apparent tendency of an increased number of publications on data mining and epidemiology in recent years

Study objectives

The summary statistics for the study objectives are as follows: sixty percent of the study objectives were to forecast or predict epidemiological values/outcomes, such as the AQI or increases in emergency room visits. Thirty percent performed hypothesis generation. This included objectives, such as learning from the data, in which combinations of variables are associated with increases in hospitalization, and understanding which combination of meteorological variables are associated with a degradation in air quality due to emissions from neighbouring cities. Finally, ten percent of the studies focused on source apportionment.

Data mining paradigm

We identified that classification, regression, clustering and association mining algorithms have been applied. Classification and regression relate to prediction and forecasting objective, whereas clustering and association mining generally apply to hypothesis generation and source apportionment.

Table  4 includes the articles in a list sorted according to the objectives of the research. Data mining methods for performing numeric predictions, such as regression and classification, were most widely applied. This area encompassed 59% of the research. Clustering algorithms were applied in 26% of the work, and 15% of the articles employed association mining.

Detailed analysis

Source apportionment.

Table  5 summarizes source apportionment studies employing data mining techniques. These studies explore the impact of chemical emissions and other airborne agents in conjunction with climatological factors [ 22 – 24 ]. They focus on apportioning particular airborne pollutants to potential sources, such as industrial sites, regions and major intersections. These studies have mainly focused on outdoor and urban air pollution as it is the most widely known issue. In particular, principal component analysis (PCA) has been applied to identify correlations and the importance of particular meteorological parameters, traffic, fuel fired equipment and industries in causing air pollution [ 23 – 25 ]. Alternative approaches have utilized clustering-based solutions with correlation analysis to accomplish the task of source apportionment [ 22 , 23 ].

Strengths : The work presented in [ 23 ] proposes to perform enhanced source apportionment and classification. The authors claim that the key to achieving this is in the use of clustering algorithms developed for data mining. The advantage of these is that they are intended for rich, high-dimensional datasets that may include outliers. These factors can be problematic for conventional methods of source apportionment, such as principle component analysis and positive matrix factorization.

Once again, we have identified that the clarity with which the authors present the problem, and then juxtapose the limitation of conventional methods with the potential of data mining approaches, to be a very strong point in this paper. In addition, we appreciate that the authors have gone beyond simply applying standard clustering algorithms, and rather, employed their domain knowledge in order to refine the method in order to develop a superior clustering algorithm for the domain. The authors describe their algorithm, how to set the threshold parameter and the data pre-processing in detail. Crucially, this makes the proposed solution easily implementable by others.

Forecasting and prediction

Tables  5 and 6 summarizes 18 studies which applied machine learning techniques. We observed that investigators are primarily interested in predicting a) the distribution of ambient pollutant concentrations or related measures such as the air quality index (AQI), b) human exposure or c) risk of a health outcome.

According to the above points, the first category consists of 15 studies which either focus on predicting the distribution of particular air pollutants or predicting the quality of air in general. Twenty-seven percent of the above 15 studies (i.e. 4) focus on predicting or forecasting the air quality or air pollution in general. Sixty-six percent of the studies (i.e. 10) are interested in well-known specific air pollutants such as nitrogen oxides (NOx), particulate matter (PM), sulfur dioxide (SO2), carbon monoxide (CO), ozone (O3) and Volatile Organic Compounds VOCs. An interesting, and potentially fruitful, data source utilized by some of the studies that focus on air quality prediction comes from social media posts; social media offers a very rich source of information, which is not typically utilized in scientific analyses of this type [ 26 , 27 ].

Fifty percent of the studies focusing on specific air pollutants use artificial neural networks. Other well-known data mining techniques used include decision trees and support vector machines. In addition, some studies have used ensemble models which are composed of multiple models. The final outcome is determined based on the consensus of the outcome of each model, or some other method of arbitration. Ensemble models have been demonstrated to outperform the base classifiers from which they are composed in a variety of settings. They can be applied to both discrete and continuous value prediction [ 28 , 29 ].

The second category, namely human exposure, consists of 4 studies. These studies focus on identifying regions or exposures, predicting the activities of humans to help understand the exposure better and quantifying the exposure levels.

Interestingly, some of the studies also focus on building better infrastructure to collect data, or on ways to improve the quality of the collected data. This investment can be interpreted as a level of confidence in the application of data mining, and its potential to help shape future research and understanding. The interest in more fundamental problems like data collection and the accuracy of the collected data, in addition to a single focus on building a model based on the available data is very important. Work based on primary questions as such these will ensure high-quality datasets are available in the future, and thus, that better data mining and machine learning models will be possible.

Strengths : The authors in [ 30 ] propose a hybrid system that incorporates a variety of machine learning methods to produce more accurate forecasts and evaluations of air pollution. The authors note that data driven approaches are often more accurate and less complex than model-based approaches, such as chemical transport models, for predicting air quality. Data mining and machine learning-based approaches are data driven methods that are recognized as being powerful forecasting tools. This motivates them as a good choice for the authors. Although we do not see this as being as strong as the previous motivations for applying data mining and machine learning, it is certainly a sufficient reason to consider machine learning.

Hypothesis generation

We observed that many studies (i.e. above 60% of the studies that we have considered) have predominantly applied association rule mining—a primary class of data mining techniques—to generate new hypotheses regarding potential connections between air pollution and adverse health conditions.

From the identified articles, we observed that respiratory disease is an adverse health outcome of interest in these studies. Many studies focusing on respiratory disease are interested in finding out any potential connection between the disease and particulate matter or other airborne pollutants such as SO2 and NOx.

Our results demonstrated that there is a growing interest in generating new hypotheses explaining the connection between a combination of air pollutants and a particular adverse health impact. In [ 31 ], for exampled, the authors used the Bayesian Kernel Machine Regression (BKMR) method, which was recently introduced by epidemiologists. This illustrates the benefit of applying data mining methods to modern epidemiological datasets.

Strengths : The authors in [ 20 ] are interested in generating hypotheses about the joint effect of multiple airborne chemicals on pediatric asthma. Their work demonstrates that classification and regression trees can be used to overcome the challenge presented by multiple chemical interactions when identifying complex joint effects.

We have identified this as a noteworthy paper because the authors are studying a problem that is difficult to solve using conventional epidemiological methods. The paper is strengthened by the fact that the authors clearly justify the ML/DM solution to the problem. In addition, the authors explain why they selected the specific ML algorithm. Finally, this work is an excellent example of how ML/DM algorithms can be augmented and combined with knowledge and practices from the target domain in order to make an accurate and appropriate joint methodology. In particular, the authors demonstrate a refinement to the standard CART algorithm to control for confounding variables. This is important to note because in some applications, data mining practitioners can lose sight of useful, and often necessary, domain knowledge, which hampers the final results. Table  7 summarizes 15 studies which applied data mining techniques to generate new hypotheses to better understand the relationship between air pollution and health.

Challenges and limitations

We have identified a few reoccurring challenges in the surveyed papers. A major theme revolves around data. Many articles, for example, report results from data collected over a short period of time, and from one, or only a few, locations [ 32 , 33 ]. As a result, the findings cannot necessarily be generalized to new locations. This is particularly the case for prediction models trained on local data.

Most real-world data requires preprocessing to combine data sources, remove noise and properly structure the data. The necessities of this may be challenging for domain practitioners. Moreover, certain decisions that must be made during preprocessing can have an impact on the effectiveness of the trained model. Decision trees and association mining algorithms, for example, take categorical variable inputs, whilst continuous variables are common in epidemiology and atmospheric science. Thus, variables, such as temperature, must be converted to discrete categories (low, medium, high) for example [ 33 ]. In many cases, the ideal split points may be unclear. In general, the current literature does not focus on how to best preprocess air pollution epidemiological datasets.

Given the volume of social media data, and the fact that the vast majority of it is irrelevant to the data mining objective, it often has to be filtered. In [ 27 ], for example, keyword filtering is applied to gather relevant micro-blogs from Sina Weibo. How exactly to filter, or process the data, is an open question. A potential new direction here is to apply feature selection or feature extraction [ 2 ].

As noted by [ 34 ], it is important to recognize the limits of your data. Issues, such as granularity and representativeness, can limit what can be discovered from the data. Likewise, when generating association rules to predict outcomes, such as hospitalization or an increase in respiratory disease from weather and pollution data, the training data may not account for all relevant factors. In [ 35 ], it is noted that their data does not account for the accumulative nature of health outcomes.

Other challenges in applying data mining methods include the selection of user-specified parameters for the algorithms. Choosing the ideal number of clusters, for example, is important for performance of clustering algorithms [ 22 ]. In addition, metrics must be used that are appropriate for the target domain. In some cases, suitable evaluation metrics may not exist within the data mining literature, in which case new metrics may be required [ 36 ].

Finally, in many cases practitioners prefer data mining models that produce predictions in a manner that can be easily analyzed and understood [ 32 , 36 ]. This limits the choice of algorithms to rule learners and decision trees, and thus, many of the strongest algorithms are omitted. Perhaps, research focused on making the predictions of artificial neural networks and support vector machines more interpretable could be helpful for the health sciences community [ 37 ].

From an application perspective, we found the discussion of data mining related choices to be limited. Whilst the majority of the articles surveyed contain sufficient details about the algorithms implemented, readers could benefit from a similar level of detail in regards to other key design and implementation decisions. With few exceptions, such as [ 38 ], most of the surveyed articles report results for a single data mining algorithm. Readers would benefit from understanding how and why the specific choice of algorithm was made. As we noted in the overview section on Paradigms of Data Mining Algorithms, it is standard within the data mining community to run trials using a diverse set of algorithms. We often missed a discussion of details regarding which other algorithms were considered, and how they were evaluated. Finally, details regarding how the software was implemented and which data mining packages were used would be valuable to other readers from the air pollution epidemiology community.

Future directions

Deep learning.

Traditional artificial neural networks have proven to be accurate predictors for classification and regression problems. Within this survey, we have found them to be used for predicting global ground level PM 2.5 [ 4 ], predicting air pollution indicators and modeling personal exposure [ 39 ]. In recent years, however, deep learning has elevated the potential for learning with artificial neural networks to new heights. Thus, deep learning methods may also be very fruitful within air pollution epidemiology.

Deep learning is based on standard artificial neural network algorithms but utilizes much larger and deeper networks trained on big datasets. The training process in conjunction with the depth of the networks enables the learning of data abstractions at the different depths. This is found to disentangle complex features. Deep learning methods have been highly effective in areas such as image classification, speech recognition, and other complex problems [ 2 ].

Model selection

Model selection and evaluation are very important aspects of applying machine learning algorithms to real-world applications. However, they often receive less attention than the machine learning algorithms themselves. It is important to consider the breadth of techniques when developing applications in data mining in order to select the right approach for the domain.

For a given machine learning algorithm, model selection refers to the choosing of a parameterized version of the model based on the training data. The key is to select a model that will perform well on unseen data in the future. Once a parametrized model has been selected, the evaluation process provides an estimate of how the model will perform during future application. Some common evaluation metrics are accuracy, root mean square error (RMSE), f-measure and the area under the ROC curve (AUC).

Cross-validation

In the surveyed literature, various forms of cross validation have been applied [ 24 , 26 , 38 ]. In addition to these, various other methods can be applied, each of which has strengths and weaknesses. It is important to select a method that is appropriate for your target domain. Evaluation metrics estimate performance in different ways, and thus, it is important to choose one that is consistent with the target domain. The details of model selection and evaluation are thoroughly discussed in [ 6 ].

  • Association mining

Our results demonstrated that much of the research that applied hypothesis generation utilize association mining. These studies typically relied on frequency to identify the associations. It is worth pointing out some alternatives, particularly for scientific domains. Statistical significance test-based methods, for example, have been developed to offer a better assessment of the quality of the association [ 40 , 41 ]. These could be of great benefit to future applications in air pollution epidemiology.

Class imbalance

In a related context, [ 38 ] noted the potential impact of class imbalance, or skewed class distributions, on the performance of machine learning algorithms. Class imbalance is said to occur when one class is significantly less likely, or less frequent, in the training set, than the other class. A detailed discussion of the impacts and potential solutions to class imbalance is undertaken in [ 42 ]. Given that we are often interested in less frequent, or even rare, events in air pollution epidemiology, methods developed for imbalanced learning may have great potential here.

Recent progress in technology and corresponding decreases in the price of computing power has made if possible to measure and store a wide variety of environmental health variables and form them into big datasets. Moreover, social media and other on-line resources provide an entirely new perspective from which to conduct environmental health analyses. These big datasets come with complexities that render it difficult to rely on traditional epidemiological or environmental health models to analyze them. To this end, data mining methods offer great potential to advance our understanding of the causes and impacts of air pollution.

From our survey, we have found a strong increase in the number of articles reporting to apply data mining methods to air pollution epidemiology. We attribute this to the increasing availability of large datasets and computing power, along with the growing awareness of the potential benefits of data mining. In spite of this trend and the potential benefit within the field, to the best of our knowledge, a survey of the existing state-of-the-art has not been performed.

To fill this void, we undertook a study to explore the extent to which data mining has been applied to air pollution epidemiology. This survey is intended for practitioners and researchers alike. We aim to point domain researchers to existing data mining applications within their respective areas, and related areas, as well as advance their understanding of the potential of data mining and inspire them to explore further research avenues.

Our survey illustrates that a wide variety of data mining algorithms have been applied to various sub-fields of air pollution epidemiology. Machine learning algorithms, for example, have been applied both as classifiers and regressors in forecasting and prediction problems. Clustering algorithms, such as K-Means and hierarchical clustering have been applied to knowledge discovery and source appropriation. In addition, a great number of studies have applied association mining for hypothesis generation.

1 see https://www.ncbi.nlm.nih.gov/pubmed

2 https://www.plos.org

3 https://scholar.google.ca

4 General refers to the study of air pollutants not specific to a certain setting.

Domingos P. A few useful things to know about machine learning. Commun ACM. 2012; 55(10):78–87.

Article   Google Scholar  

LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015; 521(7553):436–44.

Article   CAS   PubMed   Google Scholar  

Dietterich TG, et al. Ensemble methods in machine learning. Multiple Classif Syst. 2000; 1857:1–15.

Lary DJ, Faruque FS, Malakar N, Moore A, Roscoe B, Adams ZL, Eggelston Y. Estimating the global abundance of ground level presence of particulate matter (pm2. 5). Geospatial Health. 2014; 8(3):611–30.

Neto UMB, Dougherty ER. Error estimation for pattern recognition. Hoboken: John Wiley & Sons; 2015.

Book   Google Scholar  

Japkowicz N, Shah M. Evaluating Learning Algorithms: a Classification Perspective. Cambridge: Cambridge University Press; 2011.

Bellinger C, Amid A, Japkowicz N, Victor H. Multi-label classification of anemia patients. In: Proceedings of the IEEE 14th International Conference on Machine Learning and Applications (ICMLA).2015. p. 825–30.

Bishop C. Pattern Recognition and Machine Learning (Information Science and Statistics). 1613-9011. Cambridge: Springer; 2006.

Google Scholar  

Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and Regression Trees. Chicago: CRC Press; 1984.

Quinlan JR. Induction of decision trees. Mach Learn. 1986; 1(1):81–106.

Quinlan JR. C4. 5: Programs for Machine Learning. San Francisco: Morgan Kaufmann; 1993.

Rumelhart DE, Hinton GE, Williams RJ. Learning internal representations by error propagation. Parallel Distributed Processing: Exploration of the Micro-structure of Cognition. 1986; 1:1–34.

Maas AL, Hannun AY, Ng AY. Rectifier nonlinearities improve neural network acoustic models. In: Proceedings of the International Conference on Machine Learning.2013.

Vapnik V. The Nature of Statistical Learning Theory. New York: Springer; 1995.

Schölkopf B, Smola AJ. Learning with kernels. Cambridge: MIT Press; 2002.

Agrawal R, Srikant R. Fast algorithms for mining association rules. In: Proceedings of the 20th International Conference on Very Large Data Bases (VLDB).1994. p. 487–99.

Kitchenham B. Procedure for undertaking systematic reviews. Technical report, Computer Science Department, Keele University (TRISE-0401) and National ICT Australia Ltd (0400011T. 1), Joint Technical Report. 2004.

Moher D, Liberati A, Tetzlaff J, Altman D. Preferred reporting items for systematic reviews and meta-analyses: the prisma statement. J Clin Epidemiol. 2009; 62(10):1006–12.

Article   PubMed   Google Scholar  

Runge-Ranzinger S, Horstick O, Marx M, Kroeger A. What does dengue disease surveillance contribute to predicting and detecting outbreaks and describing trends?Trop Med Int Health. 2008; 13(8):1022–41.

Gass K, Klein M, Chang HH, Flanders WD, Strickland MJ. Classification and regression trees for epidemiologic research: an air pollution example. Environ Health. 2014; 13(1):17.

Article   PubMed   PubMed Central   Google Scholar  

Hall M, Frank E, Holmes G, Pfahringer B, Reutemann P, Witten IH. The weka data mining software: An update. SIGKDD Explor Newsl. 2009; 11(1):10–18.

Chen HW, Tsai CT, She CW, Lin YC, Chiang CF. Exploring the background features of acidic and basic air pollutants around an industrial complex using data mining approach. Chemosphere. 2010; 81(10):1358–67.

Chen M, Wang P, Chen Q, Wu J, Chen X. A clustering algorithm for sample data based on environmental pollution characteristics. Atmos Environ. 2015; 107:194–203.

Article   CAS   Google Scholar  

Singh KP, Gupta S, Rai P. Identifying pollution sources and predicting urban air quality using ensemble learning methods. Atmos Environ. 2013; 80:426–37.

Thurston GD, Spengler JD. A quantitative assessment of source contributions to inhalable particulate matter pollution in metropolitan boston. Atmos Environ (1967). 1985; 19(1):9–25.

Jiang W, Wang Y, Tsou MH, Fu X. Using social media to detect outdoor air pollution and monitor air quality index (aqi): a geo-targeted spatiotemporal analysis framework with sina weibo (chinese twitter). PloS ONE. 2015; 10(10):0141185.

Wang S, Paul MJ, Dredze M. Social media as a sensor of air quality and public response in china. J Med Internet Res. 2015; 17(3):22.

Freund Y, Schapire RE. Experiments with a new boosting algorithm. In: Proceedings of the Thirteenth International Conference on Machine Learning: 1996. p. 148–56.

Friedman JH. Greedy function approximation: a gradient boosting machine. Ann Stat. 2001; 29(5):1189–232.

Xu Y, Yang W, Wang J. Air quality early-warning system for cities in china. Atmos Environ. 2017; 148:239–57.

Bobb JF, Valeri L, Henn BC, Christiani DC, Wright RO, Mazumdar M, Godleski JJ, Coull BA. Bayesian kernel machine regression for estimating the health effects of multi-pollutant mixtures. Biostatistics. 2014; 16(3):058.

Payus C, Sulaiman N, Shahani M, Bakar AA. Association rules of data mining application for respiratory illness by air pollution database. Int J Basic Appl Sci. 2013; 13(3):11–16.

Kukkonen J, Partanen L, Karppinen A, Ruuskanen J, Junninen H, Kolehmainen M, Niska H, Dorling S, Chatterton T, Foxall R, et al. Extensive evaluation of neural network models for the prediction of no 2 and pm 10 concentrations, compared with a deterministic modelling system and measurements in central helsinki. Atmos Environ. 2003; 37(32):4539–550.

Reid CE, Jerrett M, Petersen ML, Pfister GG, Morefield PE, Tager IB, Raffuse SM, Balmes JR. Spatiotemporal prediction of fine particulate matter during the 2008 northern california wildfires using machine learning. Environ Sci Technol. 2015; 49(6):3887–96.

Zhu W, Wang J, Zhang W, Sun D. Short-term effects of air pollution on lower respiratory diseases and forecasting by the group method of data handling. Atmos Environ. 2012; 51:29–38.

Toti G, Vilalta R, Lindner P, Lefer B, Macias C, Price D. Analysis of correlation between pediatric asthma exacerbation and exposure to pollutant mixtures with association rule mining. Artif Intell Med. 2016; 74:44–52.

Setiono R, Liu H. Understanding neural networks via rule extraction. In: Proceedings of the International Joint Conference of Artificial Intelligence (IJCAI).1995. p. 480–5.

Pandey G, Zhang B, Jian L. Predicting submicron air pollution indicators: a machine learning approach. Environ Sci Process Impacts. 2013; 15(5):996–1005.

Aquilina NJ, Delgado-Saborit JM, Gauci AP, Baker S, Meddings C, Harrison RM. Comparative modeling approaches for personal exposure to particle-associated pah. Environ Sci Technol. 2010; 44(24):9370–6.

Hämäläinen W. Statapriori: an efficient algorithm for searching statistically significant association rules. Knowl Inform Syst. 2010; 23(3):373–99.

Hämäläinen W. Kingfisher: an efficient algorithm for searching for both positive and negative dependency rules with statistical significance measures. Knowl Inform Syst. 2012; 32(2):383–414.

He H, Ma Y. Imbalanced Learning: Foundations, Algorithms, and Applications. Piscataway: John Wiley & Sons; 2013.

Min KD, Kwon HJ, Kim K, Kim SY. Air pollution monitoring design for epidemiological application in a densely populated city. Int J Environ Res Public Health. 2017; 14(7):686.

Article   PubMed Central   Google Scholar  

Keller JP, Drton M, Larson T, Kaufman JD, Sandler DP, Szpiro AA. Covariate-adaptive clustering of exposures for air pollution epidemiology cohorts. Ann Appl Stat. 2017; 11(1):93.

Li D, Liu J, Zhang J, Gui H, Du P, Yu T, Wang J, Lu Y, Liu W, Cheng Y. Identification of long-range transport pathways and potential sources of pm2.5 and pm10 in beijing from 2014 to 2015. J Environ Sci. 2017; 56:214–29.

Stingone JA, Pandey OP, Claudio L, Pandey G. Using machine learning to identify air pollution exposure profiles associated with early cognitive skills among u.s. children. Environ Pollut. 2017; 230:730–40.

Liu BC, Binaykia A, Chang PC, Tiwari MK, Tsao CC. Urban air quality forecasting based on multi-dimensional collaborative support vector regression (svr): A case study of beijing-tianjin-shijiazhuang. PLOS ONE. 2017; 12(7):1–17.

Mirto M, Fiore S, Conte L, Bruno LV, Aloisio G. A spatial data analysis infrastructure for environmental health research. In: Proceedings of the International Conference on High Performance Computing and Simulation (HPCS): 2016. p. 435–42.

Lewis AC, Lee JD, Edwards PM, Shaw MD, Evans MJ, Moller SJ, Smith KR, Buckley JW, Ellis M, Gillot SR, et al. Evaluating the performance of low cost chemical sensors for air pollution research. Faraday Discuss. 2016; 189:85–103.

Reid CE, Jerrett M, Tager IB, Petersen ML, Mann JK, Balmes JR. Differential respiratory health effects from the 2008 northern california wildfires: a spatiotemporal approach. Environ Res. 2016; 150:227–35.

Fernández-Camacho R, Cabeza IB, Aroba J, Gómez-Bravo F, Rodríguez S, de la Rosa J. Assessment of ultra-fine particles and noise measurements using fuzzy logic and data mining techniques. Sci Total Environ. 2015; 512:103–13.

Lary D, Lary T, Sattler B. Using machine learning to estimate global pm2. 5 for environmental health studies. Environ Health Insights. 2015; 9(Suppl 1):41.

CAS   PubMed   PubMed Central   Google Scholar  

Qin S, Liu F, Wang C, Song Y, Qu J. Spatial-temporal analysis and projection of extreme particulate matter (pm 10 and pm 2.5) levels using association rules: A case study of the jing-jin-ji region, china. Atmos Environ. 2015; 120:339–50.

Chen J, Chen H, Zheng G, Pan JZ, Wu H, Zhang N. Big smog meets web science: smog disaster analysis based on social media and device data on the web. In: Proceedings of the 23rd International Conference on World Wide Web. ACM: 2014. p. 505–10. http://www.sciencedirect.com/science/article/pii/S1352231015303484 .

Dias D, Tchepel O. Modelling of human exposure to air pollution in the urban environment: a gps-based approach. Environ Sci Pollut Res. 2014; 21(5):3558–71.

Philibert A, Loyce C, Makowski D. Prediction of n 2 o emission from local information with random forest. Environ Pollut. 2013; 177:156–63.

Padula AM, Mortimer K, Hubbard A, Lurmann F, Jerrett M, Tager IB. Exposure to traffic-related air pollution during pregnancy and term low birth weight: estimation of causal associations in a semiparametric model. Am J Epidemiol. 2012; 176(9):148.

Ma Y, Richards M, Ghanem M, Guo Y, Hassard J. Air pollution monitoring and mining based on sensor grid in london. Sensors. 2008; 8(6):3601–623.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Niska H, Hiltunen T, Karppinen A, Ruuskanen J, Kolehmainen M. Evolving the neural network model for forecasting air pollution time series. Eng Appl Artif Intell. 2004; 17(2):159–67.

Kolehmainen M, Martikainen H, Ruuskanen J. Neural networks and periodic components used in air quality forecasting. Atmos Environ. 2001; 35(5):815–25.

Challoner A, Pilla F, Gill L. Prediction of indoor air exposure from outdoor air quality using an artificial neural network model for inner city commercial buildings. Int J Environ Res Public health. 2015; 12(12):15233–53.

Pegoretti S, Verdi L. Machine learning for the analysis of indoor radon distribution, compared with ordinary kriging. Radiat Prot Dosim. 2009; 137(3-4):324–8.

Bell SM, Edwards SW. Identification and prioritization of relationships between environmental stressors and adverse human health impacts. Environ Health Perspec. 2015; 123(11):1193–9.

Bell SM, Edwards SW. Building associations between markers of environmental stressors and adverse human health impacts using frequent itemset mining. In: Proceedings of the SIAM International Conference on Data Mining: 2014. p. 551–9.

Hu M, Li W, Li L, Houston D, Wu J. Refining time-activity classification of human subjects using the global positioning system. PloS ONE. 2016; 11(2):0148875.

Beckerman BS, Jerrett M, Martin RV, van Donkelaar A, Ross Z, Burnett RT. Application of the deletion/substitution/addition algorithm to selecting land use regression models for interpolating air pollution measurements in california. Atmospheric Environ. 2013; 77:172–7.

Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, Samet JM. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. Jama. 2006; 295(10):1127–34.

Corani G. Air quality prediction in milan: feed-forward neural networks, pruned neural networks and lazy learning. Ecol Model. 2005; 185(2):513–29.

Ghanem M, Guo Y, Hassard J, Osmond M, Richards M. Sensor grids for air pollution monitoring. In: Proc. 3rd UK e-Science All Hands Meeting.Nottingham: 2004.

Download references

Acknowledgments

The authors are grateful for the financial support of CIHR/NSERC.

This work was supported by a Collaborative Health Research Grant for CIHR/NSERC, and the Alberta Institute for Machine Intelligence (AMII).

Availability of data and materials

The datasets generated or analyzed during our survey are available from the corresponding author upon reasonable request.

Author information

Authors and affiliations.

Department of Computing Science, University of Alberta, Edmonton, Canada

Colin Bellinger, Mohomed Shazan Mohomed Jabbar & Osmar Zaïane

Department of Paediatrics, University of Alberta, Edmonto, Canada

Alvaro Osornio-Vargas

You can also search for this author in PubMed   Google Scholar

Contributions

CB and MSMJ performed the literature review and gather the data used in this survey. CB and MSMJ performed the initial analysis and paper verification. All authors took part in the final analysis of the articles selected for inclusion in this survey. AOV examined and summarized the objectives of the surveyed articles from a biomedical perspective. Based on this analysis, OZ, CB and MSMJ designed the outline and structure of the survey. CB and MSMJ wrote the paper. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Colin Bellinger .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Consent for publication

Competing interests.

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Cite this article.

Bellinger, C., Mohomed Jabbar, M., Zaïane, O. et al. A systematic review of data mining and machine learning for air pollution epidemiology. BMC Public Health 17 , 907 (2017). https://doi.org/10.1186/s12889-017-4914-3

Download citation

Received : 26 April 2017

Accepted : 14 November 2017

Published : 28 November 2017

DOI : https://doi.org/10.1186/s12889-017-4914-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Epidemiology
  • Air pollution
  • Data mining
  • Machine learning

BMC Public Health

ISSN: 1471-2458

systematic literature review on air pollution

Accessibility Links

  • Skip to content
  • Skip to search IOPscience
  • Skip to Journals list
  • Accessibility help
  • Accessibility Help

Click here to close this panel.

ERL graphic iopscience_header.jpg

Purpose-led Publishing is a coalition of three not-for-profit publishers in the field of physical sciences: AIP Publishing, the American Physical Society and IOP Publishing.

Together, as publishers that will always put purpose above profit, we have defined a set of industry standards that underpin high-quality, ethical scholarly communications.

We are proudly declaring that science is our only shareholder.

The health impacts of waste-to-energy emissions: a systematic review of the literature

Tom Cole-Hunter 1,2,3,4 , Fay H Johnston 1,5 , Guy B Marks 1,6,7,8 , Lidia Morawska 1,2,3 , Geoffrey G Morgan 1,9 , Marge Overs 1 , Ana Porta-Cubas 1 and Christine T Cowie 1,6,7,8

Published 1 December 2020 • © 2020 The Author(s). Published by IOP Publishing Ltd Environmental Research Letters , Volume 15 , Number 12 Focus on Energy Transitions and Health Citation Tom Cole-Hunter et al 2020 Environ. Res. Lett. 15 123006 DOI 10.1088/1748-9326/abae9f

You need an eReader or compatible software to experience the benefits of the ePub3 file format .

Article metrics

17499 Total downloads

Share this article

Author e-mails.

[email protected]

Author affiliations

1 Centre for Air pollution, energy, and health Research (CAR), University of New South Wales, Sydney, Australia

2 International Laboratory for Air Quality and Health, Institute of Health and Biomedical Innovation, Queensland University of Technology (QUT), Brisbane, Australia

3 Science and Engineering Faculty, Queensland University of Technology (QUT), Brisbane, Australia

4 Section of Environmental Health, Department of Public Health, University of Copenhagen, Copenhagen, Denmark

5 Menzies Institute for Medical Research, University of Tasmania, Hobart, Australia

6 South West Sydney Clinical School, University of New South Wales, Sydney, Australia

7 Ingham Institute for Applied Medical Research, Sydney, Australia

8 Woolcock Institute of Medical Research, University of Sydney, Sydney, Australia

9 Sydney School of Public Health, and University Centre for Rural Health, Faculty of Medicine and Health, University of Sydney, Sydney, Australia

Tom Cole-Hunter https://orcid.org/0000-0002-6827-6084

Fay H Johnston https://orcid.org/0000-0002-5150-8678

  • Received 11 July 2019
  • Accepted 12 August 2020
  • Published 1 December 2020

Peer review information

Method : Single-anonymous Revisions: 4 Screened for originality? Yes

Buy this article in print

Waste-to-energy (WtE) processes, or the combustion of refuse-derived fuel (RDF) for energy generation, has the potential to reduce landfill volume while providing a renewable energy source. We aimed to systematically review and summarise current evidence on the potential health effects (benefits and risks) of exposure to WtE/RDF-related combustion emissions.

We searched PubMed and Google Scholar using terms related to health and WtE/RDF combustion emissions, following PRISMA guidelines. Two authors independently screened titles, abstracts and then full-texts of original, peer-reviewed research articles published until 20th March 2020, plus their relevant references. Overall quality of included epidemiological studies were rated using an amended Navigation framework.

We found 19 articles from 269 search results that met our inclusion criteria, including two epidemiological studies, five environmental monitoring studies, seven health impact or risk assessments (HIA/HRA), and five life-cycle assessments. We found a dearth of health studies related to the impacts of exposure to WtE emissions. The limited evidence suggests that well-designed and operated WtE facilities using sorted feedstock (RDF) are critical to reduce potential adverse health (cancer and non-cancer) impacts, due to lower hazardous combustion-related emissions, compared to landfill or unsorted incineration. Poorly fed WtE facilities may emit concentrated toxins with serious potential health risks, such as dioxins/furans and heavy metals; these toxins may remain problematic in bottom ash as a combustion by-product. Most modelling studies estimate that electricity (per unit) generated from WtE generally emits less health-relevant air pollutants (also less greenhouse gases) than from combustion of fossil fuels (e.g. coal). Some modelled estimates vary due to model sensitivity for type of waste processed, model inputs used, and facility operational conditions.

We conclude that rigorous assessment (e.g. HRA including sensitivity analyses) of WtE facility/technological characteristics and refuse type used is necessary when planning/proposing facilities to protect human health as the technology is adopted worldwide.

Export citation and abstract BibTeX RIS

Original content from this work may be used under the terms of the Creative Commons Attribution 3.0 license . Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

1. Introduction

Global waste generation has been estimated to double in the decade from 2015 to 2025, from 3 to over 6 million tonnes of waste per day; this rate is expected to continue into the next century, when the estimate increases to 11 million tonnes per day (World Energy Council 2016 ). In parallel, the world is facing an energy sustainability crisis. Heightened electricity consumption increases energy demand, while conversely, greenhouse-gas emissions must be curbed to mitigate climate change. Sustainable energy and waste management requires policies that promote a 'circular economy', balancing product life cycles (from production to disposal), and that minimise adverse economic, environmental, and societal impacts (Beyene et al 2018 , IEA Bioenergy 2018 ). A circular economy reuses and recycles goods, where possible, restoring and regenerating products, components and materials to be at their highest utility and value at all times (IEA Bioenergy 2018 ). The process of waste-to-energy (WtE; also known as 'energy-from-waste') supports a circular economy by reducing landfill volume from municipal solid waste (MSW) by up to 80%, while also generating energy such as through combustion for turbine-driven electricity (Beyene et al 2018 , U.S. Energy Information Administration 2018 ).

Combustion of MSW is the most established method of energy recovery through WtE worldwide, accounting for nearly 90% of the WtE sector (Clean Energy Finance Corporation 2015 , World Energy Council 2016 ). MSW includes domestic, commercial and institutional waste such as plastics, rubbers, wood, metals and paper, which may be combustible or recyclable. The combustible component of MSW is known as refuse-derived fuel (RDF), and is used in a thermal process (incineration, pyrolysis, or gasification) to generate electricity, or heat, fuel gases, and solids as primary recovery products (Beyene et al 2018 ). From a health perspective, the WtE process may have advantages compared to waste management practices that solely rely on landfill sites that are associated with contamination of the air (e.g. volatile organic compounds) alongside water and soils (Vrijheid 2000 ). However, the WtE process may emit higher concentrations of carbon dioxide (CO 2 ), sulfur dioxide (SO 2 ), and nitrogen oxides (NO x ) per unit electricity produced compared to other forms of energy such as natural gas or renewables (O'Brien 2006 ). The WtE process involves the combustion of RDF components for which emissions may also include persistent organic pollutants such as dioxins (Albores et al 2016 ). This concern is offset, to some extent, in modern, well-run WtE plants that emit lower concentrations of these pollutants compared to coal and oil-fired power plants or traditional incineration of MSW (US EPA 2016 ). Hence, WtE processes may have both beneficial and adverse impacts on the emission of airborne toxins, and consequently on health, relative to alternative waste disposal and energy generation processes.

The WtE sector is already well established in Europe and provides up to 8% of electricity and up to 15% of domestic heating needs (World Energy Council 2016 , Zafar 2018 ). As of 2008, 475 European WtE plants processed an average of 59 million tonnes of MSW creating revenue of US$4.5 billion each year (Zafar 2018 ). In Scandinavia, Denmark repurposes 54% of its MSW as RDF (Zafar 2018 ). Meanwhile Sweden, which has employed WtE since the 1940s, is aiming to match the repurposing of 99% of its local MSW (two million tonnes annually) with an equivalent amount of imported MSW as RDF (Fredén 2018 ). In 2012, approximately 600 WtE plants across 35 different countries were estimated to combust 130 million tonnes of MSW (Hoornweg and Bhada-Tata 2012 ), with the sector growing at a compounded annual rate of nearly 10% (World Energy Council 2016 ). Outside of Europe, the process is being adopted with eagerness, using the established Waste Incineration Directive (WID 2000/76/EC) of the European Commission as a guide for monitoring and regulating WtE emissions (Clean Energy Finance Corporation 2016 ). In 2016, the USA alone operated 71 WtE plants generating approximately 14 billion KWh of electricity from 30 million tonnes of RDF (U.S. Energy Information Administration 2018 ). In the Asia-Pacific region, China is the fastest growing adopter of WtE, recently planning 125 new plants to double national capacity (World Energy Council 2016 , Zafar 2018 ). China, one of the major importers of MSW, has restricted imports of certain materials (e.g. plastics, paper) to reduce local widespread environmental contamination (Retamal et al 2019 ), challenging major exporters of MSW such as Australia (Cheng and Hu 2010 , Downes and Dominish 2018 ). Responding to this challenge, Australia has estimated that a national shift towards WtE presents an opportunity to repurpose 20+ million tonnes of MSW otherwise going to landfill annually and avoid 9 million tonnes of CO 2 (equivalent) emissions by replacing fossil-fuel combustion while meeting 2% of national baseload electricity demand (Clean Energy Finance Corporation 2016 ). Hence, it is timely to consider the place of WtE in the energy transitions landscape and, in particular, to consider its impact on air quality and health.

Despite the growing global interest in WtE, the public health implications of combusting RDF remains little studied. There has been no previous systematic literature review of the health impacts associated with WtE, although several reviews on municipal waste incineration have been published. In 2019, a systematic review on the evidence of health effects from waste incineration (2002 to 2017) was published in response to several new incinerators proposed for use within Australia (Tait et al 2020 ). The literature review, which did not include WtE facilities, concluded that the available evidence likely under-estimated the health effects of exposure to incineration emissions due to most studies being of low quality and only examining a limited subset of potential exposure and disease pathways (Tait et al 2020 ). Other earlier reviews on the health impacts or risks of incineration and resulting emissions have focused on hazardous (industrial) or unsorted (municipal) solid waste, rather than sorted RDF for WtE. These reviews concluded that the evidence is insufficient to support an association between a specific waste incineration process and adverse health effects (Vrijheid 2000 , Hu and Shy 2001 , Giusti 2009 , Porta et al 2009 , Cordioli et al 2013 ). Associations between exposure to emissions and health outcomes such as increased risk of lung/throat cancer or ischaemic heart disease (Hu and Shy 2001 ), as well as non-Hodgkin's lymphoma and soft-tissue sarcoma (Giusti 2009 ), have been reported, however, the findings are inconsistent. The reason for this has been suggested to be due to poor methods of exposure characterisation which have relied on distance from source or self-reporting exposure, rather than measured or modelled pollutant concentrations (Hu and Shy 2001 , Cordioli et al 2013 , Hoek et al 2018 , Tait et al 2020 ). More consistent associations have been reported between exposure to emissions and elevated biomarkers of organic chemicals or heavy metals in urine and blood (Hu and Shy 2001 ).

In their review on MSW incineration without energy recovery (i.e. not WtE), Tait et al ( 2020 ), recommended future studies be conducted on the health impacts of WtE, including studying content and volume of feedstock (waste), combustion specifications, consideration of multiple exposure pathways, reporting of a larger array of health outcomes, and controlling for potential confounding factors (Tait et al 2020 ). Other reviews have suggested that previous limitations of incineration studies could be addressed by large, prospective, multi-site cohort studies with personal measurements of exposure, based on knowledge of biological pathways and toxicological effects of specific compounds (Giusti 2009 , Porta et al 2009 , Hoek et al 2018 ), however such studies can be expensive and sample size (of the study population) can be a limiting factor. Clearly, the expanded interest in WtE facilities yet current lack of evidence on health impacts with their operation requires stringent oversight to safeguard environmental and health outcomes.

Our aim was to conduct a systematic review on the potential health effects associated with exposure to airborne emissions from WtE processes (including RDF combustion). The primary motivation for the current review was the perceived lack of data on the potential for health impacts of WtE processes and emissions, and the increasing growth in demand in regions where WtE has not yet been adopted on a widespread basis. As WtE has been promoted worldwide as a potentially sustainable form of both waste management and electricity generation, we considered it timely to ascertain the extent and breadth of evidence from published studies with health-related data or information associated with airborne emissions from WtE processes. The different types of study designs for studies included in our review include epidemiological, environmental monitoring, health risk assessments/health impact assessments, and life-cycle analyses (detailed below in Methods).

2.1. Literature search strategy

We conducted a systematic search of PubMed and Google Scholar, supplemented by a hand search of bibliographies of the articles included for full text screening. We used PubMed as the primary database source given our review was focused on health outcomes and PubMed is considered to be the most comprehensive health database. We used Google Scholar as a secondary source to identify relevant literature that PubMed does not catalogue, as done previously for hazardous waste reviews (Cordioli et al 2013 ).

Search terms and the Boolean operators (string) that we used were as follows:

"air" AND "health" AND "energy" AND "waste" AND "energy from waste" OR "waste to energy" OR "incineration" OR "refuse derive$ fuel" AND "air pollution" OR "air quality" OR "emission"

'Incineration' was chosen as it is the industrial term that represents 'combustion' and 'burning'. In addition, 'air pollution', 'air quality' or 'emission' terms were used to avoid pollutants or hazards associated with other emissions. Two investigators (TCH, CC) independently screened titles, abstracts and full-texts for inclusion or exclusion of articles. Where there was variation between the two investigators, this was resolved by reviewing the the full-text article a second time until agreement was reached.

The inclusion criteria used for selection of eligible articles were as follows:

  • (a)   Published in English.
  • (b)   Published up to and including the 20th March 2020.
  • (c)   Included an abstract and be full-text accessible.
  • (d)   Reported original research.
  • (e)   Published in a peer-reviewed journal.
  • (f)   Related to anthropogenic waste, municipal solid waste, air pollution emissions, and relevant to human health.

Exclusion criteria included articles that related to:

  • (a)   Hospital or medical waste, composting of waste, or agricultural waste.
  • (b)   Combustion of biomass fuel for cooking and heating in low-income settings.
  • (c)   Review papers.

2.2. Literature review and synthesis

We followed the approach (criteria) suggested by the PRISMA guidelines for performing and reporting the flow of a literature review process (e.g. figure 1 ) (Moher et al 2009 ). We synthesised study findings by grouping the articles by different study designs (methods). We used the following groupings: epidemiological (examining direct associations between exposure and health risk); environmental monitoring (emissions or exposure assessments or modelling); health risk assessment (focused, standard methodology to estimate risks related to a single or a mix of pollutants; applying health risk estimates from epidemiological studies to quantify the health burden due to the exposure of interest in a defined population), or health impact assessment (broader methodology that assesses the public health impacts to inform decision making; often including HRA methods or other health risk findings) (Gulis 2017 ); and, life-cycle analyses (LCA; quantifying carbon-related impacts and indirect health impacts, with some LCAs also addressing direct health impacts). We used a standardised series of tables to summarise the studies and to list exposure assessment methods, health outcomes, summary results, and risk of bias. We provided an overall quality rating for epidemiological studies, similar to the Navigation framework previously developed (Woodruff and Sutton 2014 ) and demonstrated (Johnson et al 2014 ). The Navigation framework was developed in recognition that usual quality frameworks used for reviewing health studies, such as Cochrane, do not necessarily translate well to studies of environmental exposures, due to the nature of the exposure and difficulty in conducting randomised trials. As there were few relevant epidemiological studies, the criteria were amended slightly to ensure relevance depending on the study design. For example, we included mention of sensitivity analyses in modelling studies or explicit statements about assumptions used in the analyses. However, we did not critically appraise or scrutinise the assumptions or the software used in the LCA models as this was beyond the scope of our study.

Figure 1.

Figure 1.  Systematic literature review flow diagram.

Download figure:

3. Results and discussion

3.1. literature search results.

The PubMed literature search identified 258 relevant primary records (articles) for review. The Google Scholar search identified 11 unique records relevant for review. As such, the complete search gave a combined total number of 269 unique records for consideration.

After two investigators independently reviewed the titles of these 269 records, 137 records were identified to be appropriate for abstract screening (which removed 74 records). Sixty-three records were subsequently identified as eligible for full-text review, leading to the exclusion of 46 records. Finally, 17 full-texts were selected for our review synthesis, plus two of their references to give a total of 19 full-texts to be synthesised (figure 1 ).

The 19 included articles all related to combustion of MSW as RDF or in WtE facilities or processes. MSW incineration studies were included if they presented information or data related to emissions that were relevant to WtE processes, such as incineration of RDF. All included articles were published in the past 15 years, reflecting the increasing interest in WtE. Most studies comprised health impact or risk assessments/risk modelling (n = 7), followed by life-cycle assessments (n = 5), environmental monitoring studies (n = 5), with only two being epidemiological studies.

3.2. Synthesis and discussion of findings

To our knowledge, this is the first systematic literature review focused primarily on studies of the health effects associated with WtE-related air emissions. We found that while implementation of WtE technologies is increasing, the majority of incineration-health studies to date do not specifically address the combustion of sorted waste (RDF) for WtE (shown to be different than MSW due to waste composition characteristics by environmental monitoring studies—reported on below). Previous reviews have focused on the health impacts of waste incinerators (Cordioli et al 2013 , Tait et al 2020 ) the economic implications of WtE technologies (Beyene et al 2018 ), exposure assessment methods in epidemiological studies of industrially contaminated sites (Hoek et al 2018 ) or waste incinerators, and the health impacts of general waste management practices (Giusti 2009 ). There are numerous epidemiological (e.g. cohort) studies on the health effects of other waste management risks including landfill leaching, sewage contamination and ionising radiation, yet few on air pollution emissions from RDF combustion. Due to the small number (n = 2) of epidemiological studies that directly measured health outcomes associated with WtE processes we believed it was not appropriate to meta-analyse the evidence for WtE health effects. However, we reviewed studies of environmental monitoring and health risk assessments in order to contribute to the evidence base for decision making. The following synthesis details the contributions to this evidence base, from studies detailing process emissions to health risk assessments.

3.3. Epidemiological studies of health outcomes

The direct health effects of exposure to emissions from combustion of RDF for WtE have been little studied. This is likely to be partly due to the difficulty of quantifying population health effects from generally inaccurate or low levels of exposure (Vrijheid 2000 ). This is despite previous recommendations that large prospective cohort studies with direct exposure and biomarker measurements be preferentially funded and performed (Giusti 2009 ).

We found only two epidemiological studies relevant to exposures to WtE facilities or RDF emissions. One epidemiological before/after cohort study was performed in Italy among 380 individuals residing near a new WtE facility, with exposure assessed before and one year after operation began (Ruggieri et al 2019 ). In this biomonitoring study, chromium (but not other heavy metal) concentrations were higher in the urine of participants predicted to be exposed to WtE emissions compared to unexposed but otherwise comparable participants (Ruggieri et al 2019 ). However, this finding was applicable in both the baseline and follow-up year, and so the result cannot be directly attributed to operation of the WtE facility. Interestingly, concentrations of other heavy metals were higher in the control subjects, and so were attributed to other sources of personal exposure such as fish intake (arsenic) and tobacco smoke (cadmium) (Ruggieri et al 2019 ). Hence, residing near the WtE plant was not associated with greater exposure to heavy metals. We considered the study to be of good quality, having used dispersion modelling to assign exposures and conducting before and after health outcome measurements in an 'exposed' and 'unexposed' group. Validation of the emissions modelling by environmental sampling could have improved exposure assessment. There is further follow-up planned for this cohort which is expected to provide additional data (Ruggieri et al 2019 ).

A recently published birth cohort study was conducted in Taiwan and investigated childhood social development in children residing near an incinerator (Lung et al 2020 ). The study of nearly 20 000 subjects (for which approximately five percent were considered exposed), reported a transitory negative effect on childhood social development, for children living within 3 km of a MSW incinerator, although this effect was apparent at six months but not evident at 18 months. A limitation of this study was considered to be the coarse exposure assessment applied to subjects which has the potential to lead to exposure misclassification. Exposure assessment ('whether there were incinerators within 3 km of their place of residence') and health outcome reporting were both coarse and subjective, with both being self-reported by parents (Lung et al 2020 ).

We conclude that the results from the two epidemiological studies provide little evidence of an adverse impact of WtE air emissions on health outcomes. See table 1 for further details of included epidemiological studies.

Table 1.  Epidemiological studies of health outcomes of waste-to-energy/RDF processes.

a Exposure assessment; Outcome assessment; Control for confounding; Study sample size; Conflict of interest statement b Low, moderate or high potential for concern based on quality and outcomes of study

3.4. Environmental monitoring

While studies of emissions inventory profiles do not include health outcomes, they may provide valuable information on the potential pathways and hazards posed by incineration of MSW components comprising RDF, with the potential for carcinogenic or toxic emissions relevant to WtE processes. Our review found five articles which reported on emissions testing and environmental monitoring of WtE facilities. In general, we found that the articles related to emissions monitoring predominantly fell into three categories: (1) the first related to estimating pollutant emissions of concern; (2) the second related to the need for monitoring to ensure efficacy of treatment technologies in removing/reducing pollutants; and (3) the third related to the need for appropriate monitoring to determine the influence of the feedstock on pollutant formation.

Of greatest concern for health is the combustion of plastic MSW (composed of hydrocarbon/oil-products), which is concentrated in RDF for WtE, and which emits organic and chlorinated/fluorinated compounds (e.g. dioxins), polychlorinated biphenyls, furans, chlorophenols, and mono- and polycyclic aromatic hydrocarbons (Karunathilake et al 2016 ). Notwithstanding, two environmental monitoring studies reported that after WtE upgrades to an Italian incinerator facility which included stricter emission-control measures primarily aimed at reducing dioxin emissions, particulate matter (PM) emissions also declined (Buonanno et al 2010 , 2011 ). This indicates that controlling emissions for critical contaminants such as dioxins and furans may also have a beneficial effect of leading to a reduction in PM, a standard air pollutant. The health risk of toxics predominantly relate to cancer, neurological and adverse birth outcomes and are considered to pose a greater risk to health than the standard regulated air pollutants such as PM and gaseous compounds. However, exposure to even low levels of PM is not benign and many epidemiological studies point to a range of risks associated with PM including increased risk of mortality, cardiovascular morbidity, lung cancer and more (Hime et al 2018 ). Thus changes to existing treatment facilities that improve emissions controls for both types of pollutants are beneficial from the standpoint of exposure minimisation.

The review also reports on articles which compared or discussed monitoring campaigns and/or trials of varying technologies. In one study a two-stage dry treatment system was shown to remove harmful acid gases (hydrogen chloride, SO 2 ) from WtE emissions even with a widely-varying (potentially highly chlorinated) waste stream (Dal Pozzo et al 2016 ). This is an example of a monitoring program which can help provide evidence of efficacy of treatment technologies. Two articles reported on the influence of feedstock on pollutant concentration emissions.

The articles indicated that rather than combusting RDF directly for electricity generation, experimentation suggested that mixing certain proportions of RDF components (e.g. certain plastics, wood chips) with traditional fuels (e.g. coal) for combustion and to replace electricity for industrial heating applications (e.g. cement kilns), has the potential to reduce sector or total emissions of health-relevant chemicals (e.g. dioxins, mercury) (Chen et al 2014 , Richards and Agranovski 2017 ).

We conclude from the results of the environmental monitoring studies that there is a need for regulation of the feedstock used (e.g. removing food waste) for RDF and WtE facilities to maximise complete combustion and minimise carcinogenic/contaminant emissions (e.g. volatile organic compounds), more so than the treatment technology used. See table 2 for further details of included environmental monitoring studies.

Table 2.  Environmental monitoring of waste-to-energy/RDF processes.

a Exposure assessment; Study sample size; Conflict of interest statement b Low/decreased, moderate or high potential for concern based on quality and outcomes of study

3.5. Health risk/impact assessment studies

We found seven studies comprising HRAs or HIAs of WtE facilities or RDF emissions. In table 3 we outline the health outcome assessed in a majority of the HRAs, including the hazard index (HI), hazard quotient (HQ), lifetime cancer risk (LCR) and other indices (4th column). These represent indices where cancer and non-cancer risks are considered for various chemicals of concern (3rd column, table 3 ), e.g. heavy metals, VOCs, organic compounds such as dioxins and furans, and so on. Some of the HRAs/HIAs also considered air pollutant emissions such as NO x , PM, and sulfur oxides (SO x ). The risk of exposure is based on modelled estimates of the chemicals/pollutants emissions from each WtE facility or alternative waste disposal method. Some of the studies have used proprietary software which includes the exposure-response functions for the chemical/pollutant of concern, which we list in table 3 (3rd column).

Table 3.  Health risk/impact assessment/modelling studies of waste-to-energy/RDF processes.

a Exposure assessment; Study sample size; Appropriate assumptions; Conflict of interest statement b Low, moderate or high potential for concern based on quality and outcomes of study

These studies generally showed that the risk to or impact on health from exposure to WtE and RDF incineration emissions are not substantially elevated above 'background' risk levels (Mindell 2005 , Roberts and Chen 2006 , Krajčovičová and Eschenroeder 2007 , Rovira et al 2010 , Ollson and Whitfield Aslund et al 2014 ). They also point to lower emissions from well-run WtE facilities compared to landfill (Paladino and Massabò 2017 ) and traditional incineration (Krajčovičová and Eschenroeder 2007 ) or when RDF is substituted for fossil fuel for incineration (Rovira et al 2010 ).

Six of the HRA studies estimated that exposure to WtE emissions was unlikely to increase incremental LCR or HQ for cancer risk (Roberts and Chen 2006 , Krajčovičová and Eschenroeder 2007 , Rovira et al 2010 , Ollson and Knopper et al 2014 , Li et al 2015 , Paladino and Massabò 2017 ). Two HRAs reported lower cancer risk for exposure to WtE emissions compared with incineration emissions (Karunathilake et al 2016 ) or as substitution of RDF for fossil fuels in cement production (Rovira et al 2010 ). One HRA estimated that cancer risk from exposure (all pathways) to WtE emissions (mainly dioxin) would be lower than for exposure to landfill emissions, and estimated that agricultural (milk and meat) product ingestion was a more important exposure pathway than for inhalation of WtE emissions (Paladino and Massabò 2017 ).

A health risk assessment conducted in Slovakia compared a traditional open-air (uncontained) MSW incinerator with a modern WtE plant, and found that the former increased the cancer risk 10–80 times above the background level, while the WtE plant presented a less than one-in-a-million excess risk of cancer (Krajčovičová and Eschenroeder 2007 ). That HRA estimated a substantially decreased cancer risk when MSW is sorted for RDF and its incineration emissions are properly controlled (contained) as advocated in modern, well-run WtE. In China, a more recent HRA estimated that under normal conditions, operational levels of emissions from WtE are unlikely to cause adverse health (incremental lifetime cancer) risks among nearby residents, with risks estimated for lifelong exposure through direct inhalation of ambient emissions and landfilling of bottom/solid ash residues (Li et al 2015 ). The exception to this was risk of chromium exposure which slightly exceeded the tolerance value (Li et al 2015 ). However, Li et al ( 2015 ) reported that all scenarios tested were sensitive to the model inputs and estimated that during abnormal operation (e.g. malfunction of control systems) the WtE facility could also carry an elevated risk due to inhalation of acid gas (hydrogen chloride).

Four HRAs assessed non-cancer risks (Mindell 2005 , Roberts and Chen 2006 , Ollson and Knopper et al 2014 , Li et al 2015 ). Ollson et al ( 2014a ) estimated that abnormal operation of a Canadian WtE facility could lead to infant consumption of breast milk contaminated with dioxins and furans (Ollson and Knopper et al 2014 ). Two of the HRAs (Li et al 2015 , Karunathilake et al 2016 ) estimated no increased risk of non-cancer health effects from operation of their respective WtE facilities. Modelling studies of UK WtE plants estimated premature (total non-traumatic) deaths and respiratory-related hospital admissions to be less than or equal to one-in-a-million above background rates (Mindell 2005 ), and overall risk of dying to be 1 in 4 million for any year (Roberts and Chen 2006 ). It should be noted that these UK studies were either funded by the proponent company for the WtE facility, or written by previous employees of related boards/companies.

It is clear from these studies that the choice of scenarios and model inputs can influence the risk findings, and so it is important that sensitivity analyses be conducted. Of note, Li et al ( 2015 ) reported that all of the scenarios studied in their analyses (WtE, landfill, and material recovery and composting) were sensitive to the inputs used for the reference concentrations and the landfill gas collection rates. In sensitivity analyses, the HI for the WtE option increased the most, indicating that there needs to be careful selection of the reference criteria values and that sensitivity analyses are crucial for better understanding operational limitations of WtE facilities and to avoid abnormal operations or malfunction events.

Together, we conclude that the HRA results show that under normal operating conditions there is little to no evidence of an increased risk of cancer or non-cancer effects in humans, as WtE facilities are capable of lower emissions (except for a predicted potential higher emission of chromium) than existing waste management practices of landfill and traditional incineration. However, close attention is required to ensure operational limits are not exceeded, as such conditions are estimated to be associated with increased risk of dioxin exposure (one HRA) and potentially hydrogen chloride gas exposure (one HRA). This highlights the need for appropriate sensitivity analyses to be conducted during the HRA process, along with careful selection of reference health criteria and consideration of the fuel used for combustion. See table 3 for further details of included health risk/impact assessment studies .

3.6. Life-cycle analyses

A total examination of the environmental, social, and economic impact associated with all stages of a product's life, from raw material extraction to final product disposal (e.g. landfill or WtE), is termed a life-cycle analysis/assessment (LCA) (Muralikrishna and Manickam 2017 ). An LCA is distinct from a HRA in that an LCA considers the full life-cycle of a product, from production to disposal, while a HRA typically only considers one stage of the life-cycle while focusing on a health impact. For example, an LCA for WtE will consider the impacts of not only the resulting toxic contaminants in ash and air emissions, but also emissions which have a greenhouse gas impact such as carbon dioxide, as well as the impact of the fuel used for the WtE facility. Besides health impacts, LCAs can determine equitability of a product's environmental impact, and can determine if overall impact (both to health and the environment) of one waste management process is more favourable than another. For example, one may ask if exposure to atmospheric emissions from WtE is less harmful to health than from unsorted (mass) waste incineration or landfill leachate, also taking into account health impacts of climate change due to greenhouse gas emissions from each technology—however, none of the studies considered climate change in relation to health outcomes. Future LCA studies of new energy technologies could be important in estimating direct and immediate health impacts (due to a change in pollutant emissions) balanced with estimating the potential for indirect and delayed health impacts due to increased greenhouse gas emissions from climate change.

Nevertheless, our review reports the results of five LCAs. As with the HRAs reported above, two of the LCAs predict lower pollutant emissions from combustion of RDF (sorted for WtE), compared to incineration of unsorted MSW (still producing electricity). A Canadian LCA for a WtE facility estimated lower cancer and non-cancer health risk per unit of electricity generated with RDF than for unsorted MSW incineration (Karunathilake et al 2016 ). Similarly, a lower health risk was attributable to the sorting and use of relatively high calorie, low toxicity waste for RDF (e.g. wood, paper, plastics, textiles, and rubbers; sorted, treated, shredded, and combusted to produce approximately 4 MWh of energy per tonne) (Reza et al 2013 ) compared to the use of coal. Although Reza et al estimated lower heavy metal emissions for RDF used in WtE facilities, an exception was an estimated increase in lead emissions. This LCA was conducted with a focus of comparing environmental benefits of the two feedstocks for use in cement kilns.

Two of the LCAs estimated greater impacts from WtE processes compared with other waste management processes. Scipioni et al ( 2009 ) predicted lower emissions of respiratory related pollutants but greater potential for exposure to carcinogens, climate change pollutants (mainly CO 2 ) and radiation, when comparing dry and wet fly gas scrubbing, with and without WtE processing. Tan and Khoo's ( 2006 ) LCA analysis comparing landfill, WtE incineration, and recycling and composting stated that 'energy gained from incineration of waste materials is outweighed by the air pollution generated' and estimated that recycling and composting would result in the least ecosystem impact. The authors acknowledged the generally 'wetter' conditions of their MSW (in Singapore) which they suggested might be more suitable for composting. Although the article mentioned modelling of disability adjusted life years (DALYs; as a health measure) we could not see where these results were calculated or presented. In addition, the assumptions used in this LCA were not explicitly mentioned, so it is difficult to determine how a change in model inputs might influence these findings.

In Passarini et al's ( 2014 ) LCA which compared various upgrades of an incinerator to enable functioning as a WtE plant, they estimated that concentrations of heavy metals in the fly and bottom ash were the main contributor to carcinogen endpoints and these remained constant over time. However, they estimated decreases in carcinogens and particulate matter in airborne emissions during operation as a WtE facility. The LCA concluded that human health improvements were expected with WtE operations due to both the lowered emissions and the predicted improvements associated with greenhouse gas mitigation.

Most of the LCAs reviewed used accepted international methods for LCA, such as ISO standards along with the use of specialist software such as SimaPro and impact assessment methods such as Ecoindicator99. As with the HRA studies, some of the LCA studies highlighted the variability in calculated health risks to be dependent on the reference criteria and dose and other model inputs, thus indicating the necessity for sensitivity analyses to be conducted.

In general, we conclude that the predictions from the majority of LCA studies indicate that emissions from, and therefore health risks associated with, WtE plants are lower than for landfill and traditional incineration. However, an increased potential for health risk is highlighted for lead (Reza et al , 2013 ) and other heavy metals in the bottom and fly ash (Passarini et al , 2014 ) that may be emitted in later stages of the life cycle (following combustion of RDF for WtE). See table 4 for further details of included life-cycle analyses .

Table 4.  Life cycle analyses of waste-to-energy/RDF processes.

3.7. Implications

Our review indicates that there is a dearth of studies on the potential health impacts of WtE-related emissions, even in countries where WtE facilities have been in operation for some years (such as Sweden); however, some practical implications can be drawn from the limited research done. This has implications for the emerging WtE sector.

As a consequence of the lack of health studies related to WtE facilities, inference is often drawn from exposure studies to health-related emissions common to combustion of MSW. These studies might provide some indication of potential impacts from WtE process emissions, albeit newer technologies and tighter restrictions of feedstock appear to be implemented in WtE facilities. An example of this is exposure to dioxin emissions from older MSW incinerators, where past epidemiological studies have reported weak to moderate associations between dioxin emissions and an increased incidence of cancers including non-Hodgkin's lymphoma (Viel et al 2008 ) and sarcoma (Zambon et al 2007 ) among nearby residents and incinerator workers. These studies were conducted prior to the lowering of incinerator emission volumes through introduction of stricter regulations, and so the findings cannot be directly extrapolated to current WtE technologies. Furthermore, it should be acknowledged that due to the varying waste streams in different geographic regions and for different facilities, research evidence from one country may not accurately or wholly inform policy or practice in other countries/regions. Older studies have also tended to study the incineration of unsorted MSW.

The extent to which the existing evidence base reviewed here can support a causal association between exposures to airborne emissions from WtE facilities and adverse health impacts, is very limited. While the evidence base, as a whole, is weak and there is little evidence of effects under normal operating conditions of WtE plants, the review has highlighted some potential areas for further study. There is clearly a place for more studies of the potential for health impact from WtE facilities, using the various study types included in this review: epidemiological studies; HRAs; LCAs; and, environmental monitoring. However, given the cost of completing well designed and adequately powered epidemiological studies, and the difficulty in ensuring sufficient sample size or a non-exposed control group, it is likely that other methods such as health risk assessment, along with exposure modelling, with or without LCAs, will prove to be useful in assessing new WtE facilities.

Notwithstanding the above, there is a need for well-designed epidemiological studies of exposure to WtE emissions. Such studies could provide empirical data for subsequent HRAs and LCAs, but need to address issues of exposure misclassification potential which has occurred in the past (Forastiere et al 2011 ) such as using distance based measures as an exposure proxy. The collection of environmental monitoring data of environmental media, e.g. air and soil, in the vicinity of WtE facilities, along with emissions monitoring, would also facilitate validation of the exposure models used in epidemiological and HRA studies.

There is an argument to be made for more standardised exposure assessment methods and standardised measurement units, reference criteria and models, in studying the health impacts of WtE emissions, especially for more harmful components such as dioxins, given the variety of methods presented in the LCAs and HRAs reviewed for this paper. Further, we agree with previous researchers that modelling studies, such as HRAs and LCAs, should explicitly outline model input assumptions and associated uncertainties given their influence on model outcomes (Scipioni et al 2009 ).

Some of the studies included in this review have highlighted the need for special consideration of the feedstock used for RDF and WtE facilities, given that it is one of the critical issues affecting contaminant emissions, over and above the treatment technology used. For instance, the World Energy Council ( 2016 ) stated that dioxin (and other toxins such as furan) emissions from RDF can be reduced by nearly 100% with the implementation of regulatory emission-control strategies within the WtE sector, such as controlling the nature of the feedstock. This can result in emission volumes which are lower, per equivalent energy unit, than for coal or gas-powered power plants (World Energy Council 2016 ). Regulating the pre-sorting of waste for WtE processes can help to maximise complete combustion and minimise carcinogenic emissions (Reza et al 2013 , Karunathilake et al 2016 ). Others state that food waste, for example, should be removed from the RDF stream as it yields little exportable energy in WtE processes (Diggelman and Ham 2003 ). Some researchers state that the higher calorific value of RDF results in more complete (higher temperature) combustion, resulting in less emissions of other potentially toxic pollutants such as volatile organic compounds (Friege and Fendel 2011 ).

Although not strictly airborne emissions, there is a need for increased scrutiny of the use/disposal of bottom and fly ash given at least two studies estimated increased concentrations of chromium and dioxin in bottom/fly ash. Others have advocated that WtE residuals should be re-purposed (isolated) as construction 'filler' rather than go to landfill (Tan and Khoo 2006 , Passarini et al 2014 , Malakahmad et al 2017 ). While toxins such as dioxins and furans found in breast milk, or heavy metals found in urine, are not themselves measured health impacts, they could cause health impacts with accumulation over time. The WHO recognises that, due to the omnipresence of dioxins, the whole population has background exposure which is not expected to affect human health (such as the levels found in our included studies); however, as these toxins have a high toxicity, efforts need to be taken to reduce additional exposure such as from waste incineration (WHO 2020 ). As such, we can best prevent or reduce this exposure by continuing to measure directly at the source.

More broadly, LCAs including health impacts of MSW stream management should not only consider direct pollutant emissions, but also the potential effects of repurposing waste such as reducing and recycling, and the impact on greenhouse gas production and transport emissions (Giusti 2009 ). A fair and full LCA may show that the most benefit from RDF/WtE processes may come from fuel substitution for industrial processes such as cement manufacturing (Reza et al 2013 , Richards and Agranovski 2017 ), within which combustion ash could be isolated to further reduce the environmental impact from landfill. While WtE may have a larger carbon footprint (CO 2 emissions) compared with recycling of materials (e.g. plastic) (Tan and Khoo 2006 ), it generally emits lower concentrations of greenhouse gases (CO 2 , methane) than landfill (Giusti 2009 , Clean Energy Finance Corporation 2015 , Malakahmad et al 2017 , Beyene et al 2018 , Murray 2018 , Orru et al 2019 ). Furthermore, WtE technology (e.g. dry treatment of flue gas) has the ability to offset traditional (fossil fuel) combustion for electricity generation and thereby potentially reduce total emissions of greenhouse gases or criteria air pollutants (Scipioni et al 2009 ). These are all important considerations from a broader public health perspective.

LCA methodology appears to be well suited to provide useful information for the planning and design stage of waste management facilities, as they allow identification of alternative processes and treatment requirements, and so can enable decisions on long term infrastructure investments which benefit health not only locally, and more broadly. We recommend that in regions where WtE has not yet been fully adopted, that LCA incorporating HRA, should be undertaken using local data inputs and with local conditions in mind.

As many regions of the world are needing to manage unprecedented volumes of waste, and at the same time are also experiencing slow implementation of cleaner/safer technologies, the risks related to waste management are likely to remain a challenge for years to come. Using RDF for WtE may address a gap in the circular economy for recovering energy from waste, and while seen as a renewable resource (Natural Resources Canada 2015 ), decision-makers should appropriately assess applications for new WtE facilities, taking a precautionary but not inhibitory approach, in light of the lack of rigorous health evidence.

3.8. Conclusion

We have found a dearth of well-conducted epidemiological studies investigating the health risks of exposure from WtE processes. The limited evidence from the two epidemiological studies, along with HRAs, LCAs and emissions monitoring studies suggests that the risks to human health from emissions of appropriately designed, properly managed (including feedstock), state-of-the-art WtE incineration plants are relatively lower compared to prevailing alternative waste management practices, including incineration of unsorted waste (without energy recovery) and land fill. Importantly, the waste management hierarchy recommends an emphasis on the reduction of material going to waste before it is re-purposed or recycled, as it is clear that the input waste stream can substantially influence pollutant emissions.

While WtE practice might be a reasonable option for mitigating waste management and energy security issues, its implementation requires proper design, operation, and emissions management (monitoring) and control, as well as ongoing environmental and health monitoring and surveillance to maximise both economic and environmental benefits while minimising health impacts or risks. With respect to planning and design of WtE facilities, it is important that health risk assessments supported by comprehensive exposure monitoring, and robust modelling (e.g. detailed emissions modelling plus atmospheric modelling and real population data) be conducted for proposed WtE facilities to ensure that protective measures are optimally designed and emissions criteria appropriately implemented. Furthermore, close attention to health data used and assumptions made for reference doses, exposure duration and frequency, and concentration-response functions, is needed. It is equally important for HRAs and LCAs to include sensitivity analyses to test such assumptions. Future reviews will be reliant on additional well conducted epidemiological studies or HRAs and LCAs and, exposure modelling and monitoring, to further our knowledge in this area.

Acknowledgments

At the time of submission, Tom Cole-Hunter received funding in the form of a fellowship from the Centre for Air pollution, energy, and health Research (CAR), a National Health and Medical Research Council (NHMRC) Centre for Research Excellence. NHMRC did not influence the decision to publish this manuscript. We would like to acknowledge Martine Dennekamp of the Victoria Environmental Protection Authority for guidance on policy and regulatory requirements. We would also like to acknowledge the constructive comments of three anonymous journal peer reviewers.

Data availability statement

No new data were created or analysed in this study.

February 8, 2024

Air Pollution Threatens Millions of Lives. Now the Sources Are Shifting

As EPA tightens air pollution standards for particulate matter, new research suggests some components of that pollution could worsen with climate change

By Virginia Gewin

Hairdresser applies hair care product with spray

Sergii Kolesnikov/Getty Images

Particle-based ambient air pollution causes more than 4 million premature deaths each year globally, according to the World Health Organization. The tiniest particles—2.5 microns or smaller, known as PM 2.5 —pose the greatest health risk because they can travel deep into the lungs and may even get into the bloodstream.

Although total PM 2.5 levels have decreased 42 percent in the U.S. since 2000 as a result of clean air regulations, scientists are concerned about the health impacts of even low levels of such pollution. The U.S. Environmental Protection Agency lowered the annual national air quality standard for PM 2.5 from 12 to nine micrograms per cubic meter (µg/m 3 ) this week. EPA administrator Michael Regan said in a press conference that officials estimate the new standard will save up to $46 billion dollars in avoided health care and hospitalization costs by 2032. “Health benefits will include up to 800,000 avoided cases of asthma symptoms, 4,500 avoided premature deaths, and 290,000 avoided lost workdays,” he said. The World Health Organization adopted an even lower 5 µg/m 3 standard in 2021, citing the growing evidence of deadly harm.

Beyond investigating their size, scientists are also digging into the chemistry of airborne particles, which, unlike other regulated pollutants such as lead and ozone, encompass a wide array of solid and liquid particles from soot to nitrate. Some airborne particles are directly emitted from car tailpipes or industrial sources; others form in the atmosphere. And the balance of those is shifting. To help states meet the tougher air standards, scientists will need more detailed studies of particle sources.

On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing . By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

In July 2022, for the first time in more than a decade, teams of scientists conducted an intensive campaign to characterize what’s in the summertime soup of particles that New York City residents breathe. The researchers measured the chemical makeup of PM 2.5 over the course of a month.

The team found that the PM 2.5 was 80 to 83 percent organic, or carbon-based —up from roughly 50 percent in 2001, according to the study, which was published January 22 in ACS ES&T Air . “Over the past 20 years, summertime particulate matter has shifted to organic aerosols due largely to the successful reductions of sulfate and other inorganic compounds,” says Tori Hass-Mitchell, the study’s lead author and a doctoral student at Yale University.

Roughly 76 percent of the total organic aerosols measured by the study in New York City were not directly emitted from a source but rather formed in the atmosphere. These so-called secondary organic aerosols are produced when gases, including volatile organic compounds (VOCs), oxidize in the atmosphere. VOCs are produced by a wide range of sources such as cars, vegetation and household chemicals, including cosmetics and cleaners , which complicates efforts to identify the most impactful sources.

Hass-Mitchell and colleagues’ paper is the first to include data from the Atmospheric Science and Chemistry Measurement Network ( ASCENT)—a network of 12 sites around the U.S. that is the first long-term monitoring system able to chemically characterize distinct particle types. Sally Ng, who led the design of the $12-million, National Science Foundation–funded network, says Europe has had similar measurement capabilities for more than five years. “It’s time for the U.S. to modernize its air quality measurement infrastructure,” says Ng, an aerosol scientist at the Georgia Institute of Technology and a co-author of the New York City study.

Recent studies have shown that secondary organic aerosols may be linked to serious health problems—especially cardiovascular disease. A study published last September in Environmental Science & Technology found that as organic aerosols oxidize, they produce highly reactive molecules that can break down human cells and cause tissue damage . Oxidized organic aerosols are the most toxic organic component of PM 2.5 , Ng says. And her work suggests that secondary organic aerosols become more toxic the longer they oxidize in the atmosphere.

Havala Pye, an EPA research scientist, co-authored a separate 2021 Nature study that found that secondary organic aerosols are strongly associated with county-level heart and lung disease death rates in the U.S. Secondary organic aerosols were associated with a 6.5 times higher mortality rate than PM 2.5 .

“There’s a good chance the aerosols are becoming more toxic on a per mass basis, and secondary organic aerosols would be part of the reason why,” says Allen Robinson, an atmospheric scientist at Colorado State University, who was not involved in the new research or Pye’s study. In other words, breathing more oxidized aerosols may be more toxic to humans. But the literature looking at health effects of individual components of PM 2.5 is messy, Robinson notes. More work is needed to unravel the impact of complex combinations of different particle sizes and chemistries in PM 2.5 , he explains. Pye also cautions that consistent results from repeated experiments are needed to verify whether secondary organic aerosols carry significantly greater health risks than other particles that make up PM 2.5 .

Will a warming climate worsen air pollution health risks?

Previous studies have found that warmer temperatures can lead to greater production of these secondary organic aerosols. Hass-Mitchell and colleagues found in the new study that secondary organic aerosol production increased by 60 percent and 42 percent in Queens and Manhattan, respectively, during a sweltering five-day heat wave in July 2022. “We should expect higher health burdens as temperatures rise in a warming climate, with potentially more frequent extreme heat events in the future,” Hass-Mitchell says.

“Secondary organic aerosols are an increasingly important contributor to particulate matter in the summertime and urban air quality, and [they have] a temperature sensitivity that is really important to keep in mind in the context of future climate scenarios,” says Drew Gentner, a chemical and environmental engineer at Yale University and senior author of the new paper. These compounds “are becoming more oxidized at higher temperatures,” he adds, and increased temperatures can cause greater emissions of reactive volatile organic compounds.

And as temperatures increase amid climate change, more frequent and severe wildfires have already begun to chip away at air quality gains in western states. Although Hass-Mitchell and colleagues didn’t observe smoke from wildfires in the summer of 2022, they expect that organic aerosols from wildfires—such as those in the smoke that choked much of the Northeast and Midwest last summer—will also play a major role as the climate changes.

Many other cities, such as Los Angeles, Atlanta and Seoul, have also documented an increasing proportion of PM 2.5 from secondary organic aerosols. But the exact mix of natural versus human-produced sources varies widely from city to city. To continue reducing PM 2.5 , “we need to understand the underlying sources and chemistry contributing to secondary organic aerosol production,” Gentner says.

Until the early 2000s, both the tools to measure secondary organic aerosols and the understanding of their formation were limited, says Benjamin Nault, a co-author of the New York City study and a research scientist at Johns Hopkins University. Currently, most instruments are designed to measure either the size or the chemistry of aerosols but not both, he says. Scientists rely on models to determine how much secondary organic aerosol comes from, for example, live vegetation, asphalt or cooking. But it’s unclear whether some sources are more harmful than others. “There are different signatures for the chemicals that come from taking a shower versus painting [a house],” he says. “Now we’re trying to understand how they come together in an urban environment.”

And that improved understanding is leading to more nuanced pollution research. “As aerosol studies advance, with increasing capabilities to examine the various chemical components of aerosols, we can ask important questions about the relative impact of those components on air quality, human health and the environment,” Gentner says. “It may be less straightforward to address secondary organic aerosol sources compared to primary sources of pollution, but studies [like ours] demonstrate that secondary organic aerosols are the biggest contributor in some urban areas.”

Reporting for this piece was supported by the Nova Institute for Health.

The association between ambient air pollution and migraine: a systematic review

Affiliations.

  • 1 Safety Promotions and Injury Prevention Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Islamic Republic of Iran.
  • 2 Department of Clinical Epidemiology, School of Public Health and Safety, Shahid Beheshti University of Medical Sciences, Tehran, Islamic Republic of Iran.
  • 3 Road Traffic Injury Research Center, Tabriz University of Medical Sciences, Tabriz, Iran.
  • 4 Road Traffic Injury Research Center, Tabriz University of Medical Sciences, Tabriz, Iran. [email protected].
  • 5 Tabriz Health Services Management Research Center, Tabriz University of Medical Sciences, Tabriz, Iran.
  • 6 Department of Epidemiology, School of Public Health and Safety, Shahid Beheshti University of Medical Sciences, Tehran, Iran.
  • 7 Department of Environmental Health Engineering, School of Public Health and Safety, Environmental and Occupational Hazards Control Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran.
  • PMID: 38363415
  • DOI: 10.1007/s10661-024-12376-w

Some studies have shown the effect of air pollution on migraine. However, it needs to be confirmed in larger-scale studies, as scientific evidence is scarce regarding the association between air pollution and migraine. Therefore, this systematic review aims to determine whether there are associations between outdoor air pollution and migraine. A literature search was performed in Scopus, Medline (via PubMed), EMBASE, and Web of Science. A manual search for resources and related references was also conducted to complete the search. All observational studies investigating the association between ambient air pollution and migraine, with inclusion criteria, were entered into the review. Fourteen out of 1417 identified articles met the inclusion criteria and entered the study. Among the gaseous air pollutants, there was a correlation between exposure to nitrogen dioxide (NO 2 ) (78.3% of detrimental relationships) and carbon monoxide (CO) (68.0% of detrimental relationships) and migraine, but no apparent correlation has been found for sulfur dioxide (SO 2 ) (21.2% of detrimental relationships) and ozone (O 3 ) (55.2% of detrimental relationships). In the case of particulate air pollutants, particulate matter with a diameter of 10 μm or less (PM 10 ) (76.0% of detrimental relationships) and particulate matter with a diameter of 2.5 μm or less (PM 2.5 ) (61.3% of detrimental relationships) had relationships with migraine. In conclusion, exposure to NO 2 , CO, PM 10 , and PM 2.5 is associated with migraine headaches, while no conclusive evidence was found to confirm the correlation between O 3 and SO 2 with migraine. Further studies with precise methodology are recommended in different cities around the world for all pollutants with an emphasis on O 3 and SO 2 .

Keywords: Ambient air pollution; Gaseous pollutant; Migraine; Particulate matter; Systematic review.

© 2024. The Author(s), under exclusive licence to Springer Nature Switzerland AG.

Publication types

  • Systematic Review
  • Air Pollutants* / analysis
  • Air Pollution* / adverse effects
  • Air Pollution* / analysis
  • Environmental Exposure / analysis
  • Environmental Monitoring
  • Migraine Disorders* / epidemiology
  • Nitrogen Dioxide / analysis
  • Ozone* / adverse effects
  • Ozone* / analysis
  • Particulate Matter / analysis
  • Sulfur Dioxide / analysis
  • Nitrogen Dioxide
  • Air Pollutants
  • Particulate Matter
  • Sulfur Dioxide

U.S. flag

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

  • Publications
  • Account settings
  • Advanced Search
  • Journal List

Logo of plosone

The effect of air-pollution and weather exposure on mortality and hospital admission and implications for further research: A systematic scoping review

Mary abed al ahad.

1 School of Geography and Sustainable Development, University of St Andrews, Scotland, United Kingdom

Frank Sullivan

2 School of Medicine, University of St Andrews, Scotland, United Kingdom

Urška Demšar

Maya melhem.

3 Department of Landscape Design and Ecosystem Management, American University of Beirut, Beirut, Lebanon

Associated Data

All relevant data are within the paper and its Supporting Information files.

Air-pollution and weather exposure beyond certain thresholds have serious effects on public health. Yet, there is lack of information on wider aspects including the role of some effect modifiers and the interaction between air-pollution and weather. This article aims at a comprehensive review and narrative summary of literature on the association of air-pollution and weather with mortality and hospital admissions; and to highlight literature gaps that require further research.

We conducted a scoping literature review. The search on two databases (PubMed and Web-of-Science) from 2012 to 2020 using three conceptual categories of “environmental factors”, “health outcomes”, and “Geographical region” revealed a total of 951 records. The narrative synthesis included all original studies with time-series, cohort, or case cross-over design; with ambient air-pollution and/or weather exposure; and mortality and/or hospital admission outcomes.

The final review included 112 articles from which 70 involved mortality, 30 hospital admission, and 12 studies included both outcomes. Air-pollution was shown to act consistently as risk factor for all-causes, cardiovascular, respiratory, cerebrovascular and cancer mortality and hospital admissions. Hot and cold temperature was a risk factor for wide range of cardiovascular, respiratory, and psychiatric illness; yet, in few studies, the increase in temperature reduced the risk of hospital admissions for pulmonary embolism, angina pectoris, chest, and ischemic heart diseases. The role of effect modification in the included studies was investigated in terms of gender, age, and season but not in terms of ethnicity.

Air-pollution and weather exposure beyond certain thresholds affect human health negatively. Effect modification of important socio-demographics such as ethnicity and the interaction between air-pollution and weather is often missed in the literature. Our findings highlight the need of further research in the area of health behaviour and mortality in relation to air-pollution and weather, to guide effective environmental health precautionary measures planning.

Introduction

Air-pollution and weather exposure beyond region-specific thresholds have serious effects on the public health [ 1 , 2 ]. Worldwide, population growth, increased urbanization, economic and industrial growth, intense energy consumption, high usage of transportation vehicles, improved living standards, and changing lifestyles and consumption patterns for at least the last 100 years have resulted in increased emissions of air pollutants including greenhouse gases; and fluctuations in ambient temperature and other weather variables [ 3 , 4 ].

Ambient air-pollution consists of a range of pollutants including particulate matters with diameters of less than 10 μm (PM10) and less than 2.5 μm (PM2.5), nitrogen oxides (NOx) including nitrogen dioxide (NO2), Sulphur dioxide (SO2), Carbon monoxide (CO), and Ozone (O3) that have been associated with a range of different acute and chronic health conditions [ 5 , 6 ].

Weather exposure in terms of changing temperature, relative humidity, rainfall and other weather patterns can cause a wide range of acute illness and result in deaths especially among vulnerable populations who lack adequate physiological and behavioural responses to weather fluctuations [ 7 , 8 ]. Age (elderly and children vs adults), sex, socioeconomic factors (poverty, education, and ethnicity among others), pre-existing chronic diseases, use of certain medications, and environmental conditions such as the absence of central heating increase individual’s susceptibility to environmental exposures [ 1 , 9 , 10 ]. Research has shown that hospital admissions and mortality increase when weather exposure exceed certain thresholds with lags up to 20 days [ 11 – 14 ].

Most of the literature has shown positive correlations of air-pollution and/or exposure to weather variables beyond region-specific thresholds with all-cause and cause-specific mortality and/or hospital admission especially related to respiratory and cardiovascular diseases [ 14 – 21 ]. Though, there is a lack of information on wider aspects including the role of some effect modifiers such as ethnicity and the interaction between air-pollution and weather factors. Literature has shown that ethnic minorities often live in more disadvantaged, highly populated urban communities with poor housing conditions and higher levels of air pollution exposure [ 22 – 24 ]. This results in poorer health and higher risk for chronic health problems with time. Similar to ethnicity, the interaction between air-pollution and weather variables in relation to health outcomes is often missed in the literature despite its importance in minimizing biased estimations. Air pollutants are highly reactive, and their formation is either catalysed or slowed down based on the existing weather conditions. For example, the presence of sunlight catalyses the formation of ozone pollutant resulting in higher ozone concentrations during the summer [ 25 ].

In this context, a thorough literature review is needed to map the available literature and highlight areas that require further research and investigation. Not to mention that further understanding of the effect of air-pollution and weather exposure on mortality and hospital admission is needed to achieve better environmental and health system planning, organization, resources allocation, and interventions. This article aims to provide a comprehensive review and narrative summary (not numerical estimate) of literature on the association of air-pollution and weather with mortality and hospital admissions; and to shed the light on areas that require further research. As far as we are aware, this is the first literature review examining the effect of multiple exposures (air-pollution and weather) on multiple outcomes (mortality and hospital admissions). We chose to focus our scoping literature review on countries that are part of the single European Union (EU) market (Austria, Belgium, Bulgaria, Croatia, Republic of Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain and Sweden, Norway, and Switzerland) and United Kingdom (UK) because these countries exhibit similar socio-economic, environmental, and health policies; minimizing the contextual differences in the effect of air-pollution and weather on mortality and hospital admission. Literature examining the effect of air-pollution and/or weather on mortality and hospital admissions in countries outside the EU and UK will be used for comparison purposes.

Materials and methods

Search strategy and database sources.

To ensure methodological reliability, we carried out our scoping literature review according to the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews” (PRISMA-ScR) guidelines ( S1 Checklist ) [ 26 ].

A literature search was performed on the 6 th of February 2020 using “PubMed” and “Web of Science” database sources that cover health, medical, and environmental literature. We attempted to assess the effects of air-pollution and weather events on mortality and hospital admission in Europe by searching original research articles published in peer-reviewed journals in the last 8 years (between 06/02/2012 and 06/02/2020 inclusive). We chose to review research published in the last 8 years because in March 2007, the European Union (EU) Heads of State and Government endorsed an “integrated climate change and energy strategy” that will come into action post the expiry of Kyoto Protocol targets in 2012 and that aims to combat climate change and weather fluctuations and cut air-pollution emissions to 30% below the 1990 levels [ 27 ].

Our search strategy was divided into three conceptual categories: “environmental factors”, “health outcomes”, and “Geographical region”. The “Environmental factors” refers to air-pollution, including PM10, PM2.5, NO2, SO2, CO, and O3 air pollutants and to weather variables, including air temperature, rainfall, wind, relative humidity, and vapour pressure. The “health outcomes” include hospital admissions and mortality and the “Geographical region” refers to the EU countries and UK. For each conceptual category, a set of “MeSH” and “All Fields” terms joined by the Boolean operator “OR” were developed. Later, the three conceptual categories’ search terms were joined using the Boolean operator “AND”. Our search strategy excluded the “influenza infections”, as these are considered confounders rather than outcomes for air-pollution and weather exposure. For more details about the search codes used to navigate PubMed and Web of Science search engines, please refer to S1 Table .

To minimize finding irrelevant literature, our search was limited to the following categories in the “Web of Science” search engine: environmental sciences, public environmental occupational health, medicine general internal, environmental studies, multidisciplinary sciences, geosciences multidisciplinary, respiratory system, geography physical, geography, cardiac cardiovascular systems, urban studies, healthcare sciences services, peripheral vascular disease, medicine research experimental, emergency medicine, critical care medicine, health policy services, primary healthcare, social sciences biomedical, and demography. Grey literature, non-English language articles, conference abstracts, books, reports, masters and PhD dissertations, and unpublished studies were excluded from this review.

Inclusion and exclusion criteria

To determine the studies that would be included in this scoping review, a set of inclusion and exclusion criteria were developed for the procedure of title, keyword, and abstract screening.

The inclusion criteria involved original quantitative research studies conducted in the EU and UK; that included at least one analysis where mortality and/or hospital admission was the outcome and where one or more of the following exposures were investigated: 1) ambient air pollutants including PM10, PM2.5, CO, NO2/NOx, SO2, and O3; 2) weather exposures including temperature, rainfall, wind, humidity, and pressure; and 3) extreme weather events including heat waves, cold spells, and droughts. Due to the large amount of literature on this topic and to allow comparable results between the studies, this review was limited to cohort, time-series, and case-crossover/self-controlled quantitative study designs where hazard ratios (HR), relative risks (RR), odd ratios (OR), or percentage increase were reported for quantifying the factors associated with mortality and hospital admission. These three study designs allow a temporal follow up to evaluate the effect of time varying exposures (air-pollution and weather) on the mortality and hospital admission health outcomes.

The exclusion criteria included the following:

  • Methodological studies
  • Original data studies that investigated the effect of ambient air-pollution and/or weather on mortality and/or hospital admission in countries outside the EU market and UK
  • Articles studying the effect of indoor air-pollution on mortality and hospital admission
  • Studies examining air-pollution and weather exposure on animals and plants
  • Studies on occupational air-pollution exposure
  • Non-English language articles
  • Mortality and/or hospital admission projections and forecasting studies
  • Protocol and letter to editor papers
  • Qualitative research studies
  • All types of literature reviews including but not limited to narrative, scoping, and systematic literature reviews

Screening and data abstraction

Our search strategy revealed 487 articles from the “PubMed” database and 517 articles from the “Web of Science” database. These articles were exported to the citation manager software “Endnote” where 53 duplicates were identified and removed resulting in a total of 951 articles ( Fig 1 ). Using the titles, key words, and abstracts, the 951 articles were screened for relevance according to the inclusion and exclusion criteria, explained in the previous section only by first author (MA). To ensure a rigorous and reliable application of the inclusion and exclusion criteria in the screening process, a second researcher (MM) screened independently a sample of 20% of the titles and abstracts of the 951 identified records. Disagreements between the two researchers were resolved through discussion until consensus was reached. All the studies that met the inclusion criteria (n = 149 articles) were retrieved for full text screening by MA. Following the full text screening phase, an additional 37 articles were excluded by MA resulting in a total of 112 articles to be included in the final narrative synthesis ( Fig 1 ).

An external file that holds a picture, illustration, etc.
Object name is pone.0241415.g001.jpg

For narrative synthesis, the following information was retrieved from the 112 articles:

  • Study design
  • Location of the study population
  • The outcome of interest
  • Sample size
  • Exposure variables
  • The confounders adjusted for
  • The assessed exposure time and the lags considered
  • The exposure assessment method
  • The statistical/modelling approach
  • The relative risks (RR)/incident relative risks (IRR)/odd ratios (OR)/hazard ratios (HR) with their respective confidence intervals or the percentage increase that quantify the association between the outcome of interest (mortality and/or hospital admission) and the exposures (air-pollution and/or weather events).

Ethical approval

Not applicable for this scoping literature review as it only includes descriptive narrative analysis of 112 published articles.

A total of 112 studies ( S2 Table ) were included in the final narrative review from which 70 involved the mortality outcome, 30 the hospital admission outcome, and 12 studies included both health outcomes ( Table 1 ). Most of the studies used the time-series study design (n = 74, 66%) with Poisson models for data analysis, while minority of the reviewed studies employed the case-crossover design (n = 19, 17%) with conditional logistic regression for data analysis, and the cohort design (n = 18, 16%) with Cox hazard regression for data analysis ( Table 1 ).

a Percentages do not add up to 100% as categories are not mutually exclusive.

Most of the studies examined all-cause, cardiovascular and respiratory disease mortality and hospital admission outcomes while some studies tried to focus more directly on certain types of specific diseases such as psychiatric disorders including mania and depression, pulmonary embolism, myocardial infarction, stroke, ischemic heart disease, arrhythmias, atrial fibrillation, heart failure, cerebrovascular disease, chronic obstructive pulmonary disease (COPD), lung cancer, and diabetes ( Table 1 ).

Table 1 shows the descriptive statistics of the included articles. S2 Table summarise the characteristics of the included studies in more details by the type of investigated health outcome. S3 Table demonstrate the included article’s reported associations in terms of coefficients with 95% confidence intervals between air-pollution and/or weather exposure and mortality and/or hospital admission outcomes.

The effect of air-pollution on mortality and hospital admission

In this review, six air pollutants (PM2.5, PM10, O3, CO, SO2, and NO2/NOx) were identified as causes of increased rates of mortality and hospital admissions. Each pollutant affects a range of diseases, most commonly, cardiovascular, respiratory, and cerebrovascular diseases. Some of the health effects can be immediate while others might appear after several days of initial exposure ( Fig 2 ).

An external file that holds a picture, illustration, etc.
Object name is pone.0241415.g002.jpg

The effect of particulate matter pollutants on mortality and hospital admission

Particulate matter is a heterogeneous mixtures of liquid droplets and solid particles suspended in the air that can result either from natural resources (windblown Saharan and non-Saharan dust, volcano ashes, forest fires, pollen, etc…) or from man-made activities including industrial processes, transportation vehicle smoke, burning of fossil fuels, extensive energy usage, combustion processes, and grinding and mining industries [ 28 ]. Due to its size, mass composition, and chemical components, particulate matter with larger diameter (PM10) will be deposited in nasal cavities and upper airways while particulate matter with smaller diameter (PM2.5) may penetrate more deeply the respiratory system reaching the alveoli and blood stream, carrying with them various toxic substances [ 29 ]. This in turn will cause health problems in humans such as asthma, irregular heartbeat, nonfatal heart attacks, decreased lung function, coughing and difficulty breathing symptoms [ 30 ].

Our review showed that PM10 air-pollution is positively associated with a range of cardiovascular and respiratory diseases mortality and hospital admission outcomes ( Fig 3A and S3 Table ). Fischer et al. (2015) showed an elevated hazard of 1.06 (95% CI = 1.04 to 1.08) for cardiovascular disease mortality for every 10 μg/m3 increase in PM10 pollution in the Netherlands [ 31 ]. Likewise, PM10 pollution acted as a risk factor for respiratory diseases mortality (HR = 1.11, 95%CI = 1.08 to 1.15; RR = 1.056, 95%CI = 1.043 to 1.069) [ 21 , 32 ] and hospital admission (%increase = 0.69, 95% CI = 0.20 to 1.19) [ 20 ].

An external file that holds a picture, illustration, etc.
Object name is pone.0241415.g003.jpg

Air-pollution with PM2.5 exhibited a similar effect on human health as that of PM10 ( Fig 3B and S3 Table ). Nevertheless, PM2.5 was shown to have a greater risk on human health as compared to PM10 due to its smaller diameter size allowing more deep penetration into the respiratory system [ 33 ]. In France, Sanyal et al. (2018) showed an increased risk of 1.11 and 1.02 for all-cause hospital admission and moratality respectively per 10 μg/m3 increase in PM2.5 pollutant [ 32 ].

The effect of ozone pollution on mortality and hospital admission

Contrary to particulate matter pollution, the effect of ozone on mortality and hospital admission did not show a consistent effect. In some studies, ozone acted as a protective factor agianst mortality and hospital admission, while in other studies it showed increased risk or no association with mortality and hospital admissions ( Fig 3C and S3 Table ). This is related to the fact that ozone is a highly reactive pollutant and its formation is related to the presence of sunlight [ 25 ]. In a cohort study conducted by Carey et al. (2013) in England, ozone acted as a protective factor agianst all-cause mortality (HR = 0.96, 95%CI = 0.93 to 0.98), cardiovascular mortality (HR = 0.96, 95%CI = 0.94 to 0.98), respiratory mortality (HR = 0.93, 95%CI = 0.90 to 0.96), and lung cancer mortality (HR = 0.94, 95%CI = 0.90 to 0.98) [ 21 ]. However, ozone acted as a risk factor in some of the reviewed studies leading up to 2% increase in all-cause mortality per interquartile range increase of ozone concentration [ 34 – 37 ].

The effect of nitrogen oxides pollution on mortality and hospital admission

Similar to other air pollutants, this review showed that exposure to nitrogen dioxide and nitrogen oxides pollution can cause many types of diseases resulting in increased risk for all-cause mortality and hospital admission [ 25 , 32 , 38 , 39 ] ( Fig 3D and S3 Table ). A study conducted in Belgium showed a 3.5% increase in cardiovascular hospital admission as well as 4.5% and 4.9% increase in ischemic stroke and haemorrhagic stroke hospital admissions respectively for each 10 μg/m3 increase in NO2 [ 40 ].

The effect of sulphur dioxide pollution on mortality and hospital admission

Sulphur dioxide air-pollution is mainly caused from industrial processes and power plants that involve burning of fossil fuel. Exposure to SO2 pollution can cause mild health effects including eyes, nose, and throat irritations as well as severe health effects such as bronchial spasms and deaths due to respiratory insufficiency [ 41 ].

The effect of sulphur dioxide (SO2) on mortality and hospital admission was investigated in only 12 out of the 112 reviewed studies. Exposure to SO2 air-pollution was found to increase the risk for all-cause, cardiovascular, and respiratory mortality [ 21 , 39 , 42 ] ( Fig 3E and S3 Table ).

The effect of carbon monoxide pollution on mortality and hospital admission

Carbon monoxide results from incomplete combustion of fossil fuels. Carbon monoxide is dangerous for human beings since it possess the ability to bind to haemoglobin resulting in reduction of the red blood cells to carry oxygen to cells [ 41 ].

Only 10 out of the 112 reviewed studies investigated the association of carbon monoxide (CO) with mortality and hospital admission. The majority of these studies showed that carbon monoxide exposure can cause a number of cardiovascular and respiratory health problems ( Fig 3F and S3 Table ). Exposure to carbon monoxide pollution resulted in increased odds for pulmonary embolism hospital admission [ 43 ]. Additionally, Renzi et al. (2017) showed that all-cause mortality increases by 0.12% for every 1 mg/m3 increase in CO [ 39 ]. On the contrary, carbon monoxide acted as a protective factor against chest disease hospital admission among patients with sickle cell anaemia in one of the reviewed studies [ 44 ]. This association was explained by the fact that carbon monoxide can bind to haemoglobin which enhances the affinity of other binding sites for oxygen in addition to reducing vasoconstriction and inflammation; suggesting a beneficial effect rather than risk factor for patients with sickle cell disease [ 44 ].

The effect of air temperature on mortality and hospital admission

Exposure to hot or cold temperature beyond region-specific thresholds exhibits a range of direct and indirect effects on human health. The direct effects include hyperthermia or heat stress during hot temperature exposures and hypothermia and ischemic stroke during cold temperature exposures [ 45 ]. Besides the direct effects, small fluctuations in temperature across time can result in indirect effects on the respiratory and cardiovascular systems of the body [ 45 ].

Most of the reviewed articles that studied the effect of weather exposure on mortality and hospital admission focused on air temperature exposure with lags ranging from 0 days up to 5 weeks for cold temperatures and from 0 days up to 25 days for hot temperatures. The reviewed studies examined the effect of cold temperature, hot temperature, and air temperature increase on a range of diseases, most commonly, cardiovascular, respiratory, and psychiatric disorders. Table 2 below shows the definitions of “cold temperature”, “hot temperature”, and “air temperature increase” classifications derived from the reviewed studies.

Cold temperature acted as a risk factor for several types of mortality and hospital admission outcomes ( Fig 4A and S3 Table ). Nevertheless, cold temperature was a protective factor only in one of the reviwed studies for all-cause mortality at lag 0 (RR = 0.99, 95%CI = 0.985 to 0.995); yet cold temperature acted as a risk factor for all-cause mortality in the same study at lag of 14 days with a relative risk of 1.003 emphasizing the delayed effect of cold temperature on mortality [ 46 ].

An external file that holds a picture, illustration, etc.
Object name is pone.0241415.g004.jpg

Similar to cold temperature, hot temperature also acted as a risk factor in most of the reviewed studies for a number of mortality and hospital admission outcomes ( Fig 4B and S3 Table ). On the other hand, hot temperatures were associated with reductions in hospital admission rates for ischemic heart disease (RR = 0.74, 95%CI = 0.55 to 0.99) in a study conducted by Bijelovic et al. (2017) and for all-cause hospital admissions (RR = 0.961, 95% CI = 0.956 to 0.967) and cardiovascular hospital admissions (RR = 0.975, 95% CI = 0.957 to 0.993) in a study conducted by Monteiro et al. (2013) [ 47 , 48 ].

Some studies examined the effect of increasing temperature across the whole year on mortality and hospital admission. More than half of these studies showed a significant positive association between the increasing temperature and the mortality and hospital admission outcomes ( Fig 4C and S3 Table ).

The effect of other weather exposures on mortality and hospital admission

Similar to the temperature, weather exposures that include humidity, rainfall, sunshine, snowcover, air pressure, daylight, wind speed and wind direction with lags ranging from 0 up to 7 days were found to affect a range of diseases, most commonly, cardiovascular, respiratory, and psychiatric disorders ( S3 Table ).

Weather variables that showed significant positive assciations with hospital admission included: a rainfall effect on psychiatric hospital admission [ 49 ], sunshine and daylight effects on hospital trauma [ 50 ] and psychaitric admissions [ 51 ], wind speed effects on chest disease hospital admission [ 44 ], and air pressure effects on mania and depression hospital admission [ 49 ].

It is worth mentioning that sunshine showed inconsistency in its effect on psychaitric hospital admission, acting as a risk factor in a Danish study [ 52 ] while acting as a protective factor in a study conducted in Ireland [ 49 ].

The adjustments and effect modifications for the association of air-pollution and weather exposure with mortality and hospital admission

Most of the reviewed studies stratified and adjusted their analysis by age and gender [ 25 , 40 , 51 , 53 – 61 ]. Socio-economic deprivation, education attainment, income level, marital status, and occupational class were considered as confounders or effect modifiers in some of the reviewed studies [ 25 , 58 , 62 – 66 ]. However, only one study considered ethnicity to act as an effect modifier in the association between all-cause mortality and “summer smog” days defined as having maximum temperature of 25°C and PM10 pollutant oncentration of 50 μg/m3 [ 62 ]. And only two studies investigating the effect of air-pollution on all-cause and cardiovascular mortality in England adjusted for ethnicity in their multivariate regresison models [ 67 , 68 ].

Some of the studies that examined the effect of air-pollution on mortality and hospital admission accounted for air temperature effect in their analysis [ 35 , 40 , 46 ]. Likewise, some of the reviewed articles that studied the association of weather exposure to mortality and hospital admission considered the effect of air-pollution in their analysis [ 34 , 65 , 69 – 72 ].

Other variables considered to affect the relationship of air-pollution and/or weather exposure with mortality and/or hospital admission included: weekend and holiday effect, population decrease during the summer, influenza epidemics, season, day of the week, and tobacco smoke [ 25 , 40 , 46 , 54 , 56 , 73 – 79 ].

In this scoping review of 112 articles, we aimed to examine the effect of (1) air pollution, (2) temperature, and (3) other weather exposures on mortality and hospital admission outcomes.

The first part of the review showed that air-pollution acted consistently as a risk factor for all-cause, cardiovascular, respiratory, cerebrovascular and cancer mortality and hospital admission in the EU and UK which is in line with the findings of studies conducted in other regions of the world [ 80 – 84 ]. For instance, elevated risks of cardiovascular and respiratory diseases mortality were reported in Istanbul-Turkey for every 10 μg/m3 increase in PM10, SO2 and NO2 pollutants [ 85 ]. An exception was ozone (O3) air-pollution which showed inconsistent association with mortality and hospital admission. Two explanations were offered in the literature for the negative association between health outcomes and ozone pollution. The first explanation is related to the fact that ozone is a highly seasonal pollutant since its formation is catalysed by sunlight rendering higher ozone concentrations in the summer as compared to winter season. Thus, ozone effect on health outcomes should be analysed by accounting for the season effect [ 21 ]. In the continental United States, a 49% higher risk in all-cause mortality was shown for every 10 ppb increase in ozone during the warm-season [ 86 ]. The second explanation is related to the high reactivity of ozone leading to the formation of other pollutants such as NO2 and particulate matter. Therefore, ozone is negatively correlated with other air pollutants and its effect on health outcomes should be analysed as a combined effect of O3 and NO2 (known as Ox effect) [ 25 ].

Additionally, our scoping review showed that the effect of particulate matter (PM10 and PM2.5) pollution on mortality and hospital admission is more studied in the literature as compared to the other air pollutants. This could be related to the more pronounced effects of particulate matter exposure on health which is corroborated by many studies across the world [ 81 , 87 ]. Despite the fact that PM10 particles are deposited in the nasal cavities and upper airways, PM2.5 may penetrate deep into the lung tissues (reaching the alveoli and bloodstream) and irritate the respiratory airways causing various respiratory and cardiovascular problems [ 29 , 30 , 88 ].

Similar to air pollution, the second part of this review showed that hot and cold temperature exposures beyond region-specific thresholds are risk factors for a wide range of respiratory, cardiovascular (including: ischemic heart disease, myocardial infarction, pulmonary embolism, stroke, heart failure, and COPD), and psychiatric (including: mania and depression) illness in the EU and UK. These findings are corroborated by a wide body of literature from across the world [ 89 – 97 ]. In India, cold temperatures below 13.8°C were associated with increased risk of 6.3% for all-cause mortality, 27.2% for stroke mortality, 9.7% for ischemic heart disease mortality, and 6.5% for respiratory diseases mortality [ 92 ]. In Istanbul-Turkey, 23 days of exposure to hot temperature above 22.8°C was associated with a total of 419 excess deaths [ 90 ]. In Korea, hot temperature days of 25°C compared to 15°C were significantly associated with a 4.5% increase in cardiovascular hospitalizations [ 98 ].

It is worth to point out that the effect of cold temperature on health is more delayed (up to 5 weeks) in comparison to the more immediate effects of hot temperature (up to 25 days). Similar study in Northeast-Asia showed a delayed risk of cold temperature on mortality after 5 to 11 days, yet a more immediate effect of hot temperature on mortality after 1 to 3 days in each of Taiwan, Korea, and Japan countries [ 91 ].

Although exposure to hot or cold temperature can affect the health negatively, our scoping review showed that in few studies, the increase in temperature reduced the risk of hospital admissions for some types of cardiovascular diseases; mainly for pulmonary embolism, angina pectoris, chest, and ischemic heart diseases. This could be explained by the fact that hot temperature can cause immediate increase in cardiovascular mortality rates; whereby many cases might pass directly to the death state without passing through the hospital admission state resulting in lower hospital admission rates [ 47 ].

The third part of this scoping review presented the studies that examined the effects of other weather exposures such as relative humidity, barometric pressure, rainfall, and wind speed on mortality and hospital admission outcomes. These weather exposures were found to affect significantly only hospital admission. No significant effect was noted with respect to the mortality outcome. The weather exposures acted as a risk factor for psychiatric disorders (including depression and Mania), chest disease, and trauma hospital admissions. This was corroborated by evidence from countries outside the EU as well [ 99 – 101 ]. Yet in some of the reviewed studies, weather exposures acted as a protective factor for some types of psychiatric and cardiovascular disorders. The significant negative association between ischemic heart disease hospital admission and humidity in one of the reviewed studies was explained by the fact that people in general and the elderly specifically reduce their activities during high humidity and temperature periods. This is mainly due to the lack of the body’s ability to perspire, which in turn reduces their risk of cardiovascular complications [ 48 ]. As for the protective effect of some weather exposures on psychiatric hospital admissions, similar findings were presented in Iran; with a negative association between barometric pressure and schizophrenia hospital admissions and rainy days and bipolar hospital admissions [ 101 ].

In addition to the association of air pollution and weather exposure with mortality and hospital admission outcomes, our review aimed to present the individual, socio-economic, and environmental factors that play an important role in modifying the latter association. The effect modifiers identified in this scoping review included: pre-existing health conditions, age, gender, educational attainment, wealth or income or socio-economic deprivation, occupation, marital status, tobacco smoking, season, day of the week, holidays, and influenza epidemics.

Individuals with pre-existing chronic health conditions face increased susceptibility toward air-pollution and weather exposure related mortality and hospital admission [ 52 , 54 , 78 , 102 ].

Older people are more vulnerable to the health effects associated with air-pollution, hot or cold temperatures, and other weather variables [ 54 , 56 , 103 – 105 ]. This is due to the physiological degeneration of the human body with increasing age. Aging affects the normal function of the body organs resulting in many chronic cardiovascular, urinary, and respiratory health conditions. This reduces the ability of older people to adapt to increased concentrations of air pollutants and changing weather conditions [ 103 , 105 , 106 ]. Moreover, old age people have lower immunity and antioxidant defence as compared to young people placing them at a higher risk [ 107 ]. Many older people also have reduced mobility and mental abilities which delay their access to healthcare leading to severe health complications and death [ 108 ].

As for gender, our review revealed inconsistency regarding its modification effect on the association between air-pollution and weather exposure and mortality and hospital admission health outcomes. Nevertheless, most of the reviewed studies have found that females have higher risks of mortality and/or hospital admission after exposure to air-pollution and/or weather fluctuations beyond region-specific thresholds including hot and cold temperatures [ 9 , 47 , 57 , 58 , 65 , 78 , 109 – 112 ]. Whereas some studies found higher risks of mortality and/or hospital admissions among males in relation to air-pollution and/or weather exposure [ 25 , 40 , 42 , 59 , 70 , 113 , 114 ]. One explanation for this might be due to the physiological differences between males and females. Females have smaller lung size, yet higher airways reactivity making them more susceptible to air-pollution health effects as compared to males [ 42 , 115 ]. Likewise, higher pulse rates and smaller heart size relative to the human body in females as compared to males render females more vulnerable to the health effects of air-pollution and hot or cold temperature exposures [ 115 ]. Moreover, females exhibit more fluctuations in hormone levels due to pregnancy, menstrual cycle and menopause periods which may place them at a higher health risk upon exposure to air-pollution and weather variations [ 115 ]. The different lifestyle, socio-economic position, and occupation type between males and females may also lead to different levels and duration of air-pollution and weather exposure [ 62 , 109 , 116 – 118 ]. However, it is worth mentioning that the effect modification of gender in the association of air-pollution and weather exposure with mortality and hospital admission outcomes is believed to be confounded by age since in many of the reviewed studies, higher risks were found among old aged females (age>65 years old) [ 47 , 57 , 61 , 78 , 112 ] and old aged males (age >70 years old) [ 114 ]. This confounding effect could be reduced either by assessing the combined effect modification of age and gender through an interaction term or by stratifying the analysis according to both the age groups and gender.

Wealth and socio-economic deprivation were also considered by some of the reviewed studies as an effect modifier in the relationship of air-pollution and weather with mortality and hospital admission. In general, the absence of wealth and presence of socio-economic deprivation increase the risk of exposure to air-pollution and weather variations resulting in elevated mortality and hospital admission rates in Europe [ 62 , 66 , 76 ] and in other parts of the world including New Zealand [ 22 ], United States of America [ 119 ], and Chile [ 120 ].

Educational attainment was also considered by some of the reviewed studies as an effect modifier, with higher risks detected among individuals with lower educational attainment [ 25 , 63 , 64 , 121 , 122 ]. Despite the consideration of age, gender, education, and wealth effect in the association of air-pollution and weather with mortality and hospital admissions in Europe, our scoping review revealed the lack of investigation into the role of other important socio-demographics such as ethnicity. Research has extensively shown that ethnic minorities live in more disadvantaged communities and have lower socio-economic status as well as poor housing conditions. This results in higher risk for chronic health problems associated with higher exposure on one hand and with lower access to quality healthcare on the other hand [ 22 – 24 ].

Finally, it is worth to note that most of the reviewed studies with a time-series or case-crossover design adjusted their analysis for the season effect [ 40 , 46 , 52 , 54 , 60 , 71 , 73 , 74 , 123 , 124 ]. It is well established that air-pollution, temperature, and other weather variables vary with seasons [ 125 – 127 ]. Not to mention that the emission, formation, and dispersion of air pollutants is affected by seasonal weather variations which in turn affects the individual exposure levels [ 128 ]. Outdoor activities and daily habits (eg. Window ventilation of houses) might also vary depending on the season which reflect changes in the level and duration of individual exposure to air-pollution and weather changes [ 129 ].

Despite the value of this scoping literature review, it has some limitations. First, the employed search strategy was limited to original articles published in peer reviewed journals which might have led to the omission of unpublished work or articles that were published in non-indexed journals. Nevertheless, our search strategy involved navigation through two databases which enables a good catch of major published studies addressing the effect of air-pollution and weather exposure on mortality and hospital admission. Second, limiting our inclusion criteria only to English language articles might have resulted in missing some research written in other languages. However, as most of the literature worldwide is published in the English language, we believe that no major papers have been excluded. Third, this review was limited only to quantitative research which would have led to missing out other type of important research including opinion research pieces and letters to editor as well as qualitative research studies. Opinion research pieces and letters to editor provide a critical appraisal/discussion for the findings of original studies which warrant future research development. Qualitative studies provide an overview about the effect of air-pollution or weather variations on human health from the perspective of lay people rather than relying only on objective census/statistics numbers as in quantitative research. Forth, due to resources limitations, title and abstract screening as well as data abstraction were done only by one researcher (MA). Nevertheless, a second researcher performed title and abstract screening for a random sample of 20% of the retrieved records. Given the high consensus between the two researchers, we are confident of the exact application of the inclusion and exclusion criteria. Our goal from this literature review was not to produce a numerical estimate but rather to give a narrative summary on the effect of air-pollution/weather on mortality/hospital admission. Hence, missing some studies would not be a major concern for this scoping review.

Literature gaps and implications for future research

This scoping review helped us to identify literature gaps that require further research.

First, this review revealed the extensive research carried out to determine the effect of air-pollution on human health. Yet, due to the high correlation between air pollutants and the issue of collinearity in multivariate models, most of the studies examined the effect of single pollutants on mortality and hospital admission outcomes. Nevertheless, the issue of correlation between air pollutants is highly contextual and it depends on the study settings including the season and the specific geographical area. Hence, future researchers should try to examine the effect of multi-pollutants on mortality and/or hospital admission in one model, where strong correlations between the air pollutants are absent.

Second, the majority of studies examined the direct effects of air-pollution and weather exposure on mortality and hospital admission without considering the role of certain effect modifiers. The examined effect modifiers considered mostly by the literature involve age, gender, education, socio-economic deprivation, and season. Therefore, there is a lack of evidence regarding the modifying effect of other individual factors such as previous disease conditions and ethnicity which affect the person’s health vulnerability. Indeed, future research is needed to find out the reasons behind elevated individual’s susceptibility to the detrimental effects of air-pollution and weather variations in certain groups of population.

Third, our review showed that most of the studies either investigated the effect of air-pollution or the effect of weather on mortality and hospital admission. The formation and dispersion of air pollutants depends highly on the existing atmospheric conditions such as temperature, humidity, and wind speed [ 130 ]. Therefore, future studies should consider examining the effect of both, weather conditions and air-pollution, on human health through interaction terms or adjustments in the analysis models.

Fourth, although extensive research has been performed to study the effect of particulate matter and nitrogen oxides pollution on human health, there was a lack of research with respect to other air pollutants including carbon monoxide, ozone, and sulphur dioxide. This might be due to the absence of rigorous and reliable measurements of these pollutants or due to the complexity of analysing the effect of these pollutants.

Fifth, literature is more focused on examining the effect of temperature on mortality and hospital admission, placing less emphasis on other weather exposures. Hence, future research should shift the focus toward other weather exposures such as wind speed, rainfall, humidity, snow cover, daylight, and air pressure.

Sixth, there was a lack of research examining the effect of air-pollution and weather on hospital admission. Mortality was the major outcome in most of the reviewed studies due to the ease of data access governed by less ethical considerations. Additionally, analysis is more straightforward given that it occurs only once in an individual’s life. Thus, it is recommended for future research to consider the impact of air-pollution and weather variables on hospital admission on its own and in combination with mortality through multistate modelling.

Finally, the majority of studies in this field employ the time-series design which uses aggregated mortality and hospital admission data linked to environmental exposures at the local authorities or municipalities level. Research that uses aggregated data neglect the physiological and socio-economic differences among individuals. Additionally, assigning air-pollution and weather exposure based on wide geographies overlook the small geographical exposure differences biasing the drawn estimates. Therefore, there is a need for cohort research studies that utilize individual level data linked to air-pollution and weather exposure at small geographical spatial resolution (eg. Postcodes).

In summary, our scoping review showed that air-pollution and weather exposure beyond certain thresholds lead to various impacts on human health, most commonly cardiovascular and respiratory problems, resulting in increased rates of mortality and hospital admission. Yet, further research is needed given that the effect modification of important socio-demographics such as ethnicity and the interaction between air-pollution and weather is often missed in the literature. Understanding this should give enough evidence to the policy makers to plan and act accordingly aiming to reduce the effects of air pollution and weather variations on the public health. Additionally, research should focus on projecting future health behaviour and mortality patterns in relation to air pollution and weather variations, in order to guide effective environmental and health precautionary measures planning.

Supporting information

S1 checklist, acknowledgments.

The authors would like to thank the University of St Andrews Library services for helping in developing the search codes used in this scoping literature review.

Funding Statement

This review is part of a PhD project that is funded by the St Leonard’s PhD scholarship, University of St Andrews, Scotland, United Kingdom. The open access publication fees were funded by the University of St Andrews Libraries, Scotland, United Kingdom.

Data Availability

  • PLoS One. 2020; 15(10): e0241415.

Decision Letter 0

25 Aug 2020

PONE-D-20-19327

The Effect of Air-pollution and Extreme-weather on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review

Dear Dr. Abed Al Ahad,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Oct 09 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at gro.solp@enosolp . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Chon-Lin Lee, Ph.D.

Academic Editor

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Your ethics statement must appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please also ensure that your ethics statement is included in your manuscript, as the ethics section of your online submission will not be published alongside your manuscript.

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information .

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: N/A

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript entitled “The Effect of Air-pollution and Extreme-weather on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review” used the scoping review approach to summarize the literature on the association of air pollution and extreme weather with mortality and hospital admissions. The authors used the reviews to conclude that air pollution and extreme weather affect human health negatively. They also highlighted the literature gaps that require further research. This manuscript addresses an interesting environmental health issue and contains some useful information; however, I have several major concerns that need to be adequately addressed before the manuscript can be considered for publication.

1. The author should clearly define “Extreme-weather.” It is commonly understood that extreme weather is when weather is significantly different from the usual weather pattern. However, most publications discuss the effect of ambient temperature, relative humidity, or other meteorological factors on health effects. The authors need to clarify the terminology. Also, modification of the title is also suggested.

2. Lines 42, the author mention “Climate Change” and briefly described it. Nevertheless, the effects of climate change are not discussed in the manuscript. The authors could consider to delete it in the background section.

3. Line 53-55, the author state that “there is a lack of information on the role of some effect modifiers such as ethnicity and the interaction between air-pollution and extreme-weather.” I suggest that the author provide some evidence about the importance of effect modifiers (ethnicity) and the interaction between air-pollution and extreme-weather (meteorological factors) in the context of their health effects.

4. Particulate matter (PM2.5 or PM10) is heterogeneous mixtures of solid and liquid particles emitted from a variety of sources. Recently, there are Along with size, concentration, and chemical components of particulate matter are important in mediating the effects of PM on human health. Although the evidence of PM composition with those adverse health effects is limited, I believe that, in this review, this issue is worth mentioning in the background.

5. The reasons for excluding the pediatric population is not justified. For assessing health impacts, to evaluate the whole population (from pediatrics to geriatrics) is very important. Especially, the authors attempt to provide a comprehensive review of the topic. I highly suggest the author included the pediatric population.

6. Line 132, one of the exclusion criteria is “studies investigating in-hospital death.” Is in-hospital death part of overall mortality? Why exclude those publications?

7. Line 212 and 377, the author state that “particulate matter especially the small-size ones (PM10 and PM2.5) penetrate deeply the respiratory system” is not entirely correct. Most PM10 particles are deposited in the nasal cavities and upper airways. However, PM2.5 particles may penetrate the lung alveoli and enter into the bloodstream. (Möller W, Felten K, Sommerer K, Scheuch G, Meyer G, Meyer P, et al. Deposition, retention, and translocation of ultrafine particles from the central airways and lung periphery. Am J Respir Crit Care Med. 2008;177(4):426–32.)

8. Line 236-238 (This is related to the fact that ozone is a highly reactive pollutant and its formation is related to the presence of sunlight), the citation is needed for the statement.

9. In section 3.2. (The effect of air temperature on mortality and hospital admission), the cold, warm, and hot temperatures are needed to be clearly defined. Is the temperature cutoff identical among the references?

10. Line 528, the author discussed the correlation between air pollutants. In multipollutant models, because variables are commonly highly correlated, the collinearity becomes the major problem for multivariate analysis. I suggested the author describe the issue briefly.

11. In the conclusion section (line 571 and 573), the authors mention “climatic change.” However, this review does not touch upon the topic (climatic change). Overall, the sections “Discussion” is not well-organized and well-presented. It needs to be significantly revised.

Reviewer #2: This is a review paper that summarized 106 published works on air pollution and weather on mortality and hospital admission. The work followed PRISMA guideline to search and screen from literature.

Major points:

1. the goal of this manuscript is stated at line 510-512, not a numerical estimate but a narrative summary, these words should be addressed at the abstract or introduction as well.

2. a little bit confused at 4.2 literature gap part. The first suggestion stated that the exact role of individual pollutants is still unclear; but the third said that most studies examined the effect of single pollutants. The two statements contradict to each other. It might be better to merge the two gaps into one, and emphasize the interaction of the variables is the missing link.

Minor points:

1. Line 116, typo: heat.

2. Line505, missing a space between first two words.

3. Figures 3 and 4, please indicate what the y-axis is.

6. PLOS authors have the option to publish the peer review history of their article ( what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool,  https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at  gro.solp@serugif . Please note that Supporting Information files do not need this step.

Author response to Decision Letter 0

15 Sep 2020

Professor Chon-Lin Lee

Manuscript title: The Effect of Air-pollution and Extreme-weather on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review

**Response to the Editor’s and Reviewers’ Comments

We thank the editor and the reviewers for their encouraging feedback and this opportunity to improve our manuscript. Please find below a point by point answer to the reviewers’ comments and the necessary changes made to the manuscript in track-changes. We also submitted a clean version of the manuscript in addition to the track-changes version.

Thank you for considering our work and we hope that the revised version is suitable for publication and we are looking forward to hearing from you again.

Mary Abed Al Ahad, on behalf of the authors

Journal Requirements

**Authors’ response: We have abided by the journal requirements during the submission of this version of the manuscript in terms of following the Vancouver style for citation and the journal’s style for headings/subheadings and tables and figures.

**Authors’ response: Ethics statement is not applicable for this manuscript as it is a scoping review of literature that is publicly available on “Pubmed” and “Web of Science” search engines.

**Authors’ response: We included captions for the supporting information files at the end of the revised manuscript and updated the in-text citations to match accordingly.

Reviewers’ comments

Reviewer #1:

The manuscript entitled “The Effect of Air-pollution and Extreme-weather on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review” used the scoping review approach to summarize the literature on the association of air pollution and extreme weather with mortality and hospital admissions. The authors used the reviews to conclude that air pollution and extreme weather affect human health negatively. They also highlighted the literature gaps that require further research. This manuscript addresses an interesting environmental health issue and contains some useful information; however, I have several major concerns that need to be adequately addressed before the manuscript can be considered for publication.

**Authors’ response: We have replaced the term “extreme weather” by “weather exposure” in the manuscript as our objective was to provide an overview on articles discussing the effect of ambient temperature, relative humidity and other weather exposures which could not merely be extreme weather exposures. We have also provided examples to explain what we mean by “weather exposure” in the introduction section on line 43 “Weather exposure in terms of changing temperature, relative humidity, rainfall and other weather patterns can cause a wide range of acute illness …..”.

We modified the title to be in line with the new terminology of “weather exposure”. The new title is: “The Effect of Air-pollution and weather exposure on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review”.

2. Lines 42, the author mention “Climate Change” and briefly described it. Nevertheless, the effects of climate change are not discussed in the manuscript. The authors could consider deleting it in the background section.

**Authors’ response: We have deleted the “climate change” phrase from the Introduction section.

**Authors’ response: Evidence about the importance of effect modifiers (ethnicity) and the interaction between air-pollution and weather exposure in the context of their health effects was added to the introduction section of the manuscript on page 4, line 55 to 65 as follows: “Though, there is a lack of information on wider aspects including the role of some effect modifiers such as ethnicity and the interaction between air-pollution and weather factors. Literature has shown that ethnic minorities often live in more disadvantaged, highly populated urban communities with poor housing conditions and higher levels of air pollution exposure (22-24). This results in poorer health and higher risk for chronic health problems with time. Similar to ethnicity, the interaction between air-pollution and weather variables in relation to health outcomes is often missed in the literature despite its importance in minimizing biased estimations. Air pollutants are highly reactive, and their formation is either catalyzed or slowed down based on the existing weather conditions. For example, the presence of sunlight catalyzes the formation of ozone pollutant resulting in higher ozone concentrations during the summer (25)”.

**Authors’ response: We included in the manuscript on page 11, line 216 to 226 a brief description about the effect of particulate matter on human health which is related to their size, composition, and concentration as follows: “Particulate matter is a heterogeneous mixtures of liquid droplets and solid particles suspended in the air that can result either from natural resources (windblown Saharan and non-Saharan dust, volcano ashes, forest fires, pollen, etc…) or from man-made activities including industrial processes, transportation vehicle smoke, burning of fossil fuels, extensive energy usage, combustion processes, and grinding and mining industries (28). Due to its size, mass composition, and chemical components, particulate matter with larger diameter ……”.

We mentioned this on page 11, line 216-226 in the section where we are talking about the effect of particulate matter on human health as it fits more the flow of the manuscript than including it in the introduction section.

**Authors’ response: We have added the studies on pediatrics population (4 studies in total) to our scoping review which brings the total of reviewed studies from 106 to 110 studies. Below are the references for these 4 studies:

• Ghirardi, L., Bisoffi, G., Mirandola, R., Ricci, G., & Baccini, M. (2015). The Impact of Heat on an Emergency Department in Italy: Attributable Visits among Children, Adults, and the Elderly during the Warm Season. PLoS One, 10(10). doi:10.1371/journal.pone.0141054

• Janke, K. (2014). Air pollution, avoidance behaviour and children's respiratory health: Evidence from England. Journal of Health Economics, 38, 23-42. doi:10.1016/j.jhealeco.2014.07.002

• Litchfield, I. J., Ayres, J. G., Jaakkola, J. J. K., & Mohammed, N. I. (2018). Is ambient air pollution associated with onset of sudden infant death syndrome: a case-crossover study in the UK. BMJ Open, 8(4). doi:10.1136/bmjopen-2017-018341

• Piel, F. B., Tewari, S., Brousse, V., Analitis, A., Font, A., Menzel, S., . . . Rees, D. C. (2017). Associations between environmental factors and hospital admissions for sickle cell disease. Haematologica, 102(4), 666-675. doi:10.3324/haematol.2016.154245

**Authors’ response: We have added the studies addressing in-hospital death as part of overall mortality (2 studies in total) to our scoping review which brings the total of reviewed studies from 110 to 112 studies. Below are the references for these 2 studies:

• Callaly, E., Mikulich, O., & Silke, B. (2013). Increased winter mortality: the effect of season, temperature and deprivation in the acutely ill medical patient. Eur J Intern Med, 24(6), 546-551. doi:10.1016/j.ejim.2013.02.004

• Lyons, J., Chotirmall, S. H., O'Riordan, D., & Silke, B. (2014). Air quality impacts mortality in acute medical admissions. Qjm, 107(5), 347-353. doi:10.1093/qjmed/hct253

**Authors’ response: We have modified the text on line 220-226 and line 386-393 in the new version of the manuscript to reflect the fact that PM10 particles are deposited in the nasal cavities and that PM2.5 can penetrate deep the lungs reaching the alveoli and blood stream and added the above reference for the statement.

**Authors’ response: Sorry for missing out the citation for that statement. We have now added citation “citation number 25” to this statement on line 246 of the new version of the manuscript.

**Authors’ response: We have added a table “Table 2” which includes definitions for hot and cold temperature exposures with a range of cutoff (threshold) points. We took the classification of “hot” or “cold” temperature from the studies themselves and each study had its own identified cutoff point for hot and/or cold temperature.

Table 2. The definitions of air temperature exposure classifications

Classification Definition

Cold temperature Exposures to air temperature in the winter season below identified thresholds ranging from -7 ºC to 6 ºC

Hot temperature Exposures to air temperature in the summer season above identified thresholds ranging from 20 ºC to 37 ºC

Air temperature increase Exposures to increasing temperature across the whole year. Associations are interpreted per 1 ºC increase in temperature.

**Authors’ response: We have now described the issue briefly and amended the paragraph on line 525 to 531 in the revised version of the manuscript as follows: “Yet, due to the high correlation between air pollutants and the issue of collinearity in multivariate models, most of the studies examined the effect of single pollutants on mortality and hospital admission outcomes. Nevertheless, the issue of correlation between air pollutants is highly contextual and it depends on the study settings including the season and the specific geographical area. Hence, future researchers should try to examine the effect of multi-pollutants on mortality and/or hospital admission in one model, where strong correlations between the air pollutants are absent”.

**Authors’ response: We have revised the conclusion section and replaced “climatic change” with “air pollution and weather variations” on line 577 in the revised manuscript.

Additionally, we have revised extensively the “Discussion” section to be more organized and well-presented which could be viewed in the track-changes version of the revised manuscript.

Reviewer #2:

This is a review paper that summarized 106 published works on air pollution and weather on mortality and hospital admission. The work followed PRISMA guideline to search and screen from literature.

**Authors’ response: We have added that the scoping review aimed for a narrative summary of the literature in both, the introduction (on line 71) and the abstract (on line 5).

**Authors’ response: Thank you for the comment. We have now merged the first and third literature gaps into one, focusing on the correlation and interaction between the air pollution variables on line 524-531 of the revised manuscript as follows: “First, this review revealed the extensive research carried out to determine the effect of air-pollution on human health. Yet, due to the high correlation between air pollutants and the issue of collinearity in multivariate models, most of the studies examined the effect of single pollutants on mortality and hospital admission outcomes. Nevertheless, the issue of correlation between air pollutants is highly contextual and it depends on the study settings including the season and the specific geographical area. Hence, future researchers should try to examine the effect of multi-pollutants on mortality and/or hospital admission in one model, where strong correlations between the air pollutants are absent”.

**Authors’ response: The typo mistake “heath” on line 116 was corrected to “heat”.

**Authors’ response: The space between first two words on line 505 was corrected.

**Authors’ response: We have added the Y-axis and the X-axis descriptions to Figures 3 and 4

Submitted filename: Response to Reviewers.docx

Decision Letter 1

15 Oct 2020

The Effect of Air-pollution and weather exposure on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review

PONE-D-20-19327R1

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/ , click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at gro.solp@gnillibrohtua .

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact gro.solp@sserpeno .

Additional Editor Comments (optional):

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #2: (No Response)

3. Has the statistical analysis been performed appropriately and rigorously?

4. Have the authors made all data underlying the findings in their manuscript fully available?

5. Is the manuscript presented in an intelligible fashion and written in standard English?

6. Review Comments to the Author

Reviewer #1: Thank you for having me to review the article entitled “The Effect of Air-pollution and weather exposure on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review.” This manuscript addresses an interesting environmental health issue. The authors put great effort into revising the manuscript. New, the article is well-written and contained important information to the knowledge domain about the health effects of air pollution. I think it is worthy of being published.

7. PLOS authors have the option to publish the peer review history of their article ( what does this mean? ). If published, this will include your full peer review and any attached files.

Acceptance letter

19 Oct 2020

The Effect of Air-pollution and weather exposure on Mortality and Hospital Admission and implications for further research: A Systematic Scoping Review 

Dear Dr. Abed Al Ahad:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact gro.solp@sserpeno .

If we can help with anything else, please email us at gro.solp@enosolp .

Thank you for submitting your work to PLOS ONE and supporting open access.

PLOS ONE Editorial Office Staff

on behalf of

Dr. Chon-Lin Lee

Bibliometric analysis on mercury emissions from coal-fired power plants: a systematic review and future prospect

  • Review Article
  • Published: 21 February 2024

Cite this article

  • Jiajia Gao 1 ,
  • Guoliang Li 1 ,
  • Yang Zheng 1 ,
  • Rui Li 1 &
  • Tao Yue 1  

Coal-fired power plants (CFPPs) are one of the most significant sources of mercury (Hg) emissions certified by the Minamata Convention, which has attracted much attention in recent years. In this study, we used the Web of Science and CiteSpace to analyze the knowledge structure of this field from 2000 to 2022 and then reviewed it systematically. The field of Hg emissions from coal-fired power plants has developed steadily. The research hotspots can be divided into three categories: (1) emission characterization research focused on speciation changes and emission calculations; (2) emission control research focused on control technologies; (3) environmental impact research focused on environmental pollution and health risk. In conclusion, using an oxygen-rich atmosphere for combustion and installing high-efficiency air pollution control devices (APCDs) helped to reduce the formation of Hg 0 . The average Hg removal rates of APCDs and modified adsorbents after ultra-low emission retrofit were distributed in the range of 82–93% and 41–100%, respectively. The risk level of Hg in combustion by-products was highest in desulfurization sludge (RAC > 10%) followed by fly ash (10% < RAC < 30%) and desulfurization gypsum (1% < RAC < 10%). Additionally, we found that the implementation of pollution and carbon reduction policies in China had reduced Hg emissions from CFPPs by 45% from 2007 to 2015, increased the efficiency of Hg removal from APCDs to a maximum of 96%, and reduced global transport and health risk of atmospheric Hg. The results conjunctively achieved by CiteSpace, and the literature review will enhance understanding of CFPP Hg emission research and provide new perspectives for future research.

Graphical Abstract

systematic literature review on air pollution

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

systematic literature review on air pollution

Data availability

The data supporting the findings of this study are publicly available in the paper cited in the text.

Badry A, Palma L, Beja P, Ciesielski T, Dias A, Lierhagen S, Jenssen B, Sturaro N, Eulaers I, Jaspers V (2019) Using an apex predator for large-scale monitoring of trace element contamination: associations with environmental, anthropogenic and dietary proxies. Sci Total Environ 676:746–755. https://doi.org/10.1016/j.scitotenv.2019.04.217

Article   ADS   CAS   PubMed   Google Scholar  

Brown TD, Smith DN, Hargis RA, O’Dowd WJ (1999) Mercury measurement and its control: what we know, have learned, and need to further investigate. J Air Waste Manage Assoc 49:1–97. https://doi.org/10.1080/10473289.1999.10463841

Article   CAS   Google Scholar  

Burmistrz P, Kogut K, Marczak M, Zwozdziak J (2016) Lignites and subbituminous coals combustion in Polish power plants as a source of anthropogenic mercury emission. Fuel Process Technol 152:250–258. https://doi.org/10.1016/j.fuproc.2016.06.011

Cai B, Li Q, Zhang X (2021) China carbon dioxide capture utilization and storage (CCUS) annual report 2021. Chinese Academy oF Environmental Planning. https://www.caep.org.cn/sy/dqhj/gh/202107/t20210725_851241.shtml

Cao X, Zhang J, Wang Y, Liu Y, He Z, Luxiang Jiang, Na Guo, Wang Zuwu Hu, Jiangjun, (2020) Removing mercury from flue gas of coal-fired power plants by different ultra-low emission technical routes. Environ Sci Technol 43:194–199. https://doi.org/10.19672/j.cnki.1003-6504.2020.10.026

Chang L, Zhao Y, Zhang Y, Yu X, Li Z, Gong B, Liu H, Wei S, Wu H, Zhang J (2021) Mercury species and potential leaching in sludge from coal-fired power plants. J Hazard Mater 403:123927. https://doi.org/10.1016/j.jhazmat.2020.123927

Article   CAS   PubMed   Google Scholar  

Chao W, Yang M, Cao X (2021) Coal-fired power plants and metal research based on CiteSpace. Yunnan Chem Technol 48:115–117. https://doi.org/10.3969/j.issn.1004-275X.2021.02.40

Article   Google Scholar  

Chen CM (2006) CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literature. J Am Soc Inf Sci Technol 57:359–377. https://doi.org/10.1002/asi.20317

Chen C (2017) Science mapping: a systematic review of the literature. J Data Inform Sci 2:1–40. https://doi.org/10.1515/jdis-2017-0006

Chen C, Hu Z, Liu S, Tseng H (2012) Emerging trends in regenerative medicine: a scientometric analysis in CiteSpace. Expert Opin Biol Ther 12:593–608. https://doi.org/10.1517/14712598.2012.674507

Chen L, Duan Y, Zhao S, Li Y (2020) Mercury Co-removal by the air pollutant control devices in a 350 MW ultra-low emission coal-fired power plant. J Eng Therm Energy Power 35(187–193):200

Google Scholar  

Cheng Z, Tang Y, Li E, Wu Q, Wang L, Liu K, Wang S, Huang Y, Duan L (2020) Mercury accumulation in soil from atmospheric deposition in temperate steppe of Inner Mongolia, China. Environ Pollut 258. https://doi.org/10.1016/j.envpol.2019.113692

Chmielniak T, Słowik K, Sajdak M (2017) Mercury removal by mild thermal treatment of coal. Fuel 195:290–298. https://doi.org/10.1016/j.fuel.2017.01.073

Dabrowski JM, Ashton PJ, Murray K, Leaner JJ, Mason RP (2008) Anthropogenic mercury emissions in South Africa: Coal combustion in power plants. Atmos Environ 42:6620–6626. https://doi.org/10.1016/j.atmosenv.2008.04.032

Article   ADS   CAS   Google Scholar  

Deeds DA, Banic CM, Lu J, Daggupaty S (2013) Mercury speciation in a coal-fired power plant plume: an aircraft-based study of emissions from the 3640 MW Nanticoke Generating Station, Ontario. Canada J Geophys Res-Atmos 118:4919–4935. https://doi.org/10.1002/jgrd.50349

Diao X, Yuan C-G, Wu J, Zhang K, Zhang C, Gui B (2018) Mercury fractions in gypsum and estimation of mercury emission from coal-fired power plants. Fuel 226:298–306. https://doi.org/10.1016/j.fuel.2018.04.002

Dranga B-A, Lazar L, Koeser H (2012) Oxidation catalysts for elemental mercury in flue gases—a review. Catalysts 2:139–170. https://doi.org/10.3390/catal2010139

Dronen LC, Moore AE, Kozliak EI, Seames WS (2004) An assessment of acid wash and bioleaching pre-treating options to remove mercury from coal. Fuel 83:181–186. https://doi.org/10.1016/S0016-2361(03)00251-5

Fu J, Zhong Z, Xu Y, Xue J (2022) Effect of chlorine addition on mercury speciation transformation and mercury-chlorine reaction mechanism in flue gas. Elect Power Technol Environ Protect 38:27–35. https://doi.org/10.19944/j.eptep.1674-8069.2022.01.004

Gao L, Wang Y, Huang Q, Guo S (2017) Emission of mercury from six low calorific value coal-fired power plants. Fuel 210:611–616. https://doi.org/10.1016/j.fuel.2017.09.001

Gao X, Wang Q, Xiong D, Lu X, Jianbo L, Rongdi Z, Xiaofei L, Jiabao Z, Xingdong Z (2022) Migration characteristics of mercury and mercury adsorption characteristics of fly ash in CFB boiler firing high mercury coal. Thermal Power Generation 51:120–126. https://doi.org/10.19666/j.rlfd.202110207

George J, Masto R, Ram L, Das T, Rout T, Mohan M (2015) Human exposure risks for metals in soil near a coal-fired power-generating plant. Arch Environ Contam Toxicol 68:451–461. https://doi.org/10.1007/s00244-014-0111-x

Giang A, Stokes LC, Streets DG, Corbitt ES, Selin NE (2015) Impacts of the Minamata Convention on mercury emissions and global deposition from coal-fired power generation in Asia. Environ Sci Technol 49:5326–5335. https://doi.org/10.1021/acs.est.5b00074

Goodarzi F (2004) Speciation and mass-balance of mercury from pulverized coal fired power plants burning western Canadian subbituminous coals. J Environ Monit 6:792–798. https://doi.org/10.1039/b401827c

Guffey FD, Bland AE (2004) Thermal pretreatment of low-ranked coal for control of mercury emissions. Fuel Process Technol 85:521–531. https://doi.org/10.1016/j.fuproc.2003.11.006

Guo J, Liu S, Zhang Y, Sun R (2020) Impacts of ultra-low emission in coal-fired power plants on the distribution characteristics of mercury. Environ Monitor China 36:55–59. https://doi.org/10.19316/j.issn.1002-6002.2020.01.07

Guo Y, Xiao L, Chen B, Wu Z, Chen H, Li J (2021) Rapidly changing coal-related city-level atmospheric mercury emissions and their driving forces. J Hazard Mater 411:125060. https://doi.org/10.1016/j.jhazmat.2021.125060

Han D, Xu L, Wu Q, Wang S, Duan L, Wen M, Li Z, Tang Y, Li G, Liu K (2021) Potential environmental risk of trace elements in fly ash and gypsum from ultra–low emission coal–fired power plants in China. Sci Total Environ 798:149116. https://doi.org/10.1016/j.scitotenv.2021.149116

Hassett-Sipple B, Swartout J, Schoeny R 1997 Mercury study report to Congress. Volume 5. Health effects of mercury and mercury compounds (No. PB-98–124779/XAB; EPA-452/R-97/007). Environmental Protection Agency, Research Triangle Park, NC (United States). Office of Air Quality Planning and Standards. https://doi.org/10.2172/575119

Hou J, Yang X, Chen C (2018) Emerging trends and new developments in information science: a document co-citation analysis (2009–2016). Scientometrics 115:869–892. https://doi.org/10.1007/s11192-018-2695-9

Hua X, Zhang L, Song Y, Lu H (2016) Influence of the ultra low emission modification of a coal-fired unit on the mercury emissions. J Eng Therm Energ Power 31:110–116. https://doi.org/10.16146/j.cnki.rndlgc.2016.07.018

Huang H, Luo J (2010) Effect of various fly ash compositions on mercury speciation transformation. Zhongguo Dianji Gongcheng Xuebao/Proc Chin Soc Elec Eng 30:70–75. https://doi.org/10.13334/j.0258-8013.pcsee.2010.s1.003

Hylander L, Goodsite M (2006) Environmental costs of mercury pollution. Sci Total Environ 368:352–370. https://doi.org/10.1016/j.scitotenv.2005.11.029

Jiang J, Hao J, Wu Y, Streets D, Duan L, Tian H 2005 Development of mercury emission inventory from coal combustion in China. Environ Sci 34–39. https://doi.org/10.13227/j.hjkx.2005.02.007

Kang H, Liu X, Guo J, Wang B, Xu G, Wu G, Kang S, Huang J (2019) Characterization of mercury concentration from soils to needle and tree rings of Schrenk spruce (Picea schrenkiana) of the middle Tianshan Mountains, northwestern China. Ecol Indic 104:24–31. https://doi.org/10.1016/j.ecolind.2019.04.066

Kim J-H, Pudasainee D, Yoon Y-S, Son S-U, Seo Y-C (2010) Studies on speciation changes and mass distribution of mercury in a bituminous coal-fired power plant by combining field data and chemical equilibrium calculation. Ind Eng Chem Res 49:5197–5203. https://doi.org/10.1021/ie901361q

Kimakova T, Nasser B, Issa M, Uher I (2019) Mercury cycling in the terrestrial, aquatic and atmospheric environment of the Slovak Republic — an overview. Ann Agr Env Med 26:273–279. https://doi.org/10.26444/aaem/105395

Konkler M, Hammerschmidt C (2012) Methylmercury in mosquitoes around a large coal-fired power plant in central Ohio. Environ Toxicol Chem 31:1657–1661. https://doi.org/10.1002/etc.1839

Lee S-S, Wilcox J (2017) Behavior of mercury emitted from the combustion of coal and dried sewage sludge: the effect of unburned carbon, Cl, Cu and Fe. Fuel 203:749–756. https://doi.org/10.1016/j.fuel.2017.04.104

Lee S, Seo Y, Jang H, Park K, Baek J, An H, Song K (2006) Speciation and mass distribution of mercury in a bituminous coal-fired power plant. Atmos Environ 40:2215–2224. https://doi.org/10.1016/j.atmosenv.2005.12.013

Lewandowski T (2010) Evolving understanding of the relationship between mercury exposure and autism, in: Simeonov, L., Kochubovski, M., Simeonova, B. (Eds.), City University of New York (CUNY) System. Presented at the Environmental Heavy Metal Pollution and Effects on Child Mental Development: Risk Assessment and Prevention Strategies, pp. 65–84. https://doi.org/10.1007/978-94-007-0253-0_4

Li J (2015) Technology of flue gas ultra-low emission for coal-fired power plant. China Electric Power Press, Beijing

Li R, Wu H, Ding J, Fu W, Gan L, Li Y (2017) Mercury pollution in vegetables, grains and soils from areas surrounding coal-fired power plants. Sci Rep 7:46545. https://doi.org/10.1038/srep46545

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Li N, Wei H, Duan Y, Wang C, She M, Hu P, Wang S (2018) Effects of SO 2 and O 2 on Hg~0 removal capacity of sulfur modified petrol coke adsorbents. Chem Indust Eng Progr 37:1908–1915. https://doi.org/10.16085/j.issn.1000-6613.2017-1399

Li G, Wu Q, Xu L, Wen M, Liu K, Tang Y, Zou J, Wang F, Wang Y, Wang S (2019a) A review on adsorption technologies for mercury emission control. Bull Environ Contam Toxicol 103:155–162. https://doi.org/10.1007/s00128-019-02648-4

Li X, Li Z, Fu C, Tang L, Chen J, Wu T, Lin C-J, Feng X, Fu X (2019b) Fate of mercury in two CFB utility boilers with different fueled coals and air pollution control devices. Fuel 251:651–659. https://doi.org/10.1016/j.fuel.2019.04.071

Li F, Ma C, Zhang P (2020a) Mercury deposition, climate change and anthropogenic activities: a review. Front Earth Sci 8:316. https://doi.org/10.3389/feart.2020.00316

Article   ADS   Google Scholar  

Li J, Zhou S, Wei W, Qi J, Li Y, Chen B, Zhang N, Guan D, Qian H, Wu X, Miao J, Chen L, Feng K, Liang S (2020b) China’s retrofitting measures in coal-fired power plants bring significant mercury-related health benefits. One Earth 3:777–787. https://doi.org/10.1016/j.oneear.2020.11.012

Li H, Huang J, Yang J, Yang Z, Qu W, Xu Z, Shih K (2021) Reduction of oxidized mercury over NOx selective catalytic reduction catalysts: a review. Chem. Eng. J. 421. https://doi.org/10.1016/j.cej.2020.127745

Li Y, Yu J, Liu Y, Huang R, Wang Z, Zhao Y (2022) A review on removal of mercury from flue gas utilizing existing air pollutant control devices (APCDs). J Hazard Mater 427 https://doi.org/10.1016/j.jhazmat.2021.128132

Liming C, Huang Z, Mo H, Zhu J, Jiang A, Huang R (2017) Test and study on synergic mercury removal performance of environmental protection facilities at ultra-low pollutants emission. Electric Power 50(136–139):143

Lipfert F, Morris S, Sullivan T, Moskowitz P, Renninger S (2005) Methylmercury, fish consumption, and the precautionary principle. J Air Waste Manage Assoc 55:388–398. https://doi.org/10.1080/10473289.2005.10464637

Liu Z, Yin Y, Liu W, Dunford M (2015) Visualizing the intellectual structure and evolution of innovation systems research: a bibliometric analysis. Scientometrics 103:135–158. https://doi.org/10.1007/s11192-014-1517-y

Liu K, Wang S, Wu Q, Wang L, Ma Q, Zhang L, Li G, Tian H, Duan L, Hao J (2018) A highly resolved mercury emission inventory of Chinese coal-fired power plants. Environ Sci Technol 52:2400–2408. https://doi.org/10.1021/acs.est.7b06209

Liu H, Chen Z, Wang W, Cai X, Jin W, Gaofei G, Minchao S, Bo Z (2019) Mercury emission characteristics and research progress of its adsorption removal technology of flue gas in coal-fired power plants. Environ Eng 37:128–133

Liu X, Guo C, Wu Y, Huang C, Lu K, Zhang Y, Duan L, Cheng M, Chai F, Mei F, Dai H (2023) Evaluating cost and benefit of air pollution control policies in China: a systematic review. J Environ Sci 123:140–155. https://doi.org/10.1016/j.jes.2022.02.043

Luo G-Q, Yao H, Xu M-H (2009) Partitioning behavior of mercury during coal combustion: the influence of low-NOx burners and operation load of boiler. Asia-Pac J Chem Eng 4:480–486. https://doi.org/10.1002/apj.259

MacFarlane S, Fisher JA, Horowitz HM, Shah V (2022) Two decades of changing anthropogenic mercury emissions in Australia: inventory development, trends, and atmospheric implications. Environ Sci Process Impacts 24:1474–1493. https://doi.org/10.1039/D2EM00019A

Martin J, Gutierrez C, Escuer M, Martin-Dacal M, Ramos-Miras J, Roca-Perez L, Boluda R, Nanos N (2021) Trends in soil mercury stock associated with pollution sources on a Mediterranean island (Majorca, Spain). Environ Pollut 283. https://doi.org/10.1016/j.envpol.2021.117397

Martin J, Nanos N (2016) Soil as an archive of coal-fired power plant mercury deposition. J Hazard Mater 308:131–138. https://doi.org/10.1016/j.jhazmat.2016.01.026

Moral-Muñoz JA, Herrera-Viedma E, Santisteban-Espejo A, Cobo MJ (2020) Software tools for conducting bibliometric analysis in science: an up-to-date review. Profesional de la información 29. https://doi.org/10.3145/epi.2020.ene.03

Nelson PF (2007) Atmospheric emissions of mercury from Australian point sources. Atmos Environ 41:1717–1724. https://doi.org/10.1016/j.atmosenv.2006.10.029

Pavlish JH, Sondreal EA, Mann MD, Olson ES, Galbreath KC, Laudal DL, Benson SA (2003) Status review of mercury control options for coal-fired power plants. Fuel Process Technol 82:89–165. https://doi.org/10.1016/S0378-3820(03)00059-6

Qi H, Xu W, Wang J, Tong L, Zhu T (2015) Hg-O removal from flue gas over different zeolites modified by FeCl3. J Environ Sci 28:110–117. https://doi.org/10.1016/j.jes.2014.05.050

Qian L, Xu Z, Li Z, Pan S (2016) Research of mercury emissions and removal effects after ultra-low emission modification of coal-fired units. Environ Sci Manag 41(64):67

Shen M-Q (2009) Study on the effect of NH_4Cl addition on Hg and NO produced by coal combustion. Proceedings of the CSEE 29: 41–46

Sherman LS, Blum JD, Keeler GJ, Demers JD, Dvonch JT (2012) Investigation of local mercury deposition from a coal-fired power plant using mercury isotopes. Environ Sci Technol 46:382–390. https://doi.org/10.1021/es202793c

Song C, Liu Z, Wang T, An L, Zhang Y (2017a) Emission characteristics of PM, SO 2 , NOx and Hg of ultra-low emission coal-fired power plant. J North China Elect Power Univ (natural Science Edition) 44:93–99

Song C, Yi Z, Jian HAO, Gengsheng LIU, Jiawei W, Liansuo AN, Yongsheng Z (2017b) Mercury emission characteristics from coal-fired power plant before and after ultra-low emission retrofitting. Res Environ Sci 30:672–677. https://doi.org/10.13198/j.issn.1001-6929.2017.01.96

Song Z, Wang C, Ding L, Chen M, Hu Y, Li P, Zhang L, Feng X (2021) Soil mercury pollution caused by typical anthropogenic sources in China: evidence from stable mercury isotope measurement and receptor model analysis. J Clean Prod 288:10. https://doi.org/10.1016/j.jclepro.2020.125687

Su S, Liu L, Wang L, Syed-Hassan, Syed SA, Kong F, Hu S, Wang Y, Jiang L, Xu K, Zhang A, Xiang J (2017) Mass flow analysis of mercury transformation and effect of seawater flue gas desulfurization on mercury removal in a full-scale coal-fired power plant. Energy Fuels 31(11109):11116. https://doi.org/10.1021/acs.energyfuels.7b02029

Subhavana K, Qureshi A, Roy A (2019) Mercury levels in human hair in South India: baseline, artisanal goldsmiths and coal-fired power plants. J Expo Sci Environ Epidemiol 29:697–705. https://doi.org/10.1038/s41370-018-0107-0

Sunderland EM, Driscoll CT, Hammitt JK, Grandjean P, Evans JS, Blum JD, Chen CY, Evers DC, Jaffe DA, Mason RP, Goho S, Jacobs W (2016) Benefits of regulating hazardous air pollutants from coal and oil fired utilities in the United States. Environ Sci Technol 50:2117–2120. https://doi.org/10.1021/acs.est.6b00239

Sung J-H, Back S-K, Jung B-M, Kang Y-S, Lee C-G, Jang H-N, Seo Y-C (2017) Speciation and capture performance of mercury by a hybrid filter in a coal-fired power plant. Int J Coal Geol 170:35–40. https://doi.org/10.1016/j.coal.2016.10.008

Sung J-H, Oh J-S, Mojammal AHM, Back S-K, Lee E-S, Kim S-H, Seo Y-C (2018) Estimation and future prediction of mercury emissions from anthropogenic sources in South Korea. J Chem Eng Jpn 51:800–808. https://doi.org/10.1252/jcej.17we129

Thomas Klasson K, Borole AbhijeetP, McKeown CatherineK, Hamilton ChooY (2006) Biomodification of coal to remove mercury. Appl Biochem Biotechnol 131:897–908. https://doi.org/10.1385/ABAB:131:1-3:897

Article   PubMed   Google Scholar  

Tian H, Wang Y, Xue Z, Qu Y, Chai F, Hao J (2011) Atmospheric emissions estimation of Hg, As, and Se from coal-fired power plants in China, 2007. Sci Total Environ 409:3078–3081. https://doi.org/10.1016/j.scitotenv.2011.04.039

Tian H, Liu K, Zhou J, Lu L, Hao J, Qiu P, Gao J, Zhu C, Wang K, Hua S (2014) Atmospheric emission inventory of hazardous trace elements from China’s coal-fired power plants—temporal trends and spatial variation characteristics. Environ Sci Technol 48:3575–3582. https://doi.org/10.1021/es404730j

Trasande L, Schechter C, Haynes KA, Landrigan PJ (2006) Applying cost analyses to drive policy that protects children — mercury as a case study, in: Mehlman M, Soffritti M, Landrigan P, Bingham E, Belpoggi F (Eds.), Living In A Chemical World: Framing The Future In Light Of The Past. pp. 911–923. https://doi.org/10.1196/annals.1371.034

Wang W (2020) Distribution and emission characteristics, migration and release regulation of mercury in a 600 MW coal-fired power plant. Environ Eng 38:195–201. https://doi.org/10.13205/j.hjgc.202006032

Wang M, Keener TC, Khang SJ (2000) The effect of coal volatility on mercury removal from bituminous coal during mild pyrolysis. Fuel Process Technol 67:147–161. https://doi.org/10.1016/S0378-3820(00)00098-9

Wang Y, Duan Y, Yang L, Jiang Y (2008) An analysis of the factors exercising an influence on the morphological transformation of mercury in the flue gas of a 600 MW coal-fired power plant. J Eng Therm Energy Power 23:399–403

Wang SX, Zhang L, Li GH, Wu Y, Hao JM, Pirrone N, Sprovieri F, Ancora MP (2010) Mercury emission and speciation of coal-fired power plants in China. Atmos Chem Phys 10:1183–1192. https://doi.org/10.5194/acp-10-1183-2010

Wang J, Cai B, Zhang L, Cao D, Liu L, Zhou Y, Zhang Z, Xue W (2014) High resolution carbon dioxide emission gridded data for China derived from point sources. Environ Sci Technol 48:7085–7093. https://doi.org/10.1021/es405369r

Wang S, Yu X, Gu Y, Yuan J, Zhang Y, Chen Y, Chai F (2018) Discussion of emission limits of air pollutants for ‘near-zero emission’ coal-fired power plants. Res Environ Sci 31:975–984

Wang T, Lou Y, Jiang S, Wang J, Zhang Y, Pan W (2022) Distribution characteristics and environmental risk assessment of trace elements in desulfurization sludge from coal-fired power plants. Fuel 314. https://doi.org/10.1016/j.fuel.2021.122771

Wen M, Wu Q, Li G, Wang S, Li Z, Tang Y, Xu L, Liu T (2020) Impact of ultra-low emission technology retrofit on the mercury emissions and cross-media transfer in coal-fired power plants. J Hazard Mater 396:8. https://doi.org/10.1016/j.jhazmat.2020.122729

Wiener JG, Evers DC, Gay DA, Morrison HA, Williams KA (2012) Mercury contamination in the Laurentian Great Lakes region: introduction and overview. Environ Pollut 161:243–251. https://doi.org/10.1016/j.envpol.2011.08.051

Wu Y, Streets DG, Wang SX, Hao JM (2010) Uncertainties in estimating mercury emissions from coal-fired power plants in China. Atmos Chem Phys 10:2937–2946. https://doi.org/10.5194/acp-10-2937-2010

Xu W, Hussain A, Liu Y (2018) A review on modification methods of adsorbents for elemental mercury from flue gas. Chem Eng J 346:692–711. https://doi.org/10.1016/j.cej.2018.03.049

H Xun C Leming C Yi H Wenhui Z Jinsong (2014) Mercury Migration between Flue Gas and Fly Ash in Circulating Fluidized Bed https://doi.org/10.3969/j.issn.0438-1157.2014.04.032

Yanbing Z, Yingzan C, Zheng G, Mengxi Z, Jinhe H, Chenyang C, Xianglong C, Haibo R (2021) Research progress of catalyst for simultaneous denitration and mercury removal. Low-Carbon Chem Chem Eng 46:9–16. https://doi.org/10.3969/j.issn.1001-9219.2021.06.002

Yang X, Wang L (2008) Spatial analysis and hazard assessment of mercury in soil around the coal-fired power plant: a case study from the city of Baoji, China. Environ Geol 53:1381–1388. https://doi.org/10.1007/s00254-007-0747-1

Yang H, Xu Z, Fan M, Bland AE, Judkins RR (2007a) Adsorbents for capturing mercury in coal-fired boiler flue gas. J Hazard Mater 146:1–11. https://doi.org/10.1016/j.jhazmat.2007.04.113

Yang X, Zhuo Y, Duan Y, Chen L, Yang L, Zhang L, Jiang Y, Xu X (2007b) Mercury speciation and its emissions from a 220 MW pulverized coal-fired boiler power plant in flue gas. Korean J Chem Eng 24:711–715. https://doi.org/10.1007/s11814-007-0031-9

Yang Y, Xu W, Wu Y, Xiong J, Zhu T, Zhou X, Tong L (2016) Effect of HBr formation on mercury oxidation via CaBr 2 addition to coal during combustion. RSC Adv 6:59009–59015. https://doi.org/10.1039/C6RA11468G

Yu Y, Jin X, Li J, Jia L (2021) Research on path of atmospheric mercury emission control in China’s power industry. J Central South Univ Sci Technol 52:228–238. https://doi.org/10.11817/j.issn.1672-7207.2021.01.023

Zhang X, Zhang J, Cai T, Zhang B (2015) Factor analysis of mercury form transformation in coal-fired flue gas. China Resour Comprehen Utiliz 33:53–56. https://doi.org/10.3969/j.issn.1008-9500.2015.02.029

Zhang W, Zhen G, Chen L, Wang H, Li Y, Ye X, Tong Y, Zhu Y, Wang X (2017a) Economic evaluation of health benefits of mercury emission controls for China and the neighboring countries in East Asia. Energy Policy 106:579–587. https://doi.org/10.1016/j.enpol.2017.04.010

Zhang Y, Yang J, Yu X, Sun P, Zhao Y, Zhang J, Chen G, Yao H, Zheng C (2017b) Migration and emission characteristics of Hg in coal-fired power plant of China with ultra low emission air pollution control devices. Fuel Process Technol 158:272–280. https://doi.org/10.1016/j.fuproc.2017.01.020

Zhao Y, Han L (2019) Distribution characteristics of mercury in 660 MW coal-fired power plant with ultra-low emission. Acta Sci Circums 39:853–858. https://doi.org/10.13671/j.hjkxxb.2018.0221

Zhao Y, Mann MD, Olson ES, Pavlish JH, Dunham GE (2006) Effects of sulfur dioxide and nitric oxide on mercury oxidation and reduction under homogeneous conditions. J Air Waste Manage Assoc 56:628–635. https://doi.org/10.1080/10473289.2006.10464483

Zhao S, Duan Y, Chen L, Li Y, Yao T, Liu S, Liu M, Lu J (2017a) Study on emission of hazardous trace elements in a 350 MW coal-fired power plant. Part 1. Mercury Environ Pollut 229:863–870. https://doi.org/10.1016/j.envpol.2017.07.043

Zhao S, Duan Y, Yao T, Liu M, Lu J, Tan H, Wang X, Wu L (2017b) Study on the mercury emission and transformation in an ultra-low emission coal-fired power plant. Fuel 199:653–661. https://doi.org/10.1016/j.fuel.2017.03.038

Zhao S, Duan Y, Lu JC, Gupta R, Pudasainee D, Liu S, Liu M, Lu JH (2018) Thermal stability, chemical speciation and leaching characteristics of hazardous trace elements in FGD gypsum from coal-fired power plants. Fuel 231:94–100. https://doi.org/10.1016/j.fuel.2018.05.067

Zhao S, Pudasainee D, Duan Y, Gupta R, Liu M, Lu J (2019) A review on mercury in coal combustion process: content and occurrence forms in coal, transformation, sampling methods, emission and control technologies. Prog Energy Combust Sci 73:26–64. https://doi.org/10.1016/j.pecs.2019.02.001

Zhu J, Hua W (2017) Visualizing the knowledge domain of sustainable development research between 1987 and 2015: a bibliometric analysis. Scientometrics 110:893–914. https://doi.org/10.1007/s11192-016-2187-8

Zhuang Y, Thompson JS, Zygarlicke CJ, Pavlish JH (2007) Impact of calcium chloride addition on mercury transformations and control in coal flue gas. Fuel 86:2351–2359. https://doi.org/10.1016/j.fuel.2007.02.016

Download references

The authors received financial support from the National Nature Science Foundation of China (no. 52200121).

Author information

Authors and affiliations.

School of Energy and Environmental Engineering, University of Science and Technology Beijing, Beijing, 100083, China

Qi Liu, Jiajia Gao, Guoliang Li, Yang Zheng, Rui Li & Tao Yue

You can also search for this author in PubMed   Google Scholar

Contributions

Qi Liu: conceptualization, data curation, writing — original draft. Jiajia Gao: methodology, writing — review and editing. Guoliang Li: funding acquisition. Yang Zheng: supervision. Rui Li: methodology. Tao Yue: supervision, conceptualization.

Corresponding author

Correspondence to Tao Yue .

Ethics declarations

Ethics approval.

Not applicable.

Consent to participate

Consent for publication.

All authors have consented to publish this work.

Competing interests

The authors declare no competing interests.

Additional information

Responsible Editor: Philippe Garrigues

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Liu, Q., Gao, J., Li, G. et al. Bibliometric analysis on mercury emissions from coal-fired power plants: a systematic review and future prospect. Environ Sci Pollut Res (2024). https://doi.org/10.1007/s11356-024-32369-z

Download citation

Received : 30 May 2023

Accepted : 03 February 2024

Published : 21 February 2024

DOI : https://doi.org/10.1007/s11356-024-32369-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Mercury emissions
  • Coal-fired power plants (CFPPs)
  • Emission characteristics
  • Environmental impact
  • Emission control
  • Find a journal
  • Publish with us
  • Track your research

IMAGES

  1. (PDF) Air Pollution in South Asia and its impact (A literature review)

    systematic literature review on air pollution

  2. systematic review on air pollution

    systematic literature review on air pollution

  3. Literature Review Air Pollution Dispersion Modelling

    systematic literature review on air pollution

  4. (PDF) Air Pollution and Otitis Media in Children: A Systematic Review

    systematic literature review on air pollution

  5. Book review: Fundamentals of Air Pollution

    systematic literature review on air pollution

  6. (PDF) A Review of Epidemiological Research on Adverse Neurological

    systematic literature review on air pollution

VIDEO

  1. Systematic Literature Review, by Prof. Ranjit Singh, IIIT Allahabad

  2. Systematic Literature Review Paper

  3. Systematic Literature Review Paper presentation

  4. Article on pollution/essay on pollution in english/pollution trick

  5. Webinar Systematic Literature Review sebagai Alternatif Artikel Ilmiah

  6. Introduction Systematic Literature Review-Various frameworks Bibliometric Analysis

COMMENTS

  1. Air pollution: A systematic review of its psychological, economic, and social effects

    This review (178 published articles) is the first to systematically examine the psychological (affective, cognitive, behavioral), economic, and social effects of air pollution beyond its physiological and environmental effects.

  2. Effects of air pollution on health: A mapping review of systematic

    Abstract Background: There has been a notable increase in knowledge production on air pollution and human health. Objective: To analyze the state of the art on the effects of air pollution on human health through a mapping review of existing systematic reviews and meta-analyses (SRs and MAs).

  3. Environmental and Health Impacts of Air Pollution: A Review

    Abstract One of our era's greatest scourges is air pollution, on account not only of its impact on climate change but also its impact on public and individual health due to increasing morbidity and mortality. There are many pollutants that are major factors in disease in humans.

  4. Review of epidemiological studies on air pollution and health effects

    Numerous epidemiologic studies have investigated the respiratory effects of air pollution on children, including lung function and respiratory symptoms/diseases. The systematic reviews and meta-analyses on ambient air pollution and children's respiratory health examined here are shown in Table 1 [ 8 - 21 ].

  5. Systematic reviews and metaanalyses of air pollution epidemiological

    This chapter provides basic definitions and methods used to synthesize epidemiological studies of air pollution effects in systematic reviews and metaanalyses and describes the use of systematic reviews and metaanalyses in policy- and decision-making and burden of disease/health impact assessments.

  6. Urban air pollution control policies and strategies: a systematic review

    This systematic review comprehensively appraises the policies and strategies on air pollutants controls enacted in different countries, worldwide. Three databases, Web of Science, PubMed and Scopus, were used for the search. After screening, a total of 114 eligible manuscripts were selected from 2219 documents for further analysis.

  7. Effects of air pollution on health: A mapping review of systematic

    Effects of air pollution on health: A mapping review of systematic reviews and meta-analyses - ScienceDirect Volume 201, October 2021, 111487 Effects of air pollution on health: A mapping review of systematic reviews and meta-analyses Fábio Hech, Joaquim Henrique Lorenzetti Branco, Giorgio Buonanno b, Stabile, Gameiro da Silva, Alexandro

  8. A systematic review and meta-analysis on the association between

    Abstract There is inconclusive evidence on the association between ambient air pollution and pulmonary tuberculosis (PTB) incidence, tuberculosis-related hospital admission and mortality. This...

  9. A systematic review of data mining and machine learning for air

    We conducted a systematic literature review on the application of data mining and machine learning methods in air pollution epidemiology. We carried out our search process in PubMed, the MEDLINE database and Google Scholar. Research articles applying data mining and machine learning methods to air pollution epidemiology were queried and reviewed.

  10. A systematic literature review and critical appraisal of

    Abstract Ambient air pollution is the leading environmental risk factor for disease globally. Air pollutants can increase the risk of some respiratory infections, but their effects on tuberculosis (TB) are unclear.

  11. The health impacts of waste-to-energy emissions: a systematic review of

    1 Centre for Air pollution, energy, and health Research (CAR), University of New South Wales, Sydney, Australia. ... To our knowledge, this is the first systematic literature review focused primarily on studies of the health effects associated with WtE-related air emissions. We found that while implementation of WtE technologies is increasing ...

  12. Air pollution and human cognition: A systematic review and meta

    This systematic review summarises and evaluates the literature investigating associations between exposure to air pollution and general population cognition, which has important implications for health, social and economic inequalities, and human productivity. Methods

  13. The health impacts of ambient air pollution in Australia: a systematic

    Following PRISMA guidelines, we conducted a systematic literature review. Broad search terms were applied to two databases (PubMed and Web of Science) and Google Scholar. Quality assessment and risk of bias were assessed using standard metrics. Included studies were summarised by tabulating key study characteristics, grouped by health outcomes.

  14. Air Pollution and Temperature: a Systematic Review of Ubiquitous

    Of 5138 studies identified by our literature search, this review included 30 studies on air pollution, 42 studies on temperature, 6 studies on both air pollution and temperature, and 1 study on altitude exposure and OHCA/SCD. Particulate matter air pollution, ozone, and both hot and cold temperatures are associated with increased risk of OHCA/SCD.

  15. A systematic review and meta-analysis on the association between

    Abstract There is inconclusive evidence on the association between ambient air pollution and pulmonary tuberculosis (PTB) incidence, tuberculosis-related hospital admission and mortality. This review aimed to assess the extent to which selected air pollutants are associated to PTB incidence, hospital admissions and mortality.

  16. Long-term exposure to traffic-related air pollution and selected health

    The Panel used a systematic approach to search the literature, select studies for inclusion in the review, assess study quality, summarize results, and reach conclusions about the confidence in the evidence. An extensive search was conducted of literature published between January 1980 and July 2019 on selected health outcomes.

  17. Urban air pollution control policies and strategies: a systematic review

    Search strategy. The present systematic review was performed according to standard methods the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) [].A comprehensive screening approach to find the urban air pollution control strategies and policies published in peer-reviewed journal was performed up to 19 January 2021 in three scientific databases: PubMed, Web of Science ...

  18. Air pollution interventions and respiratory health: A systematic review

    Abstract and Figures. BACKGROUND: Indoor and ambient air pollution exposure is a major risk to respiratory health worldwide, particularly in low- and middle-income countries (LMICs ...

  19. The impact of air quality on tourism: a systematic literature review

    The objective of this paper is to review what is known and has been published, about the impact of outdoor AQ on tourism demand, using a systematic literature review method. To date, there is no literature review study on this topic. It is of note that the impact of tourism on AQ was also not addressed.

  20. Air Pollution Threatens Millions of Lives. Now the Sources Are Shifting

    Public Health. Particle-based ambient air pollution causes more than 4 million premature deaths each year globally, according to the World Health Organization. The tiniest particles—2.5 microns ...

  21. Effects of air pollution on health: A mapping review of systematic

    The systematic mapping review was based on the recommendations for this type of scientific approach in environmental sciences. The search was performed using PubMed, Web of Science, Scopus, Cinahl, and Cochrane Library databases, from their inception through June 2020. Results

  22. The association between ambient air pollution and migraine: a

    Some studies have shown the effect of air pollution on migraine. However, it needs to be confirmed in larger-scale studies, as scientific evidence is scarce regarding the association between air pollution and migraine. Therefore, this systematic review aims to determine whether there are associations between outdoor air pollution and migraine. A literature search was performed in Scopus ...

  23. A systematic review and meta-analysis of intraday effects of ambient

    A systematic review and meta-analysis of intraday effects of ambient air pollution and temperature on cardiorespiratory morbidities: First few hours of exposure matters to life ... searched for relevant literature by screening the title and abstract, and eligible literature underwent a full-text review. In addition, the reference lists of the ...

  24. Effects of ambient air pollution on obesity and ectopic fat deposition

    Introduction Globally, the prevalence of obesity tripled from 1975 to 2016. There is evidence that air pollution may contribute to the obesity epidemic through an increase in oxidative stress and inflammation of adipose tissue. However, the impact of air pollution on body weight at a population level remains inconclusive. This systematic review and meta-analysis will estimate the association ...

  25. The association between ambient air pollution and migraine: a ...

    Therefore, this systematic review aims to determine whether there are associations between outdoor air pollution and migraine. A literature search was performed in Scopus, Medline (via PubMed), EMBASE, and Web of Science. A manual search for resources and related references was also conducted to complete the search.

  26. The effect of air-pollution and weather exposure on mortality and

    We conducted a scoping literature review. The search on two databases (PubMed and Web-of-Science) from 2012 to 2020 using three conceptual categories of "environmental factors", "health outcomes", and "Geographical region" revealed a total of 951 records.

  27. Travellers' exposure to air pollution: A systematic review and future

    The findings show that (a) air pollution exposure is higher in open than close transport modes, (b) pedestrians and cyclists suffer the most due to higher respiration rates and proximity to the streets, (c) air pollution exposure causes both short and long-term changes in travel behaviour (d) despite the poor air quality, many developing nations...

  28. Sustainability

    Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

  29. Bibliometric analysis on mercury emissions from coal-fired ...

    In conclusion, using an oxygen-rich atmosphere for combustion and installing high-efficiency air pollution control devices (APCDs) helped to reduce the formation of Hg0. ... Through a systematic literature review, it can be concluded that the speciation transformation patterns of Hg in coal combustion and APCD control processes have been ...