Article Text
PDF
Abstract
Objectives Climate change is a major global issue with significant consequences, including effects on air quality and human well-being. This review investigated the projection of non-communicable diseases (NCDs) attributable to air pollution under different climate change scenarios.
Design This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 flow checklist. A population-exposure-outcome framework was established. Population referred to the general global population of all ages, the exposure of interest was air pollution and its projection, and the outcome was the occurrence of NCDs attributable to air pollution and burden of disease (BoD) based on the health indices of mortality, morbidity, disability-adjusted life years, years of life lost and years lived with disability.
Data sources The Web of Science, Ovid MEDLINE and EBSCOhost databases were searched for articles published from 2005 to 2023.
Eligibility criteria for selecting studies The eligible articles were evaluated using the modified scale of a checklist for assessing the quality of ecological studies.
Data extraction and synthesis Two reviewers searched, screened and selected the included studies independently using standardised methods. The risk of bias was assessed using the modified scale of a checklist for ecological studies. The results were summarised based on the projection of the BoD of NCDs attributable to air pollution.
Results This review included 11 studies from various countries. Most studies specifically investigated various air pollutants, specifically particulate matter <2.5 µm (PM2.5), nitrogen oxides and ozone. The studies used coupled-air quality and climate modelling approaches, and mainly projected health effects using the concentration–response function model. The NCDs attributable to air pollution included cardiovascular disease (CVD), respiratory disease, stroke, ischaemic heart disease, coronary heart disease and lower respiratory infections. Notably, the BoD of NCDs attributable to air pollution was projected to decrease in a scenario that promotes reduced air pollution, carbon emissions and land use and sustainable socioeconomics. Contrastingly, the BoD of NCDs was projected to increase in a scenario involving increasing population numbers, social deprivation and an ageing population.
Conclusion The included studies widely reported increased premature mortality, CVD and respiratory disease attributable to PM2.5. Future NCD projection studies should consider emission and population changes in projecting the BoD of NCDs attributable to air pollution in the climate change era.
PROSPERO registration number CRD42023435288.
- PUBLIC HEALTH
- EPIDEMIOLOGIC STUDIES
- Climate Change
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
The effect of climate change is a catalyst that affects human health and reduces air quality. Globally, non-communicable diseases (NCDs) caused 74% of deaths. This review is relevant as air pollution increases the mortality risk due to NCDs.
This review described comprehensive findings on the projected NCD burden of disease (BoD) under various climate change scenarios enhanced with air quality and climate modelling.
This review underwent a systematic critical appraisal using a modified evaluation tool specific to ecological study designs.
One limitation is the lack of studies that use the climate change scenarios based on the Intergovernmental Panel on Climate Change (IPCC) report to project the BoD of NCDs attributable to air pollution. Using the IPCC standardised scenarios (Special Report on Emissions Scenario, Representative Concentration Pathways, Shared Socioeconomic Pathways) would facilitate the comparison of BoD across studies.
Another limitation is the lack of studies that measured health outcomes using the years of life lost, years lived with disability or disability-adjusted life years. Most studies focused on the mortality aspect of NCDs.
Introduction
In the next decade, climate change will emerge as a significant environmental challenge pivotal in influencing air pollution levels.1 The Intergovernmental Panel on Climate Change (IPCC) has projected that if climate change continues at the current rate, global warming is likely to increase by 1.5°C between 2030 and 2052.2 Furthermore, rising temperatures and extreme weather events, such as heat waves, cause air quality deterioration. The WHO determined that six air pollutants (particulate matter (PM), including PM10 and PM2.5, ozone (O3), carbon monoxide (CO), nitrogen dioxide (NO2), lead, sulphur dioxide (SO2)) are harmful to both humans and ecosystems.3 A global temperature increases human exposure to air pollution, escalating the risk of climate-related diseases4 particularly among vulnerable populations, such as children and elderly people.
Much research has assessed the health effects of air pollution worldwide, including short-term5 6 and long-term exposure.7–10 Compelling epidemiology results link PM2.5 to an increased mortality risk from non-communicable diseases (NCDs), such as cardiovascular disease (CVD),11 12 respiratory disease13–16 and lung cancer.17–19 The effects of other pollutants such as NO2 and surface O3 concentration on respiratory disease20 21 and CVD22 23 have also been extensively studied.
The WHO reported that around 12.6 million NCD deaths annually are linked to environmental factors, with air pollution being the main cause of NCDs.24 The global deaths from NCDs attributable to air pollution increased between 1990 and 2019 (Institute of Health Metric and Evaluation).25 From 1990 to 2019, the top five causes of death due to environment-caused NCDs were ischaemic heart disease (IHD, 12.2% to 16.17%), stroke (9.81% to 11.59%), diabetes (1.42% to 2.74%), chronic obstructive pulmonary disease (COPD, 5.4% to 5.8%) and lung cancer (2.28% to 3.61%).25 In addition to NCD-related death, premature death is also connected to air pollution, resulting in an annual premature death rate of up to 7 million.26 A global assessment estimated a 4% increase in premature mortality (100 000 deaths) and a 5% increase (6300 deaths) in the later part of the 21st century due to elevated PM2.5 and O3 levels, respectively.27
Currently, there is a growing body of knowledge worldwide on the projection of air pollution concentration changes driven by future emissions and population change.1 This knowledge includes significant progress in quantifying the projected effect of air pollution-related diseases under climate change scenarios, such as the Representative Concentration Pathways (RCPs) and Shared Socioeconomic Pathways (SSPs), using air quality6 28 and climate models.29 30
Previous reviews on climate change, air quality and health effects did not address the RCP-based and SSP-based scenario.31 32 Two reviews were conducted approximately a decade ago.31 32 Both reviews included short-term and long-term projections, but the long-term projection studies were only linked to the Special Report on Emission Scenarios (SRES).
Thus, a comprehensive and extensive understanding of current research on NCDs attributable to air pollution is imperative. The improved understanding would provide valuable insight for future research and effective strategies to mitigate and adapt to climate change. In the present study, we reviewed scientific studies to assess the projection of the burden of disease (BoD) of NCDs attributable to air pollution in the climate change era.
Method
Study design
This systematic review is registered with PROSPERO and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement.33
Research question formulation
The review question was developed based on the population-exposure-outcome (PEO) concept.34 Based on this concept, population referred to the general global population of all ages, the exposure of interest was air pollution and its projection and the outcome was the occurrence of NCDs attributable to air pollution and BoD based on the health indices of mortality, morbidity, disability-adjusted life years (DALY), years of life lost (YLL) and years lived with disability (YLD). The PEO concept guided the formulation of the main review question: What is the projection of the BoD of NCDs (ie, CVDs, respiratory disease) attributable to air pollution?
Search strategy and selection criteria
For this systematic review, we conducted a systematic search of EBSCOhost, Ovid MEDLINE and the Web of Science for studies evaluating the correlation between exposure to air pollution and NCD measures based on the BoD and to assess its projection. We included all studies of any design published from 2005 to 2023. The full search was conducted in May to June 2023 and is depicted in online supplemental file. We identified 709 potentially relevant articles from the three databases. A total of 15 duplicate articles were removed, leaving 694 articles for title screening. The articles were exported from the databases and arranged for screening in an Excel sheet.
Supplemental material
Inclusion and exclusion criteria
The inclusion criteria were: (1) publication from 2005 to June 2023; (2) original article; (3) in the English language; (4) related to NCDs attributable to air pollution projection as a health outcome; and (5) reported the health indices (mortality/morbidity/DALY/YLL/YLD). The exclusion criteria were: (1) non-original articles, such as conference proceedings, reports, letters to the editor, systematic reviews and meta-analyses; (2) experimental studies (animal or cell culture-based studies) and non-ecological studies; (3) studies on air quality projection or climate projection only without health outcomes.
Study selection
NK and RH screened the titles and abstracts independently according to their relevance based on the review question. We removed 516 articles during the screening. The remaining 64 articles proceeded to full-text retrieval for further assessment and eligibility. NK and RH also assessed the eligibility of the articles according to the inclusion and exclusion criteria. Disagreements during any stage of the review were resolved by discussion with a third researcher (NA) to reach a consensus. 53 articles were excluded as they focused only on air quality projection or climate projection (n=19), had a non-ecological study design (n=11), the research focus was not related to ambient air pollution (n=49), were non-original articles (n=14), did not report health outcomes (n=28) and did not report projection (n=43). Some studies were excluded due to duplication (n=1) and publication beyond the target period (n=3). Subsequently, the quality of the remaining 11 articles was appraised.
Quality assessment
NK and RH assessed the quality of the 11 remaining articles using a modified scale of a checklist designed for assessing the quality of ecological studies.35–37 Disagreements were resolved by discussion with a third researcher (NA) to reach a consensus. We used this modified scale as we believe it is the best approach to assess the quality of ecological studies, where it incorporates methodological characteristics such as sampling-based on the ecological unit, data aggregation level and analytical method. The most concerning issue was using and adjusting covariates in the regression analyses to reduce the ecological bias. The quality assessment was based on 15 items, with a maximum overall score of 20 points. The article quality was graded as low (≤5 points), medium (6–14 points) or high (≥15 points).
Data extraction and synthesis
NK and RH extracted the data independently using a standardised data extraction form and organised it in a standard Microsoft Excel 2019 spreadsheet. Any disagreement was resolved by discussion with NA to reach a consensus. The following data were collected: (1) authors, (2) publication year, (3) country, (4) time frame, (5) statistical analysis and climate or air quality modelling, (6) findings related to air quality modelling and scenario, climate projection and scenario, health impact projection and scenario and (7) adjustment for confounding and cross-validation. Figure 1 depicts the PRISMA flow diagram. All studies that were chosen followed an ecological design.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. ICD-10, International Classification of Diseases 10th Revision.
Results
Study characteristics
This systematic review involved 11 studies. Table 1 presents a descriptive summary of the included studies. The studies were conducted in the USA (27.3%, n=3), the UK (18.2%, n=2), China (36.4%, n=4), the Republic of Cyprus (9.1%, n=1) and one study was a global study (9.1%).
Descriptive summary of included studies (n=11)
Up to 36.4% (n=4) and 63.6% (n=7) of the studies were conducted in 2015–2017 and 2018–2022, respectively. Two studies (18.2%) focused on projections within a <10-year time frame.6 38 Three studies (27.3%) were within the 10–20 years time frame30 39 40 and six studies (54.5%) examined projections ≥21 years.15 27 28 41–43
Four studies assessed one disease (CVD, n=1)41 and respiratory disease (n=3) only6 15 27 while five studies (45%) focused on multiple diseases.28 30 38 39 All studies determined the BoD based on mortality. Only one study used the DALY42 and YLL27 in the BoD.
Baseline and projections of NCDs and air pollutants
Table 2 describes the selected studies based on the pollutants, air quality modelling, climate modelling, BoD, projection year and analysis. The air pollutants investigated included PM10, PM2.5, NO2, SO2, CO, elemental carbo and volatile organic compounds (VOC). Some studies specifically focused on individual air pollutants such as O3 28 and PM2.5.30 40 Three studies (27%) estimated the projection using air quality modelling and climate modelling.6 15 28 These studies include Coupled Model Intercomparison Phase 5, SSP and RCP. Another two studies used only air quality modelling, such as the Weather Research and Forecast Model-Community Multiscale Air Quality Model (WRF-CMAQ)41 and the Graz Lagrangian dispersion model.39 Two studies used a statistical prediction model.40 43 Only two studies used the Geophysical Fluid Dynamics Laboratory chemistry–global coupled climate change model (GRDL-global coupled CCM) as climate modelling with scenarios based on Special Report on Emissions Scenarios A1B (SRES A1B),27 RCPs and SSPs.28
Study characteristics
The BoD demonstrated that mortality was the main health outcome. The selected studies targeted projecting the BoD based on a single year (ie, 2015,40 2030,39 2035,41 20506 42 43) and multiple years.15 27 28 30 38 This review determined that six of the studies conducted health modelling using the concentration-response function (CRF, n=6).6 15 27 41 43 Two studies used the Global Exposure Mortality Model (GEMM),38 39 while other modelling approaches included the exposure-response function (ERF, n=2)30 40 and dose-response function (DRF, n=1).40
The confounding factors included in the selected studies (n=3) were temperature,6 socioeconomic deprivation (Jobseeker’s Allowance and median property price,6 age28 and air pollutants.43 These factors were adjusted and validated for the modelling. One study validated the result with actual air pollutant measurements (PM2.5, PM10, CO, nitrogen oxides (NOx)) from the national air quality station.39
Quality appraisal of studies
The included studies were evaluated using a modified version of a checklist designed to assess ecological research quality.35–37 The assessment scale determined that 64% of the study (n=7) were highly relevant and 36% (n=4) were moderately (medium) relevant (table 3).
Critical appraisal of selected studies; modified scales are from Baharom et al 35, Betran et al 36 and Dufault and Klar.37 Details information on the score is available in the online supplemental file 1
Regarding the data aggregation level, six studies (54.5%) were conducted at the national level,6 38 39 41–43 while five studies (45.5%) were conducted at the regional or state level15 27 28 30 40 (online supplemental file). Three studies (27.3%) used advanced statistical analysis techniques,28 38 43 including linear or Poisson regression and generalised linear model. Another six studies (55.5%) used basic Spearman’s rank and Pearson correlation as the analytical methods.6 15 39–42 Regarding reporting quality, all studies explicitly justified study design, and eight studies (73%) discussed the risk of ecological bias and its limitation (online supplemental file).6 15 27 30 39–42
Projection of NCDs attributable to air pollution
Increase in the BoD of NCDs attributable to air pollution
Five studies projected an increased BoD of NCDs attributable to air pollution (table 4). The mortality from NCDs caused by PM2.5 and O3 demonstrated a notable increase across various scenarios. At the global scale, Fang et al,27 projected increased all-cause mortality attributable to PM2.5 (+4.1%) and respiratory disease mortality attributable to O3 (+5%) in 2100. A Chinese study projected increased all non-accidental mortality (+10.7%), CVD mortality (+60.1%) and respiratory disease mortality (+6.9%) attributable to O3 will increase under the RCP8.5 scenario by 2053–2055.28 The same study projected that, under the RCP4.5–5 SSPs and RCP8.5–5 SSPs scenario, acute excess mortality attributable to O3 will increase by 111% and by 159%, respectively.
Summary of future projection of all studies (increase in burden of disease of non-communicable diseases attributable to air pollution)
Another Chinese study projected that IHD and stroke mortality attributable to PM2.5 will increase in 2020 (IHD: +2.77%, stroke:+2.35%) and 2030 (IHD: +6.16%, stroke: +5.89%).30 This increment is based on the scenario of 100% air improvement under SSP4. The same study also projected that premature mortality attributable to PM2.5 will increase in both years (2020: +8.8%, 2023: +25%). The Cyprus study projected that the CVD and respiratory disease mortality attributable to PM2.5 will increase up to 91%. The same study also projected that CVD mortality attributable to NOx will increase up to 70% during 2030.39 A US study projected that coronary heart disease (CHD) attributable to PM2.5 will increase to 29% in 2035.38 41
Decrease in BoD of NCDs attributable to air pollution
Nine studies reported a decreased BoD of NCDs attributable to air pollution like PM2.5 and NO2 (table 5). Two UK studies projected reduced all-cause mortality (nuclear power scenario: 508 700 YLL, low-greenhouse gas (GHG) scenario: 717 800 YLL)15 and respiratory disease hospital admission (RHA) attributable to NO2 (RCP2.6: 1.7%, RCP6.0: 1.4%, RCP8.5: 2.4%) under different scenarios.6
Summary of future projection of all studies (decrease in burden of disease of non-communicable diseases attributable to air pollution)
Wang et al 30 projected the BoD of NCDs under the scenarios of 100% air improvement with unchanged population and SSP4. The unchanged population scenario highlighted that during 2020 and 2030, premature death (2020: 3.5%, 2030: 22.5%), IHD mortality (2020: 8.5%, 2030: 8.8%), stroke mortality (2020: 27%%, 2030: 27.8%) and lower respiratory infections (LRI) mortality (2020: 0.26%, 2030: 0.42%) attributable to PM2.5 will be reduced. Under the SSP4 scenario,27 COPD (2020: 2.19%, 2030: 5.8%), lung cancer (2020: 1.58%, 2030: 3.59%) and LRI (2020: 1.34%, 2030: 2.83%) mortality attributable to PM2.5 are projected to reduce during the same period.
Cui et al 38 projected reduced CHD (−49.87%), stroke (−74.9%), lung cancer (−78.2%) and COPD (−72.8%) morbidity attributable to reduced PM2.5 under a scenario where PM2.5 is reduced to 15 µg/m3. Two US studies projected 36.8% reduced premature mortality attributable to PM2.5 in 205043 and 16 400 DALY (under the full electrification of light-duty passenger vehicles scenario (EV100%, detailed in online supplemental file)),39 respectively. Zhixiang et al 40 also projected reduced mortality attributable to PM2.5. Under RCP4.5, Chen et al 28 projected reduced non-accidental (−24%), CVD (−50.6%) and respiratory mortality (−10%) attributable to O3 in 2055. In Cyprus, reduced premature mortality attributable to PM2.5 (-30%) and NOx (−70%) was projected.39
Discussion
Basic characteristics
The 11 included studies from China, the USA, the UK, Cyprus and at a global scale explored the effect of air pollution on NCDs under different climate change scenarios. Most studies assessed air pollutants (PM2.5, O3, NO2) and a wide range of NCDs, including CVD (IHD, CHD, stroke) and respiratory diseases (COPD, LRI, lung cancer). The study projections mainly used premature mortality as the primary health outcome.
The studies included in this review were deemed highly relevant and were primarily conducted at the national and global levels. Nevertheless, there remains markedly sparse research on estimating NCD at smaller scales, such as the province or county levels. The lack of adequate, comprehensive and efficient air quality monitoring capabilities, which can be used to track and manage specific instances of air pollution, is the main cause of this lack of pertinent data.44
Urban economic activity increases air pollution.45 Consequently, a significant number of local monitoring stations have been installed in urban areas to regulate air quality levels and combat air pollution, and rural areas received less attention. Thus, the uneven, non-random distribution of monitoring stations between urban and rural regions will influence the spatial distribution of pollutants.46
Some studies used gridded data sets to enhance the spatial coverage of the monitoring network for more accurate and finer-scale air quality information, particularly in rural areas.47 Other than that, the daily mortality and morbidity rates at the city level must be available in open-access and accurate to increase the reliability in estimating the health effect.48
Modelling
Most of the included studies estimated the potential health effects of air pollution using CRF, ERF and DRF. These three models depict the relationships between the pollutant concentrations and health effects. DRF refers to the absorption and internal distribution of the pollutant in the body, while CRF and ERF refer to the exposure faced in a population study.49 A more recent study reported the BoD using Integrated Exposure Response (IER) for each cause of death.50–52 IER pools risk estimates from ambient and non-ambient PM2.5 assuming equal toxicity per unit dose.
Data alone do not determine the choice of modelling in CRF, and CRF might not necessarily represent a stable cause reliable for risk estimation. Rather, the CRF reflects modelling preferences. The existing CRF data are generally insufficient to accurately predict how future concentration changes will affect responses. Using model ensembles, causal graph modelling and time series methods can improve these challenges and predictions.53
The health effects of ambient air pollution on a global scale were estimated using the GEMM.8 This approach was parallel to a study in China, where the long-term mortality effects of outdoor PM2.5 and non-accidental mortality were examined using data from 41 cohorts of 16 countries.38 The GEMM-estimated number of deaths caused by PM2.5 exposure globally was approximately twice as high as that predicted by the IER model.52 Two factors contributed to this result. First, the GEMM accounted for all-natural cause mortality. In contrast, the IER model focused on specific health outcomes related to PM2.5 exposure, estimating a lower number of deaths. Second, the IER model considered additional types of exposure, such as active smoking, which has lower relative risks per unit of PM2.5 compared with ambient air pollution.52
Climate research has progressed to coupled modelling for enhanced understanding of past climate change and projecting future climate scenarios by integrating air quality modelling. The Geophysical Fluid Dynamics Laboratory (GFDL)-global coupled CCM integrates various climate system components, such as the atmosphere, oceans, land surface and sea ice, into a unified framework.27 This model is designed to simulate and investigate the interactions and feedback between its components on a global scale.
In the UK, the North American/Globally Coupled Model-Air Quality Unified Model (NA/GCM-AQUM) represents a coupled modelling system that combines a regional climate model with an air quality modelling system.6 The regional climate model simulates atmospheric conditions over the NA region, capturing regional-scale climate processes and interactions. The AQUM component focuses on simulating air quality processes, including emissions, atmospheric chemistry and pollutant transport.6
Another study implemented Global Climate Model-WRF (GCM-WRF).15 The GCM simulates the global climate system, including large-scale atmospheric circulation, oceanic processes and interactions between different climate components. The WRF model is a region-scale weather model that simulates atmospheric conditions at high spatial resolutions. The GCM-WRF coupling enables the downscaling of climate information from the global scale to the regional scale.
Worldwide research showcased advances in air quality projection. In China, GFDL–climate model 3 outputs were used as input data for air quality models or to assess the effects of climate change on air pollution.28 The state-of-the-art WRF-CMAQ air quality model has gained interest in the USA as it is a coupled modelling system for regional scale that simulates meteorological conditions and air quality.41 The CMAQ-urban is another type of CMAQ designed to simulate air quality in high-resolution urban environments considering the unique characteristics of urban emissions, land use patterns and atmospheric processes within cities. The CMAQ-urban uses fine-scale emission inventories and urban-specific parameterisations to capture urban air pollution dynamics in the UK.15
Climate change, integrated and mitigation scenarios
We categorised the scenarios in the studies into (1) climate change scenarios, (2) integrated scenarios and (3) mitigation scenarios. The climate change scenarios were inclusive of RCPs, SSPs and SRES. The integrated scenarios included combination scenarios, such as SSPs with different percentages of air quality improvement30 and RCPs and SSP integration.28 The mitigation scenarios involved reduced carbon emissions, air pollution and land use.
Of the 11 studies, 2 studies used climate change scenarios6 27 2 studies used integrated scenarios28 30 and 7 studies used mitigation scenarios15 41–43 (online supplemental file).
Climate change scenarios
A UK study6 projected the NCDs based on RCP scenarios53 where RCP2.6 represents the lowest radiative forcing where GHG emission is strongly reduced. RCP4.5 and RCP6.0 represent a moderate emission scenario also known as the best-case scenario. RCP8.5 represents an unmitigated high GHG emission scenario.54 The global warming and emission scenario is usually evaluated based on RCPs that include well-mixed GHG and emissions of air pollutants and their precursors.54
In a global study, Fang et al 27 estimated the projection based on the SRES A1B scenario. The A1 scenario envisions a world of rapid economic growth, peak global population around the mid-century followed by a decline and the swift adoption of efficient technologies.55
Integrated scenarios
Wang et al reported an integrated scenario involving air quality improvement under different population changes depicted by the SSP.30 Chen et al estimated the BoD based on integrating the RCP with five SSPs (SSP1–SSP5).28 SSP1 assumes a future progressing toward a more sustainable path. SSP2 is a ‘middle of the road’ scenario that corresponds to the medium variant of the projections. SSP3 represents a fragmented world emphasising security at the expense of international development. SSP4 represents a world involving high inequality between and within countries. SSP5 represents a world that prioritises technological progress and where rapid human capital development fosters economic growth, reflected in high educational attainment assumptions and low mortality across all countries.30 56–58
Mitigation scenarios
A UK study projected a health outcome that aligns with the UK Climate Change Act of 2008, targeting a reduction of at least 80% in GHG emissions by 2050.15 The authors used four scenarios: a baseline scenario, a nuclear power scenario, a low-GHG scenario and a constant scenario. The baseline is a scenario that does not meet the UK Climate Change Act target. The nuclear power scenario involves 80% GHG emission with a limited increase in nuclear power. The low-GHG scenario involves 80% GHG emission without policy constraint on nuclear building. The constant scenario envisions keeping 2011 air pollutant concentrations constant until 2050.
Likewise, two Chinese studies projected future scenarios where the PM2.5 concentration dropped to 15 µg/m3,38 43 corresponding to the China Air Pollution Prevention and Control Action Plan. The study in Jinan city assessed local changes regarding the exhaust emissions, air quality, mortality and morbidity of associated specific diseases and related economic benefits from 2013 through to 201738 (online supplemental file).
The USA aims to achieve carbon neutrality by 2050 using transportation scenarios.42 This scenario contrasts optimum levels of physical activity from active travel and minimal air pollution from electric cars depicted by active travel 100% (AT100%) and electric vehicle (EV) scenarios (EV100%).
Similarly, several scenarios and cases were formulated in Cyprus to address emissions and minimise human mortality risks in the city of Nicosia.39 Nine different traffic scenarios and four distinct cases were analysed for 2030. The scenarios and cases included banning diesel passenger vehicles (PV), light-duty vehicles (LDV), non-Euro six standards vehicles, speed limits, ±20% traffic fluctuation, prohibition of diesel PV and LDV, 80% electric PV and banning fireplaces. The Southern California study focused on regional transportation plan control that involves traffic density and proximity to a major road.41
Projection of NCDs attributable to air pollution
Increase in the BoD of NCDs attributable to air pollution
In the late 21st century, PM2.5 is projected to increase in all regions, with a global mean of 0.28 mg/m3. Populated areas such as East Asia, Eastern USA, Northern India and Africa are some areas projected to face increased PM2.5.27 The source apportionment of PM2.5 pollution is mainly from residential areas (India),59 industrial sectors (India and China),60 buildings (Africa),61 transportation (Africa and the USA).62 The sources, chemical composition, formation, transformation and fate of PM2.5 are markedly different across regions due to the variations in emissions and meteorological conditions.59
In the included studies, PM2.5 was associated with increased CVDs, such as CHD, IHD and stroke. Previous studies showcased the long-term effect of PM2.5 for 12–14 hours, was strongly associated with atherosclerosis, the key to a central mechanism for IHD and stroke.63–65 Cardiovascular parameters such as heart rate variability are strongly associated with PM2.5 exposure.64 66 A population study demonstrated an association between air pollution and respiratory infection severity, causing irritation, cough, phlegm and bronchial hyper-responsiveness.67 PM2.5 is highly associated with increased emergency department (ED) visits for asthma64 68 69 and COPD68 and respiratory infection.70
Apart from PM2.5, O3 also is projected to increase in south China, North India, Northeast USA and Central Africa and decrease in other remote Oceanic regions. Various studies reported that climate change poses a significant O3-related health burden globally and regionally.27 31 71–73 Global and regional health impact assessments, including China, have reported adverse effects of climate change on BoD attributable to O3.31 72 74 75
Future O3-related health burdens in China indicate increased acute excess mortality across the five SSPs under RCP4.5 and RCP8.5, respectively.28 However, the increased premature mortality attributable to O3 exposure is smaller than the premature mortality attributable to PM2.5 exposure as O3 is not as harmful as PM2.5.27
NCDs are projected to increase with the increase in air pollutant concentration. Nevertheless, population size and population ageing increase the projected NCD mortality and morbidity. The population size is projected to increase by up to 3 million people by 2035, leading to population ageing and an increased risk of CHD.41 76 Emerging evidence suggests that near-roadway air pollution is linked to CHD mortality and morbidity.77–81 The proportion of the population living near major roads is also projected to increase, which could increase exposure to traffic-related air pollution and CHD risk.
The increased O3 concentrations under RCP8.5 accompanied by the ageing population will lead to increased acute excess mortality attributable to O3. It was projected that the population size would reduce in 2050, but the steep increase in the elderly population would increase O3-related acute excess mortality.28 Furthermore, RCP8.5 suggests more O3-related deaths in colder months (November to April) due to increased O3 concentrations.28 The population ageing under SSP1 and SSP5 will offset the reduced deaths due to the decreases in age group-specific mortality rates. This will subsequently decrease the acute excess mortality attributable to O3.28
Wang et al were the first in China to estimate the future BoD at the county level.30 Under the scenario of 100% air quality improvement, SSP4 and population growth, premature mortality attributable to PM2.5 was projected to increase in 2020 and 2030.30 Chinese environmental policies can achieve health benefits with an annual PM2.5 concentration of 35 µg/m3. As the population ages and grows, the overall susceptibility to PM2.5 will increase. Thus, IHD and stroke are expected to increase in 2020 and 2030. Apart from that, highly developed regions faced a higher PM2·5-associated BoD compared with less developed regions because of the higher pollution levels and population densities. Mass migration from rural areas to more developed urban areas on the eastern coast82 and the Chinese two-child policy might differentially increase population density in different regions.30
Climate change also affects air pollutant surface concentrations, such as PM2.5 and O3.83 The global mean PM2.5 will increase due to hydrogen peroxide (H2O2). Increased H2O2 is associated with the high moisture and OH concentration in the atmosphere, which are both climate change byproducts.27 H2O2 reacts with SO2 to form sulphate (SO2 + H2O2 −> H2SO4)82 83 the main component of PM2.5.84 Among all PM2.5 components, the largest increases are in sulfate, smaller dust particles and organic matter. The global mean PM2.5 is largely driven by increased water vapour, a reactant in the chemical reactions that form O3, increasing surface O3 concentrations.27 85
In Cyprus, eight different cases were simulated to determine how traffic changes by 2030, where all vehicles will adhere to Euro six standards, will affect NOx and PM2.5 pollution. Surprisingly, the premature total mortality and cardiovascular and respiratory risks would still increase across all scenarios despite these changes. The study also investigated different pollution reduction strategies, such as setting a 30 km/hours speed limit, but this worsened matters, releasing 9% more NOx and 3% more PM2.5.39
Decrease in BoD of NCDs attributable to air pollution
The US studies projected a linear reduction in premature mortality attributable to PM2.5 concentrations under the AT100% and EV100% scenarios.42 43 Both scenarios present significant implementation and policy challenges.82–86 The AT100% scenario potentially improves health greater than electrifying light-duty PV due to increased physical activity, annual monetised net benefit and modest benefit of PM2.5 reduction. However, the risk of road traffic injuries impedes this benefit.87
Active travel requires a significant financial investment to provide proper facilities for pedestrian, bicycle and transit infrastructure and changes in land use that equilibrate future demand for housing and job growth. Contrarily, EVs require technology and comprehensive policy implementation. Furthermore, EVs require proper development, deployment and financing that addresses battery charging, vehicle range and cost. A few factors stimulate EV implementation; for example, voluntary pledges by vehicle manufacturers to phase out sales of gasoline-fueled cars by 2050, rebates and tax incentives for EV purchases and the banning of selling gasoline-fueled cars by 2035.42 The transition to 100% demands good governmental regulation, commitment and profit to the industry and consumer interest.
Fann et al determined that people gained more years of life as PM2.5 levels decreased over 30 years.43 Lower PM2.5 exposure led to fewer premature deaths and an increased average lifespan. The mean population-weighted annual mean PM2.5 levels decreased from 15.4 mg/m3 in 1980 to 8.8 mg/m3 in 2010 across all counties. Thus, if PM2.5 levels remained at 1980 levels, people born in 2050 could live 1-year longer.43 However, local, state and federal air quality policies are expected to decrease PM2.5 concentrations further.88–94 The estimated PM2.5-related premature mortality in 2000 and 2010 were 140 000 and 120 000, respectively, which was consistent with previous studies.95–98
In a UK study, Williams et al 15 projected decreased all-cause mortality in 2050 due to reduced NO2 and PM2.5. The low-GHG scenario highlighted a greater reduction of mortality attributable to NO2 and PM2.5. Similarly, Panullo et al 6 reported reduced NO2 concentrations under all three RCPs (RCP2.6, RCP6.0, RCP8.0). NO2 concentrations are projected to decrease in 2050, decreasing RHA. A UK study also reported stronger effects on RHA for NO2 compared with PM2.5 and PM10.99
China is the largest GHG emitter globally and the most populated country with a higher risk of climate change impact.28 The PM-related health burden has been the major focus for climate projection research in China, which mainly highlighted reduced premature mortality30 40 and other NCDs such as IHD, stroke,38 40 lung cancer, COPD38 and LRI.30 Improved air quality under the unchanged population scenario projected a decrease in premature deaths attributable to PM2.5 by 2020 and 2030.30 This result was consistent with that of previous studies80 100–102 which estimated 23% reduced mortality under the scenario of 35 µg/m3 PM2.5 (WHO Interim Target-1).
The decrease in age group-specific mortality rates will subsequently decrease the acute excess mortality attributable to O3.28 O3-related mortality was estimated to decrease under RCP4.5. This decline is due to assumed large reductions in emissions of O3 precursors, such as NOx and non-methane VOC. This result was consistent with those for East Asia in previous global studies.74 103
In Nicosia, Republic of Cyprus, the mitigation scenario that involved banning diesel PV and LDV and allowing only vehicles that abide by Euro six standards reduced up to 70% of NOx compared with 2017. Furthermore, premature mortality attributable to PM2.5 was projected to be reduced by up to 30% in 2030 under Case 4 (banning of diesel PVs and LDVs, banning fireplaces, reducing traffic flow).39
Strengths and limitations
Globally, 74% of deaths are associated with NCDs, making this review pertinent given the high mortality risk of NCDs attributable to air pollution. This review clarifies the prevailing public health concerns regarding NCDs attributable to air pollution under various climate change scenarios enhanced with air quality and climate modelling.
All studies included in this review underwent a systematic critical appraisal using a modified evaluation tool suitable for ecological study designs.35–37 The quality appraisal score determined that no articles included in this review were of low relevance.
Nevertheless, caution is advised when estimating the impact of air pollution on NCDs in the following aspects: the correlation of air quality databases between satellite-based or ground monitoring stations, ground monitoring station distribution, NCD database accessibility and accuracy and the inclusion of meteorological and weather data in the analysis.
One limitation of this review is the scarcity of studies using the IPCC standardised climate change scenarios to project future NCDs attributable to air pollution. Another limitation is that most of the studies primarily focused on the mortality aspect of NCDs. Only one study reported on the morbidity of NCDs. Studies that measured the health outcome using YLL, YLD or DALY were lacking.
The exclusion of non-English language articles could be another limitation of this review. Literature from non-English speaking countries might have enriched the outcome of this review, specifically those published in China. Due to the English-language bias, this review could have biased estimates of effect, reducing its generalisability. However, including non-English language studies might have required additional resources in terms of cost, time and non-English language proficiency.
A strength of using the IPCC standardised scenarios (SRES, RCPs, SSPs) is that they facilitate the comparison of BoD across studies.104 Furthermore, these updated scenarios encompass human factors and sustainability, extending beyond the focus solely on GHG emissions and air pollutants.
Recommendations
Research should focus on mitigating air pollution, especially PM2.5 as this review demonstrated that many studies projected increased NCD mortality attributable to PM2.5. Forest fires, agricultural waste burning and inefficient fuel combustion105 should be considered in the mitigation plan to ensure a sustainable future for humans.
It is advisable for future research on the projection of NCDs attributable to air pollution to incorporate advanced coupled modelling of air quality and climate. Additionally, it is crucial to consider population changes and social effects when estimating the projection of NCDs attributable to air pollution instead of solely focusing on carbon emission-related scenarios. These advancements will significantly contribute to the development of future projections for NCDs attributable to air pollution, assisting stakeholders in planning and implementing effective mitigation strategies for a sustainable world.
Conclusion
This review highlights the importance of the BoD of NCDs and their projections. The UK and China have progressed well in estimating the BoD projection for NCDs attributable to air pollution by using coupled-air quality and climate modelling approaches, and mainly projected health effects using the concentration–response function model. In this review, NCD projection was simulated using climate change, integrated and mitigation scenarios. The BoD of NCDs attributable to air pollution is projected to reduce in a scenario that promotes reduced air pollution, carbon emissions, land use and sustainable socioeconomics. Contrastingly, a scenario with growing population numbers, social deprivation and an ageing population is projected to increase the BoD of such NCDs. Overall, the included studies widely reported increased premature mortality, CVD and respiratory disease attributable to PM2.5.
This review provides a comprehensive evaluation of current research on NCD projection attributable to air pollution and emphasises the BoD of NCD under climate change scenarios.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Acknowledgments
This systematic review is a part of research by Ministry of Higher Education Malaysia under Long-term Research Grant Scheme project 3, grant number LRGS/1/2020/UKM–UKM/01/6/3, which is under the program of LRGS/1/2020/UKM/01/6.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @PijaMoris
Contributors Conceptualisation: NK, NA and RH. Methodology: NK and NA. Data curation: NK and RH. Writing—original draft preparation: NK. Writing—review and editing: NA, RH and NK. Supervision: NA, RH and MIAW. All authors have read and agreed to the published version of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.