https://orcid.org/0000-0002-2894-8018
Figures
Abstract
Background
Air pollution from indoor and outdoor sources constitutes a substantial health risk to young children in sub-Saharan Africa (SSA). Although some systematic reviews have assessed air pollution and children’s respiratory health in SSA, none have considered both ambient and indoor exposures.
Methods
This systematic review and meta-analysis assessed the effect of air pollution (ambient and indoor) on respiratory hospitalization and mortality among children under five years in SSA. We retrieved relevant articles from PubMed, Embase, Scopus, African Journals Online (AJOL), Web of Science, and medRxiv. The protocol was registered with Prospero (CRD42023470010). We used guidelines from the preferred reporting items for systematic review and meta-analysis (PRISMA-2020) to guide the systematic review process. Risk of bias was assessed using the Office of Health Assessment and Translation (OHAT) quality appraisal tool. For exposures where there were sufficient studies/data we conducted meta-analyses using random effects models and used Stata version 17 software for analysis.
Results
For the systematic review we screened 5619 titles and abstracts, reviewed 315 full texts, and included 31 articles involving 2,178,487 participants. Eleven studies examined exposure to solid fuel use in households and its association with all-cause mortality, while four studies explored the impact of passive smoking on mortality among children under five. Only two studies assessed ambient air pollution's effects on all-cause and respiratory-related mortality. Additionally, 13 studies reported varying associations between respiratory hospitalization and household tobacco smoke exposure. Meta-analyses on studies of solid fuel use and mortality and passive smoking and hospitalizations showed that children exposed to indoor solid fuels combustion had higher odds of mortality compared to non-exposed children (OR = 1.31; 95% CI: 1.16–1.47). The meta-analysis of exposure to second-hand smoke found an increased risk of respiratory hospitalization due to pneumonia, although the results were not significant (OR = 1.29; 95% CI: 0.45–3.68), and our certainty of evidence assessment indicated insufficient support to conclusively establish this association.
Conclusion and Recommendation
Our review reveals that solid fuel use and ambient PM2.5 exposure were associated with increased mortality risk in children under five years in SSA. The meta-analysis showed evidence of an increased risk of under-five years mortality associated with solid fuel use in households. Associations between secondhand smoke and pneumonia hospitalization were less clear. We conclude that significant research gaps remain in understanding the impact of discrete sources of air pollution on the causation of respiratory illness in young children living in SSA. Prioritizing interventions targeting indoor sources is essential, along with further studies which use standardized and objective exposure and outcome measures to study these associations.
Citation: Lake EA, Karras J, Marks GB, Cowie CT (2025) The effect of air pollution on morbidity and mortality among children aged under five in sub-Saharan Africa: Systematic review and meta-analysis. PLoS ONE 20(4): e0320048. https://doi.org/10.1371/journal.pone.0320048
Editor: Oyelola A. Adegboye,, Menzies School of Health Research: Charles Darwin University, AUSTRALIA
Received: November 1, 2024; Accepted: February 12, 2025; Published: April 10, 2025
Copyright: © 2025 Lake et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its cited references.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: DHS, Demographic and Health Survey; OHAT, Office of Health Assessment and Translation; NIEHS-NTP, National Institute of Environmental Health Science-National Toxicology Program; WHO, World Health Organization.
Introduction
Air pollution poses a multifaced environmental challenge that significantly impacts human health worldwide. It is defined by the World Health Organization (WHO) as the alteration of indoor or outdoor environments by any physical, chemical, or biological agent that disrupts the natural composition of the atmosphere [1]. Air pollution is detrimental to vulnerable populations, with children bearing a disproportionate burden. Their developing respiratory and immune systems, and proximity to ground-level pollutants render them particularly susceptible [2, 3]. Alarmingly, a recent WHO report showed that over 90% of children globally breathe polluted air that exceeds the WHO air quality guidelines [4].
Air pollution exposure is ranked as the fourth deadliest global risk factor for disease burden, surpassed only by dietary risks, tobacco, and high blood pressure. In 2019 alone, it was responsible for 6.67 million deaths and over 213 million disability-adjusted life years (DALYs) [5, 6]. Children’s exposure to air pollution reduces their life expectancy by an average of 20 months [6].
Hazardous exposure to air pollution disproportionally affects people residing in low and middle-income countries (LMIC), with over 95% of air pollution-related deaths occurring in LMICs in Africa and South Asia [7]. Africa has the highest excess mortality from air pollution from both indoor and outdoor origins, and respiratory infection was suggested to be a potential contributor [8–10]. Although household air pollution (HAP) has improved globally, it remains problematic in Africa, accounting for 697,000 air pollution-related deaths in 2019 [8]. The challenge of ambient air pollution, aggravated by rapid urbanization and industrialization, is remarkably pronounced in SSA, where infants and young children face adverse health outcomes due to their heightened vulnerability to environmental pollutants [8,11–13].
Epidemiological studies show substantial evidence of the relationship between air pollution and various adverse health consequences, including mortality, cancer, cardiovascular disease, and respiratory infections [14–18]. However, some studies in SSA also reported no or inconclusive associations [19, 20]. Existing systematic reviews on air pollution and adverse child health outcomes in SSA have mainly focussed on either a single country [21, 22] or specific sources of air pollution, such as HAP or its proxy [23], to the exclusion of ambient air pollution, which is also often an important source of hazardous exposure [24]. Others synthesized data either outside of our scope of respiratory hospitalization in under-five children [25] or in wider settings where data related to LMICs more broadly as part of global studies, both of which may not be fully relevant to the SSA region [14,26].
Household air pollution, primarily caused by the burning of solid fuels for cooking, lighting, and heating within homes, has far-reaching effects that extend beyond the immediate indoor environment. Pollutants generated from these activities contribute not only to household air pollution but also to ambient air pollution, which can infiltrate neighboring homes and communal spaces [27]. Hence, even dwellings that do not use biomass fuel for cooking and heating may be adversely affected by this practice within their communities. As a result, none of the existing systematic reviews provides a comprehensive understanding of the effect of indoor and ambient air pollution on respiratory morbidity and mortality among children under five years in SSA.
Hence, we aimed to (i) identify and collate relevant peer-reviewed epidemiological studies on air pollution and its effect on mortality among children aged under five years living in SSA; (ii) identify and collate relevant peer-reviewed epidemiological studies of air pollution and its effect on respiratory hospitalizations, such as acute respiratory infections (ARI), pneumonia, and asthma; (iii) quantify the association between air pollution and under-five mortality and respiratory hospitalization; and (iv) identify the existing research gaps in the evidence.
Methods
In conducting this systematic review, we followed the preferred reporting items for systematic review and meta-analysis (PRISMA-2020) guideline [28]. Firstly, we prepared a protocol based on the PRISMA-Protocol statement [29, 30]. We registered the protocol in the Prospero (Prospective Register Of Systematic Review) database (ID: CRD42023470010) [31], an international database of systematic reviews conducted since 2011, to prevent redundancy and reduce the potential for reporting bias [32]. Since this review solely relied on aggregating and analyzing previously published studies and the data was de-identified, it did not require ethical approval.
Source of information
We interrogated five databases and one internet site: Embase, PubMed, Scopus, Web of Science, African journals online(AJOL), and medRxiv [33]. We screened for all available literature held in each database until July 2024.
Search strategy
We employed a three-step search to find preprints and published studies. First, we did a preliminary search of PubMed and Embase to identify relevant articles on the topic and help inform the search strategy. We used the text words in the title and abstract of relevant articles and the index terms used to describe the reports to develop a complete search strategy in each database and adapted the search accordingly. We also checked the reference list of included articles to find additional relevant articles for potential inclusion. Preprints were searched primarily on medRxiv, with PubMed and Embase also included due to their inclusion of preprints.
We used several keywords for the search strategy, including “air pollution,” “under-five,” “mortality,” and “sub-Saharan Africa.” Furthermore, several synonyms for each keyword were used for the search process, as shown below. The final search strategy, including the results, is available in a supplementary file (S2).
Main keywords and alternative terms for the search strategy
- Air pollution (air quality, indoor air pollution, indoor air quality, outdoor air pollution, outdoor air quality, ambient air pollution, ambient air quality, household air pollution, household air quality, traffic-related air pollution * , particle * , particulate, PM10, PM2.5, carbon monoxide, “CO,” Sulfur dioxide, “SO2,” nitrogen dioxide, nitr* oxide*, “NO2,” “ozone,” “O3”, biomass, biofuel * , emission * , combustion, smoke)
- Under-five (baby, infant, child * , neonat * , fetal, foetal)
- Morbidity, Mortality (death, Respir * , pneumonia * , asthma, health, hospital * , “URTI”, “LRTI”)
- Sub-Saharan Africa (“SSA,” Angola, Benin, Botswana, (Burkina Faso), Burundi, (Cabo Verde) OR Cameroon, “Central African Republic,” Chad, Comoros, Congo, “Cote d'Ivoire,” “Equatorial Guinea,” Eritrea, Ethiopia, Gabon, Gambia, Ghana, Guinea, “Guinea- Bissau,” Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Somalia, “South Africa,” “South Sudan,” Sudan, Tanzania, Togo, Uganda, Zambia, Zimbabwe)
- *Acting as a wildcard character representing any sequenced search terms that could be available in the search engines
Eligibility criteria
The eligibility criteria for studies were based on participants, exposures, comparators, outcomes, and study designs to identify studies assessing the effect of air pollution on mortality and respiratory hospitalization in children under five years within SSA. Only studies published in English were included in this systematic review and meta-analysis.
Inclusion criteria
Participants.
All children under five years of age living in SSA. Studies were included only if the data for this age group were analyzed and reported separately.
Exposures.
Exposure was defined as ambient and indoor air pollution, including sources such as biomass combustion for cooking and heating, other household pollution sources, and traffic-related pollutants (e.g., proximity to major roads or traffic density). We included studies that quantified exposure through objective measurements of pollutants (e.g., particulate matter (PM2.5), nitrogen dioxide (NO2), and carbon monoxide (CO)), spatial proximity to pollution sources, and proxies such as household fuel type, stove usage, and tobacco smoke in the household.
Comparators
The comparator group consisted of children under five years of age with lower or no exposure to ambient and indoor air pollution. This group included children residing in environments with lower levels of pollution, as indicated by reduced proximity to pollution sources (e.g., major roads, industrial areas), the use of cleaner cooking technologies (e.g., electric or liquefied petroleum gas (LPG) stoves instead of biomass stoves), and no exposure to household tobacco smoke or lower levels of specific pollutants, serving as the reference group.
Outcomes.
The primary outcome of interest was all-cause mortality. However, we included studies that defined outcome as cause-specific mortality and hospitalization for respiratory diseases including pneumonia, asthma, and other respiratory infections.
Study Designs.
Observational studies (cohort, cross-sectional, case-control, and ecological) and interventional studies (experimental and quasi-experimental) that quantitatively assessed the relationship between air pollution exposure and outcome of interest were considered.
Exclusion Criteria.
Studies were excluded if they did not focus on SSA, lacked specific data on children under five years of age, or did not provide objective data on air pollution exposure or outcomes related to mortality or hospitalization due to respiratory diseases, including pneumonia, asthma, and other respiratory infections. Non-original research articles, such as reviews, commentaries, conference abstracts, and studies lacking full-text availability, were also excluded.
Study selection criteria.
Following the search, we imported all identified records into Covidence (Veritas Health Innovation, Melbourne, Australia), an automation tool for managing and streamlining systematic reviews [34]. We removed duplications, and two reviewers ((EAL and JK) piloted the screening of articles on the first 50 articles within Covidence. This included title and abstract screening, risk of bias assessment, and data extraction. Where there were discrepancies in opinion, both reviewers discussed this, and an independent decision was sought from the two senior authors (GBM, CTC), where required. Two of the initial 50 articles met the final inclusion criteria.
Following the piloting, the same two reviewers (EAL and JK) screened the titles and abstracts for all articles according to the inclusion criteria. Any conflicts that arose during the title and abstract screening were resolved by discussion and with the involvement of a third reviewer (CTC and/or GBM). We retrieved potentially relevant studies in full and read them independently. Studies meeting the final selection criteria were retained for data extraction. For those excluded, reasons for exclusion were reported.
The appraisal process consisted of three steps: assessing the Risk of Bias (RoB) in each study, assessing the quality of evidence across studies, and evaluating the strength of evidence across studies. We adapted a quality appraisal tool from the OHAT using methodology developed by the National Institute of Environmental Health Science-National Toxicology Program (NIEHS-NTP) and the navigation guide from the University of California [35, 36]. The tool is recommended for evaluating the internal validity of environmental epidemiological studies. It comprises eight domains: recruitment strategy, exposure assessment, confounding control, outcome assessment, incomplete outcome, selective reporting, conflict of interest, and other biases. With each prescribed criterion, we rated each bias domain as “low,” “probably low,” “probably high,” “high,” and “non-applicable.” The same two reviewers performed the assessment independently. In the case of disagreement, the resolution began by seeking the opinion of a third reviewer (CTC and/or GBM). If we did not reach a consensus, further discussion ensued, and the final decision rested on the judgment of the most conservative voter. For example, if one reviewer judged “low” and the other judged “probably low,” the “probably low” rating was selected.
The critical appraisal process of each study concluded with assigning tiers to articles based on key appraisal domains (exposure assessment, confounding bias, and outcome assessment), selected based on OHAT recommendations, along with other domains such as recruitment strategy, incomplete outcome data, and selective reporting ( S5 File). A study judged either ‘low’ or ‘probably low’ for all key domains and most of the other domains ranked as tier one (good quality). A study judged either ‘high’ or ‘probably high’ for all key domains, and most of the other domains were classified as tier three (poor quality), and a study that was neither tier one nor tier three was ranked as tier two (medium quality). We conducted our meta-analysis with tier one and two studies and excluded tier three studies. EAL, with CTC and/or GBM, assessed evidence certainty using the OHAT approach, which tailors GRADE methodology for environmental health [37] (detail supplementary file, S7).
Data extraction.
We developed and piloted a data extraction template to ensure consistency in the information extracted (Table 2 and S4 File). The template included the following: (1) author, (2) study and publication year, (3) study area and country, (4) study design, (5) sample size, (6) response rate, (7) participant characteristics, (8) age of study participants, (9) pollutants or proxies of air pollution, (10) exposure assessment method, (11) outcome, (12) outcome assessment method, (13) potential confounders, (14) effect estimate, and (15) main findings.
Measurements of exposure and outcomes.
In this review, air pollution exposure was defined as exposure to either ambient or indoor air pollution, or both, assessed through direct measurement of pollutants or inferred using proxy indicators. Proxies for indoor air pollution included specific markers such as the presence of biomass stoves, tobacco smoking in the household, and other household emission sources, whereas proxies for ambient air pollution were characterized using metrics such as traffic counts and proximity to major roadways.
The primary outcome of this review was all-cause mortality, while the secondary outcomes were cause-specific mortality and respiratory hospitalizations, including pneumonia, asthma, and other respiratory illnesses, all in children under five years of age. This systematic review included the following outcomes: neonatal mortality (death within the first 28 days of life), post-neonatal mortality (death between 28 days and one year), infant mortality (death before one year, encompassing both neonatal and post-neonatal mortality), early childhood mortality (death of a child aged 12–59 months), and hospitalization (defined as the admission of children under-five years to a hospital or a health facility for medical care due to respiratory infections including ARI, pneumonia, bronchitis and asthma) [38].
Data analysis and synthesis.
We employed a narrative synthesis to summarise texts, tables, and figures as necessary and used descriptive statistics to summarise and describe the key characteristics and findings of the eligible studies.
We conducted a meta-analysis of available quantitative data relevant to the objectives. The extracted data was exported to Stata version 17 (StataCorp LP, 4905, Lakeway Drive, College Station TX77845, USA) for analysis. We extracted the effect estimates reported in each study (that is, hazard ratio (HR), relative risk (RR), or odds ratio (OR)) and their 95% confidence intervals (95% CI). To estimate the pooled effect size, we applied the following formula to convert different effect measures to a common measure, the OR, which was the most commonly reported in the studies [39].
RR was a relative risk, r was the event rate for the reference or control group, HR was the hazard ratio, OR was the odds ratio, and P0 was the event rate in the unexposed population.
We assessed the variation across the studies using I2 statistics [40, 41]. Since we found considerable heterogeneity (I2 = 78.6%), we used the random effects model to estimate the weighted inverse-variance pooled effect estimate [40]. Odds ratios (ORs) were log-transformed to stabilize variances and create symmetrical confidence intervals. The log-transformed ORs were combined, and the final combined estimate was exponentiated for interpretation [42]. To check publications bias, we produced funnel plots and contour-enhanced funnel plots for graphical diagnosis of small study effects and, more objectively using the Egger regression test with a P-value < 0.05 as the test statistic to indicate the presence of publication bias [43, 44]. We also conducted a trim-and-fill analysis to further detect and adjust for publication bias by imputing missing studies [45]. We conducted subgroup analysis by study setting and year and also performed sensitivity analysis based on study quality to determine the source of heterogeneity. We also ran a leave-one-out sensitivity analysis to see the effect of a single study on the overall meta-analysis estimates [46].
Results
Systematic review
We identified a total of 8307 records through the electronic search of online databases: Embase (n = 3185); PubMed (n = 1166); Scopus (n = 1745); Web of Science (n = 1717); AJOL (n = 10); medRxiv (n = 481), and citation search (n = 3). We removed duplicates (n = 2688) and screened 5619 articles by title, abstract, and keywords against the inclusion criteria. Of these, 315 articles were retained for full-text screening. During full-text screening, we excluded 284 articles for the following reasons: 1) the outcome was unrelated to our inclusion criteria (n = 206); 2) the study setting was beyond our geographic scope (n = 14); 3) both young and older children were included in the study population, and data was not provided separately for children under-five years (n = 34); 4) the study did not report air pollution data or household fuel use or any measure of air pollution exposure (n = 12); 5) the study was qualitative or a review (n = 7); 6) the study was a duplicate (n = 10); and 7) the full text was unavailable (n = 1). Finally, we included the 31 articles in this systematic review (Fig 1, supplementary file S3 and S4).
Risk of bias assessment
We assessed the risk of bias for each of the 31 studies using eight predefined criteria (Table 1 and S5 File). The majority of study components (43.55%, 108 of 248) were rated as having ‘probably low’ risk of bias (ROB), followed by ‘low’ (30.65%, 76 of 248), and ‘probably high’ (24.19%, 60 of 248) ROB. Four (1.6%) study components (all related to confounding) scored a ‘high’ ROB judgment. However, the overall quality for those four studies was defined as tier 2 (moderate quality). Therefore, we did not exclude any studies from our meta-analysis based on the ROB assessment. All 31 studies except one were rated as Tier Two (moderate quality) in the overall ROB assessment.
General characteristics of eligible studies
Studies which met the inclusion criteria were from Nigeria (n = 8; 25.81%)[47–54], South Africa (n = 7; 22.58%)[55–61], SSA as a whole (n = 4; 12.91%) [18,19,62,63], Ethiopia (n = 3; 9.68%) [20,64,65], Kenya (n = 2; 6.45%) [66, 67], Tanzania (n = 2; 6.45%) [68, 69], Uganda (n = 2, 6.45%) [70, 71], Cameroon (n = 1; 3.23%)[72], Gambia (n = 1; 3.23%) [73], and Niger (n = 1; 3.23%) [74]. Most studies (n = 19) were cross-sectional in design, followed by case-control (n = 7), cohort (n = 4), and semi-ecological (n = 1) study designs. The studies were conducted between 1985 and 2020, but only two were published before 2010. The sample size ranged from 114 to 404,254 people. The majority of the studies analyzed population-based secondary data, with authors specifically selecting complete information, resulting in a 100% response rate for most studies (n = 28). However, it should be noted that minor instances of non-response may still exist within the original datasets.
Exposure assignment in studies of household air pollution
Most (n = 27) of the studies related to household air pollution exposure. Eleven of the studies used the Demographic and Health Survey (DHS) data to assign proxies of air pollution indirectly [18,19,47,48,53–55,62,65,75,76]. The DHS is a multi-country representative survey that has provided essential data on population, health, and nutrition in more than 90 LMICs since the early 1980’s. It recruits participants, usually women aged 14–49 years, using multi-stage cluster sampling and asks about the type of fuel used for cooking and the place of cooking [77], and so serves as a proxy only for household PM exposure. Fuel options include “electricity, liquid petroleum gas (LPG)/biogas, kerosene, cool/ignite, charcoal, wood, agricultural crops/straw, shrubs/grass, animal dung, and no cooking” in the households. Additionally, DHS data on smoking habits were used in one study [62] to assess the impact of maternal smoking on infant and child mortality.
Fifteen of the eligible household studies, including DHS studies (n = 11), administered structured questionnaires through interviews to investigate nuanced aspects of cooking practices. Four studies [49,68,69,71] gathered data on the types of cooking fuel used in households. Three studies [52,72,78] specifically scrutinized the use of wood biofuels. Another five studies [50,64,67,69,70] probed into the nature of the kitchen or the location of the cooking area. One study [73] surveyed the practice of mothers carrying their children while engaged in cooking activities. The remaining twelve [50,52,56–60,62,70,72–74] studies conducted interviews to gather data on secondhand smoke exposure from parents and neighbors.
The type of fuel used for cooking was a key variable used as a proxy for household air pollution exposure in the majority of HH studies, with exposed groups defined by the use of: ‘solid fuels’ [19,20,48,62,66]; ‘biomass fuels’[63] (coal, lignite, charcoal, wood, straw, shrubs, grass, agricultural crop, animal dung); being ‘exposed’ (wood, charcoal, dung, crop residues, shrubs, coal, kerosene)[18]; ‘firewood’[49]; ‘mixed fuels’ (using both clean and solid fuels)[20]; ‘household pollutants’ (smoking, nature and location of kitchen)[50]; ‘kitchen smoke’[51]; ‘indoor cooking’ [67,69,70]; ‘kerosene’[19]; ‘gas stove cooking fuel’[71]; ‘unclean cooking fuels’ (biomass, firewood, charcoal, kerosene)[68]; ‘charcoal fuel’[63] or ‘other pollutant cooking fuels’ (coal, lignite, paraffin, kerosene)[63]; ‘wood smoke’[72]; ‘wood biofuels’[52] or ‘polluting fuels’ (wood, dung, coal, paraffin)[55]. The comparator or non-exposed groups were defined as using: ‘non-solid fuels’ (electricity, LPG, natural gas, biogas, kerosene) [47,48,62]; ‘clean fuels’ (natural gas, biogas, LPG, electricity)[18–20,55,63,68]; having a ‘non-smoker in the house’ [52,57,58,62,70–74]; living in ‘low polluted areas (ambient PM2.5 levels < 25 µg/m³, measured with a DustTrak II Model 8532 air sampler)[66]; having ‘non-firewood fuels’[49]; ‘never carrying a child while cooking’[73]; ‘outdoor cooking’ [18,67,69,70]; ‘non-gas stove cooking fuel’[71]; ‘non-biomass cooking fuels’[53]; ‘separate kitchen in the household’ [18,64]; ‘no exposure to wood smoke’ [51,72] or ‘no exposure to wood biofuels’[52]. Kerosene was categorized as both ‘polluting fuels’ [19,63,68] and ‘clean /non-solid’ fuels [20,47,48,62] across different studies Gas cooking was also classified as a “clean” cooking fuel [20] (Table 2 and S4 File).
Exposure assignment used in ambient air pollution studies.
For studies that measured or modelled ambient air pollution, one study [20] employed a pre-established land use regression model (LUR) developed for NO2 and NOx which explained 68% and 75% of NOx and NO2 variances, respectively, and assigned it to the study participants’ residential addresses [79]. Another study [65] was conducted by matching the DHS enumeration area (the lowest geographic cluster in the DHS survey) with global satellite-estimated PM2.5 concentrations modeled by the Atmospheric Composition Analysis Group at Washington and Dalhousie University of the United States and Canada [80]. Three studies [59,61,66] used a hand-held sampler to collect real-time PM2.5 measurements at villages in the designated enumeration area (EA). The mean average PM2.5 within the assigned EAs was then used to examine the impact of PM2.5 on morbidity and mortality for children under five years (Table 2 and S4 File).
Potential confounders and modifiers.
Potential confounders adjusted for in the studies related to household, parental, child, and other environmental and contextual factors, and are presented in the supplementary file (S4 File). In our review, we found that none of the studies fully controlled for all key potential confounding factors like socioeconomic status, household condition, behavioural factors, and child nutritional condition, which can significantly influence air pollution's impact on child respiratory health, alongside factors like physical activity, season, and time trends. Additionally, four studies failed to report any adjustments for confounding [50,51,56,58]. In most studies (n = 15) that underwent confounder assessment, data collection on potential confounders, including birth size, relied on self-report questionnaires, although two studies [56,58] utilized health registration records for data acquisition. Five studies conducted a sensitivity analysis, showing a relatively smaller exposure-outcome association when potential confounders (e.g., season, breastfeeding, and birthweight) were excluded [18,20,62,65,67].
We observed that most studies reported differences in prevalence between exposed and unexposed groups in potential confounders such as maternal education, type of residence, birth size, breastfeeding status, and household factors but did not match or adjust for these differences in the analysis. For those studies that did adjust for confounders, we noted that this reduced the effect of air pollution on child respiratory hospitalization and mortality [19,59].
Summary of findings.
The findings of this systematic review indicated that the majority of studies have focused on proxies for air pollution, such as the use of solid fuels for cooking and household tobacco smoking, with limited emphasis on rigorous, objective assessments of both household and ambient air pollution exposures and their associated health impacts in SSA.
All-cause mortality
Sixteen studies evaluated the effect of air pollution on all-cause mortality in children under five years (Table 2). 14 of these studies reported on household air pollutants (i.e., solid fuels used for cooking or light; tobacco smoking at home) and two studies reported on ambient PM2.5 concentrations as the exposure. Ten of the fourteen studies reported the effects of using solid fuels on all-cause mortality, with seven of these studies reporting a significant association between solid fuel use and risk of all-cause mortality [18,47,48,53,55,62,63]. Conversely, three studies [19,20,54] reported that solid fuel use was not significantly associated with all-cause under-five mortality. Three studies [18,53,63] specifically reported a statistically positive association between biomass fuel use and increased risk of all-cause mortality, while two studies reported a similar adverse association with charcoal fuel use [18,63].
Three studies [18,63,64] highlighted the critical role of kitchen location, indicating that in-house kitchens were associated with an increased risk of all-cause mortality in children under five years. However, one study [63] reported that statistically significant association (HR = 0.98; 95% CI: 0.95–1.01).
Three studies reported that indoor tobacco smoking significantly contributed to all-cause mortality in children under five years [56,63,67] and one study [62] observed that maternal smoking, in conjunction with the use of solid fuels in households, resulted in a 59% increased risk of infant mortality.
Two studies reported that ambient PM2.5, estimated through satellite-based data from the Washington Atmospheric Composition Analysis Group (ACAG) and measured with DustTrak II Model 8532 air samplers (TSI Inc.), was an independent predictor of all-cause mortality in children under five years of age [65, 66], while one study reported that neither NOx nor NO2 were associated with neonatal mortality [20].
Respiratory mortality
Three studies related to respiratory mortality [50,66,73] reported varying findings, with one reporting a five-fold increased odds of ALRI-related mortality in children who were always carried by their mother while cooking compared to those who never had [73], and an increased but non-significant risk associated with a higher composite index of indoor air pollution (scored based on whether the kitchen was inside the house, the type of fire or fuel used, if the mother carried the child while cooking, and parental smoking habits)[73], and outdoor PM2.5 measured with DustTrak II Model 8532 air samplers (TSI Inc.) [66]. While our analysis showed a significant association between solid fuel exposure and under-five mortality, the limited availability of cause-specific mortality data restricted a deeper analysis of other air pollution-related causes of death.
Respiratory hospitalisations
Respiratory hospitalization outcomes reported in eligible studies included pneumonia (n = 14 studies), wheezing illness(n = 1 study), ARI (n = 4 studies), asthma (n = 1 studies), bronchitis (n = 2 studies), and pulmonary tuberculosis (n = 1 studies) (Table 2). A total of eight studies observed the relationship between parental tobacco smoking and pneumonia in children under five years, with four studies [52,57,59,70] reporting statistically significantly adverse association, while the other four studies [61,68,69,74] showed an increased but statistically insignificant association. Additionally, three studies [49,50,72] reported a significant adverse association between exposure to household biomass fuels and ARI, while one study did not [52].
Three studies [49,50,52] observed no statistically significant association between passive smoking and the odds of ARI in children under five, whereas one study [72] reported a significant association. One study [71] found no significant association between comorbidity with asthma and bronchitis and the use of gas stoves for cooking (OR = 1.4; 95% CI: 0.5–4.5), although the likelihood of asthma was three times higher in children from households using gas stoves (OR = 3.8; 95% CI: 1.2–13.3) compared to those from households without gas stoves. Our review showed variability in the association between parental smoking and bronchitis, with one study [71] reporting an increased, but non-significant association (OR = 1.5; 95% CI: 0.8–2.7), and another [52] reporting a significant protective association between passive smoking and bronchitis (RR = 0.35; 95% CI: 0.13–0.99).
Fig 2 depicts the type and number of outcomes reported in the 31 studies by exposure type. The number of outcomes reported in each study ranged from 1 to 12, with all-cause under-five mortality associated with household fuel use (n = 12) and pneumonia associated with second-hand smoking (n = 7) being the most reported outcomes.
Geographic disparity
The result of the systematic review showed the impact of air pollution exposure showed significant geographical disparities across SSA where indoor air pollution from solid fuel use in West Africa was a predominant factor, increasing the risk for under-five mortality. In Nigeria, two studies [47,54] reported that solid fuels were associated with high neonatal and child mortality rates. Specifically, 82% of neonatal deaths and 90% of post-neonatal deaths occurred in households where solid fuel use was prevalent, particularly in rural areas. These findings were further corroborated by Imo et al.,[48], where household solid fuel use was associated with over a two-fold increase in the risk for U5M.
However, studies in East Africa have identified a dual burden of indoor and outdoor pollution. In Kenya, exposure to ambient PM2.5, exceeding 25 µg/m³, elevated the risk of all-cause mortality by 22%[66], while in Ethiopia, each 10 µg/m³ increase in ambient PM2.5 corresponded to a 129% increase in U5M risk [65]. Indoor cooking with solid fuels exacerbated the risk of mortality, as demonstrated in a study in Kenya reported a 91% increased risk of U5M associated with indoor cooking using solid fuel [67], and in Ethiopia, where the absence of separate kitchen facilities increased infant mortality risk by 77% [64].
Two studies in South Africa reported households utilizing polluting fuels such as wood, coal, and paraffin face a nearly two-fold increase in U5M risk, with tobacco smoke further exacerbating the risk of sudden child death [55,58]. A paucity of studies in Central Africa mirrored the trends observed in other regions of SSA, indicating the significance of household energy sources across SSA in influencing mortality outcomes. Regionally, cause-specific mortality further revealed the variability in the impact of air pollution. Exposure to outdoor PM2.5 in East Africa [66] significantly associated with respiratory mortality, while in West Africa [50,51,73], indoor pollution from wood smoke and tobacco was implicated in ALRI and pneumonia-related deaths, with studies from the Gambia [73] and Nigeria [50, 51] demonstrating the significant contribution of exposure to wood smoke and maternal tobacco smoking on child respiratory mortality.
The result of systematic review further demonstrated that exposure to household air pollution from solid fuels and tobacco smoke in East Africa was a significant risk factor for respiratory infections in children under five years of age. In Uganda, exposure to tobacco smoke increased the risk of pneumonia threefold [70], although outdoor cooking was not significantly associated with pneumonia risk. Similarly, in Tanzania, unclean cooking fuels from firewood and charcoal fuels were significantly associated with severe pneumonia, whereas passive smoking showed no significant association [68]. A study in Uganda [71] also demonstrated the role of gas cooking, which increased asthma risk, further emphasizing that household energy sources contributed significantly to childhood respiratory morbidity. The inconsistent association between tobacco smoking and ARI outcomes in some East African studies underscored the importance of both traditional cooking practices and modern fuels in shaping respiratory health outcomes.
In Central Africa, particularly in Cameroon, wood smoke exposure was significantly linked to acute respiratory infections [72], with passive tobacco smoke being a major contributor to hospitalization due to respiratory conditions (OR = 3.58). In South Africa, studies have reported significant risks posed by both outdoor and indoor air pollution from cooking fuel and tobacco smoke exposure. Tobacco smoke consistently elevated the risk of ARI, pneumonia, and bronchitis across studies in South Africa, with maternal smoking playing a particularly prominent role [57,59,60]. Roux et al.,[60] found maternal smoking was associated with an increased risk of pneumonia in infants, while exposure to indoor PM10 and tobacco smoke heightened the risk of pneumonia hospitalization [61]. Toluene was also linked to an elevated risk of LRTIs [59].
Studies in West Africa have reported hazards of indoor air pollution from traditional cooking fuels, such as firewood and kerosene, and maternal tobacco risks. In Nigeria, three studies [49–51] identified the use of firewood and latrine as significant predictors of ARIs, with additional evidence linking maternal smoking to under-five pneumonia in a 30-month prospective study [51]. However, a study in Niger [74] reported no significant association between passive smoking and respiratory pathogen carriage, indicating regional variations in risk factors.
The results of the systematic review showed diverse, region-specific air pollution-related health risks in children under five years in SSA. Studies from West Africa indicated a primary impact of indoor air pollution due to solid fuels, while East African studies reported significant risks from both indoor and outdoor pollutants (proxies thereof). In South Africa, maternal tobacco smoke was the primary concern, with Central Africa facing mirrored risks from biomass and tobacco smoke exposure.
Urban-rural distinction
The finding of two studies showed that, in rural areas, reliance on solid fuels (wood, charcoal, and biomass) increases exposure to indoor air pollutants. A study in Nigeria found that rural households use solid fuels three times more than urban households [54]. Another study [63] conducted using DHS data from 23 Sub-Saharan African countries found that 91% of rural children were exposed to biomass smoke, with an increased under-five mortality risk (HR 1.21 for charcoal, 1.20 for other biomass fuels).
Urban-based study in Nairobi, Kenya [66] found exposure to outdoor PM2.5 raised children’s all-cause mortality risk by 22%, while in Adama, Ethiopia [20], a 10-unit increase in ambient NO2 was linked to a non-significantly higher neonatal mortality risk (OR = 0.91; 95% CI: 0.32–2.58). In a study in Ibadan Nigeria [51,81], indoor cooking with wood and charcoal significantly increased ARIs and pneumonia in children under five. In South Africa, maternal tobacco smoking in urban areas exacerbates these risks, with studies from Cape Town [58] and Johannesburg [57] linking maternal smoking to higher rates of sudden infant death and a fivefold increased risk of bacterial pneumonia in children.
In summary, this systematic review found significant adverse associations in 16 of 25 (64%) cross-sectional studies of exposure to solid cooking fuels and all-cause mortality, and 2 of 5 (40%) cross-sectional studies of exposure to second-hand/passive tobacco smoke during cooking and respiratory-related mortality. Additionally, 13 out of 31 studies (5 cross-sectional, 5 case-control, and 3 cohort) reported significant adverse associations between exposure to household cooking fuels (n = 8), household parental tobacco smoking (n = 6), indoor PM10 (n = 2), and an increased risk of respiratory hospitalizations in children aged under-five. However, seven (33.3%) findings were negative or reported non-significant associations.
Meta-analyses
We conducted a meta-analysis of fourteen findings from ten studies [18–20,47,48,53–55,62,63] to estimate a pooled OR for all-cause mortality in children under five years exposed to solid fuels in the home. In Fig 3, the forest plot and the pooled result from the random effect model illustrated that children exposed to solid fuel had a 1.31 times higher odds of all-cause mortality compared to those non-exposed to solid fuels (OR = 1.31; 95% CI: 1.16–1.47; τ² = 0.0319; X2 = 60,7%, df = 13(P < 0.0001); I2 = 78.6. The I2 statistic indicated considerable heterogeneity between the studies, although the random effects meta-analytic statistic was significantly different from 1 (P < 0.0001), indicating a high confidence in the finding.
The funnel plot in Fig 4 indicated some asymmetry, with more studies on the right-hand side of the center (the pooled OR estimate) than on the left-hand side and very few studies at the bottom of the funnel, suggesting potential publication bias with perhaps a lack of publication of smaller studies. The contour-enhanced funnel plot (supplementary file , S6) indicates perceived missed studies in the bottom two-thirds (apart from one study down at the bottom) and to the left of the plot, containing regions of high and low statistical significance, suggesting that funnel asymmetry cannot be solely attributed to publication bias. The regression-based Egger test with REML estimation to address residual heterogeneity indicated that the slope coefficient was not significantly different from zero (p = 0.69), and the bias coefficient was also not statistically significant (p = 0.11). The overall p-value for the test of the null hypothesis (no small-study effects) was 0.11, suggesting no significant evidence of small-study effects. This implies that the observed heterogeneity cannot be concluded to arise from small-study effects.
To further explore the observed heterogeneity, we conducted three sub-group analyses: by study year, study setting (single-country vs multi-country study), and by geographic region (Supplementary file, S6). The first subgroup analysis to examine the potential impact of the introduction of the Sustainable Development Goals (SDGs) in 2015 showed that the estimated pooled OR for studies conducted before 2015 was 1.34 (95% CI: 1.16–1.54; I2 = 82.9; p < 0.0001), and for those during and after 2015 was 1.24 (95% CI: 1.10–1.39; I2 = 78,6%, p = 0.343). The subgroup analysis by study setting found the estimated pooled OR for multi-country studies to be 1.18 (95% CI:1.06–1.31; I2 = 64.4%, p = 0.015) and for single-country studies was 1.46 (95% CI; 1.16–1.83; I2 = 80.0%, p < 0.0001). The sub-group analysis based on geographic region showed the pooled OR for West Africa as 1.47 (95% CI: 1.16–1.86; I² = 84.2%, p < 0.000), for East Africa as 0.61 (95% CI: 0.20–1.85; p < 0.000), and for South Africa as 1.99 (95% CI: 1.04–3.68; p < 0.000). The I2 test result revealed there was still significant heterogeneity in all subgroup analyses. In our sensitivity analysis based on study quality, excluding one study with a “high risk” of confounding bias did not significantly alter the heterogeneity, which remained high (OR = 1.29, 95% CI: 1.15–1.45; I² = 79.4%, p < 0.0001) (Supplementary file, S6). The leave-one-out sensitivity analysis to explore the influence of a single study on the overall effect size estimate showed the 95% CI of each omitted study contained the overall meta-analytic effect size estimated using all studies (OR = 1.31) (Supplementary file, S6 File). Hence, the source of the between-study heterogeneity remains unexplained based on the observed data.
Our second meta-analysis, drawing from seven studies, investigated the association between exposure to indoor tobacco smoke and the odds of pneumonia in children aged under five years [57,68–70,74], for which we report an increased, but non-significant, odds ratio (OR) of 1.32 (95% CI; 0.67–2.62) (Fig 5). The I2 statistic indicated that there was high heterogeneity between the studies (I2 = 87.3%, P < 0.0001). The wide confidence interval and high heterogeneity mean there is substantial variability between the results of the seven original studies. A funnel plot was not produced in this instance because of the minimal number (<10) of studies [82].
Discussion
This systematic review provides evidence on the relationship between exposure to sources of household air pollution (or proxies thereof) and both mortality and hospitalization for respiratory illness among children aged under five years living in SSA. Our review found that there was a dearth of studies that used measured air pollution measurements to assign exposure. The two studies [65, 66] that reported exposure to ambient PM2.5, reported that higher levels of ambient PM2.5 were associated with an increased risk of all-cause mortality in children under five years. These were consistent with a cross-country panel review of sixteen Asian countries conducted by Anwar et al.[83] which concluded that each one µg/m3 increase in annual PM2.5 resulted in a 14.5% increase in the number of children dying before the age of five.
The systematic review indicated that exposure to solid fuel has a strong, significant increased risk of under-five mortality, especially when combined with parental smoking. Suryadhi et al.[84] supported this finding, noting a synergistic effect of combined exposure that warrants further consideration in public health interventions. However, as this conclusion stems from a single study, it may be subject to overestimation due to small sample size and quality issues.
The meta-analysis showed that exposure to solid fuels in the home, increased the odds of all-cause mortality in children aged under five years. This finding is in agreement with the recent meta-analysis by Lee et al.[85] who observed an increased risk of under-five years mortality in children exposed to polluting fuels used for household cooking or heating (RR = 1.25 (95% CI: 1.18–1.33)) compared with our study’s OR of 1.31 (95% CI: 1.16–1.47). The subgroup analysis by study period, pre- and post-2015, observed continued impact of solid fuel use in studies conducted in post-2015, the era of SDGs, with an increased odds of mortality of 24%, suggesting air pollution from indoor household sources remains a fundamental pollution source to control to prevent childhood mortality.
In our systematic review, only four studies assessed the relationship between HAP from cooking fuels (n = 3) and ambient PM2.5 (n = 1) with ARI and CAP-related mortality in children under five years of age in SSA [50,51,66,73]. These studies showed a significant association between household air pollution from cooking fuels and ambient PM2.5, and ARI and/or CAP-related mortality. This finding was in line with a recent modeling study by Frosta et al.[86], who observed that 49.8% of respiratory mortality among children under five years in SSA in 2018 was attributable to total exposure to ambient and indoor PM2.5. The systematic review indicated that consistently carrying a child while cooking increased the odds of ARI mortality in children under five by fivefold, likely due to greater exposure to harmful pollutants [73]. However, this finding is from a single study and may be confounded by uncontrolled factors like fuel type and ventilation. It could also be influenced by small sample size, raising concerns about overestimation. While this highlights a potential intervention area, caution is needed to avoid unintended harm, such as neglect or increased injury risk if discouraging child-carrying during cooking.
Our systematic review found an inconclusive association between indoor air pollution exposure (mainly measured using proxies such as passive smoking and cooking fuel type) and hospitalisations for ARI, pneumonia, bronchitis, and asthma. This uncertainty may arise from poorly measured exposure or incomplete adjustment for confounders. A similar review has shown that reliance on proxies can obscure true relationships [25]. The variation in epidemiological findings might be attributable to population characteristics, cultural differences, and the type of fuels used in a wide range of LMICs.
Our meta-analysis of seven studies of children exposed to second-hand tobacco smoke indicated they were underpowered and did not allow a definite conclusion to be drawn. A previous meta-analysis by Riestiyowati et al.[87] using data from 12 primary studies from four continents (Africa (Tanzania, Ethiopia, and Morocco), South America (Brazil), Asia (Indonesia, Bangladesh, Nepal, and Vietnam) and Australia), reported results from two separate analyses which showed a 66% and about a twofold increase in odds of pneumonia in under-five years children exposed to second-hand smoke.
The potential mechanisms for adverse effects of air pollution on children’s health have been proposed in recent studies [88–90]. It has been widely suggested that children’s physiological immaturity makes them inherently vulnerable to the adverse effects of air pollution. In addition, there are many demographic and behavioural factors that influence the impact of air pollution on children, such as this age group being more likely to spend much of their time inside homes [91, 92].
The strengths of this systematic review include the focus on young children living in SSA and the inclusion of both household and ambient air pollution as exposures of interest, in an effort to study the impacts of total air pollution exposure. We employed a systematic, rigorous, and replicable approach for the review, including a comprehensive search strategy across a number of databases, retrieving published and preprinted articles to be able to incorporate the most recent articles. The quality appraisal process adhered to best practice guidelines tailored specifically for environmental health research and included screening by two reviewers throughout the whole process to minimize reviewer bias. The majority of studies used population-based large datasets such as DHS data, which serves as the primary and sometimes only source of essential data in Africa and other LMICs [26], allowing analysis with a large sample size and complete data.
However, there are limitations to our systematic review and meta-analysis. Our search focused on peer-reviewed articles in English, limiting inclusivity of articles published in other languages. Although efforts were made to encompass all studies from SSA countries, included studies were limited to eight countries, cautioning against broad extrapolation of results to all of SSA.
Exposure misclassification is likely to have occurred in some studies. For instance, kerosene fuel was inconsistently classified variably as ‘clean’ and ‘polluted’ [20,63,68], potentially leading to misclassification of exposure [93]. Furthermore, a major limitation is the reliance on proxy exposure measures, particularly from household surveys and indirect data. Relying only on information on fuels used for cooking and/or location of the kitchen within the property are relatively crude measures of exposure and do not account for time spent cooking, or nature and frequency of the biomass emissions. This introduces uncertainty when interpreting the findings and highlights the necessity for future research to use more precise exposure assessment methods. Consequently, our understanding of the causal association between household air pollution and health outcomes remains limited.
Methodological considerations including measurement bias, recall bias, and confounding, which all warrant consideration when interpreting findings. Importantly, low socioeconomic status (SES) can significantly confound the relationship between air pollution exposure and health outcomes. However, most studies in our review are cross-sectional and often inadequately adjusted for SES. This variability in study quality can lead to misleading associations. Studies that appropriately considered SES confounders demonstrated more robust associations between air pollution and adverse health effects [59,61]..
Another potential limitation relates to the assignment of hospitalization outcomes, which necessitated a broad search approach without specific disease codes. In addition, between-study heterogeneity was detected in our meta-analysis and was not reduced in both subgroup analyses and sensitivity analyses based on study quality, with inherent differences between the studies in aspects such as population structure and study designs. However, our leave-one-out sensitivity analysis showed the robustness of our estimation. Future studies could benefit from conducting an Individual Participant Data (IPD) meta-analysis using DHS data to enable granular analyses and explore detailed geographical disparities
The findings of this review underscore the critical need for effective public health strategies to mitigate the harmful effect of air pollution on young children living in SSA. However, in order to develop such strategies it is vital that better designed studies are conducted to provide an evidence base for actions. Significant gaps remain in understanding causation due to the nature of the studies in our systematic review, meaning that causality cannot be implied. To better understand the impact of air pollution on health outcomes, particularly in vulnerable populations such as children under five years, there is an urgent need for studies that incorporate objective measurements of air pollution for both indoor and outdoor environments. Such studies will provide more accurate and reliable data, enabling us to identify areas where interventions could be targeted most effectively.
PM2.5 is a mixture of various pollutants, and the chemical composition and source of particles may have different toxicities and health effects among children [94], underscoring the need to better determine the impacts of differing sources of air pollution by improved measurement of not only total pollution concentrations but also source specific contributions. Therefore, more studies are needed using valid and reliable measurement techniques to estimate exposure levels more accurately and so to minimise exposure misclassification or bias.
Further investigation is needed on the effect of solid fuel use on under-five years mortality and morbidity, including behavioural, and contextual drivers of use of fuel. Although some evidence exists for the adverse effects of solid fuel use, these may not capture the diversity and complexity of fuel sources and types in SSA, and findings from one country or region may not be applicable to another due to differences in type, availability, and affordability of cooking fuels as well as behavioural factors [63,85]. Further, the effect of demographic and other contextual variables such as socioeconomic conditions, nutrition, immunization remains unclear in SSA [62,67]. These factors may confound or exacerbate the relationship between air pollution and child health and may vary across different regions and populations in SSA.
This review found that most studies were cross-sectional in design, which can only show the association between air pollution and child health at a given point in time and may not capture the cumulative and chronic effect of air pollution exposure [85,95]. Therefore, better designed studies are needed, such as randomised trials or cohort studies, that can better assess the causal relationship between air pollution and adverse child health outcomes.
The effect of air pollution on child health depends on the location and movement or activity of the child, such as whether they are indoors or outdoors, near or far from the source, or engaged in different levels of activity. Simultaneous measurement of indoor and outdoor air is rarely performed in studies from SSA [11]. The region lacks evidence on the effect of the total burden of indoor and ambient air pollution on under-five years mortality. Therefore, studies are needed to assess the interactions and synergies between indoor and outdoor sources and pollutants, so to be able to develop and implement policies and regulations that address all sources of air pollution, especially given exposure is affected by factors beyond the household or individual level, such as community emissions, industrial emissions, traffic, or agricultural sources.
Conclusion
In summary, our meta-analysis found that the use of solid fuel in homes was associated with an increased risk of mortality in children under five years in SSA. However, a second meta-analysis provided insufficient evidence to establish a significant association between second-hand smoke and an increased risk of hospitalization due to pneumonia in children under five years of age. While there were very few studies that investigated the effects of ambient air pollution, and so we were unable to meta-analyse the results, the two eligible studies reported associations between increases in ambient PM2.5 and risk of all-cause mortality. We recommend that future studies conducted in SSA address previous study flaws to ensure that: 1) both indoor and ambient air pollution exposures are more accurately assessed, preferably by collecting measurements; 2) that potential confounders are accounted for in the statistical analyses; and, 3) that stronger epidemiological study designs better suited to infer causality, such as randomised trials of interventions taking into account total exposure, or cohort studies, are conducted.
Supporting information
S3 File. Irrelevant studies which were excluded during the title and abstract screening process (N=5304).
https://doi.org/10.1371/journal.pone.0320048.s003
(CSV)
S4 File. Studies screened for full-text review and data extraction.
https://doi.org/10.1371/journal.pone.0320048.s004
(DOCX)
References
- 1. World Health Organization. Air quality and health. 2023; Available from: https://www.who.int/teams/environment-climate-change-and-health/air-quality-and-health/health-impacts
- 2. Brumberg HL, Karr CJ, Council on environmental health. Ambient Air Pollution: Health Hazards to Children. Pediatrics. 2021;147(6):e2021051484. pmid:34001642
- 3. Makri A, Stilianakis NI. Vulnerability to air pollution health effects. Int J Hyg Environ Health. 2008;211(3–4):326–36. pmid:17719845
- 4.
World Health Organization, Air pollution and child health: prescribing clean air: summary. 2018. Geneva: World Health Organization
- 5.
Pant P. State of Global Air 2020-Release Webinar. 2020.
- 6. Global Burden of Disease Collaborative Network. Global Burden of Disease Study 2019 (GBD 2019) Results. Lancet. 2021;392:1736–88.
- 7. The World Bank. Pollution. 2023; Available from: https://www.worldbank.org/en/topic/pollution; [accessed 4 Februray 2024].
- 8. Fisher S, Bellinger DC, Cropper ML, Kumar P, Binagwaho A, Koudenoukpo JB, et al. Air pollution and development in Africa: impacts on health, the economy, and human capital. Lancet Planet Health. 2021;5(10):e681–8. pmid:34627472
- 9. Bourzac K. Air pollution kills 780,000 people in Africa each year. Chem Eng News. 2019;97(17):4.
- 10. Abbasi-Kangevari M, Malekpour MR, Masinaei M, Moghaddam SS, Ghamari SH, Abbasi-Kangevari Z, et al. Effect of air pollution on disease burden, mortality, and life expectancy in North Africa and the Middle East: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Planetary Health, 2023;7(5):e358–e369.
- 11. Heft-Neal S, Burney J, Bendavid E, Burke M. Robust relationship between air quality and infant mortality in Africa. Nature. 2018;559(7713):254–8. pmid:29950722
- 12. Amegbor PM, Sabel CE, Mortensen LH, Mehta AJ, Rosenberg MW. et al. Early-life air pollution and green space exposures as determinants of stunting among children under age five in Sub-Saharan Africa. J Exposure Science and Environmental Epidemiology. 2023;
- 13. Agyepong IA, Sewankambo N, Binagwaho A, Coll-Seck AM, Corrah T, Ezeh A, et al. The path to longer and healthier lives for all Africans by 2030: the Lancet Commission on the future of health in sub-Saharan Africa. Lancet. 2017;390(10114):2803–59. pmid:28917958
- 14. Karimi B, Shokrinezhad B. Air pollution and mortality among infant and children under five years: A systematic review and meta-analysis. Atmospheric Pollution Res. 2020;11(6):61–70.
- 15. Cohen AJ, Brauer M, Burnett R, Anderson HR, Frostad J, Estep K, et al. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015. Lancet. 2017;389(10082):1907–18. pmid:28408086
- 16. Mehta S, Shin H, Burnett R, North T, Cohen AJ. Ambient particulate air pollution and acute lower respiratory infections: a systematic review and implications for estimating the global burden of disease. Air Qual Atmos Health. 2013;6(1):69–83. pmid:23450182
- 17.
Lwin KS. Attributable risk of household air pollution associated with under-5 mortality in 56 low-and-middle-income countries, 2010-2020. 2020.
- 18. Bickton FM, Ndeketa L, Sibande GT, Nkeramahame J, Payesa C, Milanzi EB. Household air pollution and under-five mortality in sub-Saharan Africa: an analysis of 14 demographic and health surveys. Environ Health Prev Med. 2020;25(1):67. pmid:33148165
- 19. Kleimola LB, Patel AB, Borkar JA, Hibberd PL. Consequences of household air pollution on child survival: evidence from demographic and health surveys in 47 countries. Int J Occup Environ Health. 2015;21(4):294–302. pmid:25843087
- 20. Flanagan E, Oudin A, Walles J, Abera A, Mattisson K, Isaxon C, et al, Ambient and indoor air pollution exposure and adverse birth outcomes in Adama, Ethiopia. Environ. Int. 2022. 164.
- 21. Enyew HD, Mereta ST, Hailu AB. Biomass fuel use and acute respiratory infection among children younger than 5 years in Ethiopia: a systematic review and meta-analysis. Public Health. 2021;193:29–40. pmid:33713984
- 22. Tefera W, Asfaw A, Gilliland F, Worku A, Wondimagegn M, Kumie A, et al. Indoor and Outdoor Air Pollution- related Health Problem in Ethiopia: Review of Related Literature. Ethiop J Health Dev. 2016;30(1):5–16. pmid:28890631
- 23. Bede-Ojimadu O, Orisakwe OE. Exposure to wood smoke and associated health effects in sub-saharan africa: a systematic review. Ann Glob Health. 2020;86(1):32. pmid:32211302
- 24. Katoto PDMC, Byamungu L, Brand AS, Mokaya J, Strijdom H, Goswami N, et al. Ambient air pollution and health in Sub-Saharan Africa: current evidence, perspectives and a call to action. Environ Res. 2019;173:174–88. pmid:30913485
- 25. Adaji EE, Ekezie W, Clifford M, Phalkey R. Understanding the effect of indoor air pollution on pneumonia in children under 5 in low- and middle-income countries: a systematic review of evidence. Environ Sci Pollut Res Int. 2019;26(4):3208–25. pmid:30569352
- 26. Odo DB, Yang IA, Knibbs LD. A systematic review and appraisal of epidemiological studies on household fuel use and its health effects using demographic and health surveys. Int J Environ Res Public Health. 2021;18(4):1411. pmid:33546363
- 27.
Ma YLH, Zhang N, Luo X. A systematic literature review on indoor PM2. 5 concentrations and personal exposure in urban residential buildings. 2022.
- 28. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Int J Surg. 2021;88:105906. pmid:33789826
- 29. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015;4(1):1. pmid:25554246
- 30. Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ. 2015;350:g7647. pmid:25555855
- 31. Eyasu Alem Lake GBM., Christine TC. The effect of air pollution on morbidity and mortality among children aged under 5 years in sub-Saharan. Afr Syst Rev. PROSPERO. 2023. Available from: https://www.crd.york.ac.uk/prospero/export_details_pdf.php; [accessed 04 November 2023]
- 32. Page MJ, Shamseer L, Tricco AC. Registration of systematic reviews in PROSPERO: 30,000 records and counting. Syst Rev. 2018;7(1):1–9.
- 33. Gusenbauer M, Haddaway NR. Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of Google Scholar, PubMed, and 26 other resources. Res Synth Methods. 2020;11(2):181–217. pmid:31614060
- 34. Harrison H, Griffin SJ, Kuhn I, Usher-Smith JA. Software tools to support title and abstract screening for systematic reviews in healthcare: an evaluation. BMC Med Res Methodol. 2020;20(1):7. pmid:31931747
- 35.
Ohat N. Handbook for conducting a literature-based health assessment using OHAT approach for systematic review and evidence integration. 2019;5(8):2019.
- 36. Woodruff TJ, Sutton P. The Navigation Guide systematic review methodology: a rigorous and transparent method for translating environmental health science into better health outcomes. Environ Health Perspect. 2014;122(10):1007–14. pmid:24968373
- 37. Rooney AA, Boyles AL, Wolfe MS, Bucher JR, Thayer KA. Systematic review and evidence integration for literature-based environmental health science assessments. Environ Health Perspect. 2014;122(7):711–8. pmid:24755067
- 38. United Nations Inter-agency Group for Child Mortality Estimation (UN IGME). Under-five mortality. 2021; Available from: https://web.archive.org/web/20220901091013/https://data.unicef.org/topic/child-survival/under-five-mortality/#more--1603[accessed [accessed 01 March 2024].
- 39. Stare J, Maucort-Boulch D. Odds ratio, hazard ratio and relative risk. Adv Meth Stat. 2024;13(1):59–67.
- 40. Ryan R. Heterogeneity and subgroup analyses in Cochrane Consumers and Communication Group reviews: planning the analysis at protocol stage. Cochrane Consumers Communic Rev Group. 2016:2–9.
- 41. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–60. pmid:12958120
- 42. Jann B. Making Regression Tables Simplified. Stata J Promoting Communications on statistics and Stata. 2007;7(2):227–44.
- 43. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–34. pmid:9310563
- 44. Dear KB, Begg CB. An approach for assessing publication bias prior to performing a meta-analysis. Statistical Science. 1992:237–45.
- 45. Shi L, Lin L. The trim-and-fill method for publication bias: practical guidelines and recommendations based on a large database of meta-analyses. Medicine (Baltimore). 2019;98(23):e15987. pmid:31169736
- 46. Deeks JJ, Higgins JP, Altman DG, Gøtzsche PC, Juni P, Moher D, et al. Analysing data and undertaking meta‐analyses. Cochrane handbook for systematic reviews of interventions. 2019:241–84.
- 47. Ezeh OK, Agho KE, Dibley MJ, Hall JJ, Page AN. The effect of solid fuel use on childhood mortality in Nigeria: evidence from the 2013 cross-sectional household survey. Environ Health. 2014;13:113. pmid:25514998
- 48. Imo CK, Wet-Billings ND. Socio-ecological determinants of under-five mortality in Nigeria: Exploring the roles of neighbourhood poverty and use of solid cooking fuel. J Biosoc Sci. 2021.
- 49. Fakunle GA, Ana GR, Ayede AI. Environmental risk factors for acute respiratory infections in hospitalized children under 5 years of age in Ibadan, Nigeria. Paediatr Int Child Health. 2014;34(2):120–4. pmid:24621237
- 50. Johnson AW, Aderele WI. The association of household pollutants and socio-economic risk factors with the short-term outcome of acute lower respiratory infections in hospitalized pre-school Nigerian children. Ann Trop Paediatr. 1992;12(4):421–32. pmid:1283673
- 51. Johnson A-WBR, Osinusi K, Aderele WI, Gbadero DA, Olaleye OD, Adeyemi-Doro FAB. Etiologic agents and outcome determinants of community-acquired pneumonia in urban children: a hospital-based study. J Natl Med Assoc. 2008;100(4):370–85. pmid:18481475
- 52. Ujunwa F, Ezeonu C, Risk factors for acute respiratory tract infections in under‑five children in enugu Southeast Nigeria. Annals Med Health Sci Res. 2014;4(1):95–99.
- 53. Latona OO, Yusuf OB, Adebowale SA. Patterns and predictors of mortality among children under-five years in Nigeria. Afr J Biomed Res. 2017;20(1):37–43.
- 54. Samuel OW, Oni GA, KC S, Wurzer M, Akinyemi AI. Household use of solid fuel for cooking and under-five mortality in Nigeria. African Population Studies. 2018;32(1).
- 55. Wichmann J, Voyi KVV. Influence of cooking and heating fuel use on 1-59 month old mortality in South Africa. Matern Child Health J. 2006;10(6):553–61. pmid:16758332
- 56. Winterbach M, Hattingh C, Heathfield LJ. Retrospective study of sudden unexpected death of infants in the Garden Route and Central Karoo districts of South Africa: Causes of death and epidemiological factors. S Afr J CH. 2021;15(2):74.
- 57. Verani JR, Groome MJ, Zar HJ, Zell ER, Kapongo CN, Nzenze SA, et al. Risk Factors for Presumed Bacterial Pneumonia Among HIV-uninfected Children Hospitalized in Soweto, South Africa. Pediatr Infect Dis J. 2016;35(11):1169–74. pmid:27276181
- 58. Heathfield LJ, Martin LJ, Ramesar R. A 5-year retrospective analysis of infant death at Salt River Mortuary, Cape Town. S Afr J CH. 2020;14(3):148.
- 59. Vanker A, Barnett W, Workman L, Nduru PM, Sly PD, Gie RP, et al. Early-life exposure to indoor air pollution or tobacco smoke and lower respiratory tract illness and wheezing in African infants: a longitudinal birth cohort study. Lancet Planet Health. 2017;1(8):e328–36. pmid:29167839
- 60. le Roux DM, Myer L, Nicol MP, Zar HJ. Incidence and severity of childhood pneumonia in the first year of life in a South African birth cohort: the Drakenstein Child Health Study. Lancet Glob Health. 2015;3(2):e95–103. pmid:25617203
- 61. Le Roux DM, Nicol MP, Vanker A, Nduru PM, Zar HJ. Factors associated with serious outcomes of pneumonia among children in a birth cohort in South Africa. PLoS One. 2021;16(8):e0255790. pmid:34388194
- 62. Akinyemi JO, Adedini SA, Wandera SO, Odimegwu CO. Independent and combined effects of maternal smoking and solid fuel on infant and child mortality in sub-Saharan Africa. Trop Med Int Health. 2016;21(12):1572–82. pmid:27618434
- 63. Owili PO, Muga MA, Pan W-C, Kuo H-W. Cooking fuel and risk of under-five mortality in 23 Sub-Saharan African countries: a population-based study. Int J Environ Health Res. 2017;27(3):191–204. pmid:28552005
- 64. Shifa GT, Ahmed AA, Yalew AW. Socioeconomic and environmental determinants of under-five mortality in Gamo Gofa Zone, Southern Ethiopia: a matched case control study. BMC Int Health Hum Rights. 2018;18(1):14. pmid:29486757
- 65. Shiferaw AB, Kumie A, Tefera W. Fine particulate matter air pollution and the mortality of children under five: a multilevel analysis of the Ethiopian Demographic and Health Survey of 2016. Frontiers in Public Health. 2023;11.
- 66. Egondi T. Exposure to outdoor particles (PM2.5) and associated child morbidity and mortality in socially deprived neighborhoods of Nairobi, Kenya. ATMOSPHERE. 2018;9(9).
- 67. Starnes JR, Rogers A, Wamae J, Okoth V, Mudhune SA, Omondi A, et al. Childhood mortality and associated factors in Migori County, Kenya: evidence from a cross-sectional survey. BMJ Open. 2023;13(8):e074056. pmid:37607788
- 68. Ngocho JS, de Jonge MI, Minja L, Olomi GA, Mahande MJ, Msuya SE. et al. Modifiable risk factors for community-acquired pneumonia in children under 5 years of age in resource-poor settings: a case–control study. Trop Med Int Health, 2019;24(4):484–492.
- 69. PrayGod G, Mukerebe C, Magawa R, Jeremiah K, Török ME. Indoor air pollution and delayed measles vaccination increase the risk of severe pneumonia in children: results from a case-control study in Mwanza, Tanzania. PLoS One. 2016;11(8):e0160804. pmid:27508389
- 70. Kiconco G, Turyasiima M, Ndamira A, Yamile OA, Egesa WI, Ndiwimana M, et al. Prevalence and associated factors of pneumonia among under-fives with acute respiratory symptoms: a cross sectional study at a Teaching Hospital in Bushenyi District, Western Uganda. Afr Health Sci. 2021;21(4):1701–10. pmid:35283986
- 71. Nantanda R, Ostergaard MS, Ndeezi G, Tumwine JK. Factors associated with asthma among under-fives in Mulago hospital, Kampala Uganda: a cross sectional study. BMC Pediatr. 2013;13:141. pmid:24024970
- 72. Tazinya AA, Halle-Ekane GE, Mbuagbaw LT, Abanda M, Atashili J, Obama MT. et al., Risk factors for acute respiratory infections in children under five years attending the Bamenda Regional Hospital in Cameroon. BMC Pulm. Med. 2018. 18(1).
- 73. Francisco AD. Risk factors for mortality from acute lower respiratory tract infections in young gambian children. Int J Epidemiol. 1993;22(6):1174–82.
- 74. Dano ID, Ousmane S, Moumouni K, Lagare A, Iss I, Test J. Risk factors associated with Streptococcus pneumonia carriage in children under five years old with acute respiratory infection in Niger. Pan African Medical J. 2019;33.
- 75. Wichmann J, Voyi KVV. Impact of cooking and heating fuel use on acute respiratory health of preschool children in South Africa. Southern Afr J Epidemiol Infect. 2006;21(2):48–54.
- 76. Owili PO, Muga MA, Pan W-C, Kuo H-W. Indoor secondhand tobacco smoke and risk of under-five mortality in 23 sub-Saharan Africa countries: a population based study and meta-analysis. PLoS One. 2017;12(5):e0177271. pmid:28542166
- 77. Williams KN, Northcross AL, Graham JP. Health impacts of household energy use: indicators of exposure to air pollution and other risks. Bull World Health Organ. 2015;93(7):507–8. pmid:26170510
- 78. Auld AF, Tuho MZ, Ekra KA, Shiraishi RW, Mohamed F, Kouakou JS, et al. Temporal trends in mortality and loss to follow-up among children enrolled in Côte d’Ivoire’s national antiretroviral therapy program. Pediatr Infect Dis J. 2014;33(11):1134–40. pmid:25093975
- 79. Abera A, Malmqvist E, Mandakh Y, Flanagan E, Jerrett M, Gebrie GS, et al. Measurements of NOx and Development of Land Use Regression Models in an East-African City. Atmosphere. 2021;12(4):1–10.
- 80. van Donkelaar A, Martin RV, Brauer M, Hsu NC, Kahn RA, Levy RC, et al. Global estimates of fine particulate matter using a combined geophysical-statistical method with information from satellites, models, and monitors. Environ Sci Technol. 2016;50(7):3762–72. pmid:26953851
- 81. Johnson PI, Sutton P, Atchley DS, Koustas E, Lam J, Sen S, et al. The Navigation Guide - evidence-based medicine meets environmental health: systematic review of human evidence for PFOA effects on fetal growth. Environ Health Perspect. 2014;122(10):1028–39. pmid:24968388
- 82. Lau J, Ioannidis JPA, Terrin N, Schmid CH, Olkin I. The case of the misleading funnel plot. BMJ. 2006;333(7568):597–600. pmid:16974018
- 83. Anwar A, Ullah I, Younis M, Flahault A. Impact of Air Pollution (PM2.5) on Child Mortality: Evidence from Sixteen Asian Countries. Int J Environ Res Public Health. 2021;18(12):6375. pmid:34204659
- 84. Suryadhi MAH, Abudureyimu K, Kashima S, Yorifuji T. Effects of household air pollution from solid fuel use and environmental tobacco smoke on child health outcomes in Indonesia. J Occup Environ Med. 2019;61(4):335–9.
- 85. Lee KK, Bing R, Kiang J, Bashir S, Spath N, Stelzle D, et al. Adverse health effects associated with household air pollution: a systematic review, meta-analysis, and burden estimation study. Lancet Glob Health. 2020;8(11):e1427–34. pmid:33069303
- 86. Nguyen QP, Baumann MM, Blacker BF, Marczak LB, Deshpande A. Mapping development and health effects of cooking with solid fuels in low-income and middle-income countries, 2000-18: a geospatial modelling study. Lancet Glob Health. 2022;10(10):e1395–411. pmid:36113526
- 87. Riestiyowati MA, Rahardjo SS, Murti B. A meta-analysis of the effects of secondhand smoke exposure toward the incidence of pneumonia in children under five. J Epidemiol Public Health. 2020;5(4):410–9.
- 88. Johnson NM, Hoffmann AR, Behlen JC, Lau C, Pendleton D, Harvey N, et al. Air pollution and children’s health-a review of adverse effects associated with prenatal exposure from fine to ultrafine particulate matter. Environ Health Prev Med. 2021;26(1):72. pmid:34253165
- 89. Wu W, Jin Y, Carlsten C. Inflammatory health effects of indoor and outdoor particulate matter. J Allergy Clin Immunol. 2018;141(3):833–44. pmid:29519450
- 90. Kim D, Chen Z, Zhou L-F, Huang S-X. Air pollutants and early origins of respiratory diseases. Chronic Dis Transl Med. 2018;4(2):75–94. pmid:29988883
- 91. Aithal SS, Sachdeva I, Kurmi OP. Air quality and respiratory health in children. Breathe (Sheff). 2023;19(2):230040. pmid:37377853
- 92.
Perera FP, Children's Health and the Peril of Climate Change. 2022: Oxford University Press.
- 93.
World Health Organization, WHO guidelines for indoor air quality: household fuel combustion. Geneva: World Health Organization, http://www.who.int/indoorair/guidelines/hhfc/en/ 2014.
- 94. Adams K, Greenbaum DS, Shaikh R, van Erp AM, Russell AG. Particulate matter components, sources, and health: Systematic approaches to testing effects. J Air Waste Manag Assoc. 2015;65(5):544–58. pmid:25947313
- 95. Glenn BE, Espira LM, Larson MC, Larson PS. Ambient air pollution and non-communicable respiratory illness in sub-Saharan Africa: a systematic review of the literature. Environ Health. 2022;21(1):40. pmid:35422005