Skip to main content
Download PDF
Download ePub

Quantifying the association between PM2.5 air pollution and IQ loss in children: a systematic review and meta-analysis

Environmental Health volume 23, Article number: 101 (2024) Cite this article

Abstract

Background

A growing body of epidemiologic and toxicologic literature indicates that fine airborne particulate matter (PM2.5) pollution is neurotoxic and threatens children’s neurobehavioral development, resulting in reduced cognitive function. Understanding the magnitude of this effect is critical for establishing public health policies that will protect children’s health, preserve human capital, and support societal progress.

Objective

To quantify the association between ambient PM2.5 air pollution and loss of cognitive function in children, as measured by Intelligence Quotient (IQ) scores, through a systematic literature review and meta-analysis.

Methods

Following PRISMA guidelines, we conducted a systematic literature search across seven databases: Agricultural and Environmental Science, BIOSIS Citation Index, Embase, GreenFILE, PubMed, Scopus, and Web of Science to identify original scientific studies that investigated the impact of PM2.5 exposure during pre-and postnatal periods on IQ loss during childhood. Using data from studies included for final review, we conducted a meta-analysis, using a random effects model to compute a beta coefficient that quantifies the overall effect of PM2.5 exposure on Full-Scale IQ (FSIQ), Performance IQ (PIQ), and Verbal IQ (VIQ).

Findings

Of the 1,107 unique publications identified, six studies met the inclusion criteria for final review, representing 4,860 children across three continents (North America, Europe, and Asia). The mean PM2.5 concentration across all studies was 30.4 ± 24.4 µg/m3. Exposure timing ranged from the prenatal period to mid-childhood. Children were an average of 8.9 years old at the time of cognitive testing. We found that each 1 µg/m3 increase in PM2.5 concentration is associated with a -0.27 point change in FSIQ (p < 0.001), 0.39 point change in PIQ (p = 0.003), and -0.24 point change in VIQ (p = 0.021).

Conclusion

Through a systematic review and meta-analysis, we identified a statistically significant relationship between increased exposure to PM2.5 air pollution and reduced cognitive function in children, with the most pronounced impact on PIQ. This analysis will enable estimation of the burden of adverse neurobehavioral development attributable to PM2.5 in pediatric populations and will inform local and global strategies for exposure prevention.

Peer Review reports

Background

Ambient air pollution is a severe and pervasive hazard to population health. It is of particular concern for the health of children. The World Health Organization (WHO) estimates that 99% of the world’s population breathes air in which levels of fine airborne particulate matter pollution with a mass median diameter of 2.5 µm (μm) or less (PM2.5) exceed the WHO guideline of 5 µg/m3 air. Due to their extremely small diameter, PM2.5 particles can penetrate directly into the brain via the olfactory bulb and deeply into the lungs upon inhalation. The smallest inhaled particles can cross the alveolar-capillary membrane and enter the bloodstream, where they translocate through systemic circulation, inducing oxidative stress and triggering the body’s immune response, leading to persistent inflammation [1,2,3].

The health consequences of PM2.5 pollution exposure include multiple adverse respiratory, cardiovascular, immune, neurological, and neonatal outcomes, including premature mortality [4]. Air pollution was estimated to have caused 6.67 million deaths worldwide in 2019. Approximately two-thirds (4.14 million) of these deaths were attributable to ambient PM2.5 pollution. This global burden of disease is inequitable, with an estimated 92% of pollution-related deaths occurring in Low- and Middle-Income Countries (LMICs) [4].

A growing body of literature has examined the neurological effects of PM2.5. Studies in adults have identified PM2.5 as a risk factor for neurodegenerative conditions, including cognitive impairment, neuronal death, neuroinflammation, and the accumulation of neuropathological markers [5]. In children, a limited but growing number of studies have identified linkages between PM2.5 exposure and decreased cognitive performance, quantified by Intelligence Quotient (IQ) loss [6,7,8,9,10,11]. Other components of air pollution such as sulfur dioxide, nitrogen dioxide, polycyclic aromatic hydrocarbons (PAH), and PM10 have also been linked to decreased cognitive performance on verbal and math tests [12, 13]. Two recent reviews have examined associations between air pollution and a suite of child health outcomes, including cognitive and behavioral outcomes [14, 15]; however, neither produced a pooled beta coefficient linking PM2.5 exposure and changes in children’s IQ.

Children are uniquely susceptible to PM2.5 exposure due to their higher minute ventilation rate, greater oxygen consumption per unit body weight, permeable biological membranes (e.g., blood–brain barrier, airway epithelium), and immature immune/detoxification systems. Children’s vulnerability is further enhanced by the extraordinary complexity of brain development in early life. Damage done to the developing brain in utero and in early postnatal life can result in permanent injury and increase the risk for adverse neurobehavioral outcomes across the life course. At highest risk are children born prematurely and children with pre-existing health conditions.

Until now, a key impediment to quantifying the impact of PM2.5 pollution on IQ loss in children has been the lack of an exposure‐response function linking pollution to IQ loss [16]. This study aims to close this knowledge gap and develop concentration‐response functions quantifying the relationship between airborne PM2.5 concentrations and IQ loss. Such functions can be applied in future epidemiologic studies and utilized to estimate the global burden of disease attributable to air pollution (Fig. 1).

Fig. 1
figure 1

IQ Types Included in Meta Analyses

Full size image

Methods

PECO Statement

The Population, Exposure, Comparator, and Outcomes (PECO) strategy was used to define our research question: In children aged 0–18 years (Population), how does exposure to higher levels of PM2.5 during the prenatal, postnatal, and childhood period (Exposure) compared to lower levels of PM2.5 (Comparator) affect neurodevelopment, as measured by IQ scores (Outcome)?

Information Sources and Search Strategies

We conducted a systematic literature search following PRISMA guidelines to characterize the correlation between PM2.5 exposure and IQ loss in children. The search was conducted across seven databases: Agricultural and Environmental Science, BIOSIS Citation Index, Embase, GreenFILE, PubMed, Scopus, and Web of Science on 27 October 2022 (Table 1). These databases were selected as they covered a broad range of topics related to medicine, public health, and environmental health. Additionally, all the databases met the criterion of being available at Boston College (BC), where the review was conducted. Our search strategy was guided by a BC research librarian. Following the initial systematic review, a second author conducted snowball sampling by reviewing articles that cited the identified studies on 25 September 2023. The newly identified studies were assessed according to the inclusion and exclusion criteria applied in the systematic review to ensure consistency and relevance.

Table 1 Search Strategy by Database and Additional Notes
Full size table

Eligibility Criteria

Reports identified through these databases were included in this analysis if they met the following criteria:

• Reports of original research (e.g., not review articles)

• Examined human subjects

• Studied children (persons under the age of 18 years)

• Measured PM2.5 exposure prenatally and/or postnatally (including early childhood)

• Measured cognitive performance using FSIQ, PIQ, and/or VIQ

• Provided quantitative data on the correlation between PM2.5 and IQ

• Represented the most recent report from ongoing studies with multiple publications

Screening Process

The initial search retrieved 2,064 articles, which were stored in Zotero, a citation-managing software. A total of 957 duplicate reports were removed, resulting in 1,107 unique reports. After screening the titles and abstracts of these reports, 1066 were excluded because they did not meet our eligibility criteria.

The title-abstract screen eliminated articles according to the following criteria: did not involve human subjects (n = 25), did not measure PM2.5 (n = 876), did not measure IQ (n = 153), was not an original research report (n = 9), or a more recent report was available from the same ongoing study (n = 2). After excluding these records, 41 records were retrieved for full-text screening, and an additional study was added following snowball sampling (n = 42). Full-text screening excluded an additional 36 articles because they did not meet the eligibility criteria. Reasons for exclusion at this level were that the study did not measure PM2.5 (n = 21), did not measure IQ (n = 7), was not an original scientific study (n = 3), did not report a correlation between PM2.5 exposure and IQ (n = 4), or did not provide sufficient data for analysis (n = 1). In sum, this process identified six studies for inclusion in the final review and used a pooled beta coefficient using a random effect meta-analysis (Figs. 1 and 2).

Fig. 2
figure 2

Flow Chart of Literature Search and Selection for Meta-Analysis

Full size image

Data Collection and Preprocessing

Data from the six studies included for final review were compiled in Table 2. Exposure details, including the exposure window (e.g., prenatal or postnatal), exposure location, and exposure measurement, were compiled in Table 3. Cohort details, including size, location, name, and recruitment strategy, were compiled in Tables 3 and 4.

Table 2 Studies Included for Final Review
Full size table
Table 3 Exposure Details
Full size table
Table 4 Cohort Details
Full size table

Data Synthesis

To obtain comparable values for each of our two variables, results from included studies were expressed as a linear slope (IQ points lost per 1ug/m3 increase in PM2.5 concentration) when reported otherwise. Relevant calculations for individual studies are available in Table 2. Data standardization conformed to the following formulas:

For studies that reported changes in IQ per interquartile range of PM2.5 exposure:

\({\beta }=\frac{\Delta IQ points}{\Delta { PM}_{2.5} (Q4/Q3 boundary - Q2/Q1 boundary)}\)

For studies that reported changes in IQ per X linear coefficient increase in PM2.5:

\({\beta }=\frac{\Delta IQ points}{\Delta { PM}_{2.5} (linear coefficient)}\)

For studies that reported mean differences in IQ points for multiple PM2.5 exposure groups:

\(\beta =\frac{\sum ({x}_{i}-x)({y}_{i}-y)}{\sum ({x}_{i}-x{)}^{2}}\)

Where \(x\)i are the mean PM2.5 levels for each group, and \({y}_{i}\) is the IQ differences for each group

Statistical Analysis

All meta-analytic techniques were carried out using the package metafor in R statistical software (v4.4.1; R Core Team 2024), an open-source platform. To calculate a beta coefficient relating PM2.5 and IQ, we performed a meta-analysis using a random-effects model to account for between-study variability and provide a more generalized estimate of the effect size, given the heterogeneity across studies. The beta coefficient (overall effect size) represents the pooled effect across all included studies, accounting for both within- and between-study variability. Compared to a fixed-effects model, the random-effects model provides a more conservative estimate of the effect size, as it incorporates between-study heterogeneity.

Risk of Bias (RoB)

To evaluate potential biases that may affect the validity of study findings, we used a standardized Risk of Bias (RoB) framework, assessing six domains: selection bias, confounding bias, measurement of exposure, outcome assessment bias, attrition bias, and reporting bias. If each domain in a study was ranked as “low,” its overall Rob was determined as “low”. If the study had domains ranked as “low" and “moderate,” its overall Rob was determined as “moderate”. Details are specified in Supplemental Table S1.

Heterogeneity

Given that our three meta-analyses included a small number of studies, we assessed for heterogeneity both quantitatively and qualitatively. Using the metafor package in RStudio we calculated the I2 statistic, which represents the proportion of variation in effect sizes due to heterogeneity rather than chance, tau (the estimated standard deviation of the actual effect sizes e.g. between-study variability), and p-value (heterogeneity) from the Cochran’s Q-test (Supplemental Table S3). Following this quantitative analysis, we visually inspected the forest plots, paying attention to the degree of overlap of confidence intervals and position of study estimates (e.g., if they were clustered or staggered).

Results

Six epidemiological studies (five cohort and one cross-sectional) were included in our final analysis (Table 2), representing data from 4,860 children across three continents (North America, Europe, and Asia). The mean level of PM2.5 exposure was 30.4 ± 24.4 μg/m3 (Table 3), and the mean age at IQ testing was 8.9 years. (Table 4) Without exception, each study reported a negative association between PM2.5 exposure and children’s cognitive function. We present findings by order of study publication date.

In the first study, Harris et al. 2015 [7] examined the impact of prenatal exposure to PM2.5 on PIQ and VIQ scores of children from the Project Viva Cohort (Massachusetts, USA) using a cohort study design. Pregnant individuals were recruited during prenatal visits (median 9.9 weeks of gestation) at eight locations for Atrius Harvard Vanguard Medical Associates, a multi-subspecialty group practice in urban and suburban eastern Massachusetts. Exposure assessment was conducted by spatially joining geocoded residential addresses of the birthing parent (reported at study visits and on annual questionnaires) to PM2.5 models that used satellite aerosol optical depth measurements. The average PM2.5 concentration was 12.3 ± 2.6 µg/m3 (Table 3). IQ was assessed using the Kaufman-Brief Intelligence Test, Second Edition (KBIT-2) for 1109 children at eight years of age. Estimates on the impact of PM2.5 on child IQ were based on a “minimally adjusted model,” accounting for child sex and age at cognitive tests. The authors found that an increase of 2.5 μg/m3 in PM2.5 concentration in the year before testing was associated with a -0.4 (-1.8, 1.0) point change in PIQ and a -1.0 (-2.2, 0.2) change in VIQ. To re-express this relationship as a beta coefficient, we divided the change IQ points by 2.5. Thus, each 1 µg/m3 increase in PM2.5 concentration was associated with a change in -0.16 PIQ points and -0.40 VIQ points (Table 2).

In the second study, Porta et al. 2016 [9] examined postnatal exposure to PM2.5 and children’s FSIQ, PIQ, and VIQ scores from the Gene and Environment Prospective Study on Infancy in Italy (GASPII) using a cohort study design. Newborns born at two large obstetric hospitals in Rome, Italy to individuals over the age 18 from 2003–2004 were enrolled. Exposure assessment was conducted by spatially joining geocoded residential addresses of participants (as reported at study visits on questionnaires) to PM2.5 estimates generated from land-use regression models developed within the European Study of Cohorts for Air Pollution (ESCAPE). The average PM2.5 concentration was 19.5 ± 2.2 µg/m3 (Table 3). IQ was assessed using the Wechsler Intelligence Scale for Children, Third Edition (WISC-III) for 474 children at seven years of age. Estimates modeling the impact of PM2.5 on IQ were adjusted for child age, gender, maternal and paternal education level, socioeconomic index at birth, maternal age at delivery, maternal smoking during pregnancy, number of older siblings, and the psychologist who administered the test. Attrition bias was reduced by inversely weighting for the probability of participation at baseline and follow-up. The authors determined that for each 10 μg/m3 increase in PM2.5 exposure during pregnancy, there was a change in -1.9 (-7.9, 4.1) FSIQ points, -0.44 (-5.5, 6.4) change in VIQ points, and -4.1 (-3.4, 1.2) change in PIQ points. We divided IQ points lost by a coefficient of 10 to re-express the reported relationship in IQ points lost per 1 μg/m3 increase in PM2.5. Thus, each 1 µg/m3 increase in PM2.5 concentration was associated with a change in -0.19 FSIQ points, -0.04 points, -0.04 VIQ points, and -0.41 PIQ points (Table 2).

In the third study, Wang et al. 2017 [10] examined the impact of postnatal exposure to PM2.5 on the FSIQ and PIQ scores of children from the Risk Factors for Antisocial Behavioral (RFAB) twin study (California, USA) using a cohort study design. Families were recruited from Los Angeles and surrounding counties, generating a socio-economically diverse, multi-ethnic population. Exposure assessment was conducted by spatially joining geocoded residential addresses (identified through self-reports every 2–3 years) to PM2.5 estimates generated from a spatial–temporal model from the United States Environmental Protection Agency Technology Transfer Network. The average PM2.5 concentration was 13.7 ± 6.7 µg/m3 (Table 3). IQ was assessed using the Wechsler Abbreviated Scale of Intelligence (WASI) for 1085 children aged 9–11 and 18–20. Estimates on the impact of PM2.5 on child IQ were based on “Adjusted Model IId,” which adjusts for individual/family characteristics (e.g., age, sex, race/ethnicity, socioeconomic status, parental cognitive abilities) and neighborhood characteristics (e.g., greenspace, traffic density, and parent-reported neighborhood quality). The authors divided the sample into four quartiles according to PM2.5 exposure: in quartile 1, PM2.5 exposures ranged from 2.14–16.08 μg/m3; in quartile 2, (16.09–18.67 μg/m3), quartile 3, (8.66–21.12 μg/m3), and quartile 4, (21.14–25.36 μg/m3). For each interquartile increase in PM2.5, there was a -2.00 (-4.84, 0.24) point change in FSIQ, -3.08 (0.12—6.04) point change in PIQ, and -1.42 (-4.48, 1.64) change in VIQ. To re-express this relationship as a beta coefficient, we divided the reported IQ points lost by the interquartile range for PM2.5 (i.e., the 75th minus the 25th percentile of the distribution). From this calculation, we determined that each 1 μg/m3 increase in PM2.5 concentration is associated with a change in -0.40 FSIQ points, -0.61 PIQ points, and -0.28 VIQ points (Table 2).

In the fourth study, Seifi et al. 2021 [12] examined the impact of childhood exposure to PM2.5 on the FSIQ scores of children residing in Bushehr province, Iran, using a cross-sectional study design. Children were randomly selected from schools in three low-privileged geographic areas between 2019 and 2020. Exposure assessment was conducted using real-time measurements of PM2.5 mass concentrations from environmental dust monitors based on an optical scattering method. Indoor and outdoor exposures were simultaneously measured using direct reading equipment. The average PM2.5 concentrations were 39.0 \(\pm\) 16.9 µg/m3 for the low-exposure group, 58.0 \(\pm\) 23.9 µg/m3 for the intermediate-exposure group, and 84.2 \(\pm\) 32.2 µg/m3 for the high-exposure group (Table 3). IQ was assessed using the Raymond B. Cattle Scale I-A for 369 children at six to eight years of age. Estimates on the impact of PM2.5 on child IQ adjusted for e adjusted for age, gender, economic conditions, maternal education, and type of delivery. The authors determined that children's IQ in the area with high pollution was 7.48, lower than that in moderate pollution, and 16.628, lower than that in the region with low pollution. To re-express these data as a linear relationship, we created a scatter plot of IQ change vs. PM2.5 levels for the three respective exposure groups and fitted a regression line using Stata statistical software. Using this method, each 1 µg/m3 increase in PM2.5 concentration was associated with a -0.36 change in FSIQ points (Table 2).

In the fifth study, Ni et al. 2022 [11] examined the impact of prenatal exposure to PM2.5 on FSIQ of children from the Environmental Influences on Child Health Outcomes (ECHO) Cohort, with representation from California, New York, Minnesota, and Washington USA. The ECHO Cohort was created by pooling participant data from three individual prospective cohort studies, detailed in Table 4. Exposure assessment was conducted by spatially joining geocoded residential addresses of participants (collected at enrollment and updated at each subsequent point of contact) to PM2.5 estimates from spatial–temporal models using data reported on a two-week scale. Additional cohort-specific monitors enhanced these measurements. The average PM2.5 concentration was 8.75 \(\pm\) 2.0 µg/m3 (Table 3). IQ was assessed using the Stanford-Binet Intelligence Scales, Fifth Edition (SB-5), the Wechsler Intelligence Scale for Children, Fifth Edition (WISC-V), and the Wechsler Preschool & Primary Scale of Intelligence, Fourth Edition (WPPSI-IV) for 1311 children at four to six years of age. Estimates on the impact of PM2.5 on child IQ were based on the “primary model,” which controlled for child sex, age, study site, child race, maternal education, log-transformed region, inflation-adjusted household income, household members, an interaction between household members and income, material status, maternal age at delivery, birth order, pregnancy smoking, pregnancy alcohol consumption, maternal depression, maternal IQ, child second-hand smoke exposure, and Child Opportunity Index (domains of educational and economic opportunity) in the corresponding window of PM2.5 exposure. The authors determined that a 2 µg/m3 increase in PM2.5 during pregnancy was associated with − 0.26 (− 1.53, 1.01) points. To re-express this relationship as a beta coefficient, we divided the change IQ points by 2. Thus, each 1 µg/m3 increase in PM2.5 concentration was associated with -0.13 FSIQ points (Table 2).

In the sixth study, Sun et al. 2023 [8] assessed the impact of prenatal exposure to PM2.5 on FSIQ of children from the Shanghai-Minhang Birth Cohort (Shanghai, CN) using a cohort study design. Pregnant individuals who underwent their first prenatal examination at 12–16 weeks of gestation in the Minhang Maternal and Child Health Hospital in 2012 were enrolled. Exposure assessment was conducted by spatially joining geocoded residential addresses of the birthing parent (reported during enrollment and follow-up visits) to PM2.5 estimates developed using satellite-based monitoring and aerosol optical depth retrieval. Ground measurements from approximately 1000 monitors were used for cross-validation, which was reported to be consistent with real-time measurements (R2 = 0.78). The average PM2.5 concentration was 38.8 \(\pm\) 6.2 µg/m3 (Table 3). IQ was assessed using the Wechsler Abbreviated Scale of Intelligence (WASI) for 512 children at six years of age. Estimates on the impact of PM2.5 on child IQ adjusted for maternal factors (age, BMI before pregnancy, IQ, parity, education, intake of folic acid in early pregnancy, depression in early pregnancy), paternal factors (education), gestational weeks, and trimester-specific temperature and humidity. The authors reported a -1.34 (− 2.71, 0.04) change in FSIQ points for every 5 µg/m3 increase in PM2.5 during the first trimester of pregnancy. To re-express this relationship as a beta coefficient, we divided the change IQ points by 5. Thus, each 1 µg/m3 increase in PM2.5 concentration was associated with -0.27 FSIQ points (Table 2).

Pooling data from the studies included for final review, our meta-analyses indicated a significant negative association between PM2.5 and each IQ domain. Specifically, a 1 µg/m3 increase in ambient PM2.5 is associated with a -0.27 (-0.37, -0.17) point change in FSIQ (p < 0.001), a -0.39 (-0.65, -0.14) point change in PIQ (p = 0.003), and a -0.24 (-0.45, -0.04) point change in VIQ (p = 0.021) (Tables 5, 6 and 7). Forest plots are detailed in Tables 5, 6 and 7.

Table 5 PM2.5 and Full-Scale IQ
Full size table
Table 6 PM2.5 and Performance IQ
Full size table
Table 7 PM2.5 and Verbal IQ
Full size table

Overall, we determined that there was a high degree of heterogeneity among our studies. Through our quantitative analysis, the I2 statistic was high (> 88%), and Cochran’s Q test of heterogeneity was significant (p < 0.05) for all meta-analyses (Supplemental Table S3). Qualitatively, we observed a moderate degree of overlap in the confidence intervals and a narrow spread of the effect sizes around the pooled estimates in our FSIQ model, suggesting moderate heterogeneity (Table 5), while a slight overlap of the confidence intervals for our PIQ and VIQ models (Tables 6 and 7), suggesting high heterogeneity.

Discussion

To our knowledge, this is the first study to quantify the correlation between PM2.5 air pollution exposure and decreased cognitive function in children based on a systematic review of the world’s literature. All six of the studies included in our analysis consistently found that prenatal and childhood exposures to PM2.5 pollution are associated with declines in verbal and nonverbal cognitive abilities, as measured by IQ loss. Our findings, which include data from geographically, socially, economically, and culturally diverse populations, indicate that PIQ is the component of cognitive function most profoundly affected by PM2.5. PIQ is more severely affected than either VIQ or FSIQ, suggesting that it is an especially sensitive indicator of brain injury caused by adverse environmental exposures.

Children’s communication and language skills—the “crystallized” cognitive abilities reflected by VIQ —may be more resilient to adverse environmental exposures such as PM2.5 pollution due to the ubiquitous nature and importance of daily communication and social interaction, as opposed to PIQ, which reflects non-verbal, more “fluid” cognitive abilities such as the ability to reason and to solve novel problems and is contingent on adequate sensory input and rich play environments. Differences in critical periods of brain development may also help to elucidate the differential impact of PM2.5 on PIQ. Thus, experience-dependent synapse formation peaks around two months of age for the visual cortex [17], a brain structure supporting PIQ, whereas the receptive language/speech production area (critical for VIQ) peaks later, around eight months of age (Supplemental Figure S1). It is possible that children are more susceptible to PM2.5 exposure in the early postnatal period when the development of brain regions associated with PIQ is most rapid.

Loss of non-verbal cognitive function, as measured by PIQ loss, has significant implications for children’s health, development, and future life accomplishments. Long-term follow-up studies of children who suffered IQ loss from early-life exposures to neurotoxicants other than PM2.5 pollution have found linkages to a range of developmental and behavioral deficits. These include a shortened attention span, compromised reading and math abilities, reduced control over impulsive and aggressive behaviors, and increased rates of school failure [18]. Longer-term ramifications include increased risks of juvenile delinquency, criminal behavior, and incarceration [19,20,21]. Additional neurobehavioral consequences of early-life exposures to PM2.5 pollution include increased incidence of behavioral abnormalities, such as autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) [14, 15, 22,23,24,25,26,27,28].

Population-wide reduction in mean cognitive capacity by as little as 5 IQ points results in a more than 50% decrease in the number of children with superior intelligence (IQ above 130) and a corresponding increase in the number of children with IQ scores below 70. Such a loss of cognitive capacity in a population represents a massive erosion of human capital, reduces a society’s future leadership potential, and threatens national survival [29]. At the same time, the significant increase in the number of children with reduced cognitive capacities imposes substantial economic and social burdens on societies by reducing the lifelong productivity of future generations and increasing the need for remedial education and custodial care. For these reasons, and given the widespread nature of children’s exposure to excessive levels of PM2.5 air pollution, our finding that early-life PM2.5 exposure is associated with concentration-related loss of cognitive function in children has significant economic and policy implications [30].

The high degree of heterogeneity we observed (both quantitatively and qualitatively) across the studies we examined was expected as the studies included for final review used different methodologies to model air pollution exposure (Table 3) and examined neurobehavioral outcomes of children at various ages in diverse geographical locations (Table 4). Further, differential access to opportunities for educational and social engagement may have mediated cognitive outcomes, obscuring the direct effects of PM2.5. For these reasons, our findings should be interpreted cautiously and reexamined as additional data become available.

Our findings are consistent with an expanding body of literature on the harmful impacts of pre- and postnatal exposures to PM2.5 air pollution on children’s neurobehavioral development [14, 15]. Because many millions of children across the globe are exposed to PM2.5 pollution, the aggregate losses in cognitive function resulting from these exposures have the potential to be as large or more prominent than those caused by other exposures to other widespread neurotoxicants. A study of Full-Scale IQ (FSIQ) losses in American children less than 5 years of age reported that lead was responsible for an aggregate loss of 23,000,000 FSIQ points, methylmercury for a loss of 285,000 FSIQ points, and organophosphate pesticides for a loss of 17,000,000 FSIQ points [31].

Our study has several limitations. First, our derivation of a concentration–response function relating PM2.5 concentration to IQ loss in children is based on only six studies. This scarcity of available data underscores the early stage of research on the neurobehavioral consequences of early-life exposure to air pollution. Moreover, only one author screened all studies included in this analysis, which may introduce bias in the selection process. To mitigate this limitation, future research should consider employing multiple reviewers to ensure a more rigorous and unbiased screening.

Another limitation is that we were not able to assess the potentially synergistic impacts of air pollution and other neurotoxicants in children’s development. Although all of the studies in our analysis included comprehensive evaluations of potential confounding from demographic and socioeconomic variables, none formally evaluated the possibility of other adverse environmental exposures exacerbating the adverse impacts of increased PM2.5 on children’s IQ scores. Further, our models do not account for potential non-linear trends, an essential consideration for geographic locations with pollution exposures outside our study range. Mounting evidence from other toxicological exposures (e.g., endocrine-disrupting chemicals from plastic additives) suggests that non-monotonic dose–response relationships may be common. We lack the data to determine whether the concentration–response functions we derived for IQ loss at PM2.5 concentrations are linear, supralinear, attenuated, or flat at higher PM2.5 concentrations. Estimating IQ losses in children with higher exposures will necessitate either direct study or extrapolations that incorporate assumptions about the shape of the concentration–response function at those exposures.

Yet another limitation of our study is that it could not account for differences in the chemical constituents of PM2.5 pollution that may occur in different places around the world. For example, biomass burning in wildfires produces PM2.5 with higher concentrations of PAHs as opposed to more complete combustion processes such as vehicle emissions. Such variability could modify the concentration–response function. However, in the absence of studies that parse out the relative contributions of different constituents of PM2.5, the assumption of equitoxicity has been widely used in epidemiologic studies of the health effects of air pollution, including the Global Burden of Disease (GBD) study [32,33,34].

We note that IQ loss in children caused by harmful environmental exposures such as PM2.5 pollution is not included in the Global Burden of Disease (GBD) calculations unless exposures are so severe that they result in IQ scores below 70 (the criterion for diagnosing a child as having an “intellectual disability”) [32, 34]. The consequence of this unfortunate omission is that most cognitive impairments associated with toxic environmental exposures are not counted in the GBD study. This impedes the estimation of the actual population health impacts and human capital losses caused by air pollution and other neurotoxicants. For the many countries that utilize GBD findings to guide priority setting and resource allocation in public health, this omission underestimates the adverse impacts of air pollution on children’s cognitive function and results in opportunities for prevention being lost. We encourage future research initiatives undertaken in partnership with the GBD study collaborators to address this information gap and improve the ability of the GBD study to guide preventive policy in pediatrics [32, 33].

Future epidemiological studies should seek to diversify their patient populations and geographical locations, as we were unable to identify studies matching our search criterion that represented populations in South America, Africa, or Oceania. We note that each study included for the final review used proximal estimates to determine children’s PM2.5 exposure (e.g., geographical information systems, land-use regression models, satellite imagery). While effective and efficient for population-level cohorts, continued research should seek to quantify exposure at the individual level (e.g., biomarkers and wearable devices) to increase sensitivity and precision when generating exposure–response functions. In the setting of unprecedented planetary changes, we stress the need to evaluate the impact of PM2.5 on children’s neurodevelopment with other adverse co-exposures, such as extreme heat, stress from displacement/migration, altered nutrient availability/quality, and increased vector-borne disease.

Lastly, since children’s brains are experiencing rapid periods of tightly organized growth and development, it is necessary to track IQ over time to characterize neurodevelopmental trajectories. For example, the difference between IQ scores among children of high SES versus those of low SES was tripled at age 16, when compared to their difference at age two in a British cohort [35] – demonstrating that adverse events/exposures during the early years of childhood exacerbate existing health inequities, but also serve as opportunities for resiliency in children’s neurodevelopment.

Conclusion

Our combined effect estimates, based on data generated from six epidemiological studies representing children from three continents (North America, Europe, and Asia), use standard coefficients, supporting a negative impact of PM2.5 exposure on children’s neurocognitive development, as measured by IQ scores. Though relatively small, the estimated effect sizes for FSIQ, PIQ, and VIQ are of significant public health importance, considering the lifelong effects of adverse neurodevelopmental on children’s health, well-being, and human capital and the wide extent of children’s exposure to levels of PM2.5 air pollution that exceed WHO guidelines. Substantial gains in economic productivity have occurred in countries that reduced airborne lead pollution by removing lead from gasoline [30]. Similar benefits may be expected to result from sustained reductions in PM2.5 pollution. Developing an exposure-response function linking PM2.5 concentration to IQ loss provides the means for quantifying these benefits and translating them into public policy.

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

IQ:

Intelligence Quotient

FSIQ:

Full-Scale Intelligence Quotient

PIQ:

Performance Intelligence Quotient

VIQ:

Verbal Intelligence Quotient

PM:

Particulate Matter

PM10 :

Particulate Matter Air Pollution 10 Micrometers in Diameter

PM2.5 :

Particulate Matter Air Pollution 2.5 Micrometers in Diameter

SDI:

Socio-demographic Index

WHO:

World Health Organization

LMIC:

Low or Middle-Income Country

References

  1. Xing YF, Xu YH, Shi MH, Lian YX. The impact of PM2.5 on the human respiratory system. J Thorac Dis. 2016;8(1). Available from: https://jtd.amegroups.org/article/view/6353. Cited 2024 Mar 15.

  2. Yacobi NR, Fazllolahi F, Kim YH, Sipos A, Borok Z, Kim KJ, et al. Nanomaterial interactions with and trafficking across the lung alveolar epithelial barrier: implications for health effects of air pollution particles. Air Qual Atmosphere Health. 2011;4(1):65–78.

    Article  Google Scholar 

  3. Geiser M, Rothen-Rutishauser B, Kapp N, Schürch S, Kreyling W, Schulz H, et al. Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and in cultured cells. Environ Health Perspect. 2005;113(11):1555–60.

    Article  Google Scholar 

  4. Fuller R, Landrigan PJ, Balakrishnan K, Bathan G, Bose-O’Reilly S, Brauer M, et al. Pollution and health: a progress update. Lancet Planet Health. 2022;6(6):e535–47.

    Article  Google Scholar 

  5. You R, Ho YS, Chang RCC. The pathogenic effects of particulate matter on neurodegeneration: a review. J Biomed Sci. 2022;29(1):15.

    Article  Google Scholar 

  6. Landrigan PJ, Fisher S, Kenny ME, Gedeon B, Bryan L, Mu J, et al. A replicable strategy for mapping air pollution’s community-level health impacts and catalyzing prevention. Environ Health. 2022;21(1):70.

    Article  CAS  Google Scholar 

  7. Harris MH, Gold DR, Rifas-Shiman SL, Melly SJ, Zanobetti A, Coull BA, et al. Prenatal and childhood traffic-related pollution exposure and childhood cognition in the project viva cohort (Massachusetts, USA). Environ Health Perspect. 2015;123(10):1072–8.

    Article  CAS  Google Scholar 

  8. Sun X, Liu C, Ji H, Li W, Miao M, Yuan W, et al. Prenatal exposure to ambient PM2.5 and its chemical constituents and child intelligence quotient at 6 years of age. Ecotoxicol Environ Saf. 2023;255:114813.

    Article  CAS  Google Scholar 

  9. Porta D, Narduzzi S, Badaloni C, Bucci S, Cesaroni G, Colelli V, et al. Air pollution and cognitive development at age 7 in a prospective italian birth cohort. Epidemiol Camb Mass. 2016;27(2):228–36.

    Google Scholar 

  10. Wang P, Tuvblad C, Younan D, Franklin M, Lurmann F, Wu J, et al. Socioeconomic disparities and sexual dimorphism in neurotoxic effects of ambient fine particles on youth IQ: a longitudinal analysis. PLoS ONE. 2017;12(12):e0188731.

    Article  Google Scholar 

  11. Ni Y, Loftus CT, Szpiro AA, Young MT, Hazlehurst MF, Murphy LE, et al. Associations of pre- and postnatal air pollution exposures with child behavioral problems and cognitive performance: a U.S. multi-cohort study. Environ Health Perspect. 2022;130(6):067008.

    Article  CAS  Google Scholar 

  12. Seifi M, Yunesian M, Naddafi K, Nabizadeh R, Dobaradaran S, Ziyarati MT, et al. Exposure to ambient air pollution and socio-economic status on intelligence quotient among schoolchildren in a developing country. Environ Sci Pollut Res. 2022;29(2):2024–34.

    Article  Google Scholar 

  13. UNEP. 2018. Air pollution linked to “huge” reduction in intelligence. Available from: http://www.unep.org/news-and-stories/story/air-pollution-linked-huge-reduction-intelligence. Cited 2024 Mar 15.

  14. Perera F, Ashrafi A, Kinney P, Mills D. Towards a fuller assessment of benefits to children’s health of reducing air pollution and mitigating climate change due to fossil fuel combustion. Environ Res. 2019;1(172):55–72.

    Article  Google Scholar 

  15. Volk HE, Perera F, Braun JM, Kingsley SL, Gray K, Buckley J, et al. Prenatal air pollution exposure and neurodevelopment: a review and blueprint for a harmonized approach within ECHO. Environ Res. 2021;1(196):110320.

    Article  Google Scholar 

  16. Burnett RT, Pope CA, Ezzati M, Olives C, Lim SS, Mehta S, et al. An integrated risk function for estimating the global burden of disease attributable to ambient fine particulate matter exposure. Environ Health Perspect. 2014;122(4):397–403.

    Article  Google Scholar 

  17. Nelsen C. Chapter 8: The Developing Brain “From Neurons to Neighborhoods: The Science of Early Childhood Development” at NAP.edu. Available from: https://nap.nationalacademies.org/read/9824/chapter/12. Cited 2024 Sep 22.

  18. Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The long-term effects of exposure to low doses of lead in childhood. N Engl J Med. 1990;322(2):83–8.

    Article  CAS  Google Scholar 

  19. Needleman HL, McFarland C, Ness RB, Fienberg SE, Tobin MJ. Bone lead levels in adjudicated delinquents: a case control study. Neurotoxicol Teratol. 2002;24(6):711–7.

    Article  CAS  Google Scholar 

  20. Nevin R. Understanding international crime trends: the legacy of preschool lead exposure. Environ Res. 2007;104(3):315–36.

    Article  CAS  Google Scholar 

  21. Coulton C, Richter FGC, Cho Y, Park J, Jeon J, Fischer RL. Making the case for lead safe housing: Downstream effects of lead exposure on outcomes for children and youth. Health Place. 2023;84:103118.

    Article  Google Scholar 

  22. Sram RJ, Veleminsky M, Veleminsky M, Stejskalová J. The impact of air pollution to central nervous system in children and adults. Neuro Endocrinol Lett. 2017;38(6):389–96.

    CAS  Google Scholar 

  23. Calderón-Garcidueñas L, Engle R, Mora-Tiscareño A, Styner M, Gómez-Garza G, Zhu H, et al. Exposure to severe urban air pollution influences cognitive outcomes, brain volume and systemic inflammation in clinically healthy children. Brain Cogn. 2011;77(3):345–55.

    Article  Google Scholar 

  24. Sunyer J, Esnaola M, Alvarez-Pedrerol M, Forns J, Rivas I, López-Vicente M, et al. Association between traffic-related air pollution in schools and cognitive development in primary school children: a prospective cohort study. PLOS Med. 2015;12(3):e1001792.

    Article  Google Scholar 

  25. Thygesen M, Holst GJ, Hansen B, Geels C, Kalkbrenner A, Schendel D, et al. Exposure to air pollution in early childhood and the association with attention-deficit hyperactivity disorder. Environ Res. 2020;1(183):108930.

    Article  Google Scholar 

  26. Volk HE, Lurmann F, Penfold B, Hertz-Picciotto I, McConnell R. Traffic-related air pollution, particulate matter, and autism. JAMA Psychiat. 2013;70(1):71–7.

    Article  Google Scholar 

  27. Zhang X, Chen X, Zhang X. The impact of exposure to air pollution on cognitive performance. Proc Natl Acad Sci. 2018;115(37):9193–7.

    Article  CAS  Google Scholar 

  28. Chun H, Leung C, Wen SW, McDonald J, Shin HH. Maternal exposure to air pollution and risk of autism in children: a systematic review and meta-analysis. Environ Pollut. 2020;1(256):113307.

    Article  Google Scholar 

  29. Erickson MD. Our stolen future – are we threatening our fertility, intelligence, and survival?: A scientific detective story Theo Colborn, Dianne Dumanoski, and John Peterson Myers, Dutton, New York, NY,(1996), 306 Pages, U.S. List Price: $24.95. Environ Prog. 1997;16(1):S8–9 [ISBN No. 0–525–93982–2].

    Article  CAS  Google Scholar 

  30. Grosse SD, Matte TD, Schwartz J, Jackson RJ. Economic gains resulting from the reduction in children’s exposure to lead in the United States. Environ Health Perspect. 2002;110(6):563–9.

    Article  Google Scholar 

  31. Bellinger DC. Comparing the population neurodevelopmental burdens associated with children’s exposures to environmental chemicals and other risk factors. Neurotoxicology. 2012;33(4):641–3.

    Article  CAS  Google Scholar 

  32. Bellinger DC. Applying methods of the global burden of diseases, injuries, and risk factors study to developmental neurotoxicants: a commentary. Environ Health. 2018;17(1):53.

    Article  Google Scholar 

  33. Hu H, Landrigan PJ, Fuller R, Lim SS, Murray CJL. New initiative aims at expanding global burden of disease estimates for pollution and climate. Lancet Planet Health. 2018;2(10):e415–6.

    Article  Google Scholar 

  34. Institute for Health Metrics and Evaluation. GBD Results. Available from: https://vizhub.healthdata.org/gbd-results. Cited 2024 Mar 15.

  35. Von Stumm S, Plomin R. Socioeconomic status and the growth of intelligence from infancy through adolescence. Intelligence. 2015;48:30–6. https://doiorg.publicaciones.saludcastillayleon.es/10.1016/j.intell.2014.10.002.

    Article  Google Scholar 

Download references

Acknowledgements

The authors gratefully acknowledge the support of The Heinz Endowments.

Author information

Authors and Affiliations

  1. Boston College Global Observatory on Planetary Health, Boston, MA, USA

    Naomi C. Alter, Ella M. Whitman & Philip J. Landrigan

  2. Department of Neurology, Boston Children’s Hospital and Harvard Medical School, Boston, MA, USA

    David C. Bellinger

  3. Centre Scientifique de Monaco, Monaco, MC, Monaco

    Philip J. Landrigan

Authors
  1. Naomi C. Alter
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Ella M. Whitman
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. David C. Bellinger
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Philip J. Landrigan
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

N.C.A and E.M.W. are co-first authors and were responsible for writing the manuscript. N.C.A. performed the systematic literature search. E.M.W. performed all data analyses. D.C.B. and P.J.L. offered extensive guidance and editorial support throughout the writing process. E.M.W is the corresponding author. All authors contributed to the manuscript, approved the final version, and declare no competing interests.

Corresponding author

Correspondence to Ella M. Whitman.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

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

Supplementary Information

Rights and permissions

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

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Alter, N.C., Whitman, E.M., Bellinger, D.C. et al. Quantifying the association between PM2.5 air pollution and IQ loss in children: a systematic review and meta-analysis. Environ Health 23, 101 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12940-024-01122-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12940-024-01122-x

Keywords

Environmental Health

ISSN: 1476-069X