Systematic Review of Prenatal Exposure to PM_2.5 and Its Chemical Components and Their Effects on Neurodevelopmental Outcomes in Neonates

Abstract

Particulate matter with a diameter less than 2.5 µm (PM_2.5) and its chemical constituents—including ammonium (NH₄⁺), sulfate (SO₄²⁻), nitrate (NO₃⁻), organic carbon (OC), soil dust, and black carbon (BC)—have been increasingly recognized for their potential impact on fetal neurodevelopment. This systematic review aimed to synthesize current evidence on the relationship between prenatal exposure to PM_2.5 and its chemical components and neurodevelopmental outcomes in neonates, focusing on diagnoses such as autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD). Following PRISMA 2020 guidelines, a comprehensive literature search was conducted on PubMed and Embase databases from April to July 2025. Twenty-five studies meeting inclusion criteria were analyzed, of which sixteen addressed PM_2.5 exposure generally, and nine assessed specific chemical constituents. The findings indicate that increased exposure to PM_2.5, particularly during the third trimester, is associated with a higher risk of ASD. Additionally, prenatal exposure may adversely affect early neurodevelopmental domains including motor skills, problem-solving, and social interactions. Certain PM_2.5 components, notably sulfate ions (SO₄²⁻), were identified as important contributors to neurological health outcomes. These results underscore the importance of reducing prenatal exposure to PM_2.5 and its harmful constituents to protect neurodevelopment.

1. Introduction

Air pollution is a significant global issue, affecting over 90% of the world’s population who breathe air exceeding the guidelines recommended by the World Health Organization (WHO), particularly in low- and middle-income countries [1]. Among environmental pollutants, fine particulate matter—also known as particulate matter (PM)—is of particular concern. PM_2.5, with aerodynamic diameters less than 2.5 μm, is especially hazardous due to its small size and light weight, allowing it to remain suspended in the atmosphere for extended periods and be transported over long distances by atmospheric circulation, leading to widespread environmental contamination [2].

This particulate matter is a complex mixture of primary and secondary components with diverse physical and chemical properties, originating from both natural and anthropogenic sources [3]. The toxicity of PM_2.5 depends greatly on its chemical composition rather than solely on its total mass concentration. Globally, the main constituents include sulfates (SO₄²⁻), nitrates (NO₃⁻), ammonium (NH₄⁺), organic matter (OM), and carbonaceous components, although their proportions vary by location [4,5,6]. In addition to regional variability, the seasonal fluctuations in PM_2.5 composition due to heating, agricultural burning, and meteorological conditions can further modify its toxicity and health impact [7]. Moreover, ultrafine particles and metal-rich fractions can exert additional oxidative stress, representing an underexplored dimension of toxicity [8].

Health impacts also differ depending on PM_2.5 composition. Worldwide, PM_2.5 concentrations have gradually declined, largely due to air quality guidelines that set an annual exposure limit of 5 μg/m³ and a daily limit of 15 μg/m³ [9]. Nevertheless, some countries still experience elevated levels despite reductions; for instance, China’s annual concentration remains approximately eight times above recommended values [1,9].

Pregnant women are especially vulnerable to air pollutants such as PM_2.5. Multiple studies have demonstrated that exposure to air pollution during pregnancy can induce oxidative stress and inflammatory responses in the placenta [2,10,11], and these contaminants may even cross the placental barrier, reaching the fetus [5]. Emerging evidence also suggests that prenatal PM_2.5 exposure can disrupt placental vascular development and epigenetic programming, potentially altering fetal growth trajectories [12]. Additionally, inflammatory processes triggered in the maternal organism could indirectly affect neurodevelopment through cytokine-mediated mechanisms [13].

Fetal life and early childhood are considered periods of heightened vulnerability due to extensive cellular differentiation and growth critical for brain maturation and neural network development [14]. Epidemiological research has linked prenatal exposure to particulate matter with cognitive and motor neurodevelopmental issues in children, such as memory difficulties, attention deficits, executive function impairments, language deficits, and behavioral problems [5,15,16]. Recent cohort studies have further confirmed that even low-to-moderate levels of PM_2.5 exposure are associated with measurable decrements in IQ and increased risk of developmental delays [17]. Moreover, some studies report that prenatal exposure to air pollution can have sex-specific effects on neurodevelopmental outcomes, with boys often displaying higher vulnerability [18]. Even subtle air pollution-induced disruptions in neurological development during this sensitive period may increase lifelong susceptibility to various neurological disorders, posing substantial medical and economic burdens on families and society [19].

Among these disorders are autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD). ASD comprises a heterogeneous group of neurodevelopmental conditions—including autistic disorder, Asperger syndrome, childhood disintegrative disorder, and pervasive developmental disorder not otherwise specified (PDD-NOS)—characterized by impaired communication and social interaction, alongside repetitive and restrictive behaviors [20]. Conversely, ADHD is a prevalent neurodevelopmental disorder beginning in childhood, often persisting into adolescence and adulthood. Its etiology is largely unknown but involves genetic and environmental risk factors that may interact to increase susceptibility to neurodevelopmental problems [21]. Globally, ASD affects approximately 1 in 100 children, while ADHD has a prevalence estimated at 5–7%, with substantial heterogeneity across regions [22].

Although numerous studies support an association between air pollution and neurodevelopmental issues, evidence specifically linking chemical components of particulate matter is limited. The various components of air pollution may act synergistically, underscoring the need for further research on this phenomenon [14].

In response to this, a systematic review was conducted to investigate the relationship between prenatal exposure to chemical compounds associated with PM_2.5 and neurological alterations in neonates. This review qualitatively examined the scientific evidence on prenatal exposure to particulate matter, focusing on PM_2.5 and its chemical constituents that may influence neonatal neurological development during gestation, potentially contributing to motor and cognitive disorders such as ASD and ADHD.

2. Materials and Methods

2.1. Study Design

This study was conducted through an exhaustive analysis of the available literature, following the guidelines of the PRISMA 2020 statement (http://www.prisma-statement.org/, accessed on 23 August 2025) [23] for systematic reviews. Articles were examined that investigated the association between different concentrations of particulate matter to which pregnant women were exposed and the effects observed on neonatal neurological development during gestation.

2.2. Search Strategy and Selection Criteria

A targeted search was performed in the PubMed and Embase databases (https://pubmed.ncbi.nlm.nih.gov/advanced/ and https://www.embase.com/, accessed on 23 August 2025) using the following English search query:

(“Air pollution” OR “Particulate matter” OR “PM_2.5” OR “Particulate matter 2.5” OR “Fine particulate matter”) AND (“Prenatal” OR “Pregnancy woman” OR “Neonates” OR “Pregnant woman”) AND (“Neurodevelopment” OR “Mental disease” OR “ASD” OR “Autism spectrum disorder” OR “ADHD” OR “Attention deficit hyperactivity”).

Filters were applied to restrict results to publications dated between 2000 and 2025 and to define inclusion criteria regarding study design (cohort, prospective cohort, and case–control studies), language (English), particulate matter size (PM_2.5), exposure period (pregnancy), and the target population.

The Population/Exposure/Comparator/Outcome (PECO) framework was adopted [24]:

• Population (P): Pregnant women, mother-child pairs, and neonates.
• Exposure (E): Environmental pollution during pregnancy, specifically particles <2.5 µm and their chemical constituents.
• Comparator (C): Defined by study design:

  ○

  Cohort/longitudinal studies: reference exposure strata (e.g., lowest category or predefined reference level), enabling dose–response comparisons across exposure gradients.

  ○

  Case–control/cross-sectional studies: control groups (pregnant women–child pairs without the adverse neurodevelopmental outcome at assessment), as defined in each study.

• Outcome (O): Studies assessing neurodevelopment and disorders such as autism spectrum disorder (ASD) or attention deficit hyperactivity disorder (ADHD).
• Study Design (S): Cohort, prospective cohort, cross-sectional, and case–control studies.

Inclusion and exclusion criteria were established following PRISMA 2020 [23] and PECO guidance [24]:

Inclusion Criteria:

• Articles published in English;
• Publication date between 2000 and 2025;
• Cohort, prospective cohort, and case–control studies;
• Studies conducted in humans;
• Studies addressing prenatal, perinatal, or neonatal exposure;
• Exposure to PM_2.5 and related chemical components;
• Outcomes focusing on neurodevelopmental effects.

Exclusion Criteria:

• Articles published in languages other than English;
• Narrative reviews or other non-systematic searches;
• Animal studies;
• Studies without prenatal exposure assessment;
• Studies not including PM_2.5 or its components;
• Studies on exposure during youth or adulthood;
• Outcomes unrelated to neurodevelopment.

These criteria were applied to ensure consistency in selecting studies relevant to the objectives and scope of this review.

2.3. Risk of Bias Assessment

The quality of each included article was assessed using a risk of bias instrument for non-randomized studies of exposures [25]. This tool was considered seven key domains:

• Confounding: Risk that differences in baseline characteristics between groups could affect the outcomes independently of the exposure.
• Selection: Bias arising from how participants are selected into the study or into analysis groups.
• Measurement of exposure: Risk that the exposure was inaccurately measured or misclassified.
• Departures from exposure: Bias due to participants not adhering to the assigned exposure or intervention.
• Missing data: Risk that incomplete outcome or exposure data could distort the results.
• Measurement of outcomes: Risk that outcomes were measured inaccurately or differently between groups.
• Reported results: Bias from selective reporting of outcomes, analyses, or subgroups.

This approach provided a detailed understanding of the validity and reliability of the selected studies, thereby strengthening the rigor of the review.

3. Results

The selection of articles was conducted in accordance with the PRISMA 2020 guidelines [23], which were refined during the process to determine which studies would be included or excluded.

From the search and organization of records in Zotero, a total of 1093 publications were identified. To facilitate data management, 442 duplicate records were removed. Subsequently, 605 articles were excluded after title and abstract screening because they did not meet the inclusion criteria (e.g., not related to PM_2.5, prenatal exposure, or neurodevelopmental outcomes).

From the initial search, 46 articles were selected for a more comprehensive assessment, which involved full-text reading of each study. An additional 21 studies were then excluded because they did not provide information on PM_2.5 or its chemical components in relation to neurodevelopment. After careful evaluation, 25 articles were ultimately included in the review.

The complete selection process is summarized in the flow diagram presented in Figure 1.

3.1. General Characteristics of the Studies

The selected studies were analyzed by assessing various variables: geographic location, study design, research period, sample size, exposure intervals, outcomes, pollutants under investigation, conclusions reached, as well as the adjustment factors considered.

Table 1 presents a summary of the main characteristics of the 25 included studies. Geographically, the studies were conducted across seven countries, with the USA and China being the most represented. Other studies were conducted in Spain, Denmark, Taiwan, Canada, and France.

Analyzing the type of studies included in the review, they were distributed as follows: 15 prospective cohort studies [5,6,21,26,27,28,29,30,31,32,33,34,35,36,37], 4 retrospective cohort studies [38,39,40,41], 5 case–control studies [42,43,44,45,46], and 1 matched case–control study [47].

Table 1. Characteristics of the Studies.

First Author, Year	Country	Type of Study	Study Period	Sample Size (Cases/Controls)	Pollutant(s)	Exposure Periods	Outcome
Becerra et al., 2013 [42]	USA	Case–control	1995–2006	7594/75,635	PM2.5	Three specific periods throughout pregnancy (first, second, and third trimester)	ASD
Volk et al., 2013 [45]	USA	Case–control	1997–2006	245/279	PM2.5	During the pregnancy trimesters up to the first year of life	ASD
Lertxundi et al., 2015 [27]	Spain	Cohort	2006–2008	438	PM2.5	During pregnancy	Cognitive and psychomotor neurodevelopment
Raz et al., 2015 [43]	USA	Case–control	1990–2002	245/1767	PM2.5	During the 9 months of pregnancy and the 9 months after birth	ASD
Talbott et al., 2015 [44]	USA	Case–control	2005–2009	217/226; 211/219	PM2.5	Prenatal and postnatal period	ASD
Thygesen et al., 2020 [21]	Denmark	Prospective cohort	1992–2007; 1997–2013	19,045/809,654	PM2.5	Prenatal and postnatal period	ADHD
P. Wang et al., 2021 [34]	China	Prospective cohort	2016–2018	4009	PM2.5	During pregnancy	Neurodevelopment
Chang et al., 2022 [26]	Taiwan	Cohort	2004–2005	9294/425,736	PM2.5	Prenatal and postnatal period	ADHD
Lei et al., 2022 [5]	China	Cohort	2013–2016	2435	PM2.5 and its primary components (CN, mineral dust, sea salts) and secondary components (NH4+, NO3−, SO42−, OM)	During the first, second, and third trimesters of pregnancy	Cognitive and psychomotor neurodevelopment
Su et al., 2022 [28]	China	Cohort	2016–2020	15,778 mother–child pairs	PM2.5	During the first, second, and third trimesters of pregnancy	Neurodevelopment
H. Wang et al., 2022 [29]	China	Cohort	2013–2014	1331 mother–child pairs	PM2.5	During pregnancy and the first two years of life	Neurodevelopment
Xu et al., 2022 [35]	China	Prospective cohort	2017–2019	1531	PM2.5 and its components (CN, mineral dust, NH4+, NO3−, SO42−, OM)	During pregnancy	Neurodevelopment
Rahman et al., 2023 [38]	USA	Retrospective cohort	2001–2014	4559/318,750 mother–child pairs	PM2.5 and its components (CE, NH4+, NO3−, SO42−, OM)	During pregnancy and the first year of life	ASD
Rahman, et al., 2023 [39]	USA	Retrospective cohort	2001–2014	4559/318,750 mother–child pairs	PM2.5 and its components (CE, CO, Fe, and Mn)	During pregnancy up to 5 years of life	ASD
X. Sun et al., 2023 [6]	China	Cohort	2012–2018	512 mother–child pairs	PM2.5 and its components (CN, soil, NH4+, NO3−, SO42−, CO)	During pregnancy	Neurodevelopment
Luglio et al., 2024 [31]	USA (California)	Retrospective cohort	2001–2014	318,750 mother-child pairs	Source-specific PM2.5 (9 sources)	Entire pregnancy (assigned by residence)	ASD
Murphy et al., 2024 [47]	Canada	Matched case–control	2012–2016	1589 ASD cases/7563 controls	PM2.5	Pre-conception, 1st, 2nd, 3rd trimesters	ASD
Yu et al., 2023 [41]	USA (California)	Population-based cohort	2001–2014	318,751 mother-child pairs (4559 ASD cases)	PM2.5 and its components (BC, OM, NO3−, and SO42−)	Pregnancy	ASD
Lin et al., 2023 [33]	Taiwan	Birth cohort	2004–2011	168,062 live term births (666 ASD cases)	PM2.5 heavy metals (As, Cd, Hg, Pb)	Prenatal and postnatal (up to 9 months after birth)	ASD
Fu et al., 2024 [30]	China	Birth cohort	2015–2019	4405 infants/mothers	PM2.5	Prenatal (whole pregnancy, first and second trimester)	Growth retardation (body length), neurodevelopmental retardation
Perera et al., 2024 [36]	USA (NYC)	Prospective cohort	1998–2006	470 mother-child pairs	PM2.5	Prenatal (weekly, trimester, and whole pregnancy)	Neurodevelopmental outcomes (MDI at ages 1 and 3; no significant associations for PDI)
Ghassabian et al., 2025 [37]	USA (44 cohorts)	Cohort meta-analysis	2000–2016	8035 mother-child pairs	PM2.5	Prenatal (pregnancy)	ASD and autism-related traits (SRS scores)
Whitworth et al., 2024 [32]	Spain	Cohort study	2003–2008	1303 mother-child pairs	PM2.5	Prenatal (gestational weeks 1–17), Postnatal (0.5–1.2 years)	Cognitive and motor function (McCarthy Scales)
Mortamais et al., 2025 [46]	France	Case–control	2008–2013	125 ASD cases/500 controls	PM2.5	Prenatal	ASD
O’Sharkey et al., 2025 [40]	USA (California)	Population-based cohort	1990–2018	13,591,003 children; 138,460 ASD cases	PM2.5, NO2, O3, benzene, 1,3-butadiene, chromium, lead, nickel, zinc	Prenatal	ASD

Table 1. Characteristics of the Studies.

First Author, Year	Country	Type of Study	Study Period	Sample Size (Cases/Controls)	Pollutant(s)	Exposure Periods	Outcome
Becerra et al., 2013 [42]	USA	Case–control	1995–2006	7594/75,635	PM2.5	Three specific periods throughout pregnancy (first, second, and third trimester)	ASD
Volk et al., 2013 [45]	USA	Case–control	1997–2006	245/279	PM2.5	During the pregnancy trimesters up to the first year of life	ASD
Lertxundi et al., 2015 [27]	Spain	Cohort	2006–2008	438	PM2.5	During pregnancy	Cognitive and psychomotor neurodevelopment
Raz et al., 2015 [43]	USA	Case–control	1990–2002	245/1767	PM2.5	During the 9 months of pregnancy and the 9 months after birth	ASD
Talbott et al., 2015 [44]	USA	Case–control	2005–2009	217/226; 211/219	PM2.5	Prenatal and postnatal period	ASD
Thygesen et al., 2020 [21]	Denmark	Prospective cohort	1992–2007; 1997–2013	19,045/809,654	PM2.5	Prenatal and postnatal period	ADHD
P. Wang et al., 2021 [34]	China	Prospective cohort	2016–2018	4009	PM2.5	During pregnancy	Neurodevelopment
Chang et al., 2022 [26]	Taiwan	Cohort	2004–2005	9294/425,736	PM2.5	Prenatal and postnatal period	ADHD
Lei et al., 2022 [5]	China	Cohort	2013–2016	2435	PM2.5 and its primary components (CN, mineral dust, sea salts) and secondary components (NH4+, NO3−, SO42−, OM)	During the first, second, and third trimesters of pregnancy	Cognitive and psychomotor neurodevelopment
Su et al., 2022 [28]	China	Cohort	2016–2020	15,778 mother–child pairs	PM2.5	During the first, second, and third trimesters of pregnancy	Neurodevelopment
H. Wang et al., 2022 [29]	China	Cohort	2013–2014	1331 mother–child pairs	PM2.5	During pregnancy and the first two years of life	Neurodevelopment
Xu et al., 2022 [35]	China	Prospective cohort	2017–2019	1531	PM2.5 and its components (CN, mineral dust, NH4+, NO3−, SO42−, OM)	During pregnancy	Neurodevelopment
Rahman et al., 2023 [38]	USA	Retrospective cohort	2001–2014	4559/318,750 mother–child pairs	PM2.5 and its components (CE, NH4+, NO3−, SO42−, OM)	During pregnancy and the first year of life	ASD
Rahman, et al., 2023 [39]	USA	Retrospective cohort	2001–2014	4559/318,750 mother–child pairs	PM2.5 and its components (CE, CO, Fe, and Mn)	During pregnancy up to 5 years of life	ASD
X. Sun et al., 2023 [6]	China	Cohort	2012–2018	512 mother–child pairs	PM2.5 and its components (CN, soil, NH4+, NO3−, SO42−, CO)	During pregnancy	Neurodevelopment
Luglio et al., 2024 [31]	USA (California)	Retrospective cohort	2001–2014	318,750 mother-child pairs	Source-specific PM2.5 (9 sources)	Entire pregnancy (assigned by residence)	ASD
Murphy et al., 2024 [47]	Canada	Matched case–control	2012–2016	1589 ASD cases/7563 controls	PM2.5	Pre-conception, 1st, 2nd, 3rd trimesters	ASD
Yu et al., 2023 [41]	USA (California)	Population-based cohort	2001–2014	318,751 mother-child pairs (4559 ASD cases)	PM2.5 and its components (BC, OM, NO3−, and SO42−)	Pregnancy	ASD
Lin et al., 2023 [33]	Taiwan	Birth cohort	2004–2011	168,062 live term births (666 ASD cases)	PM2.5 heavy metals (As, Cd, Hg, Pb)	Prenatal and postnatal (up to 9 months after birth)	ASD
Fu et al., 2024 [30]	China	Birth cohort	2015–2019	4405 infants/mothers	PM2.5	Prenatal (whole pregnancy, first and second trimester)	Growth retardation (body length), neurodevelopmental retardation
Perera et al., 2024 [36]	USA (NYC)	Prospective cohort	1998–2006	470 mother-child pairs	PM2.5	Prenatal (weekly, trimester, and whole pregnancy)	Neurodevelopmental outcomes (MDI at ages 1 and 3; no significant associations for PDI)
Ghassabian et al., 2025 [37]	USA (44 cohorts)	Cohort meta-analysis	2000–2016	8035 mother-child pairs	PM2.5	Prenatal (pregnancy)	ASD and autism-related traits (SRS scores)
Whitworth et al., 2024 [32]	Spain	Cohort study	2003–2008	1303 mother-child pairs	PM2.5	Prenatal (gestational weeks 1–17), Postnatal (0.5–1.2 years)	Cognitive and motor function (McCarthy Scales)
Mortamais et al., 2025 [46]	France	Case–control	2008–2013	125 ASD cases/500 controls	PM2.5	Prenatal	ASD
O’Sharkey et al., 2025 [40]	USA (California)	Population-based cohort	1990–2018	13,591,003 children; 138,460 ASD cases	PM2.5, NO2, O3, benzene, 1,3-butadiene, chromium, lead, nickel, zinc	Prenatal	ASD

Sample sizes ranged from 438 participants [27] to more than 13 million children [40]. Overall, the studies covered a total of over 15 million individuals, though this figure includes overlapping cohorts and population registries.

The studies analyzed different exposure windows, primarily focusing on the prenatal period, including the three trimesters of pregnancy. Several studies extended the observation to postnatal windows, such as the first year of life or up to five years of age, while one study also included pre-conception exposure [47].

They examined prenatal exposure to PM_2.5 and its chemical components in relation to alterations in cognitive and psychomotor neurodevelopment, as well as specific disorders such as ASD and ADHD. Regarding exposure to PM_2.5, different effects were identified: 9 studies addressed neurodevelopment [6,28,29,30,32,34,35,36,46], 3 studies focused on cognitive and psychomotor neurodevelopment [5,27,33], 10 studies on ASD [31,37,38,39,40,41,42,43,44,45], and 3 studies on ADHD [21,26,47].

Exposure assessment methods varied across studies. Many relied on statistical modeling combined with satellite data to estimate exposure to PM_2.5. Others used monitoring station data, or assigned exposures based on residential address. A number of studies also analyzed PM_2.5 components, including black carbon (BC), organic matter (OM), ammonium (NH₄⁺), nitrate (NO₃⁻), sulfate (SO₄²⁻), carbonaceous elements (CE), metals (e.g., Pb, As, Cd, Hg, Fe, Mn), and source-specific particles such as mineral dust and sea salts.

Finally, the studies can be grouped into two categories based on the type of pollutant analyzed: those assessing PM_2.5 as a whole, and those investigating PM_2.5 along with its components or sources, highlighting increasing efforts in recent years to refine exposure characterization.

3.2. Impact of Particulate Matter on Neurological Development

3.2.1. Prenatal Exposure to PM_2.5 and Its Effects

Table S1 (Main results and conclusions of the studies on PM_2.5) presents the studies that analyzed the effects of PM_2.5 exposure on neurological development.

Regarding specific pathologies, Becerra et al. (2013) reported a 5 to 15% increase in the probability of ASD per interquartile range (IQR) increase of 4.68 μg/m³ of PM_2.5 during pregnancy. Volk et al. (2013) [45] observed a particularly strong association between ASD and PM_2.5 exposure during pregnancy, especially in the first year of life. Lertxundi et al. (2015) [27] concluded a relationship between a 1 μg/m³ increase in PM_2.5 during pregnancy and motor and cognitive development in the children studied.

Regarding exposure during the third trimester, Raz et al. (2015) [43] established a positive association with ASD. Thygesen et al. (2020) [21] found that the association between PM_2.5 and ADHD was positive at low levels of NO₂. Additionally, P. Wang et al. (2021) [34] suggested that prenatal exposure to PM_2.5 could influence early neurological development and problem-solving skills in girls. Chang et al. (2022) [26] identified a higher risk of ADHD related to elevated PM_2.5 levels during the first trimester. Su et al. (2022) [28] showed that PM_2.5 exposure was associated with an increased risk of neurological developmental delay, particularly in the first trimester. Finally, H. Wang et al. (2022) [29] concluded that each 10 μg/m³ increase in PM_2.5 was associated with declines in child development, highlighting adverse effects both during pregnancy and after birth.

The study results demonstrated that increased PM_2.5 levels during pregnancy were associated with a higher risk of developing various adverse neurological effects. For example, an 8.7 μg/m³ increase in PM_2.5 was related to an adjusted risk of 2.08 during pregnancy (95% CI: 1.93, 2.25). Furthermore, a 10 μg/m³ increase in PM_2.5 was associated with an incidence rate ratio (IRR) of 1.38 (95% CI: 1.35, 1.42) and a decrease in BMI scores of −1.68 (95% CI: −2.48, −0.88).

These findings suggest that prenatal exposure to PM_2.5 has a significant impact on health and development.

More recently, additional studies have enriched this evidence base by exploring a broader spectrum of neurodevelopmental outcomes. For instance, X. Sun et al. (2023) and Fu et al. (2024) [6,30] further supported the association between prenatal PM_2.5 exposure and general developmental delay, while Whitworth et al. (2024) [32] emphasized its effect on both cognitive and psychomotor development. Several new studies also confirmed a significant relationship with ASD, including those by Rahman et al. (2023) and Ghassabian et al. (2025) [37,38], who found consistent associations particularly during early gestation. In terms of ADHD, Murphy et al. (2024) [47] observed that early PM_2.5 exposure may alter attentional performance in childhood. Moreover, Fu et al. (2024) [30] reported an increased risk of intrauterine growth retardation, further indicating the broad impact of particulate matter on prenatal health.

Taken together, these new contributions underline the growing body of evidence that highlights prenatal PM_2.5 exposure as a critical environmental risk factor for a variety of developmental and neurological outcomes.

3.2.2. Effects of Exposure Related to PM_2.5 Concentration and Pollutants

Table S2 (Main results and conclusions of the studies on PM_2.5 and its chemical components) presents various studies that analyzed the effects of prenatal exposure to PM_2.5 and its components on neurological development.

Regarding PM_2.5 concentration, Lei et al. (2022) [5] concluded that prenatal exposure to PM_2.5, adjusted to 10 μg/m³ during pregnancy and specifically during the third trimester, was associated with decreased scores in gross motor skills, problem-solving, and personal-social abilities in children. Adjusted compositions at 10 μg/m³ were also inversely associated with scores in these domains, with certain components such as dust, NH₄⁺, and SO₄²⁻ showing significant correlations with lower scores across multiple domains.

In particular, exposure to PM_2.5 and its chemical components was negatively linked to cognitive capacity and motor functions in children, with a notable effect on problem-solving abilities. Xu et al. (2022) [35], after adjusting for confounding variables, determined that average PM_2.5 exposure during pregnancy increased the risk of poor motor development by 31% per 10 μg/m³ increase (RR: 1.31; 95% CI: 1.04, 1.64). Furthermore, prenatal exposure to elevated levels of SO₄²⁻ in PM_2.5 was also associated with a higher risk of poor motor development (RR: 1.40; 95% CI: 1.08, 1.81). No significant associations were observed in other aspects of neurodevelopment. Prenatal exposure to PM_2.5, especially when containing high amounts of SO₄²⁻, may influence neurological development in infants.

Rahman et al. (2023) [38,39], using two-pollutant models, found that PM_2.5 had less influence when considering other components, particularly with organic matter (OM). An inverse relationship between PM_2.5 and ASD risk was observed when adjusting for elemental carbon (EC), but not for carbonaceous nanoparticles (CN). Effect estimates adjusted for components were generally equal to or greater than those in single-pollutant models. Positive associations between ASD and PM_2.5 as a sole pollutant were found in all approaches, mostly across trimesters. Effect estimates for PM_2.5 remained consistent when adjusting for EC or OM and in SO-CTM models. In single-pollutant models, associations between ASD and average exposures to PM_2.5, EC/CN, OM, and SO₄²⁻ persisted even after adjusting for variable factors. EC/CN, OM, and SO₄²⁻ remained strong after adjustment for other components. EC/CN and OM estimates were more consistent in the first and second trimesters, while SO₄²⁻ stood out in the third trimester.

Furthermore, Rahman et al. (2023) [38,39] employed various single-pollutant models and concluded that all components showed associations with increased ASD risk. Each interquartile range (IQR) increase in concentration during pregnancy was associated with hazard ratios (HRs; 95% CI) ranging from 1.08 to 1.17 for different components such as EC, CO, Cu, Fe, and Mn. This suggests that prenatal exposure to non-exhaust emissions may contribute to ASD risk.

Finally, X. Sun et al. (2023) [6] found that prenatal exposure to PM_2.5 and its main components was negatively associated with cognitive development in 6-year-old children, especially during the first trimester. Effects were stronger in boys, but physical activity and prolonged breastfeeding could mitigate these adverse impacts.

Recent additional evidence provides further insight into the nuanced impact of prenatal PM_2.5 exposure on neurodevelopment. Luglio et al. (2024) [31] found that prenatal exposure to total PM_2.5 was associated with an increased risk of ASD, with the strongest links observed for particulate matter originating from gasoline combustion (both on-road and off-road), aircraft, and other sources. Associations with PM_2.5 from sources such as food cooking and natural gas were weaker or inconsistent.

Yu et al. (2023) [41] reported that prenatal PM_2.5 exposure and maternal immune activation (MIA) independently increased the risk of ASD, but no synergistic interaction was detected between these factors.

Lin et al. (2023) [33] highlighted the role of heavy metals within PM_2.5, showing that postnatal exposure to mercury between 10 and 25 weeks significantly raised ASD risk, particularly among low-birth-weight children, emphasizing the importance of minimizing heavy metal exposure during pregnancy and early infancy.

Lastly, O’Sharkey et al. (2025) [40] observed that while prenatal PM_2.5 exposure remains linked to ASD risk, this association has weakened over time in parallel with declining pollution levels. Their findings suggest that postnatal exposure to PM_2.5 may represent a more critical window for ASD risk than prenatal exposure, with differences also influenced by infant sex and socioeconomic factors.

3.3. Risk of Bias Assessment

The overall quality of the included studies was evaluated using a structured risk of bias framework adapted to observational research, encompassing domains such as confounding, selection, exposure measurement, adherence to exposure, missing data, outcome measurement, and selective reporting. Across the studies, confounding represented a recurring concern. While most studies adjusted for a range of potential confounders, including maternal and perinatal characteristics, parental socioeconomic factors, and gestational variables, residual confounding could not be completely excluded, particularly from unmeasured environmental or genetic factors. For example, Becerra et al. (2013), Volk et al. (2013), and Raz et al. (2015) [42,43,45] accounted for multiple maternal and child characteristics, yet the possibility of residual confounding from indoor exposures or other environmental factors remained.

Selection bias was generally low, as most studies relied on well-defined population-based cohorts or registries with clear inclusion criteria and high follow-up completeness. Minor concerns arose in some studies, such as X. Sun et al. (2023) [6], where only a portion of the original cohort was included in the final analysis, potentially introducing selection bias due to loss to follow-up. Similarly, moderate selection bias was noted in P. Wang et al. (2021) and Murphy et al. (2024) [34,47] due to cohort exclusions or incomplete registry linkages, though the overall representativeness of the cohorts remained satisfactory.

Exposure assessment varied across studies. Most studies used high-resolution models, land-use regression, chemical transport models, or satellite-based estimates to quantify ambient PM_2.5 or traffic-related pollutants. While these approaches are sophisticated, moderate risk of exposure misclassification was noted in many cases, especially when individual-level factors such as time-activity patterns or indoor exposures were not captured. Lin et al. (2023) [33] was an exception, where exposure to heavy metals was estimated solely through modeling, introducing a serious risk of misclassification. In contrast, studies such as Thygesen et al. (2020) and Chang et al. (2022) [21,26] achieved low risk due to validated high-resolution exposure models and comprehensive residential histories.

Given the observational nature of the studies, departures from assigned exposure were largely not applicable, and the risk of bias in this domain was consistently low. Missing data were handled variably; most studies reported minimal exclusions or applied appropriate statistical methods, resulting in low to moderate risk of bias. However, studies with incomplete follow-up or unreported handling of missing data, such as Wang et al. (2021) and Su et al. (2022) [28,34], presented a moderate risk.

Outcome measurement generally posed low risk across the studies. Standardized diagnostic tools, registry data, or validated clinical assessments were commonly used, ensuring reliable outcome ascertainment. Nonetheless, moderate risk was observed in studies relying on parent-reported neurodevelopmental measures [5,28,37], which may introduce some measurement error. Finally, selective reporting of results was generally low, with most manuscripts presenting full analyses, sensitivity checks, and multiple models. No studies demonstrated evidence of substantial outcome-reporting bias.

In summary, as presented in

Table 2. Risk of Bias Assessment.

First Author, Year	Cofounding	Selection	Measurement of Exposure	Departure from Exposure	Missing Data	Measurement of the Outcome	Reported Results
Becerra et al., 2013 [42]
Volk et al., 2013 [45]
Lertxundi et al., 2015 [27]
Raz et al., 2015 [43]
Talbott et al., 2015 [44]
Thygesen et al., 2020 [21]
P. Wang et al., 2021 [34]
Chang et al., 2022 [26]
Lei et al., 2022 [5]
Su et al., 2022 [28]
H. Wang et al., 2022 [29]
Xu et al., 2022 [35]
Rahman et al., 2023 [38]
Rahman, et al., 2023 [39]
X. Sun et al., 2023 [6]
Luglio et al., 2024 [31]
Murphy et al., 2024 [47]
Yu et al., 2023 [41]
Lin et al., 2023 [33]
Fu et al., 2024 [30]
Perera et al., 2024 [36]
Ghassabian et al., 2025 [37]
Whitworth et al., 2024 [32]
Mortamais et al., 2025 [46]
O’Sharkey et al., 2025 [40]

Green: Low risk of bias, Yellow: Moderate risk of bias, Orange: Serious risk of bias, Red: Critical risk of bias.

, the risk of bias assessment revealed that while most studies exhibited strong methodological approaches in selection, outcome measurement, and reporting, potential limitations remained in confounding adjustment and exposure assessment. These findings underscore the importance of cautious interpretation, particularly regarding the influence of residual confounding and possible exposure misclassification on observed associations.

4. Discussion

This systematic review on air pollution selected 25 studies and highlighted the importance of understanding and addressing the detrimental effects of exposure to PM_2.5 and its components on prenatal neurological development and child health.

Air pollution is recognized as one of the leading environmental risk factors worldwide, with PM_2.5 representing a complex mixture of particles capable of crossing the placental barrier and reaching fetal circulation [15,48]. This property underlines the biological plausibility of its impact on neurodevelopment.

Exposure to urban traffic and PM_2.5 during pregnancy increased the risk of ASD [42,43,44,45] and negatively affected motor and cognitive development in children [27,28,29,34]. Additionally, higher PM_2.5 exposure, particularly during the third trimester, increased the likelihood of developing ASD [43]. An association between PM_2.5 and ADHD was also observed [21,26].

Several studies consistently demonstrated that timing of exposure plays a crucial role, with the first and third trimesters emerging as particularly vulnerable windows. This is in line with developmental neurobiology, since critical processes such as neuronal proliferation, migration, and synaptogenesis occur during these periods [49,50]. It should be noted that the included studies exhibited substantial heterogeneity in exposure assessment methods (e.g., satellite versus monitoring station data), outcome indicators, and geographical regions. Most of the included studies were conducted in the U.S. and China, with relatively few from other regions such as Africa and South America, which may limit the generalizability of the conclusions. Due to the limited number of studies in each subgroup and the high variability in study designs and outcomes, formal subgroup analyses were not feasible. This heterogeneity is acknowledged as a limitation and emphasizes the need for future research using standardized exposure and outcome measures to allow more robust comparative analyses.

Prenatal exposure to PM_2.5 was shown to affect early neurological development and increased the risk of developmental delay, particularly in problem-solving abilities. Furthermore, air pollution during pregnancy, especially during the first two trimesters, elevated the risk of delayed brain development in young children. These findings highlight that even relatively low increments of PM_2.5 can have measurable effects on developmental trajectories, reinforcing the relevance of continuous monitoring and prevention efforts.

Overall, these findings suggested that the relationship between air pollution exposure, both during the prenatal and early postnatal periods, has negative effects on offspring neurodevelopment.

The results of the studies indicated that increasing concentrations of PM_2.5 during pregnancy are associated with a higher risk of various adverse health outcomes. For example, an increase of 8.7 µg/m³ in PM_2.5 was associated with an adjusted pregnancy risk of 2.08 (95% CI: 1.93, 2.25) [45].

Moreover, an increment of 10 µg/m³ in PM_2.5 was associated with a relative risk ratio (IRR) of 1.38 (95% CI: 1.35, 1.42) and a decrease in IDM scores of −1.68 (95% CI: −2.48, −0.88) [21]. These results suggest that prenatal PM_2.5 exposure has a significant impact on child health and development.

One of the key points highlighted in this systematic review was the adverse relationship between prenatal PM_2.5 exposure and child neurological development. This effect is reflected in various impairments, including gross motor skills and personal and social abilities. This relationship has also been observed in cognitive capacity and motor functions, especially in problem-solving.

Additionally, it is worth noting that exposure disparities often correlate with socioeconomic factors, leading to environmental injustice that disproportionately affects disadvantaged communities [51]. Children born in these contexts may be doubly burdened by both higher exposures and limited access to healthcare and early intervention services.

The study by Xu et al., 2022 [35], provided a quantitative estimate of PM_2.5 exposure during pregnancy and the increased risk of poor motor development, further underscoring the relevance of this issue. Emerging evidence from multiple studies also suggests that exposure to specific PM_2.5 components, such as sulfate ions (SO₄²⁻), polycyclic aromatic hydrocarbons (PAHs), and certain metal constituents, may contribute to adverse neurodevelopmental outcomes. Although the number of studies assessing individual components is still limited and findings are heterogeneous, mechanistic studies indicate biologically plausible pathways, including oxidative stress, neuroinflammation, and epigenetic modifications, that could mediate these effects [52]. Collectively, these results highlight the potential importance of considering particle composition in addition to total PM_2.5 concentration, while also emphasizing the need for further research to strengthen the evidence base.

The studies by Rahman et al., 2023 [38,39], were noteworthy for highlighting the influence of different PM_2.5 components on the risk of autism spectrum disorder (ASD). These findings suggest that not only the total PM_2.5 concentration but also the chemical composition of these particles play a critical role in their impact on neurological health.

Furthermore, the studies by Becerra et al., 2013; Chang et al., 2022; Lertxundi et al., 2015; Raz et al., 2015; Su et al., 2022; Thygesen et al., 2020; Volk et al., 2013; H. Wang et al., 2022; P. Wang et al., 2021 [21,26,27,28,29,34,42,43,45] support the consistency of these findings across multiple contexts and populations, increasing the robustness of the evidence.

Additionally, recent meta-analyses have shown similar patterns of association between prenatal exposure to PM_2.5 and neurodevelopmental disorders, reinforcing the plausibility of these findings [53,54]. Emerging evidence also points to epigenetic mechanisms, oxidative stress pathways, and neuroinflammation as potential biological mediators linking particulate matter exposure to developmental outcomes [55,56,57].

Recent mechanistic studies have further elucidated how PM_2.5 exposure may disrupt neurodevelopment. Fine particulate matter can induce oxidative stress by generating reactive oxygen species (ROS), leading to lipid peroxidation, protein damage, and DNA strand breaks in neural cells [58,59]. Additionally, PM_2.5 can trigger neuroinflammatory responses via activation of microglia and astrocytes, resulting in the release of pro-inflammatory cytokines that may interfere with neuronal proliferation, migration, and synaptogenesis [60]. Epigenetic modifications, including DNA methylation and histone acetylation, have been reported in both animal models and human studies, suggesting that prenatal exposure can lead to long-lasting changes in gene expression relevant to neurodevelopment [61]. Some PM_2.5 components, such as polycyclic aromatic hydrocarbons (PAHs) and metal constituents, are particularly implicated in these pathways, highlighting the importance of considering particle composition in addition to total PM_2.5 concentration [62]. Collectively, these mechanistic insights provide a biologically plausible explanation for the epidemiological associations observed between prenatal PM_2.5 exposure and adverse neurological outcomes, including ASD, ADHD, and delayed motor and cognitive development.

These results underscore the urgent need to address air pollution and reduce PM_2.5 emissions and its components to protect the neurological health of developing children. They also highlight the critical importance of implementing policies and actions to improve air quality, especially during pregnancy [63]. Moreover, they emphasize the relevance of continuing to conduct research in this field to understand the underlying mechanisms and develop effective mitigation strategies. Prevention strategies should also consider public awareness campaigns and targeted interventions in highly polluted urban areas to mitigate exposures among pregnant women.

Finally, it is important to note that future research could benefit from harmonized exposure assessment methods and longitudinal cohort designs to strengthen causal inference. Studies focusing on susceptible subgroups, such as children with genetic predispositions or socioeconomically disadvantaged families, could provide deeper insights into differential vulnerabilities [64,65]. It should also be noted that this review only searched PubMed and Embase databases. While these cover a substantial portion of the medical and scientific literature, other important platforms (e.g., Web of Science, Cochrane Library) were not included, which may have led to missing relevant studies and could affect the comprehensiveness of the evidence. This limitation has been acknowledged to provide transparency regarding potential gaps in the literature coverage.

In the context of this study, it is essential to highlight the findings related to the association between PM_2.5 particulate matter and potential neurodevelopmental effects. The analysis revealed significant patterns of association between PM_2.5 exposure during pregnancy and various aspects of neurological development, including cognitive and psychomotor neurodevelopment, as well as the incidence of ASD and ADHD. These findings are consistent with previous research suggesting links between air pollution exposure and neurological health in early life stages [2,10,11].

Ultimately, this review aimed to emphasize the imperative need to implement policies and actions addressing air pollution exposure during pregnancy as a potential risk factor for offspring neurodevelopment. This study highlights the critical importance of continuing to conduct research linking specific compounds associated with particulate matter to better understand the mechanisms of action of these compounds in pregnant women and across different stages of neonatal neurological development, as current evidence on individual components is still emerging but mechanistically plausible. Such research has significant implications for both public health and environmental policy.

5. Conclusions

This systematic review, encompassing studies on prenatal exposure to PM_2.5 and its effects on neurological development, highlights several key findings. Evidence indicates that prenatal PM_2.5 exposure is associated with increased risks of autism spectrum disorder (ASD), attention-deficit/hyperactivity disorder (ADHD), and impairments in motor and cognitive development, particularly in studies conducted in the U.S. and China. Higher exposure levels during the third trimester appear to further elevate the likelihood of developing ASD. Specific components of PM_2.5, such as sulfate (SO₄²⁻), have been suggested to contribute to adverse neurological outcomes; however, evidence is still emerging and further studies are needed to clarify the strength and mechanisms of these associations. While these findings are consistent across multiple contexts, heterogeneity in exposure assessment methods, outcome indicators, and geographical regions suggest that the conclusions should be interpreted with caution. Targeted interventions and further research in underrepresented regions are needed to clarify context-specific effects and to strengthen the evidence base.