Relationship Between Autism Spectrum Disorder and Maternal Exposure to Passive Smoking and Environmental Factors: A Case‐Control Study in Bangladesh

Abstract

ABSTRACT

Background and Aims

Pre and postnatal environmental exposures can influence gene expression and significantly contribute to the development of autism spectrum disorder (ASD). Bangladesh, recognized as the world's most polluted country, faces a dearth of comprehensive studies focusing on environmental factors associated with ASD. In this observational case‐control study, an exploration was conducted into the relationships between ASD risk and exposure to maternal passive smoking and environmental factors within Bangladesh.

Methods

Twenty‐four out of 103 “Protibandhi Sheba O Shahajjo Kendro” (PSOSK, the disability support and service centers) were selected using a simple random sampling method, ensuring a representative distribution across each division of Bangladesh. A structured questionnaire was used about exposure to maternal passive smoking and related environmental factors. The questionnaire was filled out by face‐to‐face interviews with parents of 310 ASD individuals and 310 healthy controls from January 2020 to June 2021. IBM SPSS version 23 was used for uni‐variate, bi‐variate, and multivariate logistic regression analyses. The significance level was p ≤ 0.05, and the odds ratio (OR) was within 95% confidence intervals (CIs), to determine whether the variable is a risk.

Results

Exposure to maternal passive smoke, living within a mile of a highway, and using household mosquito aerosol (repellents) during pregnancy and early childhood were all significantly linked to an increased risk of ASD (n = 310 each) in this study. Conversely, maternal residence in an urban and brick kiln within one mile of their residence during pregnancy to 3 years postnatal life was significantly associated with a decrease in the adjusted odds of ASD in offspring.

Conclusion

Exposure to maternal passive smoke and household aerosols, along with proximity to highways within one mile during pregnancy and early childhood, increased the odds of ASD. Further research is imperative to overcome the inconsistency and to observe and generalize the association.

Introduction

Autism spectrum disorder (ASD), a multifaceted neurodevelopmental disability affecting early childhood, presents with persistent deficits in social communication and interaction across multiple contexts, accompanied by restricted, repetitive patterns of behavior, interests, or activities [1, 2]. Its global prevalence has surged in recent years, with estimates indicating 1 in 100 children is diagnosed with ASD [3]. Moreover, 2.60% of Scottish primary school children identified as having ASD in 2022 [4].

While primarily considered a genetic disorder, pre and postnatal environmental factors can influence gene expression and significantly contribute to ASD development [5, 6]. Yet, these factors alone cannot fully explain the rising prevalence [7], and research in this area remains in its early stages, often yielding inconclusive results [8, 9]. Smoking is one such environmental risk factor that has attracted attention. Maternal smoking, from 6 months before conception to delivery, was consistently linked to ASD features in the meta-analysis of US cohorts [10]. The history of maternal smoking ( ≥ 20 cigarettes/day) during pregnancy consistently increased the risk for ASD in offspring [11]. The consumption of tobacco by males was 9.38 times higher than that of females in rural areas of Bangladesh [12]. Correlational analyses of one meta-analysis on maternal smoking and ASD suggest that the prevalence of male smoking mirrors exposure to passive smoke (PS) [13], underscoring the imperative to protect non-smokers, especially women and children, from household PS [14].

Some of the major sources of environmental pollution are unplanned urbanization and industrialization, emissions from highways, brick kilns, biomass combustion, mosquito repellents (pesticides), and so forth [15–19]; (). A recent study highlights the emissions and higher concentrations of particulate matter (PM) from several types of mosquito repellents (pesticides) [20]. Lee et al. [20] unequivocally showed that indoor air pollution is caused by all types of repellant. These factors might have effects on health risks, especially for pregnant mothers and their offspring [17, 21, 22]. Bangladesh is the world's most polluted country [23]. However, despite the concerning trend of ASD, comprehensive research on ASD remains under-prioritized in developing countries like Bangladesh [24–26]. In Bangladesh, the first systematic review highlighted the lack of large-scale studies on the relationships between ASD and exposure to environmental factors [25]. Identifying modifiable environmental factors that could be targeted by environmental pollution reduction efforts presents a vital step in preventing ASD and mitigating its severity [13]. This study aims to assess the relationship between ASD and maternal passive smoking and related environmental factor exposures in Bangladesh.

Methods

Study Design

This is an observational case-control survey, using structured questionnaire through face-to-face interviews with respondents.

Participants

The cases were drawn from individuals registered with ASD at the “Protibandhi Sheba O Shahajjo Kendro” (PSOSK), a disability support and service center under the Ministry of Social Welfare, spanning all eight divisions of Bangladesh. The healthy controls (HCs) were selected from relatives and friends residing within the same areas [27]. The diagnosis of ASD was conducted before this study by the relevant specialist team of PSOSK using the “Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5).” The DSM-5 was used in a consistent manner for all the ASD cases. The study included individuals with ASD and HCs aged between 2 and 24 years during the survey. Respondents exclusively comprised parents of both ASD cases and HCs, given their heightened awareness compared to other relatives, caregivers, or step/foster parents [28]. However, exclusions encompassed parents who declined consent (36 individuals), respondents beyond parents (two foster- or step-parents), and individuals with psychiatric illnesses (two) or hearing impairments (three). The healthy controls were evaluated by the relevant specialist team for this project to ensure that they did not have any developmental disabilities or other. Further details regarding participant demographics are presented in Tables 2 and 3.

Procedures

Twenty-four out of 103 PSOSK centers were randomly chosen across Bangladesh using a simple random sampling method, ensuring a representative distribution amongst all eight divisions (Table 1). Employing Krejcie and Morgan's “table for determining sample size for a given population” (1970) [29], a sample size of 310 ASD cases was determined from a population of 1498 ASD individuals. Subsequently, an equivalent number of HCs were randomly selected using simple random sampling, drawn from relatives and friends residing in the same areas as the cases, matching the same age ranges (2–24 years) [27, 30]. Therefore, the total sample size was 620 (310 for each group). Throughout the procedures, strict confidentiality measures were adhered to, ensuring the scientific integrity of all processes.

Table 1 Sampling method for case selection from the study population of ASD, recruited from Bangladesh, 2020–2021 (n = 620, 310 for each ASD case and healthy control groups).

Division wise number of selected PSOSK†	Division wise total number of PSOSK	Selected PSOSK by simple random sampling	Registered number of ASD case at selected PSOSK	Approximate % to be included from total population of ASD	Collected % of ASD case	Collected % of HC
Two: from lowest through 9 PSOSK	Mymensingh, 7	2	84	5	2.9	10.6
Barisal, 9	2	55	5	6.8	6.8
Sylhet, 8	2	134	10	8.4	9.4
Three: from 10–14 PSOSK	Rajshah, 10	3	332	20	16.8	9.7
Rangpur, 14	3	136	10	15.8	19
Khulna, 13	3	145	10	6.5	7.4
Four: from 15–19 PSOSK	Chattogram, 18	4	304	20	23.9	18.14
Five: from ≥ 20 PSOSK	Dhaka, 24	5	308	20	19	18.7
Total	103	24	1498	100	100	100

Materials and Data Collection

The selection of independent variables (IVs) for the construction of a structured questionnaire was guided by an extensive review of existing literature [5, 6, 12–15, 17–19, 25, 31–37] focusing on maternal passive smoke and related some environmental factors including some socio-demographic factors (as covariates) that might be associated with the ASD development in Bangladesh. The questionnaire was validated by experts from academics (supervisors, biostatistician), practitioners (pediatrician, physician, disability specialists), and a researcher specializing in ASD, guided by the existing literature [38]. The questionnaire was pretested on 30 parents having a child with ASD. The data were collected through face-to-face interviews with respondents from January 2020 to June 2021 after obtaining written informed consent. The data collection process was conducted by skilled data collectors, including the principal investigator. Additionally, the principal investigator himself collected most of the data to reduce the challenges.

Measures

During pregnancy to postnatal 3 years, history of maternal passive smoking exposure and related environmental factors (biomass [wood, dried dung, coal, etc.] burning for cooking; use of coil and aerosol for mosquitoes at home; and the presence of highways, industry/factory, and brick-kilns within one mile of the mother's residence) were categorized as no (reference group) and yes. Additionally, the mother's residence during pregnancy to 3 years of postnatal life was divided into rural (reference group) and urban. The postnatal exposures to the above-mentioned environmental factors were basically considered for the children. Correspondingly, some socio-demographic covariates were also assessed across various categories: (1) Monthly total household income (MTHI) during pregnancy was used to measure socioeconomic status (SES), and categorized as follows: ≤ 10,000 (reference group), 10,001–20,000, and ≥ 20,001 Bangladeshi Taka (BDT). (2) Level of education during pregnancy for both father and mother, segmented into three categories: ≤ Primary (reference group), SSC/HSC, ≥ Bachelor degree. (3) Age during childbirth is categorized into three groups {≤ 21 years (reference group), 22–35 years, and ≥ 36 years for fathers, and ≤ 18 years (reference group), 19–30 years, and ≥ 31 years for mothers}. (4). The subject's gender was categorized as female (reference group) and male. The subject's age was not a variable of interest.

Statistical Analysis

The statistical analysis was conducted utilizing IBM SPSS version 23. Active maternal tobacco smoking is very rare in Bangladesh [12, 39]. In the present study, there was only one mother who smoked actively, and this factor (active smoking by mother during pregnancy) was not included as a measure and was excluded from analysis. Thus, this study analyzed the history of maternal passive smoking exposure from tobacco and related environmental factors during pregnancy to 3 years postnatal. Chi-square (X²) tests were done for frequency distribution and to evaluate the relationship between variables. Additionally, to examine the relationship between the factors/independent variables (IVs) and expected outcomes/dependent variables (DV), binary logistic regression (BLR) analysis [for unadjusted or crude Odds Ratio (cOR)] was employed (Table 3). Finally, multiple logistic regression (MLR) analysis was performed to compute the adjusted odds ratios (aORs) accompanied by 95% confidence intervals (CIs), determining the risk association of each variable with the offspring's ASD. The MLR model was adjusted for socio-demographic covariates (monthly total household income (MTHI), parental age and education, and gender of the subjects). The threshold for significance was set at p ≤ 0.05.

Patient and Public Involvement

While this study did not directly involve patients in formulating research questions, selecting outcome measures, or designing and implementing the research, parent participants actively contributed through open-ended interview dialogs and received regular updates on the study's progress and findings.

Ethics Statement

This study was in accordance with the Declaration of Helsinki (World Medical Association, 2013) and approved by the “Higher Studies Committee” (Serial number: 45; Reg. Number: 2931; date: March 6, 2018), and “Biosafety, Biosecurity and Ethical Committee” {Ref No: BBEC. JU/M 2023/01(12)} of the Jahangirnagar University. Official permission was obtained from PSOSK to collect the data. Confidentiality of the person and the information was maintained. Written informed consent was obtained before data collection from respondents.

Results

Demographics, Frequency Distribution of Socio-Demographic Factors

The statistics pertaining to socio-demographic factors are presented in both Table 2 and Table 3. The ages of fathers and mothers at the subject's birth ranged from 15 to 63 and 12 to 48 years, respectively, while the subjects' ages varied from 2 to 24 years. The monthly total household income (MTHI) during pregnancy ranged from 1075 to 600,000 BDT. The X² test in cross-tabulation revealed a significant association between ASD and all socio-demographic factors except the mother's age group (p = 0.11) at childbirth (Table 3).

Table 2 The distribution of subject's age, parent's age and MTHI during pregnancy in the ASD cases and healthy controls, recruited from Bangladesh, 2020–2021 (n = 620).

Factors	Control (n = 310)	ASD (n = 310)	p value
Subject's age, year† (mean ± SD)	9.8 ± 6.06	10.3 ± 4.65	0.25
Father's age† (mean ± SD)	32. 83 ± 6.66	32.69 ± 7.27	0.81
Mother's age† (mean ± SD)	24.97 ± 5.79	23.88 ± 5.86	0.02
MTHI during pregnancy (median) in BDT	20,000	15,000	< 0.001

Table 3 The frequency distribution (cross tabulation with Chi-square test) between ASD and maternal passive smoking and related environmental factor exposures during pregnancy to postnatal three years in the ASD cases and healthy controls, recruited from Bangladesh, 2020–2021 (n = 620).

Factors category Sociodemographic factors	Levels/groups	Cross tabulation (n, %) with X2-test
Healthy control (310)	ASD (310)	p value
Father's education†	≤ Primary SSC/HSC ≥ Bachelor degree	113 (36.5) 47 (15.2) 150 (48.4)	130 (41.9) 71 (22.9) 109 (35.2)	0.002
Mother's education†	≤ Primary SSC/HSC ≥ Bachelor degree	122 (39.4) 78 (25.2) 110 (35.5)	143 (46.1) 98 (31.6) 69 (22.3)	0.001
Monthly total household income†	≤ 10000 BDT 10,001–20,000 BDT  ≥ 20,001 BDT	71 (22.9) 96 (31.0) 143 (46.1)	128 (41.3) 83 (26.8) 99 (31.9)	< 0.001
Age group of father‡	≤ 21 22–35  ≥ 36	3 (1.0) 228 (73.5) 79 (25.5)	10 (3.2) 203 (65.5) 97 (31.3)	0.03
Age group of mother‡	≤ 18 19–30  ≥ 31	42 (13.5) 224 (72.3) 44 (14.2)	61 (19.7) 205 (66.1) 44 (14.2)	0.11
Subject's gender	Female Male	147 (47.4) 163 (52.6)	91 (29.4) 219 (70.6)	< 0.001
Environmental factors
History of maternal passive smoking§	No Yes	223 (71.9) 87 (28.1)	169 (54.5) 141 (45.5)	< 0.001
Mother's residence§	Rural Urban	120 (38.7) 190 (61.3)	150 (48.4) 160 (51.6)	0.02
Use of biomass for cooking§	No Yes	174 (56.1) 136 (43.9)	146 (47.1) 164 (52.9)	0.02
Highway¶	No Yes	165 (53.2) 145 (46.8)	137 (44.2) 173 (55.8)	0.02
Industry/factory¶	No Yes	246 (79.4) 64 (20.6)	255 (82.3) 55 (17.7)	0.36
Brick-kiln¶	No Yes	277 (89.4) 33 (10.6)	288 (92.9) 22 (7.1)	0.12
Mosquito coil use at home§	No Yes	154 (49.7) 156 (50.3)	137 (44.2) 173 (55.8)	0.17
Aerosol use for mosquitoes at home§	No Yes	278 (89.7) 32 (10.3)	248 (80.0) 62 (20.0)	0.001

Frequency Distribution and Unadjusted Effects

The Chi-square test in cross-tabulation (Table 3) and the outcomes of the binary logistic regression (BLR) analysis (Table 4) uncovered a significant association between ASD and the maternal passive smoking including a majority of environmental factors except three (the presence of a factory/industry and brick kiln within one mile of the mother's residence, and history of coil use at home for mosquitoes). The BLR analysis presents the cORs along with their respective 95% CIs.

Table 4 Binary logistic regression (BLR) model and multiple logistic regression (MLR) model (adjusted with socio-demographic covariates) analyses between ASD and maternal passive smoking and related environmental factor exposures during pregnancy to postnatal 3 years in the ASD cases and healthy controls, recruited from Bangladesh, 2020–2021 (n = 620).

Environmental factors	Levels/groups	Binary LR	Multiple LR
cOR (95% CI)	p value	aOR (95% CI)	p value
History of maternal passive smoking†	No Yes	Reference 2.14 (1.53, 2.99)	< 0.001	Reference 1.92 (1.30, 2.82)	0.001
Mother's residence†	Rural Urban	Reference 0.67 (0.49, 0.93)	0.02	Reference 0.62 (0.41, 0.95)	0.03
Use of biomass for cooking†	No Yes	Reference 1.44 (1.05, 1.97)	0.03	Reference 1.21 (0.77,1.90)	0.41
Highway‡	No Yes	Reference 1.44 (1.05, 1.97)	0.03	Reference 2.09 (1.42, 3.07)	< 0.001
Industry/Factory‡	No Yes	Reference 0.83 (0.56, 1.24)	0.36	Reference 0.69 (0.42, 1.13)	0.14
Brick-kiln†,‡	No Yes	Reference 0.64 (0.37, 1.13)	0.12	Reference 0.52 (0.27, 0.99)	0.05
Mosquito coil use at home†	No Yes	Reference 1.25 (0.91, 1.71)	0.17	Reference 1.06 (0.74, 1.53)	0.75
Aerosol use for mosquitoes at home†	No Yes	Reference 2.17 (1.37, 3.44)	0.001	Reference 3.01 (1.76, 5.12)	< 0.001

Adjusted Effects

[IMAGE OMITTED. SEE PDF]

In the MLR model, maternal passive smoke and related environmental factors were adjusted together with socio-demographic covariates (monthly total household income, parental age, parental education, and gender) (Table 4 and Figure 1).

The risk of ASD was nearly two times higher in children whose mothers reported passive smoke exposure during pregnancy to 3 years of postnatal life than who did not expose (aOR = 1.92; 95% CI = 1.30, 2.82; p = 0.001). This risk was slightly higher than two times for the presence of a highway within one mile of the mother's residence (aOR = 2.09; 95% CI = 1.42, 3.07; p < 0.001). The history of aerosol use at home demonstrated a three-fold higher risk of ASD in offspring compared to those who did not use (aOR = 3.01; 95% CI = 1.76, 5.12; p < 0.001). The residence of mothers in urban areas was associated with a 38% reduced risk of ASD in offspring compared to rural (aOR = 0.62; 95% CI = 0.41, 0.95; p = 0.03). The presence of brick kilns within one mile of their residence was associated with a 48% reduced risk of ASD in offspring compared to those who did not (aOR = 0.52; 95% CI = 0.27, 0.99; p = 0.05). However, all other relationships were not significant after correcting for covariates (Table 4).

Discussion

This study represents the first case-control study on this topic based on the published literature, covering the entire country of Bangladesh. The findings suggest that maternal passive smoke exposure, along with several environmental factors, during either the prenatal or postnatal periods are associated with increased odds of ASD in Bangladesh. However, it is important to interpret these outcomes with caution, as some of the associations remain insignificant and a few of them reveal inconsistent findings compared to previous studies, even after adjusting for socio-demographic covariates in the MLR model.

Following the adjustment of socio-demographic covariates in the MLR analysis, specific factors emerged as significant. Since there was only one mother who actively smoked in the current study, this component (active smoking by mother during pregnancy) was excluded from the measures and was not included in the analysis. Thus, the analysis was conducted on prenatal and postnatal exposure to maternal passive smoking and revealed a significant rise in the odds of ASD in offspring (in both LRs). The history of maternal smoking during pregnancy consistently increased the risk for ASD in offspring [10, 11]. Children with ASD might encounter household environmental tobacco smoke either during gestation or early childhood [14]. Awareness development about the cessation of exposure to passive smoking and proper implementation of laws are necessary to mitigate the risk of ASD.

This study revealed that highways within one mile of the mother's residence during pregnancy to 3 years postnatal increased the odds of ASD in offspring. This aligns with research indicating that exposure to near-roadway air pollution during pregnancy increases the risk of ASD in American children [37]. Additionally, an animal (rat) model study suggested that traffic pollution during pregnancy or after birth could alter behaviors linked to offspring's ASD, underscoring early-life environmental risks as crucial for epigenetic perturbations, despite the elusive nature of the environmental insult [19]. However, in a 2021 Egyptian study, no association was found between residences near highways during the postnatal period and ASD development [32]. Zhou et al. [19] highlighted in their animal study that the combined impact of traffic-related air pollution and traffic noise impacts the onset of neurological disorders in China [19]. These studies suggest that noise stemming from traffic, industries, and constructions might collectively influence the emergence of ASD.

The use of biomass for cooking demonstrated an increased risk of offspring's ASD in both logistic regression models in this study, although the aOR was not statistically significant. Thus, this finding is inconsistent with previous studies based on adjusted effect. The combustion of biomass for cooking purposes has been identified as a significant source of indoor air pollution in Bangladesh [18]. Similar investigations in China have linked exposure to indoor air pollution during pregnancy from cooking smoke with the risk of ASD, while using natural gas for cooking showed potential for risk reduction [40].

From pregnancy to 3 years of postnatal periods in this study, aerosol (pesticide) use for mosquitoes significantly increased the odds of ASD in offspring, while use of the mosquito coil did not, in both LR analyses. Compared to liquid (aerosol) and mat versions, coil repellents for mosquitoes' exhibit noticeably higher particulate matter (PM) concentrations, faster burning rates, and greater emission rates [20]. Nabgha-E-Amen et al., [15] found that pesticides led to increased mercury (Hg) exposure among children with ASD [15]. The Utah physicians [41] (retrieved on 16th June 2022 from ) stated that the use of pesticide spray for mosquitoes may increase the risk of ASD. However, Hicks et al. [33] found that aerial pesticide exposure did not increase the individual risk of autism [33]. In contrast with present study, the indoor air pollution exposure during pregnancy from mosquito coil incense was associated with risk of ASD in Chinese children [22]. Exposure to high concentrations of mosquito repellents (containing heavy to light particles) in early childhood has harmful effects [21]. Some inconsistency is present in the current study regarding the types of mosquito repellents used in Bangladesh and their relation to ASD. Thus, coil repellents are mostly concerning and warrant closer scrutiny and potential reevaluation of their safety for widespread use, while all repellent types contribute to indoor air pollution [20].

Surprisingly, in this study, the presence of industries/factories within one mile of a mother's residence during pregnancy to 3 years postnatal showed an insignificant reduction in the odds (aOR) of ASD in offspring. The finding is consistent with a Chinese study that found no association between residing near a factory/industry during the postnatal period and ASD risk [19]. However, it is important to note that industrialization is one of the major sources of environmental pollution in Bangladesh [42, 43]. Similarly, the presence of brick kilns within one mile of a mother's residence during pregnancy to 3 years postnatal showed significantly decreased odds (aOR) of ASD in the offspring of this study. However, this finding is inconsistent, as brick kilns are also a major source of environmental pollution in Bangladesh and pose a risk to human health and the environment [17]. Saha et al. [17] indicated that various air pollutants from brick kilns in Bangladesh—such as SO2, NOx, PM_2.5, PM₁₀, SPM—exceeded the National Ambient Air Quality Standard (NAAQS) limits [17]. The risk of ASD in children was linked to greater PM_2.5 exposure throughout the first two trimesters and O₃ exposure in the late third trimester of pregnancy in the American cohort [34]. Therefore, implementing safe and eco-friendly technologies in brick kilns could potentially mitigate pollution levels and consequently reduce the risk of ASD in Bangladesh. Furthermore, this study revealed that maternal residence in urban areas was significantly associated with a lower risk of ASD compared to rural. This finding contrasts with the previous studies [44–46] that have observed an increased risk of ASD associated with rapid urbanization. The intersection of biogeochemical trace element emissions into human habitats and disruption of the fragile environment, exacerbated by industrialization and urbanization in developing nations, poses a significant health risk [15]. This complex scenario reflects the interplay between genetic and environmental factors in the onset of ASD [8, 47–49], where environmental influences notably impact frontal lobe development, potentially amplifying susceptibility among individuals with ASD compared to their neurotypical counterparts [47]. The inconsistencies observed in this study regarding residence in urban areas, proximity to industries/factories, or brick kilns may be due to a lack of matched control or cohort design.

In terms of parental age and socioeconomic status (SES), the study found that the 22–35-year-old age group of fathers and the 19–30-year-old age group of mothers, as well as higher SES levels, were negatively associated with ASD in the Chi-square test. This aligns with previous research that has shown a decrease in ASD risk with same paternal age group and SES levels of education in the relevant project of this study, published in 2024 [26]. Thus, it's plausible that younger fathers may have lower levels of education, whereas lower-income families may be more inclined to expose smoking, which can lead to passive smoking for women and children at home (based on the Chi-square test). The advanced parental ages were associated with a decrease in the odds of offspring's ASD [35]. This association might stem from the assumption that younger parents might be less inclined towards higher educational attainment, while older parents, more aware of the impact of exposure to environmental factors, might take precautions. The odds of tobacco use was 3.62 times higher for individuals without any education than for those with at least 6-years of schooling, and it was nearly twice as high for the impoverished as for the wealthy in rural areas of Bangladesh [12]. Therefore, ensuring safety from exposure to household passive tobacco smoke, especially for women and children, is imperative [14]. Furthermore, the strong association between ASD and pollutants in recent years indicates the increasing environmental pollution in the developed world [50], and it would be seriously alarming for developing countries like Bangladesh. Bangladesh is the world's most polluted country, and Dhaka is the world's second most air-polluted city in 2020 [23], and this position has not been changed. The findings of the present study emphasize the relationship between ASD and maternal passive smoking and a few environmental factors (pollution from highways and household aerosol use for mosquitoes) among children in Bangladesh. However, some major sources of environmental pollution, such as residence in urban areas (which leads to urbanization) and pollution from brick kilns and industry, showed inconstancy in this study. Thus, it is necessary to take precautions against passive smoking exposure in the house for nonsmokers, especially women and children [14]. Hence, ensuring cessation of passive smoking and related environmental safety, coupled with a keen focus on socio-demographic considerations, becomes an urgent imperative to mitigate the risk of ASD in Bangladesh. Strategic interventions, such as avoidance or modification of passive smoking and associated environmental factors during pregnancy and early childhood, hold promise in mitigating ASD risk. These insights should strongly inform policymakers, stakeholders, governmental bodies, and private agencies in crafting effective interventions and policies aimed at reducing ASD risk in the country.

Strengths and Limitations

This is the first case-control study on this topic in Bangladesh based on published literature, with a relatively large sample size covering all divisions, which is considered a strength of this study. It provides compelling and consistent evidence with previous studies linking specific environmental factors (history of maternal passive smoking exposure and aerosol pesticide exposure to control mosquitoes at home, and highway within one mile from their residence during pregnancy to 3 years of postnatal life) to an increased the odds of ASD.

However, this study does have its limitations. Notably, the inability to recruit matched controls posed a challenge due to feasibility constraints. This limitation was mitigated by adjusting for potential covariates in the MLR analysis. Furthermore, the absence of an “Autism Diagnostic Observation Schedule” score or a similar severity index among the studied parameters is another limitation. This absence stemmed from the recruitment method, as the cases were drawn from individuals registered with ASD (diagnosed by DSM-5) at the PSOSK. Moreover, to mitigate the recall bias, a random sampling method was used and ensured the questions' wording did not influence participants' answers. Additionally, the principal investigator himself collected most of the data to reduce challenges. However, it is worth notting that there was a deviation in the collected data for the case group from the anticipated proportion in the Mymensingh and Khulna divisions. In specific divisions like Mymensingh and Rajshahi, the ratio of collected controls did not closely align with the cases due to the unavailability of respondents. Finally, some sources of environmental factors (residence in urban areas, which leads to urbanization and pollution from brick kilns and industry) showed inconstancy in this study. Some of the confounders related to maternal and child health were not included in this study. Thus, the MLR model was adjusted with some of the socio-demographic covariates to see the adjusted effects of environmental factors. Consequently, it is necessary to ensure environmental safety and further study.

Conclusions

This study investigated the link between passive smoke exposure, environmental exposures during pregnancy and early childhood, and the risk of autism spectrum disorder (ASD) in children in Bangladesh. The findings provide strong evidence that maternal passive smoke exposure, household aerosol pesticide use for mosquito control, and residence within one mile of a major highway during pregnancy and the first 3 years of life significantly increased the odds of ASD. However, inconsistencies were found for a few major sources of environmental pollution, such as residence in an urban area and living within one mile of a brick kiln (a traditional industry in Bangladesh) or an industry/factory. These findings highlight the complex interplay between environmental factors and ASD development. Future research using larger population-based cohorts or matched case-control designs with experimental components should further explore these associations and their generalizability. Importantly, this study underscores the need for prevention strategies targeting environmental exposures during critical windows of early development to potentially reduce the risk of ASD and improve outcomes for children.

Author Contributions

Md. Shahid Khan: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, project administration, data curation, resources. Mohammad Alamgir Kabir: conceptualization, methodology, validation, writing – review and editing, supervision, formal analysis. Shafi M. Tareq: conceptualization, methodology, validation, writing – review and editing, supervision.

Acknowledgments

The authors extend their appreciation to the respondent's families, assistant data collectors, officers and staff of “Protibandhi Sheba O Shahajjo Kendro” (PSOSK), experts, supervisors, teachers and all members of the Department of Environmental Sciences (Jahangirnagar University).

Conflicts of Interest

The authors declare no conflict of interest.

Data Availability Statement

The data supporting this study's findings are available on request from the corresponding author. The data are not publicly available for reasons of privacy or ethical restrictions.

Transparency Statement

The lead author Md. Shahid Khan affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

References

H. Shimomura, H. Hasunuma, S. Tokunaga, et al., “Early Developmental Signs in Children With Autism Spectrum Disorder: Results From the Japan Environment and Children's Study,” Children (Basel, Switzerland) 9, no. 1 (2022): [eLocator: 90], [DOI: https://dx.doi.org/10.3390/children9010090].