https://orcid.org/0000-0001-5004-574X
Figures
Abstract
Background
Outdoor air pollution is increasingly recognised as a key threat to population health globally, with particularly high risks for urban residents. In this study, we assessed the association between residential nitrogen dioxide (NO2) exposure and children’s cognitive and behavioural development using data from São Paulo Brazil, one of the largest urban agglomerations in the world.
Methods
We used data from the São Paulo Western Region Birth Cohort, a longitudinal cohort study aiming to examine determinants as well as long-term implications of early childhood development. Cross-sectional data from the 72-month follow-up was analysed. Data on NO2 concentration in the study area was collected at 80 locations in 2019, and land use regression modelling was used to estimate annual NO2 concentration at children’s homes. Associations between predicted NO2 exposure and children’s cognitive development as well as children’s behavioural problems were estimated using linear regression models adjusted for an extensive set of confounders. All results were expressed per 10 μg/m3 increase in NO2.
Results
1143 children were included in the analysis. We found no association between NO2 and children’s cognitive development (beta -0.05, 95% CI [-0.20; 0.10]) or behavioural problems (beta 0.02, 95% CI [-0.80; 0.12]).
Conclusion
No association between child cognition or child behaviour and NO2 was found in this cross-sectional analysis. Further research will be necessary to understand the extent to which these null results reflect a true absence of association or other statistical, biological or adaptive factors not addressed in this paper.
Citation: Luminati O, Brentani A, Flückiger B, Ledebur de Antas de Campos B, Raess M, Röösli M, et al. (2022) Assessing the association between air pollution and child development in São Paulo, Brazil. PLoS ONE 17(5): e0268192. https://doi.org/10.1371/journal.pone.0268192
Editor: Santosh Kumar, Sam Houston State University, UNITED STATES
Received: September 23, 2021; Accepted: April 19, 2022; Published: May 13, 2022
Copyright: © 2022 Luminati et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Outdoor air pollution is increasingly recognised as one of the most critical challenges for global public health. In particular the concern arises in urban areas, where pollution levels are generally highest, since pollutants are primarily produced by the combustion of fossil fuel [1].
The most studied pollutants affecting health are particulate matter (PM), ozone (O3), sulphur dioxide (SO2) and nitrogen dioxide (NO2). Exposure to these pollutants can trigger inflammatory processes that can cause damage to various organs such as the lungs, heart, and brain [2,3] and has also been associated with a higher risk of developing illnesses such as respiratory diseases [4], cancer [2,5] and cardiovascular problems [3,6,7].
Various studies have reported adverse health effects of air pollution for children. Prenatal and childhood exposure to NO2 has been associated with asthma [8,9], failing lung development and consequent respiratory difficulties [10] and lower birth weight [10]. Existing research also suggests a positive association between traffic related air pollutants, and delays in brain maturation [11,12], memory [13,14] and cognition [12,15] among school children. Epidemiological and animal studies [16,17] have attributed these associations to changes in the central nervous system (CNS) which can be caused by an activation of the systemic [16] or local [18] immune systems and by oxidative stress [17,19].
Given that in Brazil over 80% of the population live in dense urban settings [20], average exposure to air pollution among Brazilian citizens is high [21]. São Paulo constitutes the largest urban area in Brazil and ranks among the largest cities in the world, with an estimated metropolitan population of 22 million and continued rapid growth [22]. Despite municipal efforts to improve air quality [23], air pollution levels remain a concern.
Previous studies from São Paulo have linked air pollution to non-accidental and cardiovascular mortality [7], especially in the elderly population [24], hospitalisations for respiratory problems in children [25], low birth weight [26] and premature births [27].
Few studies have investigated the impact of air pollution on cognitive development in children, most of which focus on Europe. Given that children are likely to be more susceptible to air pollution because their central nervous and immune systems are still developing [2,3,16], the lack of evidence in this area is somewhat surprising.
This study aims to address this research gap by exploring the association between nitrogen dioxide exposure and children’s cognitive and behavioural development using prospectively collected data on child development and air pollution from São Paulo, Brazil.
Methods
Study population and study area
Data on child development was collected through the São Paulo Western Region Birth Cohort (SP-ROC) [28]. The SP-ROC cohort was set up as a longitudinal study to assess the impact of social and environmental factors on child development. All children born at the University Hospital of São Paulo between April 1, 2012 and March 31, 2014 and living in the Butantã- Jaguaré area were included in the original sample. Data was collected at birth from medical records and then during five visits: postpartum, at 6, 12, 36 and 72 months. The Butantã- Jaguaré district is about 70km2 large and situated in the Western part of São Paulo city, with an estimated population of 637’000 people in 2010 [29]. The study area is characterized by a dense network of streets of all sizes and includes a wide variety of different neighbourhoods, ranging from the university campus, residential neighbourhoods and green areas to favelas and industrial areas [29].
In this study, all cohort children still living in the Butantã-Jaguaré area with a complete 72-month follow-up and geocoded address were included, as shown in the flowchart in Fig 1. For the 72-month assessment, trained interviewers surveyed 1849 children. Due to the Covid-19 situation in Brazil, data collection for this follow-up had to be terminated in April 2020. A total of 163 records had invalid addresses for geocoding and were dropped from the analysis. Additionally, 128 children were living outside the study area, while for 57, the International Development and Early Learning Assessment (IDELA) was incomplete. Furthermore, 358 children moved between the 36-month assessment and the 72-month assessment and were excluded from the analysis because assigning appropriate NO2 exposure levels was not possible. This resulted in 1143 children being eligible for the main analysis. 65 children out of these 1143 moved before the 36-month assessment and 1078 never moved.
The data collection for the SP-ROC cohort was approved by the University of São Paulo’s Hospital das Clínicas ethics committee (CAPPESQ HC_FMUSP) under protocol N° 01604312.1.0000.0065. Written informed consent from the caregivers was obtained before data collection. The work presented in this paper was formally reviewed and approved by the Ethikkommission Nordwest- und Zentralschweiz (EKNZ) under protocol N° AO_2020–00024.
NO2 exposure
With the goal to investigate the effect of air pollution on child development, we focused on NO2 as our primary pollutant. NO2 primarily results from traffic related burning of fossil fuel and can be measured relatively easily and cheaply using passive gas samplers (80). Annual mean outdoor NO2 concentrations at the home address of the children were predicted by a land use regression (LUR) model based on NO2 measurements over 2 one-week periods at 80 locations in 2019. The first measurements were done in summer (February), reflecting the hot humid and rainy conditions, the second in winter (August) reflecting the colder and dry conditions. Temporal adjustments were made using weekly data collected at a reference monitoring site all year around to calculate adjusted annual mean NO2 concentrations. As described in more details elsewhere [30], a LUR model was developed, using supervised stepwise linear regression by including GIS predictor variables (e.g. roads, land use, altitude, green space etc.), explaining 66% of the variation in the adjusted annual mean NO2 concentrations. The predictions from the annual model were then mapped for the entire study area on a 25x25 meters grid in QGIS 3.4.4. Using the same software, the predicted NO2 values were assigned to each child’s geocoded home address.
Outcome measures
Based on the existing literature, we considered two primary outcomes: children’s cognitive ability and children’s behavioural difficulties.
Cognition.
Children’s overall development was assessed using the International Development and Early Learning Assessment (IDELA) tool. The IDELA includes 22 tasks divided into the following four domains: Motor (gross and fine), Communication/Emerging Literacy, Problem solving/Emergent Numeracy, and Personal-social/Social-emotional [31]. Trained interviewers explained the tasks to each child and recorded observed answers. Raw IDELA scores were standardised to z-scores within the analysed sample.
Behavioural problems.
Children’s behavioural problems were captured using the Child Behavior Checklist (CBCL) [32]. The CBCL scale contains 120 items for children aged 6 to 18 and are graded on a three-point Likert scale, where 0 means that the behaviour was not reported by the caregiver (absent), 1 means the behaviour occurs sometimes, and 2 means the behaviour occurs often. All questions ask about children’s behaviour within the six months preceding the interview. In order to facilitate interpretation, we normalised raw CBCL scores within our sample to a z-score with mean zero and standard deviation of one.
Caregiver interview and child assessments were performed during a home visit by a trained team of enumerators with previous experience in child assessments. Before assessment, caregivers signed a written consent form.
Statistical analysis
In order to describe the sample regarding demographic, socioeconomics and other variables, we firstly performed descriptive statistics. The main analysis, was a cross-sectional analyses of the outcome data collected during the 72-month follow-up. We standardised both outcomes to mean zero and standard deviation one. We executed univariate and multivariable linear regressions to check for confounders. The univariate and the multivariate models are given by:
- Univariate Model
- Multivariate Model
Where Yi is the outcome of an child i, NO2 is the estimated annual exposure, and εi is the model residuals. To address heteroscedasticity as well as potential non-linearity concerns, we used the robust sandwich estimator. For the sensitivity analysis linear regressions were performed on a sub sample. Only children living at the same address since birth were considered in the subsample. Additionally, we performed linear regressions with a categorical exposure variable to compare children exposed to NO2 concentrations in the bottom and the top decile with the rest of the sample. As a further sensitivity analysis, we repeated the main analysis by omitting the children considered being outliers. Children showing an IDELA z-score lower than -3.5 (11 children) and children showing a CBCL z-score higher than 3.75 (6 children) were considered outliers. All statistical analyses were performed using the STATA version 16.0 statistical software package [33].
Other variables
We adjusted for an extensive list of potential confounders in the regression, including child gender, child age in months, child skin-colour, birthweight, gestational length, delivery type, mother’s age at delivery, mother’s skin-colour, maternal depression, caregiver’s marital status, caregiver’s relation to the child, caregiver’s age, highest school grades of caregiver, highest school grade of household head, household size, financial support, socio-economic status (following the Brazilian economic status classifications) and home stimulation score.
Apart from gender, birthweight, gestational length and delivery type, which were collected at birth; all data were collected by the SP-ROC-Cohort 72-month follow-up questionnaire.
Results
A total of 1143 children were included in our main analysis. As shown in Table 1, 51% of children were male. The age at assessment varied between 4 to 7.5 years, with a mean age of 76 months. Most of the children were classified as white (45%) or mixed (51%). 81 children were born preterm (7%) and 6% had a birth weight < 2500 grams. Mother’s age at delivery ranged between 13 and 46 years, with a median of 26 years. 37% of children were born with a caesarean. At the 72-month follow up 65% of the caregivers lived with a partner and 68% completed a middle or higher school grade. 24% of the households received financial support. 44% of the mothers showed symptoms of mild or severe depression. Further characteristics are described in S1 Table. The comparison of the sample at birth and the sample at the 72-month follow-up is presented in S2 Table. Although we had a high rate of loss of follow-up, mostly because of the Covid-19 pandemic, the sample’s characteristics did not differ from the original sample at birth.
Mean annual NO2 concentration ranged from 31.1 μg/m3 to 128.8 μg/m3 in the study area. Estimated concentrations at the residential addresses of children ranged from 34.0 μg/m3 to 98.0 μg/m3, with a median exposure of 40.8 μg/m3 and a standard deviation of 4.9 μg/m3. All participants were exposed to NO2 concentrations higher than 10 μg/m3, the WHO air quality guideline value for annual mean NO2 [34].
IDELA and CBCL z-scores ranged between -5.98 and 1.40 (IDELA) and between -1.15 and 5.69 (CBCL).
Table 2 provides the main regressions’ results for the full sample of 1143 children, as well as for the subsample of children who never moved (N = 1078). We found no significant associations between NO2 and children’s cognition (IDELA) and behaviour (CBCL) in either sample.
Table 3 shows results for a categorical exposure variable. No indications for decreased cognitive functions or increased behavioural problems for the highest exposed group (highest decile) was found. Fig 2 provides further details on the exposure levels in the higher and lower decile.
In S3 Table we show results that exclude outliers—the results are consistent with the full sample regressions presented here.
Discussion and conclusion
This study aimed to assess association between exposure to NO2 and child development. Overall exposure levels were very high in the study settings, with all participants exposed to NO2 concentrations above the WHO annual recommended level of 10 μg/m3 [34]. Despite these high levels of exposure, we found no association between exposure to NO2, and children’s cognitive development or children’s behavioural problems in our sample.
In the literature, evidence on the association between NO2 and child development remains uncertain. Freire et al. found a negative association between NO2 exposure higher than 24 μg/m3 and general cognition compared with NO2 exposures lower than 15.4 μg/m3 [35], although these differences were not statistically significant. Forns et al. looked at air pollution in different schools in Barcelona and found a significant negative association with all measured pollutants and working memory, and largest associations for NO2 [13]. Comparing children at high and low NO2 exposure Sunyer et al. found negative associations between NO2 and working memory [14]. In general, most of the literature seems to find a negative relationship between NO2 exposure and children’s cognition, although the overall evidence was classified as insufficient in literature review from 2015 [36]. The same literature review also reported insufficient evidence for the association between NO2 and child behaviour [36]. The majority of studies published so far have focused on exposure levels at school [13,14].
This project was initiated with a strong prior of finding negative associations between NO2 exposure and children’s development and behaviour. We performed NO2 measurements directly in the study area, within the personal premises of 80 SP-ROC study participants. These measurements enabled us to build a robust LUR model to predict NO2 concentrations. Another strength of our study is the SP-ROC-Cohort dataset, which contains detailed information on families and home environments, enabling us control for a large range of potential confounders. The data set also contains both direct assessments of children’s cognitive skills and detailed reports on behavioural issues, allowing us to look at both dimensions of child development.
Despite this setup, we found no evidence of negative associations between NO2 and child’s development and behaviour. There are some limitations that may have contributed to the lack of association. First, we estimated the NO2 concentration at the residential address of the children, which describes their exposure only during a part of the day. Most of the children in the study went to kindergarten or school, and we did not collected data on NO2 concentration in these locations. Second, as with all NO2 exposure models, a certain extent of misclassification of exposure levels cannot be avoided. Given that we measured exposure levels twice across 80 different measurement locations, the scope for such measurement error seems however small. From a purely statistical perspective, we were of course also limited by the very high average exposure levels with relatively limited variability. The sample contained only 2 children with NO2 exposure levels below 35 μg/m3, and only a few with NO2 exposure levels above 50 μg/m3. Even if NO2 exposure is harmful, it is possible that differences in exposure levels in our sample were too small to result in statistically detectable differences in child outcomes. It is of course also possible that the true causal effect of NO2 exposure is smaller than the current literature suggests, or that negative effects only occur in settings where other protective factors (that may have been in place here) are not present.
Further research will be needed to better understand the diverging results in the current literature.
Supporting information
S1 Table. Participants’ characteristics at 72-month follow-up.
https://doi.org/10.1371/journal.pone.0268192.s001
(DOCX)
S2 Table. Participants’ characteristics at birth vs. at 72-month follow-up.
https://doi.org/10.1371/journal.pone.0268192.s002
(DOCX)
S3 Table. Sensitivity analysis without outliers.
https://doi.org/10.1371/journal.pone.0268192.s003
(DOCX)
Acknowledgments
We would like to thank the Eckenstein-Geigy Professorship for supporting the collection of the NO2 data.
References
- 1.
European Environment Agency. Air pollution sources 2019 [https://www.eea.europa.eu/themes/air/air-pollution-sources-1.
- 2. Schraufnagel DE, Balmes JR, Cowl CT, De Matteis S, Jung SH, Mortimer K, et al. Air Pollution and Noncommunicable Diseases A Review by the Forum of International Respiratory Societies’ Environmental Committee, Part 2: Air Pollution and Organ Systems. Chest. 2019;155(2):417–26. pmid:30419237
- 3. Babadjouni RM, Hodis DM, Radwanski R, Durazo R, Patel A, Liu Q, et al. Clinical effects of air pollution on the central nervous system; a review. Journal of clinical neuroscience: official journal of the Neurosurgical Society of Australasia. 2017;43:16–24.
- 4.
WHO WHO. Review of evidence on health aspects of air pollution—REVIHAAP Project. 2013.
- 5. Ribeiro AG, Downward GS, Freitas CU, Chiaravalloti Neto F, Cardoso MRA, Latorre M, et al. Incidence and mortality for respiratory cancer and traffic-related air pollution in São Paulo, Brazil. Environ Res. 2019;170:243–51. pmid:30594696
- 6. Künzli N, Jerrett M, Garcia-Esteban R, Basagaña X, Beckermann B, Gilliland F, et al. Ambient air pollution and the progression of atherosclerosis in adults. PloS one. 2010;5(2):e9096. pmid:20161713
- 7. Bravo MA, Son J, de Freitas CU, Gouveia N, Bell ML. Air pollution and mortality in São Paulo, Brazil: Effects of multiple pollutants and analysis of susceptible populations. Journal of exposure science & environmental epidemiology. 2016;26(2):150–61.
- 8. Khreis H, Kelly C, Tate J, Parslow R, Lucas K, Nieuwenhuijsen M. Exposure to traffic-related air pollution and risk of development of childhood asthma: A systematic review and meta-analysis. Environment international. 2017;100:1–31. pmid:27881237
- 9. Gehring U, Wijga AH, Hoek G, Bellander T, Berdel D, Brüske I, et al. Exposure to air pollution and development of asthma and rhinoconjunctivitis throughout childhood and adolescence: a population-based birth cohort study. The Lancet Respiratory medicine. 2015;3(12):933–42. pmid:27057569
- 10. Veras MM, de Oliveira Alves N, Fajersztajn L, Saldiva P. Before the first breath: prenatal exposures to air pollution and lung development. Cell and tissue research. 2017;367(3):445–55. pmid:27726025
- 11. Pujol J, Martínez-Vilavella G, Macià D, Fenoll R, Alvarez-Pedrerol M, Rivas I, et al. Traffic pollution exposure is associated with altered brain connectivity in school children. NeuroImage. 2016;129:175–84. pmid:26825441
- 12. Clifford A, Lang L, Chen R, Anstey KJ, Seaton A. Exposure to air pollution and cognitive functioning across the life course–A systematic literature review. Environmental Research. 2016;147:383–98. pmid:26945620
- 13. Forns J, Dadvand P, Esnaola M, Alvarez-Pedrerol M, López-Vicente M, Garcia-Esteban R, et al. Longitudinal association between air pollution exposure at school and cognitive development in school children over a period of 3.5 years. Environmental Research. 2017;159:416–21. pmid:28858754
- 14. Sunyer J, Esnaola M, Alvarez-Pedrerol M, Forns J, Rivas I, Lopez-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):24. pmid:25734425
- 15. 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. Epidemiology (Cambridge, Mass). 2016;27(2):228–36.
- 16. Calderón-Garcidueñas L, Calderón-Garcidueñas A, Torres-Jardón R, Avila-Ramírez J, Kulesza RJ, Angiulli AD. Air pollution and your brain: what do you need to know right now. Primary health care research & development. 2015;16(4):329–45.
- 17. Costa LG, Cole TB, Dao K, Chang YC, Coburn J, Garrick JM. Effects of air pollution on the nervous system and its possible role in neurodevelopmental and neurodegenerative disorders. Pharmacology & therapeutics. 2020;210:107523. pmid:32165138
- 18. Genc S, Zadeoglulari Z, Fuss SH, Genc K. The adverse effects of air pollution on the nervous system. Journal of toxicology. 2012;2012:782462. pmid:22523490
- 19. Hahad O, Lelieveld J, Birklein F, Lieb K, Daiber A, Münzel T. Ambient Air Pollution Increases the Risk of Cerebrovascular and Neuropsychiatric Disorders through Induction of Inflammation and Oxidative Stress. International journal of molecular sciences. 2020;21(12). pmid:32560306
- 20.
United Nations Statistics Divisions. UNdata A world of information 2021 [updated 24.03.2021. https://data.un.org/Data.aspx?q=brazil+urban&d=POP&f=tableCode%3a1%3bcountryCode%3a76.
- 21.
WHO WHO. Country profile of Environmental Burden of Disease—Brazil. 2009.
- 22.
Schneider RM, Minkel CW, Leite A. São Paulo: Encyclopaedia Briatnnica, inc.; 2020 [https://global.britannica.com/place/Sao-Paulo-Brazil.
- 23. Andrade MdF, Kumar P, de Freitas ED, Ynoue RY, Martins J, Martins LD, et al. Air quality in the megacity of São Paulo: Evolution over the last 30 years and future perspectives. Atmospheric Environment. 2017;159:66–82.
- 24. Costa AF, Hoek G, Brunekreef B, Ponce de Leon AC. Air Pollution and Deaths among Elderly Residents of São Paulo, Brazil: An Analysis of Mortality Displacement. Environmental health perspectives. 2017;125(3):349–54. pmid:27713111
- 25. Gouveia N, Corrallo FP, Leon ACP, Junger W, Freitas CU. Air pollution and hospitalizations in the largest Brazilian metropolis. Revista de saude publica. 2017;51:117. pmid:29211200
- 26. Gouveia N, Bremner SA, Novaes HM. Association between ambient air pollution and birth weight in São Paulo, Brazil. Journal of epidemiology and community health. 2004;58(1):11–7. pmid:14684720
- 27. Saldiva S, Barrozo LV, Leone CR, Failla MA, Bonilha EA, Bernal RTI, et al. Small-Scale Variations in Urban Air Pollution Levels Are Significantly Associated with Premature Births: A Case Study in São Paulo, Brazil. International journal of environmental research and public health. 2018;15(10). pmid:30322031
- 28.
Brentani A, Scoleze Ferrer AP, Brentani H, Liu CH, Grisi SJFE, Valente MH, et al. Cohort profile: São Paulo Western Region Birth Cohort (ROC). International Journal of Epidemiology. 2020.
- 29.
São Paulo P. Mapa Digital da Cidade de São Paulo 2020 [http://geosampa.prefeitura.sp.gov.br/PaginasPublicas/_SBC.aspx#.
- 30. Luminati O, Ledebur de Antas de Campos B, Flückiger B, Brentani A, Röösli M, Fink G, et al. Land use regression modelling of NO(2) in São Paulo, Brazil. Environmental pollution (Barking, Essex: 1987). 2021;289:117832.
- 31. Pisani L, Borisova I, Dowd AJ. Developing and validating the International Development and Early Learning Assessment (IDELA). International Journal of Educational Research. 2018;91:1–15.
- 32.
Achenbach TM, & Rescorla L.A. Manual for the ASEBA Preschool forms and Profiles.: Burlington, VT: University of Vermont Department of Psychiatry; 2000.
- 33.
StataCorp. Stata Statistical Software:Release 16. College Station, TX:StataCorp LLC2019.
- 34.
World Health O. WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. Geneva: World Health Organization; 2021 2021.
- 35. Freire C, Ramos R, Puertas R, Lopez-Espinosa M-J, Julvez J, Aguilera I, et al. Association of traffic-related air pollution with cognitive development in children. 2010;64(3):223–8.
- 36. Suades-González E, Gascon M, Guxens M, Sunyer J. Air Pollution and Neuropsychological Development: A Review of the Latest Evidence. Endocrinology. 2015;156(10):3473–82. pmid:26241071