Air Pollution, Autismspectrum disorders, and brain imaging in CHildrenamong Europe

Mònica Guxens1-4, Małgorzata J Lubczyńska1-4, Laura Pérez-Crespo1-3, Albert Ambrós1-3, MatteoRenzi5, Matteo Scortichini5, Maciej Strak6, Xavier Basagaña1-3, Ryan Muetzel4, Antònia Valentín1-3, ItaiKloog7, Gerard Hoek6, Joel Schwartz8, Francesco Forastiere5, Tonya White4, Jordi Sunyer1-3, HenningTiemeier4,8, Bert Brunekreef6,9, Massimo Stafoggia5, Hanan El Marroun4

1ISGlobal, Barcelona, Spain; 2Pompeu Fabra University, Barcelona, Spain; 3Spanish Consortium for Research on Epidemiology and PublicHealth, Spain; 4Erasmus University Medical Centre–Sophia Children’s Hospital, Rotterdam, The Netherlands; 5Lazio Regional HealthService, Rome, Italy; 6Institute for Risk Assessment Sciences, Utrecht, The Netherlands; 7Ben-Gurion University of the Negev, Beer Sheva, Israel; 8Harvard T.H. Chan School of Public Health, USA; 9Julius Center for Health Sciences and Primary Care, Utrecht, The Netherlands

Contact information:monica.guxens@isglobal.org

@m_guxenswww.monicaguxens.com

AUTISM SPECTRUM DISORDERS STUDY

Aim. To assess the relationship between prenatal air pollution exposure at different time windows and the development of autism spectrum disorders

BRAIN IMAGING STUDY

Aim. To assess the relationship between prenatal and postnatal air pollution exposure at different time windows and brain structural and functional changes in children

Manuscript 1. Prenatal PM2.5 exposure was associated with a thinner cortex in several brain regions in 6-10 years old children and these alterations partially mediated the association between prenatal PM2.5 exposure and impaired child inhibitory control (n=783) (Guxens et al. Biol Psychiatry. 2018; pii: S0006-3223(18)30064-7)

Manuscript 2. Fetal and childhood exposure to several traffic-related air pollutants was associated to an impaired white matter microstructure in 9-12 years old children (n=2,954) (Lubczynska et al. under review)

Manuscript 3. Association between fetal and childhood exposure to several traffic-related air pollutants and brain morphology in 9-12 years old children (n=3,133) (Lubczynska et al. in preparation)

Postnatal air pollution exposure

Conclusions

• We found an association between higher fetal and childhood exposure to pollutants representative of traffic related sources, with attenuated cortical thickness, larger cortical and subcortical volumes, ventricle enlargement, and lower volume of corpus callosum in school-age children.

• Associations with fetal life exposure to air pollution were predominantly observed in girls rather than in boys.

• Since this is the first study to find relationships with increases in the volumes of various grey matter structures and the ventricles of the brain, more studies are warranted to confirm our findings.

Table 1. Autism spectrum disorders prevalence in 2017 in Catalonia, Spain

Sex Age groups

Total Boy GirlSex-ratio

2 to 5 years old

6 to 10 years old

11 to 17 years old

N 15,466 12,647 2,819

4.5

1,579 5,621 6,340

Prevalence(95% CI)

1.23 (1.21; 1.25)

1.95 (1.92; 1.99)

0.46 (0.44; 0.48)

0.53 (0.54; 0.59)

1.67(1.63; 1.72)

0.75 (0.74; 0.76)

CI, confidence interval; N, number of children with autism spectrum disorders

Figure 2. High spatially and temporally resolved air pollution models of Spain for 2015

Figure 1. Autism spectrum disorders incidence between 2009 and 2017 in Catalonia, Spain

A. Overall incidence B. Incidence by sex C. Incidence by age groups

Walter A. Rosenblith New Investigator Award 2016

Models adjusted for parental ages, educational levels, ethnicities, psychiatric symptoms, height, and body mass index, household income, marital status, maternal prenatal smoking, prenatal alcohol consumption, parity, and intelligence quotient, and child’s age at scanning. In bold: associations that remain after effective number of tests correction. CC: corpus callosum; GM, grey matter; WM, white matter

Table 2. Association with brain morphology using a vertex-wise approachTable 1. Association with global brain morphology

Models adjusted for parental ages, educational levels, ethnicities, psychiatric symptoms, height, and body mass index, household income, marital status, maternal prenatal smoking, prenatal alcohol consumption, parity, and intelligence quotient, and child’s age at scanning. In bold: associations that remain after effective number of tests correction. LH, left hemisphere; RH, right hemisphere

Lateral view of the right hemisphere

Red, postcentral gyrus; Purple, precentral gyrus; Yellow, pars triangularis; Brown, rostral middle frontal gyrus

Medial view of the left hemisphere

Light blue, lingual gyrus; Dark blue, rericalcarine cortex; Pink, precuneus

• Fine spatial grid 1-km2

• PM monitoring sites ~ Aerosol Optical Depth + Spatial predictors (land cover, climate types, road density, population density, emission data, orography, imperious surface) + Spatio-Temporal predictors (planetary boundary layer, light-at-night, normalization difference vegetation index, Saharan dust, meteorology)

• Random forest model: “machine learning” model made of multiple regression trees

• Total cross-validation R2: 0.55 for PM10, 0.58 for PM2.5, and 0.40 for PM2.5-10

• Final models for each year between 2002 and 2016

0,07

0,10 0,10

0,12

0,16

0,180,19

0,21

0,23

0,00

0,05

0,10

0,15

0,20

0,25

2009 2010 2011 2012 2013 2014 2015 2016 2017

Au

tis

m s

pe

ctr

um

dis

or

de

rs

in

cid

en

ce

(%

)

Years

0,12

0,160,15

0,19

0,260,27

0,30

0,330,35

0,02 0,030,04 0,04

0,06 0,070,08

0,09 0,09

0,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

0,40

2009 2010 2011 2012 2013 2014 2015 2016 2017

Years

0,05

0,08 0,08

0,11

0,16 0,16

0,18

0,22

0,25

0,10

0,130,14

0,16

0,21

0,23 0,23

0,25

0,27

0,070,08 0,08

0,09

0,12 0,130,14

0,180,17

0,00

0,05

0,10

0,15

0,20

0,25

0,30

2009 2010 2011 2012 2013 2014 2015 2016 2017

Years