Cáceres et al. BMC Medicine (2023) 21:142 BMC Medicine https://doi.org/10.1186/s12916-023-02815-9 RESEARCH ARTICLE Open Access Prenatal environmental exposures associated with sex differences in childhood obesity and neurodevelopment Alejandro Cáceres1,2,3*† , Natàlia Carreras‑Gallo1†, Sandra Andrusaityte4, Mariona Bustamante1,2,5,6, Ángel Carracedo7,8, Leda Chatzi9, Varun B. Dwaraka10, Regina Grazuleviciene4, Kristine Bjerve Gutzkow11, Johanna Lepeule12, Léa Maitre1,2,5, Tavis L. Mendez10, Mark Nieuwenhuijsen1,2,5, Remy Slama12, Ryan Smith10, Nikos Stratakis1, Cathrine Thomsen11, Jose Urquiza1,2,5, Hannah Went10, John Wright13, Tiffany Yang13, Maribel Casas1,2,5, Martine Vrijheid1,2,5 and Juan R. González1,2,14* Abstract Background Obesity and neurodevelopmental delay are complex traits that often co-occur and differ between boys and girls. Prenatal exposures are believed to influence children’s obesity, but it is unknown whether exposures of preg‑ nant mothers can confer a different risk of obesity between sexes, and whether they can affect neurodevelopment. Methods We analyzed data from 1044 children from the HELIX project, comprising 93 exposures during pregnancy, and clinical, neuropsychological, and methylation data during childhood (5–11 years). Using exposome-wide interac‑ tion analyses, we identified prenatal exposures with the highest sexual dimorphism in obesity risk, which were used to create a multiexposure profile. We applied causal random forest to classify individuals into two environments: E1 and E0. E1 consists of a combination of exposure levels where girls have significantly less risk of obesity than boys, as compared to E0, which consists of the remaining combination of exposure levels. We investigated whether the association between sex and neurodevelopmental delay also differed between E0 and E1. We used methylation data to perform an epigenome-wide association study between the environments to see the effect of belonging to E1 or E0 at the molecular level. Results We observed that E1 was defined by the combination of low dairy consumption, non-smokers’ cotinine levels in blood, low facility richness, and the presence of green spaces during pregnancy (­ ORinteraction = 0.070, P = 2.59 × ­10−5). E1 was also associated with a lower risk of neurodevelopmental delay in girls, based on neuropsy‑ chological tests of non-verbal intelligence ­(ORinteraction = 0.42, P = 0.047) and working memory ­(ORinteraction = 0.31, P = 0.02). In line with this, several neurodevelopmental functions were enriched in significant differentially methylated probes between E1 and E0. † Alejandro Cáceres and Natalia Carreras-Gallo equally contributed to the paper *Correspondence: Alejandro Cáceres alejandro.caceres@isglobal.org Juan R. González juanr.gonzalez@isglobal.org Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Cáceres et al. BMC Medicine (2023) 21:142 Page 2 of 14 Conclusions The risk of obesity can be different for boys and girls in certain prenatal environments. We identified an environment combining four exposure levels that protect girls from obesity and neurodevelopment delay. The com‑ bination of single exposures into multiexposure profiles using causal inference can help determine populations at risk. Keywords Prenatal environment, Sexual dimorphism, Childhood obesity, Neurodevelopment, DNA methylation, Causal inference, Multiexposure profile Background on causal random forest, that can determine profiles of Boys and girls develop differently. For instance, their multiple exposures that are associated with high sexual immune response to infections differs from an early age, dimorphism [10]. Here, we aimed to adapt our method to their brains grow at different rates, and the prevalence of determine which combination of prenatal exposures can numerous common diseases, like obesity, is also differ- produce an environment where girls are more protected ent [1–3]. As reported by Shah et al., 65% of the coun- from obesity than boys during the preteen years. Further- tries around the world and 96% of high-income countries more, obesity in children is associated with lower cogni- reported a greater prevalence of obesity for boys than tive function, particularly inhibitory control and working girls in children aged 5–9 years old [3, 4]. Given the con- memory, critical for academic achievement [11]. Obe- trasting paths of development, it is remarkable that bio- sity often co-occurs with neurodevelopmental disorders, medical studies typically consider sex as a confounder particularly in boys [12]. Therefore, we also evaluated rather than the main effect or an effect modifier [5]. whether the environment of high sexual dimorphism in Exposome studies, in particular, are characterized by obesity also shows a significant sexual dimorphism in the acquisition of massive amounts of data at individual non-verbal intelligence, working memory, attention, and and population levels [6]. A crucial goal of these studies ADHD. is to inform the likely conditions for which a given pub- Finally, we investigated whether the protective envi- lic health intervention would be optimal, such that the ronment may be associated with epigenetic changes best intervention is applied at the right time to the right since many exposures during pregnancy are associated population [7]. However, as the main difference between with specific methylation profiles [13]. This analysis individuals is sex, exposome studies aiming at improving may provide information about the molecular pathways precision medicine and precision public health cannot that may be participating in the association between the do without considering how environmental risk factors environment and the sexual differences in obesity and affect sexual dimorphism in development and disease. neurodevelopment. From a mechanistic context, studying the factors that Here, we aimed to (1) combine multiple exposure levels increase sexual dimorphic outcomes of disease can offer to define an environment with high sexual dimorphism in important insights into its etiology and comorbidities, obesity risk; (2) given the correlation between obesity and and inform of possible interventions and targeted treat- neurodevelopmental delay in children, we also enquired ments. Important advancements have been made in if the subpopulation exposed to this environment shows studying sex-related risk factors for diseases like cancer, a significant sexual dimorphism in neurodevelopment; Alzheimer’s, and autoimmune diseases [8]. However, a and (3) we then hypothesized that the individuals who relevant component of these age-related diseases is hor- belong to such an environment can be characterized by monal regulation. Studying sex differences in preteens specific patterns of DNA methylation. offers not only the opportunity for identifying targeted treatments for early-age illnesses but also to explore dis- ease mechanisms unlikely influenced by sex hormones Methods that may also onset early in life. Previous research has, for Study population instance, underlined that maternal factors during preg- We analyzed data from The Human Early Life Exposome nancy can affect disease outcomes later in life [9] and, (HELIX). This is a multi-center study that included a therefore, motivates the question of which pregnancy total of 1301 mother–child pairs from six existing birth factors may promote later sexual dimorphism in disease. cohorts in Europe: BIB (Born in Bradford; the UK) [14], Environmental exposures likely orchestrate environ- EDEN (Etude des Déterminants pré et postnatals du ments that are more toxic to one sex than to the other développement et de la santé de l’Enfant; France) [15], one. However, methods to determine such multiple- INMA-SAB (Infancia y Medio Ambiente; Spain; subco- exposure environments are not readily available. We hort Sabadell) [16], KANC (Kaunas cohort; Lithuania) have developed a method of causal modeling, based [17], MoBa (The Norwegian Mother, Father and Child Cáceres et al. BMC Medicine (2023) 21:142 Page 3 of 14 Cohort study; Norway) [18]), and Rhea (Greece) [19]. The Clinical outcomes pairs participated in a common, completely harmonized, Height and weight measurements were measured during follow-up examination, when children were between the clinical visit performed at ages 5 to 11 years. These 5–11 years old to fully characterize the pregnancy and measurements were converted to body mass index (BMI childhood exposome [20]. During the clinical examina- in kg/m2) for age-and-sex z-scores using the interna- tion, urine (pooled spot urine samples from before bed- tional WHO reference curves to allow comparison with time and first morning void) and blood samples were other studies [21]. Obese children were defined as those collected from the children. Urine and blood samples above the age- and sex-specific 95th percentile, as recom- previously collected from mothers during pregnancy mended by WHO. were also available for biomarkers of chemical exposure Neurodevelopmental outcomes were assessed assessment. In our analyses, we selected the individuals through a battery of internationally standardized, who had data on prenatal exposures, performed the clini- non-linguistic, and culturally blind computer tests cal and neurodevelopment examination, and had methyl- also at ages 5 to 11 years. We assessed working mem- ation data (N = 1044). All studies received approval from ory, attention, and general non-verbal intelligence the ethics committees of the centers involved and writ- with the N-back test [22], the attention network test ten informed consent was obtained from all participants. (ANT) [23], and Raven’s colored progressive matri- Cohort characteristics are shown in Table 1. ces [24]; respectively. The tests were administered in Table 1 Characteristics of HELIX cohort. Clinical characteristics of children during pregnancy and follow-up Children assessed at follow-up N = 1044 Sex, male 571 (54.6%) Cohort BIB (UK) 90 (8.6%) EDEN (France) 135 (12.9%) INMA (Spain) 198 (19.0%) KANK (Lithuania) 196 (18.8%) MOBA (Norway) 239 (22.9%) RHEA (Greece) 136 (17.8%) Age in years, mean (range) 7.9 (5.4–11.9) BMI (kg/m2), median (range) 16.3 (12.2–29.5) Obesity 62 (5.9%) – F: 23 (4.9%) – M: 39 (6.8%) Raven’s matrices, median (range) 27 (9–36) Affected 189 (18.2%) – F: 78 (16.6%) – M: 111 (19.5%) N-back (2-back accuracy), median (range) 0.91 (0.36–1) Affected 104 (12.9%) – F: 46 (12.6%) – M: 58 (13.0%) ANT (accuracy), median (range) 0.97 (0.51–1) Affected 206 (20.0%) – F: 72 (15.6%) – M: 134 (23.7%) ADHD 104 (10.0%) – F: 27 (5.8%) – M: 77 (13.6%) Prenatal characteristics N = 1044 Mother’s age during pregnancy, mean (range) 30.9 (16–34) Mother’s BMI during pregnancy, median (range) 23.6 (15.8–51.4) Mother’s weight gain during pregnancy, mean (range) 13.9 (0–40) Maternal education Primary school 119 (11.3%) Secondary school 359 (34.3%) University degree or higher 566 (54.2%) Gestational age, median (range) 40 (30.8–44.1) Year of birth, median (range) 2006 (2003–2009) Number of parents native from the country cohort, mean (range) 1.9 (02) Parity 0.69 (0–2) BMI body mass index, ANT attention network test, ADHD attention deficit hyperactive disorder, F female, M male Cáceres et al. BMC Medicine (2023) 21:142 Page 4 of 14 a standardized way by trained field workers through mother’s BMI, mother’s weight gain during pregnancy, study-provided laptops. The outcomes did not dis- gestational age, mother’s age during pregnancy, moth- tribute normally. We dichotomized them, taking as er’s education, whether parents were native from the cases individuals with outcomes below the first quin- country cohort, parity, and children’s age at clinical tiles (20%), which clearly captured the long lower tail assessment (Table 1). of the outcomes’ distributions (Fig. 1). These percen- tiles were chosen because they allowed the selection Pregnancy exposome of the lowest performers in the outcomes using a sin- HELIX has collected a wide range of exposures measured gle criterion and preserving a representative number during two main windows: a prenatal window including of individuals within the groups. We thus studied as the pregnancy period and a postnatal window including clinical outcomes the events of having these cogni- the exposome data of children at the same time as omics tive abilities affected. We also considered ADHD sampling (childhood). In this study, we only considered diagnosis. the first window (pregnancy exposome) which consists We considered as common covariates (covariates of 93 exposures distributed across 17 exposure families, used in all the analyses) 10 variables, based on Maitre including the urban environment, the chemical expo- et al. [25]. These covariates are cohort, year of birth, some, and social and lifestyle factors. Fig. 1 Distributions of clinical outcomes in the HELIX study. Analyses were performed for categorized variables shown in gray (reference) and red (affected) Cáceres et al. BMC Medicine (2023) 21:142 Page 5 of 14 The urban environment includes exposure estimates subjects. We used the Combat algorithm to remove the for built environment, surrounding green and blue batch effects supported by the slide. Methylation levels spaces, ultraviolet (UV) radiation, road traffic noise were expressed as beta values corrected by surrogate var- levels, air pollution, noise, meteorology, and socioeco- iables and CpG sites were annotated to genes by Illumina nomic deprivation index [20, 26]. These exposures were HM450 manifest file (version 1.2). We discarded the assessed during pregnancy by environmental geographic subjects without exposome data and without European information systems (GIS) according to their residential ancestry based on genomic data, resulting in 993 individ- addresses. Tobacco smoke and diet were evaluated by uals for the methylome analysis. We computed blood cell questionnaires. Biomarkers of contaminant exposure, like type proportions following Houseman et al. algorithm cotinine levels, were measured in appropriate biological [35] and Reinius reference panel [36]. samples (urine or blood) collected from mothers during pregnancy. Details on the exposure assessment methods Statistical methods and exposure factors can be seen in the Additional file 1: Figure 2 shows the statistical workflow. Supplementary Methods [20, 27–33]. Missing values for all exposures were imputed using Identification of prenatal exposures with sexual the method of chained equations using the mice pack- dimorphism in obesity risk age in R [32], as described in detail elsewhere [33]. When We used exposome-wide interaction analyses to deter- possible, multiple imputation procedure was applied mine the exposures whose association with obesity was (missing values are imputed stochastically several times). significantly different between sexes. We assessed the For the imputation process, continuous variables should associations between obesity (cases and controls) and have a normal distribution. Thus, skewed exposure vari- the interactions between sex (S) and each of the prenatal ables were transformed to achieve normality or catego- exposures (Di) using the logistic regression model. rized if no transformation worked. Exposure variables with their corresponding transformation are described in Additional file 2: Table S1. Exposures with more than E(Y ) = logit−1 αi + βi (S × Di ) + γir Cri r=1...k 70% of missing values in each cohort were excluded from the imputation process. Therefore, missing values ranged where Y is the obesity status of an individual with sex S from 1.5% in traffic density to 65% in fast-food intake and ith exposure Di. γir are the regression coefficients during pregnancy. Although none of the participants had of the k covariates Cri that included sex, exposure I, and complete data on all exposures, 95% of individuals had the 10 common covariates mentioned before (cohort, missing values in less than 30% of exposures. year of birth, mother’s BMI, mother’s weight gain dur- ing pregnancy, gestational age, mother’s age during DNA methylation pregnancy, mother’s education, whether parents were One of the main goals of HELIX was to associate multiple native from the country cohort, parity, and children age environmental factors with omics biomarkers and child at clinical assessment). βi were the effects of interest that health outcomes. For these same children, multi-omics measure the association between obesity and the inter- molecular phenotyping was performed, which included action between sex and each exposure i. We adjusted measurement of blood DNA methylation (450 K, Illu- p-values using false discovery rate to correct for multiple mina), among others. comparisons. The DNA was obtained from buffy coat collected in EDTA tubes at 5–11 years of age. Briefly, DNA was Creation of a multiexposure profile (E1 and E0) extracted using the Chemagen kit (Perkin Elmer) in We calculated the residuals of the exposures with nomi- batches of 12 samples. Samples were extracted by cohort nal significant interactions adjusted by the 10 common and following their position in the original boxes. DNA covariates. Then, we used these residuals as covariates concentration was determined in a NanoDrop 1000 in causal inference modeling, using causal random for- Quant-iT™ PicoGreen® dsDNA Assay Kit (Life Technol- UV–Vis Spectrophotometer (ThermoScientific) and with est and taking sex as the treatment variable, to determine which children in HELIX had been in personal environ- ogies). DNA methylation was assessed using the Infinium ments with significant sexual dimorphism in obesity Human Methylation 450 beadchip (Illumina), following (female > male or female < male). We then aimed to deter- the manufacturer’s protocol. Preprocessing of methyla- mine whether the personal environments of the children tion data has been described elsewhere [34]. After sample with one of the significant dimorphisms (F > M or F < M) and probe quality control measures, the number of CpG could be averaged into two prenatal environments, one probes analyzed was 371,533, initially available for 1192 whose female protection against obesity was stronger Cáceres et al. BMC Medicine (2023) 21:142 Page 6 of 14 Fig. 2 Statistical workflow. The figure shows all the statistical analyses carried out along the paper than those observed for the individual exposures and the N-back test, ANT, Raven’s colored progressive matrices, other the opposite. We did not find enough children with and ADHD) and we tested the interaction between the negative dimorphism (F > M). We thus created an aver- environment (E1/E0) with sex. We adjusted the model by age environment with highly significant female protec- sex, the environment, and the 10 common covariates. tion against obesity, which hereinafter we will refer to it as E1. E1 was defined as a binary vector, with one entry Methylation differences between E1 and E0 for each level of the exposures, indicating whether a We performed an epigenome-wide association study given exposure averaged across the children with positive (EWAS) in the HELIX cohort between E1 and E0. As dimorphism was higher or lower than the average across previously, we used logistic regression models to identify all other children in the training set. We used the multi- the probes that were differentially methylated between exposure profile to classify all the individuals in the entire environments. We adjusted the analysis by the 10 com- HELIX cohort depending on whether they belong to the mon covariates used in previous analyses and counts of E1 or not (E0: F < M and F = M). To this end, we used soft different immune cells in the blood. Associations were targeting that tested whether they matched the environ- corrected for multiple comparisons using false discovery ment in at least 60% of the exposures. The causal infer- rate, as computed by limma. For the enrichment analysis, ence and the classification into the multiexposure profile we used clusterProfiler Bioconductor package (V.3156). associated with the E1 environment were performed with The commented analysis code is available in Additional the algorithm teff, taking sex as the treatment variable file 3: Supplementary Code. [10] (https://​teff-​packa​ge.​github.​io/). Results Neurodevelopment differences between E1 and E0 Sexual dimorphism of clinical outcomes We used the classification of individuals into E1 and E0 We first assessed whether obesity and the categorized to assess their relationship with sex differences in neu- neuropsychological measures were associated with dif- rodevelopment. For this analysis, we used logistic regres- ferences between sexes (Fig. 1). We fitted logistic regres- sion models on the clinical outcomes (working memory, sion models adjusting by the 10 common covariates. attention and general non-verbal intelligence with the Girls showed a lower frequency of obesity than boys, but Cáceres et al. BMC Medicine (2023) 21:142 Page 7 of 14 it was not statistically significant (OR = 0.64, P = 0.13, buffer during the pregnancy period. We categorized this see Fig. 3A). For the neuropsychological measures, we variable into low (less than 0.05%), moderate (between observed that ADHD was lower in girls than boys, con- 0.05 and 0,12%), and high abundance (more than 0.12%). sistent with girls’ higher protection in attention difficulty. Fourth, the presence of green spaces ­(ORintreraction = 0.27, Both associations were statistically significant (OR = 0.37, P = 0.029), answering the question of whether the mother P = 2.87 × ­10−5, OR = 0.54, P = 4.32 × ­10−4). For Raven’s lived within a distance of 300 m of green space during the matrices and N-back, we did not see significant associa- pregnancy period (yes/no). A stratified analysis by sex of tions with sex (OR = 0.72, P = 0.10, OR = 0.94, P = 0.78, the association between obesity and the significant expo- respectively) (Fig. 3A). sures revealed that dairy consumption and cotinine lev- els were risk factors only for girls (OR = 2.88, P = 0.0009; Exposome‑wide analysis of sex‑exposure interactions OR = 1.91, P = 0.0128) while facility richness and green on obesity spaces were protective and risk factors for boys, respec- We searched for prenatal exposures that could modulate tively (OR = 0.92, P = 0.005; OR = 5.06, P = 0.007), see the association between sex and obesity in childhood. Fig. 4. We finally asked the extent to which the four expo- Particularly, we searched for maternal exposure levels sures were correlated between each other. Interestingly, in which one sex would be more obese than the other we found weak but significant Pearson’s correlations of at 5–11 years of age. We performed logistic regressions facility richness with dairy intake (r =  − 0.11, P = 0.0002) on obesity for all 93 sex-prenatal exposures interactions, and cotinine levels (r = 0.07, P = 0.01). adjusting by the common covariates, sex, and each expo- sure (Fig. 3B, C). We did not observe any interaction that Exposure environment of high differences in obesity risk passed multiple comparison corrections. However, at the between sexes nominal level (P < 0.05), we observed four interactions We asked whether a combination of the four significant between sex (males as reference) and prenatal exposures. exposures and their levels could define specific environ- First, dairy consumption (­ ORintreraction = 2.44, P = 0.008) is ments where one sex is likely more obese than the other defined as mother’s dairy consumption during pregnancy one. The exposure residuals, adjusted by common covari- times per week and categorized as less than 18 times per ates, were used in causal inference modeling, with the week (low), between 18 and 27 (moderate), and more aim to classify individuals into environments of high sex- than 27 (high). Second, cotinine levels in mothers during ual dimorphism in obesity. We considered the multiexpo- pregnancy ­(ORintreraction = 1.92, P = 0.034) are classified sure profile defined by the mother’s dairy intake, cotinine into three categories: non-smokers (less than 18.4 µg/L), levels, living richness facilities, and green spaces during second-hand smokers (between 18.4 and 48.4 µg/L), and pregnancy. We randomly selected a set of 208 individu- smokers (more than 48.4 µg/L). Third, facility richness als from the HELIX cohort to infer their expected sex dif- ­(ORintreraction = 1.11, P = 0.013) is defined as the percent- ference in obesity risk given their personal multiexposure age of different facility types present compared to the profiles. We thus applied the causal modeling algorithm maximum potential number of facility types at a 300-m teff, taking sex as the treatment variable, and observed Fig. 3 A Association of sex with the clinical outcomes, adjusting by covariates. B Number of prenatal exposures in each family measured in HELIX. C Exposome-wide Manhattan plot. Association of obesity with 93 sex-prenatal exposure interactions (the color follows the exposure family from panel A). The dotted line marks nominal significance (P = 0.05) Cáceres et al. BMC Medicine (2023) 21:142 Page 8 of 14 Fig. 4 Sex-exposure interaction plots on obesity. A Mother’s dairy consumption during pregnancy. The figure shows the highest proportion of obesity in girls with the highest level of dairy consumption. B Mother’s cotinine levels in the blood. The highest levels of obesity were observed in girls with high cotinine levels. C Facility richness in living neighborhoods of pregnant mothers. A high abundance of facility richness is correlated with a low prevalence of obesity in boys. D Green spaces at 300 m from pregnant mothers’ homes. The highest prevalence of obesity was observed in boys with mothers living in the presence of green spaces. The bars represent the 95% confidence intervals for the estimated proportion of obesity 27 children (13 females, 14 males) living in personal environments where girls are less likely obese than boys. By contrast, we found only one boy living in a personal environment where girls are more likely obese than boys (Fig. 5). We aimed to classify all individuals into two environ- ments: E1 and E0. The first one (E1) consisted of a com- bination of exposure levels that protects girls against obesity (F < M). The second one (E0) consists of the remaining combinations of exposure levels (F > M and F = M). E1 was obtained using the personal environ- ments of the 27 children where girls are expected to be Fig. 5 The estimated sex difference on obesity risk in personal less obese than boys. E1 was defined as a binary vec- prenatal environments. The personal prenatal exposure environments tor, with one entry for each level of the four exposures, were defined by the mother’s dairy intake, cotinine levels, living indicating whether a given exposure averaged across 27 richness facilities, and green spaces. The sex of the individual living individuals was higher or lower than the average across in a particular prenatal environment is shown in blue (male) and the entire training set of 208 children. We used the mul- orange (female). The bars show the 95%CI for the effect of a personal prenatal environment on females in relation to males. The intervals tiexposure profile to classify all the individuals in HELIX were estimated using causal modeling implemented in teff. Green and observed a total of 675 (64%) individuals classified lines are significant sex differences in obesity risk given by the into E1. We found that E1 was characterized by moderate prenatal environments Cáceres et al. BMC Medicine (2023) 21:142 Page 9 of 14 Fig. 6 Characterization of the common prenatal environment where girls are more protected than boys against obesity (E1) against the reference environment (E0). Environment E1 is the common environment of individuals with personal environments where girls are significantly less obese than boys (female < male), these are the individuals with green confidence intervals in Fig. 5. E1 is defined by low mother dairy intake, non-smokers’ cotinine levels, low richness facilities, and the presence of green spaces. An individual belongs to E0 if he/she does not belong to E1 (female > male or female = male) dairy consumption, non-smokers’ cotinine levels, low Sexual dimorphism in neurodevelopment abundance of facility richness, and the presence of green We asked whether the environment of high differences in spaces (Fig. 6A–D). Therefore, the environment captured obesity between sexes was also an environment of high both obesity protection for girls and obesity risk for boys, differences in neurodevelopment. First, we assessed the as expected from the individual exposures. association between obesity and four neuropsychological We then observed a strong association of the sex- outcomes, fitting logistic regression models on obesity environment interaction on child obesity, adjusting by and adjusting by common covariates, sex, and the envi- covariates ­(ORinteraction = 0.070, P = 2.59 × ­10−5). Strati- ronment (E1/E0). We observed that low values of Raven’s fied associations by sex between the environment and matrices and N-back test tests were significant risk fac- obesity risk were also significant (girls: OR = 0.18, tors for obesity (OR = 2.42, P = 0.01; OR = 2.65, P = 0.02, P = 4.73 × ­10−4; boys: OR = 3.14, P = 0.012), suggesting see Fig. 7B), as ADHD diagnosis increased the risk stronger environment gains in the protection for girls (OR = 2.15, P = 0.03, see Fig. 7C). However, we did not than in the risk for boys (Fig. 7A). These results show find significant associations between obesity and atten- that E1 can be regarded as a prenatal environment of tion outcome. female protection against childhood obesity, with much We tested whether the subject classification into the stronger protection than those given its individual expo- environments E1 and E0 significantly interacted with sex sure components. on each of the neuropsychological outcomes, as it did Fig. 7 A Sex-environment interaction plot on obesity risk. The figure shows that E1 defines a prenatal environment across HELIX of strong female protection against childhood obesity (female < male), while E0 defines a prenatal environment without sexual dimorphism in obesity (F = M) or with male protection (F > M). B Sex-environment interaction plot on raven’s matrices underperformance. Affected individuals are those with outcomes below the first quintiles. C Sex-environment interaction plot on N-back underperformance. Affected individuals are those with outcomes below the first quintiles Cáceres et al. BMC Medicine (2023) 21:142 Page 10 of 14 with obesity. We found that the sex-environment interac- Discussion tion was associated with higher outcomes of both Raven’s We have shown in the HELIX cohort that environments matrices ­(ORinteraction = 0.42, P = 0.047) and N-back test defined by a multiexposure profile with different effects ­(ORinteraction = 0.31, P = 0.02), suggesting a higher per- on obesity for each sex can be identified with the novel formance of girls with respect to boys in these two tests, use of causal inference [10]. In a previous study on the within E1. Associations were fully adjusted by covariates. same cohort, no significant associations were observed for individual prenatal exposures with overweight and Methylation profile associated with the prenatal obesity status, while cotinine levels were associated environment of high sex differences in obesity with BMI only at nominal significance [37]. Although We aimed to investigate whether the methylome captured we observed only four nominally significant interac- the differences between individuals belonging to E1 or tions between prenatal exposures and sex on obesity, we E0. We performed an EWAS of the classification of chil- revealed a prenatal environment defined by specific lev- dren in the prenatal environment, adjusting by common els of these exposures whose effect on obesity strongly covariates and immune cell counts. Methylation data was changed between sexes, with a 93% reduction in obe- extracted from blood samples and were previously nor- sity risk for girls in relation to boys ­(ORinteraction = 0.070, malized and corrected for surrogate variation. We did P = 2.59 × ­10−5). In the environment defined by moder- not observe any significant association at a genome-wide ate dairy consumption, non-smokers’ cotinine levels, low level after correcting for multiple comparisons, see top facility richness, and the presence of green spaces, girls associations in Additional file 2: Table S2. We also per- are more protected than boys against obesity. formed an enrichment analysis for the top associations Previous studies have shown conflictive findings on (nominal P < 0.01). We tested different GO terms from dairy intake during pregnancy and its relation to long- molecular function, cellular components, and biological term body composition of children. Voerman et al. processes (Fig. 8), and observed several pathways related reported significant associations with abdominal fat to neuronal processes. Most remarkably, synapse organi- in children and strong interaction with sex on the peri- zation (P-adjusted = 0.0001) and regulation of synapse cardial fat mass index, with a higher risk for girls [38]. structure or activity (P-adjusted = 0.006) are two biologi- However, other studies have reported no significant cal processes directly related to neurodevelopment. Fig. 8 Enrichment analysis for the differentially methylated sites associated with E1. A Gene Ontology. B Cellular components. C Biological processes. Epigenome-wide analysis for the prenatal environment E1 was performed and methylation probes with associations at P < 0.001 were selected. Probes were mapped to genes that were used in enrichment analyses. The analyses mainly show pathways related to neuronal function Cáceres et al. BMC Medicine (2023) 21:142 Page 11 of 14 associations [39, 40]. Our findings suggest that part of the abilities and ADHD [52, 53]. Animal studies have shown discrepancy could be due to the interaction with sex. that mice whose mothers were on high-fat diets dur- Concerning obesity and cotinine levels in the blood of ing pregnancy have alterations in brain methylation of pregnant mothers, previous studies have shown a 50% dopaminergic and opioid genes [54, 55]. In addition, the increase in childhood overweight for smoking during neurodevelopmental delay appears to be more frequent pregnancy [41], with a dose–response relationship [42]. in obese boys [12]. A longitudinal prospective study has Cotinine levels have also been associated with low birth shown that working memory and attention performance weight but rapid gains in BMI after delivery [43]. In a are reduced by increasing BMI in children [56]. Our Japanese population, Susuki et al. observed that boys of study offers additional evidence of this relationship, since mothers who smoked during pregnancy had higher gains the environmental changes that modulate the association in BMI trajectories compared with girls [44]. We found, between sex and obesity also modulate the association however, higher obesity frequency for girls of mothers between sex and neurodevelopmental delay. Further- with smoker’s cotinine levels. In a large study of ~ 90,000 more, the environment is associated with methylation mother-children pairs, also in Japan [43], they observed probes that are enriched in neurodevelopmental path- that rapid gains in BMI of children were associated with ways, providing more evidence for this hypothesis. urinary cotinine concentration of mothers but not with The generalizability of these results is subject to certain self-reported smoking status. While their results were limitations. First, we did not observe a significant sexual not stratified by sex, it shows that cotinine is a more dimorphism in obesity as expected. In the HELIX popu- accurate assessment of pregnancy smoking. lation, 4.9% of girls were obese contrasting with 6.8% of In relation to green spaces, systematic reviews have boys. Since we found that there is a clear difference, we shown weak evidence for its relationship with children’s suspect that it was not significant due to the low number obesity [45, 46]. Associations of green spaces during of children with obesity (39 boys and 23 girls). By con- pregnancy and their differential effect on sex have not trast, we found significant differences between the sexes been previously assessed. We found that prenatal green for the ADHD diagnosis and the ANT distribution. In space is a risk factor for boys’ obesity only. A recent study this case, the number of children affected was higher than of the HELIX cohort showed significant associations in obesity (77 boys and 27 girls for ADHD and 134 boys between children’s overweight and obese status with the and 72 girls for ANT). Second, the interactions between built environment (land use mix) [47]. Children living prenatal exposures and sex in obesity were significant at in built environments in absence of green spaces could a nominal level but not after correcting by multiple com- be at higher risk of obesity (likely due to its relationship parisons. Again, this could be because of the small sam- with physical activity). However, we observed that a low ple size and the low statistical power, which is especially abundance of facility richness and the presence of green important when evaluating interactions. spaces during pregnancy are risk factors for obesity in Our study also had notable strengths. We confirmed boys. Both environmental conditions of the pregnant that the combination of exposures greatly increased the mother are consistent with less urbanized environments significance of interactions between prenatal exposures where adult obesity may be more frequent [48]. and sex. We also confirmed the importance of the rela- In this study, we observed that the combination of the tionship between the obesogenic prenatal environment specific levels for the four exposures maximizes the dif- and neurodevelopment. We not only found a significant ferences in obesity risk between girls and boys. Previ- sexual dimorphism in neurodevelopment delay when ous studies have already suggested that better prediction comparing E1 and E0, but we also found enrichment in of an outcome can be obtained from the aggregation of neurodevelopmental pathways in the methylation probes multiple environmental factors into risk scores [49, 50] or associated with the environment. This provides a possible the use of mixture models [51]. In line with this, we used molecular mechanism that could explain the association causal inference for classifying the individuals in two between the obesogenic environment and sexual dimor- environments (E0 and E1) based on the combination of phism in neurodevelopment. Moreover, this study evalu- the four exposures. ates the different effects of a prenatal environment in girls After classifying individuals in the two environments, and boys, which is very innovative and important to con- we further investigated whether the individuals belong- sider sex differences in the prenatal exposure guidelines. ing to the environment with higher sexual dimorphism in obesity presented also sexual dimorphism in neurode- Conclusions velopmental delay. Based on previous studies, prenatal We aimed to advance a novel approach to the study of factors, such as maternal obesity, have been seen asso- sexual dimorphism, based on high dimensional expo- ciated with both obesity in children and lower cognitive sure data and recent methods of causal inference. The Cáceres et al. BMC Medicine (2023) 21:142 Page 12 of 14 methodological approach can also be used to deter- EU FP6. STREP Hiwate, EU FP7 ENV.2007.1.2.2.2. Project No 211250 Escape, EU FP7-2008-ENV-1.2.1.4 Envirogenomarkers, EU FP7-HEALTH-2009- single mine the environmental landscape that promotes sexual stage CHICOS, EU FP7 ENV.2008.1.2.1.6. Proposal No 226285 ENRIECO, EU- FP7- dimorphisms in studies with high dimensional exposure HEALTH-2012 Proposal No 308333 HELIX), and the Greek Ministry of Health data. (Program of Prevention of obesity and neurodevelopmental disorders in preschool children, in Heraklion district, Crete, Greece: 2011–2014; “Rhea Plus”: In summary, girls in childhood may be protected Primary Prevention Program of Environmental Risk Factors for Reproductive against obesity if their pregnant mothers had moderate Health, and Child Health: 2012–15). dairy consumption, non-smokers cotinine levels, and This research has received funding from the Spanish Ministry of Science and Innovation through the “Centro de Excelencia Severo Ochoa 2019–2023 lived in environments with a low abundance of rich facili- (CEX2018-000,806-S) program, and support from the Generalitat de Catalunya ties and the presence of green spaces. The environment through the CERCA Program. NC and JU are supported by Spanish regional is also protective against the neurodevelopmental delay program PERIS (Ref.: SLT017/20/000061 and SLT017/20/000119, respectively), granted by Departament de Salut de la Generalitat de Catalunya. TruDiagnos‑ of non-verbal intelligence and working memory. While tics also provided funding for data analysis. female protection is measured against male risk, female protection outweighs the risk of obesity in boys. Our Availability of data and materials Any custom code or software used in our analysis is available at Additional study motivates further public health efforts to raise pub- file 3: Supplementary Code. lic awareness of moderating a high-fat diet and avoiding The HELIX data warehouse has been established as an accessible resource for smoking and second-hand smoking during pregnancy to collaborative research involving researchers external to the project. Access to HELIX data is based on approval by the HELIX Project Executive Commit‑ protect children against obesity and neurodevelopmental tee and by the individual cohorts. Further details on the content of the data delay. warehouse (data catalog) and procedures for external access are described on the project website (http://​www.​proje​cthel​ix.​eu/​index.​php/​es/​data-​inven​ tory). The data used in this analysis are not available for replication because Supplementary Information specific approvals from the HELIX Project Executive Committee and the The online version contains supplementary material available at https://​doi.​ University of Southern California Institutional Review Board must be obtained org/​10.​1186/​s12916-​023-​02815-9. to access them. Additional file 1. Supplementary Methods. Declarations Additional file 2: Table S1. Exposure variables in the HELIX cohort. Table S2. CpG sites with a nominal P-value lower than 0.001 in the Ethics approval and consent to participate epigenome-wide association study for the classification of children in the Local ethical committees approved the studies that were conducted accord‑ prenatal environment (E0 and E1). ing to the guidelines laid down in the Declaration of Helsinki. The ethical com‑ Additional file 3. Supplementary Code. mittees for each cohort were the following: BIB: Bradford Teaching Hospitals NHS Foundation Trust, EDEN: Agence nationale de sécurité du médicament et des produits de santé, INMA: Comité Ético de Inverticación Clínica Parc de Acknowledgements Salut MAR, KANC: LIETUVOS BIOETIKOS KOMITETAS, MoBa: Regional komité We are grateful to all participants and researchers who took part in this study. for medisinsk og helsefaglig forskningsetikk, Rhea: Ethical committee of the general university hospital of Heraklion, Crete. Informed consent was obtained Authors’ contributions from a parent and/or legal guardian of all participants in the study. AC1 and JRG conceived the study. AC1 designed the analysis. AC1 and NC-G performed statistical analyses and wrote the first draft of the manuscript. Consent for publication JRG supervised the study and contributed to reviewing and editing the Not applicable. manuscript. MV coordinates the HELIX project, JU is the data manager, and LM is the scientific coordinator. MB, JW, RSl, MC, JRG, and MV designed the omics Competing interests study in HELIX. The following authors participated in omics data acquisition VBD, RS, HW, and TLM are employees of TruDiagnostic. JRG has received fund‑ and quality control: MB (DNA methylation), AC2 (DNA methylation), MJN ing from TruDiagnostic as a scientific advisor. The other authors have nothing (exposome), and CT (exposome). JW, RG, MV, RSl, LC, and CT are the PIs of the to declare. cohorts. TY, SA, MC, JL, NS, and KG participated in sample and data acquisition. AC1 and NC-G wrote the original draft of the paper and JRG, MC, KG, VBD, RS, Author details 1 HW, and TLM contributed to reviewing and editing the manuscript. All authors Instituto de Salud Global de Barcelona (ISGlobal), 08003 Barcelona, Spain. 2 read and approved the final version of the manuscript. Centro de Investigación Biomédica en Red en Epidemiología Y Salud Pública (CIBERESP), Madrid, Spain. 3 Department of Mathematics, Escola d’Enginyeria Funding de Barcelona Est (EEBE), Universitat Politècnica de Catalunya, 08019 Barcelona, The study has received funding from the European Community’s Seventh Spain. 4 Department of Environmental Science, Vytautas Magnus University, Framework Programme (FP7/2007–2013) under grant agreement no 308333 44248 Kaunas, Lithuania. 5 Department of Health and Experimental Sciences, (HELIX project); and the H2020-EU.3.1.2.—Preventing Disease Programme Universitat Pompeu Fabra (UPF), Barcelona, Spain. 6 Center for Genomic Regu‑ under grant agreement no 874583 (ATHLETE project). lation (CRG), Barcelona, Institute of Science and Technology (BIST), Barcelona, BiB received core infrastructure funding from the Wellcome Trust Spain. 7 Medicine Genomics Group, Centro de Investigación Biomédica en Red (WT101597MA) and a joint grant from the UK Medical Research Council (MRC) Enfermedades Raras (CIBERER), CIMUS, University of Santiago de Compostela, and Economic and Social Science Research Council (ESRC) (MR/N024397/1). Santiago de Compostela, Spain. 8 Galician Foundation of Genomic Medi‑ INMA-SAB data collections were supported by grants from the Instituto de cine, Instituto de Investigación Sanitaria de Santiago de Compostela (IDIS), Salud Carlos III, CIBERESP, and the Generalitat de Catalunya-CIRIT. KANC was Servicio Gallego de Salud (SERGAS), Galicia, Santiago de Compostela, Spain. 9 funded by the grant of the Lithuanian Agency for Science Innovation and Department of Preventive Medicine, Keck School of Medicine, University Technology (6–04-2014_31V-66). The Norwegian Mother, Father and Child of Southern California, Los Angeles, USA. 10 TruDiagnostic, Lexington, KY, USA. 11 Cohort Study is supported by the Norwegian Ministry of Health and Care Division of Climate and Environmental Health, Norwegian Institute of Public Services and the Ministry of Education and Research. The Rhea project was Health, 0456 Oslo, Norway. 12 Institut National de La Santé Et de La Recherche financially supported by European projects (EU FP6-2003-Food-3-NewGeneris, Médicale (Inserm) and Université Grenoble-Alpes, Institute for Advanced Cáceres et al. BMC Medicine (2023) 21:142 Page 13 of 14 Biosciences (IAB), Team of Environmental Epidemiology Applied to Reproduc‑ 21. De Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann J. tion and Respiratory Health, Grenoble, France. 13 Bradford Institute for Health Development of a WHO growth reference for school-aged children and Research, Bradford Teaching Hospitals NHS Foundation Trust, Bradford, UK. adolescents. Bull World Health Organ. 2007;85:660–7. 14 Department of Mathematics, Universitat Autònoma de Barcelona, Bellaterra, 22. Vuontela V, Steenari MR, Carlson S, Koivisto J, Fjällberg M, Aronen ET. 08193 Barcelona , Spain. Audiospatial and visuospatial working memory in 6–13 year old school children. Learn Mem. 2003;10:74–81. Received: 2 November 2022 Accepted: 6 March 2023 23. Rueda MR, Fan J, McCandliss BD, Halparin JD, Gruber DB, Lercari LP, et al. Development of attentional networks in childhood. Neuropsy‑ chologia. 2004;42:1029–40. 24. Raven JC, Raven J. Progressive matrices couleur/colored progressive matrices. Paris: Centre de Psychologie Appliquée; 1998. References 25. Maitre L, Guimbaud J-B, Warembourg C, Güil-Oumrait N, Petrone PM, 1. Muenchhoff M, Goulder PJR. Sex differences in pediatric infectious Chadeau-Hyam M, et al. State-of-the-art methods for exposure-health diseases. J Infect Dis. 2014;209(Suppl 3):S120. studies: Results from the exposome data challenge event. Environ Int. 2. De Bellis MD, Keshavan MS, Beers SR, Hall J, Frustaci K, Masalehdan A, 2022;168: 107422. et al. Sex differences in brain maturation during childhood and adoles‑ 26. Wild CP. Complementing the genome with an “‘Exposome’”: the cence. Cereb Cortex. 2001;11:552–7. outstanding challenge of environmental exposure measure‑ 3. Shah B, Tombeau Cost K, Fuller A, Birken CS, Anderson LN. Sex and ment in molecular epidemiology. Cancer Epidemiol Biomarkers gender differences in childhood obesity: contributing to the research Prev. 2005;14:1847–50. agenda. BMJ Nutr Prev Heal. 2020;3:387–90. 27. Haug LS, Sakhi AK, Cequier E, Casas M, Maitre L, Basagana X, et al. In- 4. Lobstein T, Brinsden H. Atlas of childhood obesity. World Obes Fed. utero and childhood chemical exposome in six European mother-child 2019;211. https://​s3-​eu-​west-1.​amazo​naws.​com/​wof-​files/​11996_​Child​ cohorts. Environ Int. 2018;121(Pt 1):751–63. hood_​Obesi​ty_​Atlas_​Report_​ART_​V2.​pdf. 28 Serra-Majem L, Ribas L, Ngo J, Ortega RM, García A, Pérez-Rodrigo C, 5. Stachenfeld NS, Mazure CM. Precision medicine requires understanding et al. Food, youth and the Mediterranean diet in Spain. Development how both sex and gender influence health. Cell. 2022;185:1619–22. of KIDMED, Mediterranean Diet Quality Index in children and adoles‑ 6. Vrijheid M, Slama R, Robinson O, Chatzi L, Coen M, van den Hazel P, et al. cents. Public Health Nutr. 2004;7:931–5. The human early-life exposome (HELIX): project rationale and design. 29. Boyce W, Torsheim T, Currie C, Zambon A. The family affluence scale as Environ Health Perspect. 2014;122:535–44. a measure of national wealth: validation of an adolescent self-report 7. Zhang P, Carlsten C, Chaleckis R, Hanhineva K, Huang M, Isobe T, et al. measure. Soc Indic Res. 2006;78:473–87. Defining the scope of exposome studies and research needs from a 30. Kritsotakis G, Koutis AD, Alegakis AK, Philalithis AE. Development multidisciplinary perspective. Environ Sci Technol Lett. 2021;8:839–52. of the Social Capital Questionnaire in Greece. Res Nurs Health. 8. Mauvais-Jarvis F, BaireyMerz N, Barnes PJ, Brinton RD, Carrero JJ, DeMeo 2008;31:217–25. DL, et al. Sex and gender: modifiers of health, disease, and medicine. 31 White IR, Royston P, Wood AM. Multiple imputation using chained Lancet (London, England). 2020;396:565. equations: Issues and guidance for practice. Stat Med. 2011;30:377–99. 9. Sadovsky Y, Mesiano S, Burton GJ, Lampl M, Murray JC, Freathy RM, et al. 32. van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation Advancing human health in the decade ahead: pregnancy as a key by chained equations in R. J Stat Softw. 2011;45:1–67. window for discovery: a Burroughs Wellcome Fund Pregnancy Think Tank. 33. Tamayo-Uria I, Maitre L, Thomsen C, Nieuwenhuijsen MJ, Chatzi L, Am J Obstet Gynecol. 2020;223:312–21. Siroux V, et al. The early-life exposome: description and patterns in six 10. Cáceres A, González JR. teff: estimation of Treatment EFFects on transcrip‑ European countries. Environ Int. 2019;123:189–200. tomic data using causal random forest. Bioinformatics. 2022. https://​doi.​ 34 Carreras-Gallo N, Cáceres A, Balagué-Dobón L, Ruiz-Arenas C, Andru‑ org/​10.​1093/​BIOIN​FORMA​TICS/​BTAC2​69. saityte S, Carracedo Á, et al. The early-life exposome modulates the 11. Miller AL, Lee HJ, Lumeng JC. Obesity-associated biomarkers and execu‑ effect of polymorphic inversions on DNA methylation. Commun Biol. tive function in children. Pediatr Res. 2015;77:143–7. 2022;5:1–13. 12. Wentz E, Björk A, Dahlgren J. Neurodevelopmental disorders are highly 35 Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, over-represented in children with obesity: a cross-sectional study. Obe‑ Nelson HH, et al. DNA methylation arrays as surrogate measures of cell sity (Silver Spring). 2017;25:178–84. mixture distribution. BMC Bioinforma. 2012;13:1–16. 13. Perera F, Herbstman J. Prenatal environmental exposures, epigenetics, 36 Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlén S-E, Greco D, et al. and disease. Reprod Toxicol. 2011;31:363–73. Differential DNA methylation in purified human blood cells: implica‑ 14. Wright J, Small N, Raynor P, Tuffnell D, Bhopal R, Cameron N, et al. Cohort tions for cell lineage and studies on disease susceptibility. PLoS One. profile: the born in bradford multi-ethnic family cohort study. Int J Epide‑ 2012;7:e41361. miol. 2013;42:978–91. 37. Vrijheid M, Fossati S, Maitre L, Márquez S, Roumeliotaki T, Agier L, et al. 15. Heude B, Forhan A, Slama R, Douhaud L, Bedel S, Saurel-Cubizolles MJ, Early-life environmental exposures and childhood obesity: an exposome- et al. Cohort Profile: the EDEN mother-child cohort on the prenatal and wide approach. Environ Health Perspect. 2020;128:1–14. early postnatal determinants of child health and development. Int J 38. Voerman E, Gaillard R, Geurtsen ML, Jaddoe VWV. Maternal first-trimester Epidemiol. 2016;45:353–63. cow-milk intake is positively associated with childhood general and 16. Guxens M, Ballester F, Espada M, Fernández MF, Grimalt JO, Ibarluzea J, abdominal visceral fat mass and lean mass but not with other cardio‑ et al. Cohort profile: the INMA-INfancia y Medio Ambiente-(environment metabolic risk factors at the age of 10 years. J Nutr. 2021;151:1965–75. and childhood) project. Int J Epidemiol. 2012;41:930–40. 39. Hrolfsdottir L, Rytter D, Hammer Bech B, Brink Henriksen T, Danielsen I, 17. Grazuleviciene R, Danileviciute A, Dedele A, Vencloviene J, Andrusaityte S, Steingrimsdottir L, et al. Maternal milk consumption, birth size and adult Uždanaviciute I, et al. Surrounding greenness, proximity to city parks and height of offspring: a prospective cohort study with 20 years of follow-up. pregnancy outcomes in Kaunas cohort study. Int J Hyg Environ Health. Eur J Clin Nutr. 2013;67:1036–41. 2015;218:358–65. 40. Leary SD, Ness A, Emmett P, Smith GD. Maternal diet in pregnancy and 18. Magnus P, Birke C, Vejrup K, Haugan A, Alsaker E, Daltveit AK, et al. Cohort offspring height, sitting height, and leg length. J Epidemiol Community profile update: the Norwegian Mother and Child Cohort Study (MoBa). Health. 2005;59:467–72. Int J Epidemiol. 2016;45:382–8. 41. Oken E, Levitan EB, Gillman MW. Maternal smoking during pregnancy 19. Chatzi L, Leventakou V, Vafeiadi M, Koutra K, Roumeliotaki T, Chalkiadaki and child overweight: systematic review and meta-analysis. Int J Obes G, et al. Cohort profile: the Mother-Child Cohort in Crete, Greece (Rhea (Lond). 2008;32:201–10. Study). Int J Epidemiol. 2017;46:1392–3. 42. Møller SE, Ajslev TA, Andersen CS, Dalgård C, Sørensen TIA. Risk of child‑ 20. Maitre L, De Bont J, Casas M, Robinson O, Aasvang GM, Agier L, et al. hood overweight after exposure to tobacco smoking in prenatal and Human Early Life Exposome (HELIX) study: a European population-based early postnatal life. PLoS One. 2014;9(10):e109184. exposome cohort. BMJ Open. 2018;8: e021311. Cáceres et al. BMC Medicine (2023) 21:142 Page 14 of 14 43. Hirai H, Okamoto S, Masuzaki H, Murata T, Ogata Y, Sato A, et al. Maternal urinary cotinine concentrations during pregnancy predict infant BMI trajectory after birth: analysis of 89617 mother-infant pairs in the Japan environment and children’sstudy. Front Endocrinol (Lausanne). 2022;13:850784. 44. Suzuki K, Kondo N, Sato M, Tanaka T, Ando D, Yamagata Z. Gender differ‑ ences in the association between maternal smoking during pregnancy and childhood growth trajectories: multilevel analysis. Int J Obes (Lond). 2011;35:53–9. 45. Jia P, Cao X, Yang H, Dai S, He P, Huang G, et al. Green space access in the neighbourhood and childhood obesity. Obes Rev. 2021;22(Suppl 1 Suppl):1. 46. Lachowycz K, Jones AP. Greenspace and obesity: a systematic review of the evidence. Obes Rev. 2011;12(5):e183–9. 47. de Bont J, Márquez S, Fernández-Barrés S, Warembourg C, Koch S, Persavento C, et al. Urban environment and obesity and weight-related behaviours in primary school children. Environ Int. 2021;155:106700. 48. Okobi OE, Ajayi OO, Okobi TJ, Anaya IC, Fasehun OO, Diala CS, et al. The burden of obesity in the rural adult population of America. Cureus. 2021;13(6):e15770. 49. Padmanabhan JL, Shah JL, Tandon N, Keshavan MS. The “polyenviromic risk score”: aggregating environmental risk factors predicts conversion to psychosis in familial high-risk subjects. Schizophr Res. 2017;181:17–22. 50. Jeon EJ, Kang SH, Piao YH, Kim SW, Kim JJ, Lee BJ, et al. Development of the Korea-Polyenvironmental Risk Score for Psychosis. Psychiatry Investig. 2022;19:197–206. 51. Güil-Oumrait N, Cano-Sancho G, Montazeri P, Stratakis N, Warembourg C, Lopez-Espinosa MJ, et al. Prenatal exposure to mixtures of phthalates and phenols and body mass index and blood pressure in Spanish preadoles‑ cents. Environ Int. 2022;169: 107527. 52. Edlow AG. Maternal obesity and neurodevelopmental and psychiatric disorders in offspring. Prenat Diagn. 2017;37:95–110. 53. Contu L, Hawkes CA. A review of the impact of maternal obesity on the cognitive function and mental health of the offspring. Int J Mol Sci. 2017;18(5):1093. 54. Grissom NM, Herdt CT, Desilets J, Lidsky-Everson J, Reyes TM. Dissociable deficits of executive function caused by gestational adversity are linked to specific transcriptional changes in the prefrontal cortex. Neuropsy‑ chopharmacology. 2015;40:1353–63. 55. Vucetic Z, Kimmel J, Totoki K, Hollenbeck E, Reyes TM. Maternal high-fat diet alters methylation and gene expression of dopamine and opioid- related genes. Endocrinology. 2010;151:4756–64. 56. Li N, Yolton K, Lanphear BP, Chen A, Kalkwarf HJ, Braun JM. Impact of early-life weight status on cognitive abilities in children. Obesity (Silver Spring). 2018;26:1088–95. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions