Skip to main content
Download PDF

A meta-analysis of epigenome-wide association studies of ultra-processed food consumption with DNA methylation in European children

Clinical Epigenetics volume 17, Article number: 3 (2025) Cite this article

Abstract

Background/objective

There is limited knowledge on how diet affects the epigenome of children. Ultra-processed food (UPF) consumption is emerging as an important factor impacting health, but mechanisms need to be uncovered. We therefore aimed to assess the association between UPF consumption and DNA methylation in children.

Methods

We conducted a meta-analysis of epigenome-wide association studies (EWAS) from a total of 3152 children aged 5–11 years from four European studies (HELIX, Generation XXI, ALSPAC, and Generation R). UPF consumption was defined applying the Nova food classification system (group 4), and DNA methylation was measured in blood with Illumina Infinium Methylation arrays. Associations were estimated within each cohort using robust linear regression models, adjusting for relevant covariates, followed by a meta-analysis of the resulting EWAS estimates.

Results

Although no CpG was significant at FDR level, we found suggestive associations (p-value < 10–5) between UPF consumption and methylation at seven CpG sites. Three of them, cg00339913 (PHYHIP), cg03041696 (intergenic), and cg03999434 (intergenic), were negatively associated, whereas the other four, cg14665028 (NHEJ1), cg18968409 (intergenic), cg24730307 (intergenic), and cg09709951 (ATF7), were positively associated with UPF intake. These CpGs have been previously associated with health outcomes such as carcinomas, and the related genes are mainly involved in pathways related to thyroid hormones and liver function.

Conclusion

We only found suggestive changes in methylation at 7 CpGs associated with UPF intake in a large EWAS among children: although this shows a potential impact of UPF intake on DNAm, this might not be a key mechanism underlying the health effects of UPFs in children. There is a need for more detailed dietary assessment in children studies and of intervention studies to assess potential epigenetic changes linked to a reduction in UPF in the diet.

Graphical abstract

Introduction

Ultra-processed foods (UPFs) are industrially formulated products that undergo a series of physical, chemical, and biological processing. UPFs often lack intact healthy food components, contain various additives, and are energy dense and nutritionally poor (i.e. high in free sugar, salt, and saturated and trans fats but low in dietary fibre, and micronutrients) [1] and usually have softer textures [2]. Their consumption has risen dramatically in the last decades among countries worldwide. UPF intake (percentage of total calories consumed) in the adult population ranges from 15 to 57% in Romania and in the UK, respectively [3]. In school-aged children, it was estimated that UPF comprised 65.4% of daily calorie intake in a representative sample in the UK (4–10 years) [4], 66% in US children [5], while it was lower in Belgian children aged 3–9 years (33.3%) [6] and in Colombia (19%) [7].

The consumption of UPF has been associated with a range of adverse health outcomes in adults, as shown in a recent umbrella review [8]. In children, UPF intake has also been associated with an increase in energy intake and weight gain [9] and the modulation of biological pathways related to adiposity accumulation [10, 11], increasing risk of obesity-related diseases [12, 13], and insulin resistance [14]. The rise in obesity prevalence in children [15] is a significant public health concern with far-reaching ramifications, as childhood obesity tracks into adolescence and adulthood [16]. Obesity is associated with an elevated risk of the development of non-communicable diseases (NCDs) [17], in particular cardiometabolic diseases, type 2 diabetes, and some cancers [18].

Epigenetic mechanisms have been implicated in the development of obesity, with studies showing that early-life obesity risk factors, such as birthweight, maternal BMI, and rapid weight gain during infancy, are associated with variations in DNA methylation (DNAm) [19,20,21]. Given the role of epigenetics in obesity, nutritional epigenetics can play an important role in unravelling the effect of nutrition in the regulation of gene expression implicated in the aetiology of disorders such as obesity [22]. Recent epigenome-wide associations studies (EWAS) in adults found differentially methylated CpGs associated with diet quality, using measures of diet quality such as Mediterranean-style diet scores [23, 24]. A large proportion of these CpGs associated with diet quality have been linked to chronic diseases, in particular cardiometabolic diseases in adults [23, 24]. Other studies have investigated the intake of specific nutrients, such as fat and carbohydrate [25], or specific foods (coffee and tea) in relation to DNAm. However, the number of identified CpGs in these studies was comparatively smaller [26].

Existing diet-EWASs have been mainly focused on diet quality and were conducted in adults [23,24,25,26], newborns, or mothers [27, 28]. The few published EWASs in children so far have focused on glycaemic index [29] or on parental diet quality [30], but none have specifically examined UPF intake.

To explore the potential epigenetic changes associated with UPF consumption, we conducted the first meta-EWAS of UPF intake in relation to blood DNA methylation, in children aged between 5 and 11 years (middle childhood) from four European studies representing nine birth cohorts.

Materials and methods

Participating cohorts

We used data from four studies (Human Early-Life Exposome [HELIX], the Avon Longitudinal Study of Parents And Children [ALSPAC], Generation XXI, and Generation R) including a total of nine longitudinal birth cohorts that started with the recruitment of women during pregnancy or at delivery and followed up their children at different time points. The HELIX study comprises six sub-cohorts. All the children included were in overall healthy condition and did not have any congenital malformations.

The participants in cohorts from the HELIX sub-cohort were recruited between 2003 and 2010 from six birth cohorts across Europe [31]: BIB (Born in Bradford, UK [32, 33], EDEN (Study of determinants of pre- and postnatal development, France) [34], INMA (Environment and Childhood, Spain) [35], KANC (Kaunas Cohort, Lithuania) [36], MoBa (The Norwegian Mother and Child Cohort Study, Norway (Oslo region)) [37], and RHEA (Mother–Child Cohort in Crete, Greece) [38]. The follow-up period of HELIX children used in the present study was from 2013 to 2016 and included children between 5 and 11 years. All were of European ancestry, except in the BIB study which included 50% of children of South Asian ethnicity (n = 115). The Generation XXI (G21) study [39] recruited mothers living in the Porto metropolitan region, Portugal, in 2005–2006, and evaluated diet and blood DNAm of children of diverse ethnicity aged 9–11 years in 2015–2017. Mothers included in the ALSPAC study, in the UK [40, 41], were recruited between 1990 and 1992 with a total of 14,203 mothers, and we used the data of the children (G1 children) who attended examination at 7 years old. Finally, the Generation R (GenR) study was carried out in Rotterdam, included mothers recruited between April 2002 and January 2006. Child data used in the present study were from the follow-up when children were 10 years old [42].

Ethical approval for the study was granted by the Research Ethics Committees of each participating centre of each cohort. Full details are listed in the Supplementary Material S1.

Dietary assessment

Dietary data of the HELIX children were collected via a semi-quantitative food frequency questionnaire (FFQ) covering the child’s habitual diet, which was filled in by the parent attending the examination appointment. The FFQ, developed by the HELIX research group, was translated to the corresponding country’s languages for each participating sub-cohorts of the HELIX project [31]. It included 43 questions about the consumption of various food and drink items [14].

Generation XXI dietary data were obtained from 2-day or 3-day food diaries (1 or 2 weekdays and 1 weekend day) filled by the main caregiver [43]. The codification process of food diaries was conducted by a team of trained nutritionists, using an age-specific food coding manual described elsewhere [44]. Energy and nutrient intakes were estimated using the software Food Processor SQL (2004–2005 ESHA Research) linked with the Food Composition Table of the US Department of Agriculture, and adapted with information of the Portuguese Food Composition Table at PortFir (http://portfir.insa.pt), and data from the European Food Information Resource (EuroFIR) network databases.

Dietary data in ALSPAC children were obtained using a 3-day food diary to record all food and beverage items the child consumed for 2 weekdays and 1 weekend day (not necessarily consecutive) which were completed by the caregivers [45]. Dietary data were reviewed by a nutritionist, and nutrient intakes were coded using the DIDO (Diet In, Data Out) computer program and linked to the 5th McCance and Widdowson British food composition tables [46]. Validity of dietary reporting was calculated using an individualized method based on the ratio of energy intake to estimated energy requirement and its 95% confidence interval.

In GenR, dietary intake was assessed using a validated 71-item semi-quantitative FFQ, completed by the parents of the child, reflecting the last 4 weeks of intake [47]. Information on frequencies, types, and portion sizes was converted into grams of individual food items per day based on standard Dutch portion sizes, using SAS VoVris (Vovris V2.4, TNO, 1999–2006).

To determine UPF intake, we identified foods and drinks as ‘ultra-processed’ by using the Nova classification (group 4), a food classification system based on the nature, extent, and purpose of industrial food processing [48]. To better capture dietary habits in each country, food group classifications are slightly different in each cohort, see Supplemental Table S1. The consumption of UPF for each child was defined as a continuous variable describing the intake of food group 4 relative to the total consumption of all food groups, either in weight (servings) or energy. In HELIX and GenR, UPF intake was expressed as the daily proportion of UPF serving (%) to the total daily sum of all food and drink servings as described in a recent article [14, 49], whereas in ALSPAC [15] and G21 [50], we used the proportion (%) of the total daily energy intake (kcal) derived from UPFs.

DNA methylation

DNA methylation was measured in peripheral whole blood. HELIX, ALSPAC, and GenR used the Illumina Infinium HumanMethylation450 BeadChip array, which measures DNAm at 485,512 CpG sites and G21 used the Illumina Infinium MethylationEPIC BeadChip array that measures DNAm at 866,836 CpG. We used normalized beta values, ranging from 0 (fully unmethylated) to 1 (fully methylated). Further details on data normalization and quality control of DNAm in each cohort are given in Supplemental Material S2. In HELIX, ALSPAC, and G21, the array batch effect was corrected using the ComBat R package [51], while in GenR it was corrected by adjusting for sample plate. To reduce the influence of extreme methylation levels at individual samples in analysis, we winsorized methylation levels at each site. Winsorizing limits extreme values in a dataset and reduces the effect of spurious outliers by setting all data within the specified threshold. All the data below or above the 5th and 95th were set to those percentiles. White blood cell proportions (CD4 + and CD8 + T-cells, natural killer (NK) cells, monocytes, eosinophils, neutrophils, and B-cells) were estimated using the Houseman algorithm [52] and the Reinius reference panel [53].

Covariates

The basic covariates considered were child’s age at time of follow-up, sex (categorized as male or female), and ethnicity (white European or non-white European children), assessed though questionnaires [10, 31,32,33, 49]. In the case of HELIX, most of the ethnicity heterogeneity comes from the BIB sub-cohort; therefore, HELIX models were adjusted by sub-cohort instead of ethnicity.

Maternal characteristics likely to affect either DNAm or diet related outcomes of the child [54, 55] were maternal age (in years) at birth (HELIX and G21) [31, 56] or at pregnancy (ALSPAC and GenR) [41, 57], maternal early-pregnancy body mass index (BMI) (kg/m2) with weight and height being measured, and smoking status during pregnancy (active smoking yes or no) as covariates. Maternal education in HELIX was defined as low, medium, and high [14]; in G21 based on the years of completed schooling as < 9y, 9-12y, and > 12y [49]; in ALSPAC as low (Certificate of Secondary Education, Vocational or Ordinary- (O-) level, educational qualifications generally obtained at 17 years of age), intermediate (Advanced- (A-) level (subject-specific qualifications generally obtained at age 18 years and required for university entry)), and high (university degree and above) [49]; in Gen R as low (no/primary education), intermediate (secondary school, vocational training), and high (bachelor’s degree, university) categories [42]. Maternal education was used as a proxy of socioeconomic status.

We also included as covariates: child’s BMI [58] and physical inactivity. BMI was calculated in kg/m2 based on weight (kg) and height (m2) measured at examination. Sedentary behaviour [59] was assessed by asking the carers about the minutes/day their child spent watching TV, playing on the computer, or other sedentary games in HELIX [14] and in GenR [60], as minutes/day without doing sports in G21 [49], and in ALSPAC categorized as 1 h, 1–2 h or > 3 h average time per day watching TV [49]. Vegetable and fruit intake (from the FFQ or food diaries) were also used as covariates as a proxy of overall diet quality. This was done to address if there are intrinsic associations of methylation with UPF independently of diet quality [61]. Vegetable intake was categorized as low (< 6 serving/week), medium (6–8.5/week), and high (> 8.5/week) and fruit intake as low (< 7 servings/week), medium (7–14.1 servings/week), and high (> 14.1 servings/week) in HELIX [31] and G21 [62]. In ALSPAC and in Gen R, vegetable and fruit intakes were continuous in grams/week [45, 47].

Epigenome-wide association analysis (EWAS)

We included children who had available data on UPF intake (exposure) and whole-blood DNA methylation (outcome). UPF intake, covariates, and DNA methylation were assessed at the same time point, so this study has a cross-sectional design. Prior to conducting the analyses (Fig. 1), a data analysis plan including details of the variable definition, the statistical modelling, and a sample R code was distributed to the participating cohorts. All statistical analyses were carried out using R statistical software 4.3 version. To estimate the association between UPFs intake (exposure) and methylation at each CpG site (outcome), robust linear regression models (rlm() from MASS package) were employed. The models incorporated several covariates based on previous literature as described above and the Houseman 7-cell type proportions. In each cohort, nested models were fitted as follows:

  1. (1)

    Model 1: DNA methylation ~ UPF intake + HELIX sub-cohort or ethnicity + child sex + child age + cell type proportions

  2. (2)

    Model 2: model 1 + maternal age + maternal BMI + maternal smoking + maternal education

  3. (3)

    Model 3: model 2 + child BMI + child sedentary behaviour

Fig. 1
figure 1

Flowchart of meta-EWAS

Full size image

In sensitivity analysis, we further adjusted for fruit and vegetable intake. Ethnicity was also considered, and children of non-European ancestry a minority in our study population were removed from the analysis (only in HELIX and G21) to ensure that any observed associations were not confounded by this factor. Therefore, two additional models were conducted:

  1. •

    Model 3 + fruit intake + vegetable intake

  2. •

    Model 3 restricted to white European ancestry children

Quality control and meta-analysis

We performed the quality control of the cohort-specific EWAS results using the EASIER R package at ISGlobal (https://github.com/isglobal-brge/EASIER/). We excluded control probes, non-CpG probes, probes that mapped to X/Y chromosomes, probes with poor base pairing quality (lower than 40 on 0–60 scale), probes with non-unique 30 bp, subsequence (with cross-hybridizing problems), Infinium II probes with SNPs of global MAF over 1% affecting the extension base, and probes with a SNP in the extension base that causes a colour channel switch from the official annotation [63]. We also looked at the QQ plots as well as the inflation factor lambda (λ) and its confidence intervals.

To identify differentially methylated positions, EWAS results from each cohort were meta-analysed using inverse-variance weighted fixed-effects model. Heterogeneity was assessed using the I2 statistics [64] for each CpG site, considering low heterogeneity an I2 < 25%, medium 25– < 50%, and high ≥ 50%. We also performed random-effect model as sensitivity analyses. The meta-analysis was performed using GWAMA software in the EASIER package using the pcent parameter at 75% therefore only those probes that were available in at least 3 out of 4 cohorts were included in the meta-analysis. Only common probes for autosomal sites between the 450 K and EPIC arrays were considered, resulting in 408,712CpG sites to be tested in a total of 3152 children in the main analysis.

To account for multiple hypothesis testing, we applied the Benjamin–Hochberg false discovery rate (FDR) correction; an FDR-corrected p value < 0.05 denoted statistical significance [65]. A less stringent p-value threshold (pSuggestive) of P < 1 × 10−5 was used to select the top CpGs, as done in previous EWASs [66, 67]. Additionally, a shadow meta-analysis was conducted by an independent researcher using the metafor R package to confirm the obtained results. Finally, we conducted a leave-one-out analysis, in which we ran the main models with one of the cohorts removed each time, to explore if any of the studies was disproportionately influencing our results.

Differentially methylated regions (DMRs)

To address the correlated structure of DNA methylation patterns, a regional analysis was conducted to identify differentially methylated regions (DMRs) associated with UPF intake in the main model in each individual cohort. The analysis was performed using the dmrff package in R, and all probes were included irrespective of their adjusted p-values. DMRs were defined based on the following criteria:

  1. (i)

    Distance between two neighbouring probes was at most 500 base pairs.

  2. (ii)

    The FDR-adjusted p-value for the region was < 0.05.

  3. (iii)

    A DMR must include at least two CpG sites.

Genomic and biologic context

We inspected the genomic surrounding of those CpGs that surpassed the suggestive threshold (P < 1 × 10−5) by looking at their annotation (i.e. chromosome, position, nearest gene, if any) according to the human genome assembly GRCh37/hg19 provided by Illumina R package. Additionally, to assess whether methylation levels of CpGs were associated with the expression levels of nearby genes in child blood, we consulted the HELIX Expression Quantitative Trait Methylation (eQTM) catalog and considered significant CpG-eQTM pairs those with a p-value < 10–5 [68].

To identify whether there were any existing associations between these CpG sites and exposures or health outcomes, we also consulted the EWAS Catalog [69]—a manually curated databases of CpG–trait associations (with P-value < 1 × 10–4) from published EWAS—and the EWAS atlas—a manually curated knowledgebase of EWAS associations (P-value < 1 × 10–4 or adjusted P-value < 0.05) [70].

Functional enrichment analyses

To gain biological insight, we performed a series of functional enrichment analyses. To allow enough CpGs-genes to be included, a more lenient cut-off was used: p-value of 10–4 in model 3. Molecular functions and pathways that may be implicated in the associations between DNA methylation and high UPF intake were identified using Gene Ontology (GO) terms and pathways of the Kyoto Encyclopedia of Genes, Genomes (KEGG) using the missMethyl method [71], and protein complex-based sets (P) available in the Consensus Path database [72]. This was implemented in the Functional Enrichment module of the EASIER R package. We used eFORGE version 2.0 [73] to test for enrichment of CpGs (p-value < 10–4) for DNase in hypersensitive sites, 15 chromatin states, and 5 histone marks in specific tissues and cells.

Results

Cohort characteristics

The characteristics of the participating cohorts are summarized in Table 1. Additionally, HELIX individual sub-cohort’s descriptive data are provided in Table S2. Our study population consisted of 1138 children from the HELIX study, 732 children from the G21 cohort, 892 children from the ALSPAC cohort, and 390 children from the GenR Study, resulting in a total of 3152 participants.

Table 1 Characteristics of mothers and children in all participating cohorts
Full size table

The average age was between 5 and 11 years overall. Most of the children were of white European ancestry (95%). The distribution of UPFs intake varied across the study cohorts. UPF intake contributed on average 24% in HELIX and 36% in GenR to the total food consumption in servings, and 34% in G21 and 61% in ALSPAC to total energy intake. Mothers were on average 30 years old at the time of pregnancy in all cohorts and their early-pregnancy BMI ranged from ~ 22 kg/m2 in ALSPAC to ~ 25 kg/m2 in HELIX, G21, and GenR. We observed the lowest percentage of highly educated mothers in ALSPAC (23%) and the highest in HELIX (51%). Maternal smoking during pregnancy ranged from 12% in GenR to 20% in G21. All cohorts have complete case data except for G21, where there is a 10.8% of missing data on fruit and vegetable intake.

Epigenome-wide association study meta-analysis

The percentage of probes excluded in each cohort ranged between 3 and 10% (Table S3). The lambdas (measure of p-value inflation) of different models in each cohort and in the meta-analysis ranged from 0.97 to 1.01 (Table S4).

No CpG reached the FDR significance threshold in association with UPF intake (see Manhattan plot Figure S1). However, as reported in Table 2, four CpGs reached the suggestive significance level (p-value < 10–5) with low heterogeneity between cohorts (I2 < 50%) and consistent direction of effect. These results are also consistent with the random-effect model, as shown in Table S5.

Table 2 Fixed effect meta-analysis for the top CpGs associated with UPF intake in children
Full size table

Two of these suggestive CpGs were annotated to genes: (cg14665028, annotated to NHEJ1 on chromosome 2 and cg00339913 annotated to PHYHIP on chromosome 8), while the other two were in an intergenic region: cg18968409 on chromosome 2 and cg24730307 on chromosome 22. Forest plots for the associations of these 4 CpGs with UPF intake across the 4 cohorts are presented in Fig. S2.

No statistically significant (FDR-adjusted P > 0.05) or suggestive (p-value < 10–5) DMRs were found.

Sensitivity analysis

First, when restricting the meta-analysis to children of European ancestry (n = 3007), most estimates remained similar, although the magnitude was slightly stronger for cg18968409 (effect size). Three additional CpGs passed the indicative threshold of p < 10–5: cg09709951, annotated to ATF7, cg03999434 (intergenic) and cg03041696 (intergenic) (Table 2). Secondly, when further adjusting for fruit and vegetable intake, we found suggestively significant associations for 4 CpGs, two of them previously detected in model 3 ((cg18968409 and cg09709951 (ATF7)) and two CpGs identified in the ethnicity sensitivity analyses (cg03999434 and cg03041696). Therefore, additionally to the 4 CpGs identified in model 3, we detected 3 additional CpGs in the two sensitivity analyses. Forest plots of these 3 additional CpGs are also included in Figure S2.

In leave-one-out analyses, estimates remained fairly consistent and there was no strong evidence that a specific cohort disproportionately influenced findings (Table S6).

Genomic and biological context

Table 3 shows genomic characteristics of the 7 suggestion CpGs and previously identified associations of DNAm with other exposures or traits in various age populations. Focusing on the genomic context, out of 7 CpGs, only 3 were annotated to a gene, of which only one (cg09709951) was associated with two transcript clusters: TC12002893.hg.1 and TC12001553.hg.1, which encode for the nearby ATF7 gene. As for the biological context, according to the EWAS catalog, 5 of the 7 identified CpGs had previously been reported in relation to ageing from birth to adolescence [74]. Moreover, natural juice consumption has been inversely associated with methylation at cg09709951 in adult studies [75], while UPF is associated with higher methylation at this site in our study, including when adjusting for fruit and vegetable intake. Lower methylation levels in the liver have been observed at cg03999434 (transcript: long non-coding RNA, RP3-340I3.1) in foetus compared to adults [76]. In the EWAS atlas, lower methylation at cg00339913 and higher methylation at cg18968409 have been associated with different carcinomas in adults [77, 78]. Also, higher methylation at cg09709951 has been found to be related to neurodevelopmental disorder related to SETD1B gene in children [73]. The directions of these associations with methylation correspond to a higher UPF intake in our study.

Table 3 Genomic context and biological exposures and traits previously associated with suggestive CpGs
Full size table

When searching EWAS of diet at any life stages, we did not find our suggestive CpGs (p-value < 1 × 10–5) to be reported in existing studies (Table S7). However, 16 of the CpGs reported in these studies were associated with UPF at a nominal p-value (p < 0.05) in our analyses (Table S8).

Functional enrichment analyses

We found some enriched pathways in the protein complex set of Consensus Path (Table 4). Three of the identified pathways are related to thyroid hormone, two others related to cancer and one related to LKB1 gene, which encodes a serine/threonine kinase protein that regulates cell polarity and energy metabolism and functions as a tumour suppressor [79].

Table 4 Protein complex gene sets annotated to CpGs associated with UPF with a p value < 10–4
Full size table

We did not find any enriched GO or KEGG terms that surpassed the FDR significance threshold; however, a list of the top enriched terms with a nominal p value < 0.05 is shown in Table S9. The first three nominal enriched GO terms are related to transcription processes, followed by transaminase-related terms, which are involved in the GABA catabolism, an inhibitory neurotransmitter [80]. In KEGG, we identified pathways related to thyroid hormone signalling, DNA repair, and cancer, the latter aligning with previous associations found with carcinomas. Some nominal KEGG enriched terms are also related to amino acid metabolism and energy metabolism which is related to gut microbiota and also is involved in GABA metabolism.

No specific enriched tissue or cell type for any chromatin states was found significant (Table S11).

Discussion

In this exploratory study, we report the first meta-EWAS of UPF intake and DNA methylation in blood in children in middle childhood (5–11 years). No CpG reached the FDR significance threshold in association with UPF intake; however, we found 7 suggestive CpGs (p < 10–5) associated with UPF consumption. The functional analysis reveals some enriched pathways involved in thyroid and liver function.

Biological interpretation and association with health outcomes

We observed higher methylation levels at cg09709951 (ATF7) associated with UPF intake, whereas the literature reports lower methylation with juice consumption in adults [75]. Our results align that quality of the diet might modulate methylation at this CpG, with a lower diet quality associated with higher methylation, although fruit juice is not similar to fruit in terms of fibre content and glycaemic load, and on its health effects. Moreover, higher methylation of this CpG has been associated with a neurodevelopmental disorder in children [81]. Finally, ATF7 encodes a stress-responsive chromatin regulator that plays a role in various biological processes including innate immunological memory, adipocyte differentiation, and telomerase regulation [82]. Some studies have shown that ATF7 ablation prevents diet-induced obesity and insulin resistance [82].

Maternal characteristics have also been associated with some of the suggestive CpGs; for instance, maternal alcohol consumption has been associated with a higher methylation at cg03041696 [74] (opposite direction with UPF), while lower methylation of cg00339913 (PHYHIP) has been associated with higher maternal BMI and overweight [20], following the same direction as UPF consumption. Our analyses were adjusted for maternal and child BMI; therefore, the differential methylation at this CpG previously observed with maternal BMI might be at least partly explained by consumption of UPF in the diet.

Some of our results point towards implications of UPF-related CpGs in liver function pathways. First, higher methylation at cg18968409 linked to higher UPF intake was also observed in the presence of hepatocellular carcinoma [78]. Second, cg03999434 is reported to be less methylated in the liver in the foetus compared to adults [76] and methylation is also expected to increase in blood at this CpG from birth to adulthood, whereas UPF is associated with lower methylation levels in our study. Additionally, pathway analyses revealed a protein complex near the LKB1 gene, which in the liver tissue is involved in the glycogenesis pathway. In adults, UPF intake has been related to adverse liver outcomes, in particular non-alcoholic fatty liver disease [83], cirrhosis, and severe liver disease [84]. Studies in children are needed to investigate if this starts in early life, particularly with subclinical biomarkers such as liver enzymes.

We found that UPF intake was associated with higher methylation of cg14665028, located at the promoter of NHEJ1 gene. This gene is one of the main genes involved in the non-homologous end-joining (NHEJ) pathway, the major double-strand breaks (DSB) repair mechanisms in normal cells [85]. This gene has been associated with a negative correlation with DNAm in tumour tissues [85], and an increase in DNA damage has been described in mice fed with high-fat diet and knockout in NHEJ pathway [86].

UPF consumption was inversely associated with methylation at cg00339913, located at the gene body of PHYHIP, also named PAHX-AP1, a brain specific protein mainly related to early postnatal maturation of visual function [87]. Lower methylation at this CpGs has been associated with presence of papillary thyroid carcinoma in non-neoplastic adjacent tissue in comparison to PTC tissue [77].

UPF intake and health outcomes

Despite conducting our EWAS on a very large multi-cohort population, the identified associations between UPF intake and DNA methylation in our study are of weak magnitude. UPF intake is associated with a range of health outcomes [8], some of them related to the UPF-related CpGs and pathways identified here, in particular thyroid and liver function. For instance, studies have reported that some food additives are related to thyroid hormone dysregulation [88] and an increased risk of subclinical thyroid dysfunction [88]. A recent report on the HELIX data showed that metabolites associated with UPF intake were associated with C-peptide levels in children, a well-known marker of β-cell function and insulin resistance. However, a wide range of non-methylation mechanisms that include the gut microbiome, oxidative stress, and inflammation are likely to be more affected by UPF intake and by the contaminants (acrylamide, etc.) and additives (emulsifiers, sweeteners, etc.) that they contain. Although UPFs are typically high in at least one nutrient of concern (saturated fat, salt, added sugar), Nova classification classifies UPFs based on the extent of food processing and not on nutrient profiles [89].

Limitations and strengths

This study provides initial evidence for the possible association of UPF consumption with DNA methylation in children. However, these results should be interpreted considering both the limitations and strengths of the study.

A first limitation is the cross-sectional design, which limits causal inference of the observed associations.

However, we carefully selected potential confounders and provide estimates at various levels of adjustment to reduce the likelihood of confounding. Additionally, our results are confined to the 'middle childhood' period (5–11 years) [90] and are not generalizable to younger ages or to adolescents. Moreover, it is important to acknowledge that both dietary and methylation patterns are likely to evolve from the age of 5 to 11 and that the association between UPF consumption and DNA methylation may vary across different ages.

Secondly, the definition of the exposure as percentage of UPF of the total dietary intake was defined differently across cohorts: proportion in servings in HELIX and GenR, and proportion of energy intake in ALSPAC and Gen XXI. Despite these definition differences between studies (a limitation inherent to any meta-analysis), the top hits CpGs were all associated with UPF in the same direction and with a similar magnitude across cohorts. Moreover, it should be noted that UPF intake calculated using the same definition in HELIX was associated with metabolites associated with insulin resistance and inversely related with a Mediterranean diet index [14].

In addition, our study considers geographical specificities and some foods were classified as UPF in some cohorts, but not in others. In particular, bread is most commonly purchased in supermarkets (cooked from ready-made industrial frozen dough, sliced packaged bread) in the UK and was therefore classified as ultra-processed in the ALSPAC study, whereas in the other countries involved in this study (Portugal, Spain, France, Lithuania, Greece and the Netherlands), fresh bread is most commonly consumed and therefore not classified as UPF [91].

Our results report methylation levels in whole blood. However, since DNA methylation is tissue specific, our findings in whole blood may not reflect DNA methylation levels in other tissues, where such associations may be more evident. For instance, gastrointestinal tissue, responsible for food digestion nutrient extraction, absorption, and waste excretion, could be a more appropriate tissue to provide a more relevant context [92]. Considering that UPF have been linked to metabolic biomarkers [14] and that the microbiome contains genomic information shaped by factors such as age and diet [20], it would be of interest to detect associations between diet and UPF within the gastrointestinal tissue DNA methylation. This may also be true for the liver tissue—as methylation may be related to some pathways such as gluconeogenesis [93]—the thyroid or the adipose tissue [94].

Despite the effort to combine various well characterized European children cohorts to increase power, the magnitude of the associations between DNA methylation and UPF consumption observed was small. A possible explanation is that the life-long cumulative exposure to UPF is shorter compared to adults, which might explain the lower number of CpGs related to diet in previous EWASs conducted in children too [29, 30]. This absence of clear results also likely comes partly from the inaccuracy of dietary assessment preventing the discovery of larger effects. Well-designed intervention studies that aim at reducing UPF intake and increase diet quality are needed to identify more robust effects of diet on DNA methylation, in particular if this is measured in relevant tissues beyond blood.

Conclusions

This study constitutes the first meta-EWAS evaluating the association between child UPF consumption and differential DNA methylation in blood in over 3000 school-aged children across Europe. We found suggestive changes in methylation at 7 CpGs. Future research with larger sample sizes, more detailed dietary intake estimation, and additional epigenetic time points will help to clarify the epigenetic signature of UPF intake.

Availability of data and materials

The manuscript is based on the data from the 4 European birth cohort studies (HELIX, Generation XXI, ALSPAC and GenerationR). The data that support the finding of this study are available from the four cohort, but restrictions apply to the availability of these data, which were used under licence for the current study and so are not publicly available. The HELIX cohort available data can be found in the codebooks publicly available at the project website (https://www.projecthelix.eu/index.php/es/data-inventory). The ALSPAC data are available through a fully searchable data dictionary and variable search tool (http://www.bristol.ac.uk/alspac/researchers/our-data/). All data are available on request for all participating cohort. The R scripts are available here: https://github.com/joanall/EWAS_UPF.

References

  1. Monteiro CA, Cannon G, Levy RB, Moubarac J-C, Louzada ML, Rauber F, et al. Ultra-processed foods: what they are and how to identify them. Public Health Nutr. 2019;22:936–41.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Teo PS, Lim AJY, Goh AT, Janani R, Choy JYM, McCrickerd K, et al. Texture-based differences in eating rate influence energy intake for minimally processed and ultra-processed meals. Am J Clin Nutr. 2022;116:244–54.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Touvier M, DaCostaLouzada ML, Mozaffarian D, Baker P, Juul F, Srour B. Ultra-processed foods and cardiometabolic health: public health policies to reduce consumption cannot wait. BMJ. 2023. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmj-2023-075294.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Onita BM, Azeredo CM, Jaime PC, Levy RB, Rauber F. Eating context and its association with ultra-processed food consumption by British children. Appetite. 2020;2021(157): 105007.

    Google Scholar 

  5. Neri D, Martinez-Steele E, Monteiro CA, Levy RB. Consumption of ultra-processed foods and its association with added sugar content in the diets of US children, NHANES 2009–2014. Pediatr Obes. 2019;14:1–11.

    Article  Google Scholar 

  6. Vandevijvere S, De Ridder K, Fiolet T, Bel S, Tafforeau J. Consumption of ultra-processed food products and diet quality among children, adolescents and adults in Belgium. Eur J Nutr. 2019;58:3267–78.

    Article  PubMed  Google Scholar 

  7. Khandpur N, Cediel G, Obando DA, Jaime PC, Parra DC. Sociodemographic factors associated with the consumption of ultra-processed foods in Colombia. Rev Saude Publica. 2020;54:19.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Lane MM, Gamage E, Du S, Ashtree DN, Mcguinness AJ, Gauci S, et al. Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ. 2024. https://doiorg.publicaciones.saludcastillayleon.es/10.1136/bmj-2023-077310.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Hall KD, Ayuketah A, Brychta R, Cai H, Cassimatis T, Chen KY, et al. Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake. Cell Metab. 2019;30:67-77.e3.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  10. Vineis P, Handakas E, Alfano R, Millett C, Fecht D, Chatzi L, et al. The contribution to policies of an exposome-based approach to childhood obesity. Exposome. 2023. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/exposome/osad006.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Chang K, Khandpur N, Neri D, Touvier M, Huybrechts I, Millett C, et al. Association between childhood consumption of ultraprocessed food and adiposity trajectories in the Avon longitudinal study of parents and children birth cohort. JAMA Pediatr. 2021;175:1–11.

    Article  PubMed Central  Google Scholar 

  12. Griffin J, Albaloul A, Kopytek A, Elliott P, Frost G. Effect of ultraprocessed food intake on cardiometabolic risk is mediated by diet quality: a cross-sectional study. BMJ Nutr Prev Heal. 2021;4:174–80.

    Article  Google Scholar 

  13. Handakas E, Keski-Rahkonen P, Chatzi L, Alfano R, Roumeliotaki T, Plusquin M, et al. Cord blood metabolic signatures predictive of childhood overweight and rapid growth. Int J Obes. 2021;45:2252–60.

    Article  CAS  Google Scholar 

  14. Stratakis N, Siskos AP, Papadopoulou E, Nguyen AN, Zhao Y, Margetaki K, et al. Urinary metabolic biomarkers of diet quality in European children are associated with metabolic health. Elife. 2022;11:71332.

    Article  Google Scholar 

  15. Handakas E, Chang K, Khandpur N, Vamos P, Millett C, Sassi F, et al. Metabolic profiles of ultra-processed food consumption and their role in obesity risk in British children. Clin Nutr. 2022;41:2537–48.

    Article  PubMed  CAS  Google Scholar 

  16. Lister NB, Baur LA, Felix JF, Hill AJ, Marcus C, Reinehr T, et al. Child and adolescent obesity. Nat Rev Dis Prim. 2023;9:24.

    Article  PubMed  Google Scholar 

  17. Romieu I, Dossus L, Barquera S, Blottière HM, Franks PW, Gunter M, et al. Energy balance and obesity: what are the main drivers? Cancer Causes Control. 2017;28:247–58.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Chen X, Zhang Z, Yang H, Qiu P, Wang H, Wang F, et al. Consumption of ultra-processed foods and health outcomes: a systematic review of epidemiological studies. Nutr J. 2020;19:86.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Küpers LK, Monnereau C, Sharp GC, Yousefi P, Salas LA, Ghantous A, et al. Meta-analysis of epigenome-wide association studies in neonates reveals widespread differential DNA methylation associated with birthweight. Nat Commun. 2019;10:1893.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Sharp GC, Salas LA, Monnereau C, Allard C, Yousefi P, Everson TM, et al. Maternal BMI at the start of pregnancy and offspring epigenome-wide DNA methylation: findings from the pregnancy and childhood epigenetics (PACE) consortium. Hum Mol Genet. 2017;26:4067–85.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  21. Robinson N, Brown H, Antoun E, Godfrey KM, Hanson MA, Lillycrop KA, et al. Childhood DNA methylation as a marker of early life rapid weight gain and subsequent overweight. Clin Epigenet. 2021;13:8.

    Article  CAS  Google Scholar 

  22. Landecker H. Food as exposure: nutritional epigenetics and the new metabolism. BioSocieties. 2011;6:167–94.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Domínguez-Barragán J, Fernández-Sanlés A, Hernáez Á, Llauradó-Pont J, Marrugat J, Robinson O, et al. Blood DNA methylation signature of diet quality and association with cardiometabolic traits. Eur J Prev Cardiol. 2024;31:191–202.

    Article  PubMed  Google Scholar 

  24. Ma J, Rebholz CM, Braun KVE, Reynolds LM, Aslibekyan S, Xia R, et al. Whole blood DNA methylation signatures of diet are associated with cardiovascular disease risk factors and all-cause mortality. Circ Genomic Precis Med. 2020;13:E002766.

    Article  Google Scholar 

  25. Lai CQ, Parnell LD, Smith CE, Guo T, Sayols-Baixeras S, Aslibekyan S, et al. Carbohydrate and fat intake associated with risk of metabolic diseases through epigenetics of CPT1A. Am J Clin Nutr. 2020;112:1200–11.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Karabegović I, Portilla-Fernandez E, Li Y, Ma J, Maas SCE, Sun D, et al. Epigenome-wide association meta-analysis of DNA methylation with coffee and tea consumption. Nat Commun. 2021;12:2830.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Lecorguillé M, Charles M-A, Lepeule J, Lioret S, de Lauzon-Guillain B, Forhan A, et al. Association between dietary patterns reflecting one-carbon metabolism nutrients intake before pregnancy and placental DNA methylation. Epigenetics. 2022;17:715–30.

    Article  PubMed  Google Scholar 

  28. Küpers LK, Fernández-Barrés S, Nounu A, Friedman C, Fore R, Mancano G, et al. Maternal Mediterranean diet in pregnancy and newborn DNA methylation: a meta-analysis in the PACE Consortium. Epigenetics. 2022;17:1419–31.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Ott R, Stein R, Hauta-Alus HH, Ronkainen J, Fernandez-Barres S, Spielau U, et al. Epigenome-wide meta-analysis reveals associations between dietary glycemic index and glycemic load and DNA methylation in children and adolescents of different body sizes. Diabetes Care. 2023;46:2067–75.

    Article  PubMed  Google Scholar 

  30. Lecorguillé M, Teo S, Phillips CM. Maternal dietary quality and dietary inflammation associations with offspring growth, placental development, and DNA methylation. Nutrients. 2021;13:3130.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Maitre L, de Bont J, Casas M, Robinson O, Aasvang GM, Agier L, et al. Human early life exposome (HELIX) study: a European population-based exposome cohort. BMJ Open. 2018;8: e021311.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Wright J, Small N, Raynor P, Tuffnell D, Bhopal R, Cameron N, et al. Cohort Profile: the Born in Bradford multi-ethnic family cohort study. Int J Epidemiol. 2013;42:978–91.

    Article  PubMed  Google Scholar 

  33. McEachan RRC, Santorelli G, Watmuff A, Mason D, Barber SE, Bingham DD, et al. Cohort profile update: born in Bradford. Int J Epidemiol. 2024. https://doiorg.publicaciones.saludcastillayleon.es/10.1093/ije/dyae037.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Heude B, Forhan A, Slama R, Douhaud L, Bedel S, Saurel-Cubizolles M-J, et al. Cohort Profile: The EDEN mother-child cohort on the prenatal and early postnatal determinants of child health and development. Int J Epidemiol. 2016;45:353–63.

    Article  PubMed  Google Scholar 

  35. Guxens M, Ballester F, Espada M, Fernández MF, Grimalt JO, Ibarluzea J, et al. Cohort profile: the INMA–INfancia y Medio ambiente–(environment and childhood) project. Int J Epidemiol. 2012;41:930–40.

    Article  PubMed  Google Scholar 

  36. Grazuleviciene R, Danileviciute A, Nadisauskiene R, Vencloviene J. Maternal smoking, GSTM1 and GSTT1 polymorphism and susceptibility to adverse pregnancy outcomes. Int J Environ Res Public Health. 2009;6:1282–97.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  37. Magnus P, Birke C, Vejrup K, Haugan A, Alsaker E, Daltveit AK, et al. Cohort profile update: the Norwegian mother and child cohort study (MoBa). Int J Epidemiol. 2016;45:382–8.

    Article  PubMed  Google Scholar 

  38. Chatzi L, Leventakou V, Vafeiadi M, Koutra K, Roumeliotaki T, Chalkiadaki G, et al. Cohort profile: the mother-child cohort in crete, Greece (Rhea Study). Int J Epidemiol. 2017;46:1392–1393k.

    Article  PubMed  Google Scholar 

  39. Kana MA, Rodrigues C, Fonseca MJ, Santos AC, Barros H. Effect of maternal country of birth on breastfeeding practices: results from Portuguese GXXI birth cohort. Int Breastfeed J. 2018;13:1–9.

    Article  Google Scholar 

  40. Fraser A, Macdonald-wallis C, Tilling K, Boyd A, Golding J, Davey smith G, et al. Cohort profile: the avon longitudinal study of parents and children: ALSPAC mothers cohort. Int J Epidemiol. 2013;42:97–110.

    Article  PubMed  Google Scholar 

  41. Boyd A, Golding J, Macleod J, Lawlor DA, Fraser A, Henderson J, et al. Cohort profile: the ‘children of the 90s’—the index offspring of the avon longitudinal study of parents and children. Int J Epidemiol. 2013;42:111–27.

    Article  PubMed  Google Scholar 

  42. Kooijman MN, Kruithof CJ, van Duijn CM, Duijts L, Franco OH, van IJzendoorn MH, et al. The Generation R Study: design and cohort update 2017. Eur J Epidemiol. 2016;31:1243–64.

    Article  PubMed  Google Scholar 

  43. Vilela S, Severo M, Moreira T, Oliveira A, Hetherington MM, Lopes C. Association between eating frequency and eating behaviours related to appetite from 4 to 7 years of age: findings from the population-based birth cohort generation XXI. Appetite. 2019;132:82–90.

    Article  PubMed  Google Scholar 

  44. Moreira T, Severo M, Oliveira A, Ramos E, Rodrigues S, Lopes C. Eating out of home and dietary adequacy in preschool children. Br J Nutr. 2015;114:297–305.

    Article  PubMed  CAS  Google Scholar 

  45. Emmett P. Dietary assessment in the avon longitudinal study of parents and children. Eur J Clin Nutr. 2009;63:S38-44.

    Article  PubMed  Google Scholar 

  46. Holland B, Welch A, Unwin I, Buss D, Paul A. McCance and Widdowson’s the composition of foods. 1991.

  47. van der Velde LA, Nguyen AN, Schoufour JD, Geelen A, Jaddoe VWV, Franco OH, et al. Diet quality in childhood: the generation R Study. Eur J Nutr. 2019;58:1259–69.

    Article  PubMed  Google Scholar 

  48. Monteiro CA, Cannon G, Lawrence M, Laura Da Costa Louzada M, Machado PP. Ultra-processed foods, diet quality, and health using the NOVA classification system.

  49. Vilela S, Magalhães V, Severo M, Oliveira A, Torres D, Lopes C. Effect of the food processing degree on cardiometabolic health outcomes: a prospective approach in childhood. Clin Nutr. 2022;41:2235–43.

    Article  PubMed  Google Scholar 

  50. Vedovato GM, Vilela S, Severo M, Rodrigues S, Lopes C, Oliveira A. Ultra-processed food consumption, appetitive traits and BMI in children: a prospective study. Br J Nutr. 2021;125:1427–36.

    Article  PubMed  CAS  Google Scholar 

  51. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. 2007;8:118–27.

    Article  PubMed  Google Scholar 

  52. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinform. 2012;13:86.

    Article  Google Scholar 

  53. Reinius LE, Acevedo N, Joerink M, Pershagen G, Dahlén S-E, Greco D, et al. Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS ONE. 2012;7: e41361.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  54. Si J, Meir AY, Hong X, Wang G, Huang W, Pearson C, et al. Maternal pre-pregnancy BMI, offspring epigenome-wide DNA methylation, and childhood obesity: findings from the Boston Birth Cohort. BMC Med. 2023;21:317.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  55. Joubert BR, Felix JF, Yousefi P, Bakulski KM, Just AC, Breton C, et al. DNA methylation in newborns and maternal smoking in pregnancy: genome-wide consortium meta-analysis. Am J Hum Genet. 2016;98:680–96.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  56. Larsen PS, Kamper-Jørgensen M, Adamson A, Barros H, Bonde JP, Brescianini S, et al. Pregnancy and birth cohort resources in europe: a large opportunity for aetiological child health research. Paediatr Perinat Epidemiol. 2013;27:393–414.

    Article  PubMed  Google Scholar 

  57. Jaddoe VWV, van Duijn CM, van der Heijden AJ, Mackenbach JP, Moll HA, Steegers EAP, et al. The generation R Study: design and cohort update 2010. Eur J Epidemiol. 2010;25:823–41.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Rousham EK, Goudet S, Markey O, Griffiths P, Boxer B, Carroll C, et al. Unhealthy food and beverage consumption in children and risk of overweight and obesity: a systematic review and meta-analysis. Adv Nutr. 2022;13:1669–96.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  59. Martines RM, Machado PP, Neri DA, Levy RB, Rauber F. Association between watching TV whilst eating and children’s consumption of ultraprocessed foods in United Kingdom. Matern Child Nutr. 2019. https://doiorg.publicaciones.saludcastillayleon.es/10.1111/mcn.12819.

    Article  PubMed  PubMed Central  Google Scholar 

  60. Wijtzes AI, Bouthoorn SH, Jansen W, Franco OH, Hofman A, Wv Jaddoe V, et al. Sedentary behaviors, physical activity behaviors, and body fat in 6-year-old children: the Generation R Study. Int J Behav Nutr Phys Act. 2014. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12966-014-0096-x.

    Article  PubMed  PubMed Central  Google Scholar 

  61. Ramsay SA, Shriver LH, Taylor CA. Variety of fruit and vegetables is related to preschoolers’ overall diet quality. Prev Med Rep. 2017;5:112–7.

    Article  PubMed  Google Scholar 

  62. Barbosa C, Costa A, Hetherington MM, Oliveira A. Association of early feeding practices with dietary patterns of 7-year-olds from the birth cohort Generation XXI. Appetite. 2021;2022(171): 105909.

    Google Scholar 

  63. Zhou W, Laird PW, Shen H. Comprehensive characterization, annotation and innovative use of Infinium DNA methylation BeadChip probes. Nucleic Acids Res. 2016;45:gkw67.

    Google Scholar 

  64. Grant J, Hunter A. Measuring inconsistency in knowledgebases. J Intell Inf Syst. 2006;27:159–84.

    Article  Google Scholar 

  65. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B. 1995;57:289–300.

    Article  Google Scholar 

  66. Aguilar-Lacasaña S, Fontes Marques I, de Castro M, Dadvand P, Escribà X, Fossati S, et al. Green space exposure and blood DNA methylation at birth and in childhood—a multi-cohort study. Environ Int. 2024;188:108684.

    Article  PubMed  Google Scholar 

  67. Tobi EW, Juvinao-Quintero DL, Ronkainen J, Ott R, Alfano R, Canouil M, et al. Maternal glycemic dysregulation during pregnancy and neonatal blood DNA methylation: meta-analyses of epigenome-wide association studies. Diabetes Care. 2022;45:614–23.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  68. Ruiz-Arenas C, Hernandez-Ferrer C, Vives-Usano M, Mari S, Quintela I, Mason D, et al. Identification of autosomal cis expression quantitative trait methylation (cis eQTMs) in children’s blood. Elife. 2022;11:1–26.

    Article  Google Scholar 

  69. Battram T, Yousefi P, Crawford G, Prince C, Sheikhali Babaei M, Sharp G, et al. The EWAS Catalog: a database of epigenome-wide association studies. Wellcome Open Res. 2022;7:41.

    Article  PubMed  PubMed Central  Google Scholar 

  70. Li M, Zou D, Li Z, Gao R, Sang J, Zhang Y, et al. EWAS Atlas: a curated knowledgebase of epigenome-wide association studies. Nucleic Acids Res. 2019;47:D983–8.

    Article  PubMed  CAS  Google Scholar 

  71. Phipson B, Maksimovic J, Oshlack A. missMethyl: an R package for analyzing data from Illumina’s HumanMethylation450 platform. Bioinformatics. 2016;32:286–8.

    Article  PubMed  CAS  Google Scholar 

  72. Kamburov A, Wierling C, Lehrach H, Herwig R. ConsensusPathDB—a database for integrating human functional interaction networks. Nucleic Acids Res. 2009;37 suppl_1:623–8.

    Article  Google Scholar 

  73. Breeze CE, Reynolds AP, van Dongen J, Dunham I, Lazar J, Neph S, et al. eFORGE v2.0: Updated analysis of cell type-specific signal in epigenomic data. Bioinformatics. 2019;35:4767–9.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  74. Mulder RH, Neumann A, Cecil CAMM, Walton E, Houtepen LC, Simpkin AJ, et al. Epigenome-wide change and variation in DNA methylation in childhood: trajectories from birth to late adolescence. Hum Mol Genet. 2021;30:119–34.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  75. Nicodemus-Johnson J, Sinnott RA. Fruit and juice epigenetic signatures are associated with independent immunoregulatory pathways. Nutrients. 2017;9:752.

    Article  PubMed  PubMed Central  Google Scholar 

  76. Bonder MJ, Kasela S, Kals M, Tamm R, Lokk K, Barragan I, et al. Genetic and epigenetic regulation of gene expression in fetal and adult human livers. BMC Genom. 2014. https://doiorg.publicaciones.saludcastillayleon.es/10.1186/1471-2164-15-860.

    Article  Google Scholar 

  77. Beltrami CM, dos Reis MB, Barros-Filho MC, Marchi FA, Kuasne H, Pinto CAL, et al. Integrated data analysis reveals potential drivers and pathways disrupted by DNA methylation in papillary thyroid carcinomas. Clin Epigenetics. 2017;9:1–11.

    Article  Google Scholar 

  78. Sun XJ, Wang MC, Zhang FH, Kong X. An integrated analysis of genome-wide DNA methylation and gene expression data in hepatocellular carcinoma. FEBS Open Bio. 2018;8:1093–103.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  79. Sanchez-Cespedes M. A role for LKB1 gene in human cancer beyond the Peutz-Jeghers syndrome. Oncogene. 2007;26:7825–32.

    Article  PubMed  CAS  Google Scholar 

  80. Parviz M, Vogel K, Gibson K, Pearl P. Disorders of GABA metabolism: SSADH and GABA-transaminase deficiencies. J Pediatr Epilepsy. 2015;03:217–27.

    Article  Google Scholar 

  81. Krzyzewska IM, Maas SM, Henneman P, Lip KVD, Venema A, Baranano K, et al. A genome-wide DNA methylation signature for SETD1B-related syndrome. Clin Epigenetics. 2019;11:156.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  82. Liu Y, Maekawa T, Yoshida K, Furuse T, Kaneda H, Wakana S, et al. ATF7 ablation prevents diet-induced obesity and insulin resistance. Biochem Biophys Res Commun. 2016;478:696–702.

    Article  PubMed  CAS  Google Scholar 

  83. Grinshpan LS, Eilat-Adar S, Ivancovsky-Wajcman D, Kariv R, Gillon-Keren M, Zelber-Sagi S. Ultra-processed food consumption and non-alcoholic fatty liver disease, metabolic syndrome and insulin resistance: a systematic review. JHEP Rep. 2024;6: 100964.

    Article  PubMed  Google Scholar 

  84. Zhao L, Clay-Gilmour A, Zhang J, Zhang X, Steck SE. Higher ultra-processed food intake is associated with adverse liver outcomes: a prospective cohort study of UK Biobank participants. Am J Clin Nutr. 2024;119:49–57.

    Article  PubMed  CAS  Google Scholar 

  85. Singh J, Arora M, Kumari S, Verma D, Kumar Palanichamy J, Qamar I, et al. Molecular associations and clinical significance of core NHEJ pathway genes in renal clear cell carcinoma. Gene Rep. 2021;23: 101167.

    Article  CAS  Google Scholar 

  86. Yu H, Harrison FE, Xia F. Altered DNA repair; an early pathogenic pathway in Alzheimer’s disease and obesity. Sci Rep. 2018;8:5600.

    Article  PubMed  PubMed Central  Google Scholar 

  87. Ahn KY, Il NK, Kim BY, Cho CW, Jeong SK, Yang KJ, et al. Postnatal expression and distribution of Refsum disease gene associated protein in the rat retina and visual cortex: effect of binocular visual deprivation. Int J Dev Neurosci. 2002;20:93–102.

    Article  PubMed  CAS  Google Scholar 

  88. Melough MM, Maffini MV, Otten JJ, Sathyanarayana S. Diet quality and exposure to endocrine-disrupting chemicals among US adults. Environ Res. 2022;211: 113049.

    Article  PubMed  CAS  Google Scholar 

  89. Braesco V, Souchon I, Sauvant P, Haurogné T, Maillot M, Féart C, et al. Ultra-processed foods: how functional is the NOVA system? Eur J Clin Nutr. 2022;76:1245–53.

    Article  PubMed  PubMed Central  Google Scholar 

  90. Reber M. Development during middle childhood: the years from six to twelve. J Dev Behav Pediatr. 1986;7:137–8.

    Article  Google Scholar 

  91. Monteiro CA, Moubarac JC, Levy RB, Canella DS, Da Costa Louzada ML, Cannon G. Household availability of ultra-processed foods and obesity in nineteen European countries. Public Health Nutr. 2018;21:18–26.

    Article  PubMed  Google Scholar 

  92. Pinho RM, Maga EA. DNA methylation as a regulator of intestinal gene expression. Br J Nutr. 2021;126:1611–25.

    Article  PubMed  CAS  Google Scholar 

  93. Just P-A, Charawi S, Denis RGP, Savall M, Traore M, Foretz M, et al. Lkb1 suppresses amino acid-driven gluconeogenesis in the liver. Nat Commun. 2020;11:6127.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  94. Baca P, Barajas-Olmos F, Mirzaeicheshmeh E, Zerrweck C, Guilbert L, Sánchez EC, et al. DNA methylation and gene expression analysis in adipose tissue to identify new loci associated with T2D development in obesity. Nutr Diabetes. 2022;12:50.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  95. Islam SA, Goodman SJ, MacIsaac JL, Obradović J, Barr RG, Boyce WT, et al. Integration of DNA methylation patterns and genetic variation in human pediatric tissues help inform EWAS design and interpretation. Epigenetics Chromatin. 2019;12:1.

    Article  PubMed  PubMed Central  Google Scholar 

  96. Sharp GC, Arathimos R, Reese SE, Page CM, Felix J, Küpers LK, et al. Maternal alcohol consumption and offspring DNA methylation: findings from six general population-based birth cohorts. Epigenomics. 2018;10:27–42.

    Article  PubMed  CAS  Google Scholar 

  97. Nicodemus-Johnson J, Myers RA, Sakabe NJ, Sobreira DR, Hogarth DK, Naureckas ET, et al. DNA methylation in lung cells is associated with asthma endotypes and genetic risk. JCI Insight. 2016. https://doiorg.publicaciones.saludcastillayleon.es/10.1172/jci.insight.90151.

    Article  PubMed  PubMed Central  Google Scholar 

  98. Wang C, Cardenas A, Hutchinson JN, Just A, Heiss J, Hou L, et al. Short- and intermediate-term exposure to ambient fine particulate elements and leukocyte epigenome-wide DNA methylation in older men: the Normative Aging Study. Environ Int. 2022;158: 106955.

    Article  PubMed  CAS  Google Scholar 

  99. Pérez RF, Santamarina P, Ramón Tejedor J, Urdinguio RG, Álvarez-Pitti J, Redon P, et al. Longitudinal genome-wide DNA methylation analysis uncovers persistent early-life DNA methylation changes. J Transl Med. 2019;17:15.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We are grateful to participants in all cohorts included in the four studies (HELIX, Generation XXI, ALSPAC and Generation R) across countries who have generously given their time and collaborated in the studies. The HELIX cohort received funding from the European Community’s Seventh Framework Programme (FP7/2007–2013) [grant number 308333]. Joana Llauradó-Pont has received support from the European Union—NextGenerationEU and the Spanish government, under the Investigo Programme (grant ID: 2022 INV-1 00046). Camille Lassale is supported by a Ramon y Cajal Fellowship RYC2020-029599 funded by MICIU/AEI/https://doiorg.publicaciones.saludcastillayleon.es/10.13039/501100011033 and the European Social Fund 'Invest in your future'. Nikos Stratakis has been supported from the European Union’s Horizon Europe research and innovation programme under the Marie Skłodowska-Curie Actions (MSCA) Postdoctoral Fellowships (grant ID: 101059245), from the Ministry of Science and Innovation and State Research Agency through the 'Centro de Excelencia Severo Ochoa 2019–2023' Program (CEX2018-000806-S) and by a Juan de la Cierva-Incorporación fellowship (IJC2020-043630-I) financed by Ministerio de Ciencia e Innovación (MCIN)/Agencia Estatal de Investigación (AEI)/https://doiorg.publicaciones.saludcastillayleon.es/10.13039/501100011033 and the European Union 'NextGenerationEU/PRTR'. Oliver Robinson was supported by a UKRI Future Leaders Fellowship (MR/S03532X/1). This work received funding from the European Union’s Horizon 2020 research and innovation programme (774548, STOP; 874739, LongITools; 874583, EUCAN-Connect) and from the European Joint Programming Initiative 'A Healthy Diet for a Healthy Life' (JPI HDHL, NutriPROGRAM project, ZonMw the Netherlands no.529051022 and PREcisE project ZonMw the Netherlands no.529051023). Details of funding for each study and for contributions of individual studies can be found in Supplementary material S3. No funders listed here or in Supplementary material S3 influenced the study aim, design, analysis or interpretation of results. The results expressed here are those of the authors and not necessarily any listed funder.

Funding

This project received funding from the European Union’s Horizon 2020 research and innovation programme (874583, ATHLETE).

Author information

Author notes

  1. Joana Llauradó-Pont, Nikos Stratakis, Oliver Robinson and Camille Lassale have contributed equally to this work.

Authors and Affiliations

  1. ISGlobal, Barcelona, Spain

    Joana Llauradó-Pont, Nikos Stratakis, Marianna Karachaliou, Martine Vrijheid, Mariona Bustamante & Camille Lassale

  2. Clinical Bioinformatics Unit, IRCCS Instituto Giannina Gaslini, Genova, Italy

    Giovanni Fiorito

  3. Medical Research Council Centre for Environment and Health, School of Public Health, Imperial College London, London, UK

    Evangelos Handakas, Paolo Vineis & Oliver Robinson

  4. Department of Child and Adolescent Psychiatry/Psychology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

    Alexander Neumann

  5. Generation XXI Study Group, EPIUNIT/ITR- Laboratory for Integrative and Translational Research in Population Health, Institute of Public Health, University of Porto, Porto, Portugal

    Henrique Barros & Carla Lopes

  6. Department of Food Safety, Centre for Sustainable Diets, Norwegian Institute of Public Health, Oslo, Norway

    Anne Lise Brantsæter

  7. Public Health Policy Evaluation Unit, School of Public Health, Imperial College London, London, UK

    Kiara Chang, Christopher Millett & Eszter P. Vamos

  8. Department of Preventive Medicine, University of Southern California, Los Angeles, USA

    Leda Chatzi

  9. Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

    Janine F. Felix & Vincent W. V. Jaddoe

  10. Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

    Janine F. Felix & Vincent W. V. Jaddoe

  11. Department of Environmental Sciences, Vytautas Magnus University, Kaunas, Lithuania

    Regina Grazuleviciene

  12. Université Paris Cité and Université Sorbonne Paris Nord, Inserm, INRAE, Center for Research in Epidemiology and Statistics (CRESS), Paris, France

    Marion Lecorguillé

  13. NOVA National School of Public Health, Public Health Research Centre, Comprehensive Health Research Center, CHRC, NOVA University Lisbon, Lisbon, Portugal

    Christopher Millett

  14. Division of Health Service, Global Health Cluster, Norwegian Institute of Public Health, Oslo, Norway

    Eleni Papadopoulou

  15. Team of Environmental Epidemiology, IAB, Institute for Advanced Biosciences, Inserm, CNRS, CHU-Grenoble-Alpes, University Grenoble-Alpes, CNRS, Grenoble, France

    Remy Slama

  16. Universitat Pompeu Fabra (UPF), Barcelona, Spain

    Martine Vrijheid, Mariona Bustamante & Camille Lassale

  17. Consortium for Research on Epidemiology and Public Health (CIBERESP), Instituto de Salud Carlos III, Madrid, Spain

    Martine Vrijheid & Mariona Bustamante

  18. Bradford Institute for Health Research, Bradford Teaching Hospitals NHS Foundation Trust, Bradford, West Yorkshire, UK

    Rosemary R. C. McEachan & John Wright

  19. Department of Epidemiology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

    Trudy Voortman

  20. Consortium for Biomedical Research - Pathophysiology of Obesity and Nutrition (CIBEROBN), Instituto de Salud Carlos III, Madrid, Spain

    Camille Lassale

Authors
  1. Joana Llauradó-Pont
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Nikos Stratakis
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Giovanni Fiorito
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Evangelos Handakas
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Alexander Neumann
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Henrique Barros
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Anne Lise Brantsæter
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Kiara Chang
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Leda Chatzi
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Janine F. Felix
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Regina Grazuleviciene
    View author publications

    You can also search for this author inPubMed Google Scholar

  12. Vincent W. V. Jaddoe
    View author publications

    You can also search for this author inPubMed Google Scholar

  13. Marianna Karachaliou
    View author publications

    You can also search for this author inPubMed Google Scholar

  14. Marion Lecorguillé
    View author publications

    You can also search for this author inPubMed Google Scholar

  15. Carla Lopes
    View author publications

    You can also search for this author inPubMed Google Scholar

  16. Christopher Millett
    View author publications

    You can also search for this author inPubMed Google Scholar

  17. Rosemary R. C. McEachan
    View author publications

    You can also search for this author inPubMed Google Scholar

  18. Eleni Papadopoulou
    View author publications

    You can also search for this author inPubMed Google Scholar

  19. Remy Slama
    View author publications

    You can also search for this author inPubMed Google Scholar

  20. Eszter P. Vamos
    View author publications

    You can also search for this author inPubMed Google Scholar

  21. Paolo Vineis
    View author publications

    You can also search for this author inPubMed Google Scholar

  22. Martine Vrijheid
    View author publications

    You can also search for this author inPubMed Google Scholar

  23. John Wright
    View author publications

    You can also search for this author inPubMed Google Scholar

  24. Trudy Voortman
    View author publications

    You can also search for this author inPubMed Google Scholar

  25. Mariona Bustamante
    View author publications

    You can also search for this author inPubMed Google Scholar

  26. Oliver Robinson
    View author publications

    You can also search for this author inPubMed Google Scholar

  27. Camille Lassale
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

NS, OR, MB, CL designed the study; GF, EC, AN, and JLP conducted the EWAS; JLP conducted the meta-analysis and GF the shadow meta-analysis; JLP and CL drafted the manuscript; MV coordinates the HELIX project; HB, CL, JW, RG, MV, RSl, JFF, TV, ALB, EP, KC, CM, EV, PV participated in funding, sample, and data acquisition and interpretation. All authors critically revised and approved the final manuscript.

Corresponding author

Correspondence to Camille Lassale.

Ethics declarations

Ethics approval and consent to participate

Detailed information about the ethical approvals and the consent procedures for each cohort study can be found in the supplemental materials S1 accompanying this manuscript.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

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://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Llauradó-Pont, J., Stratakis, N., Fiorito, G. et al. A meta-analysis of epigenome-wide association studies of ultra-processed food consumption with DNA methylation in European children. Clin Epigenet 17, 3 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-024-01782-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13148-024-01782-z

Keywords

Clinical Epigenetics

ISSN: 1868-7083

Contact us