Breastfeeding association with DNA methylation in the pregnancy and childhood epigenetics (PACE) consortium.

Breastfeeding association with DNA methylation in the pregnancy and childhood epigenetics (PACE) consortium - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Clin Epigenetics . 2026 Feb 28;18:63. doi: 10.1186/s13148-025-02042-4 Search in PMC Search in PubMed View in NLM Catalog Add to search Breastfeeding association with DNA methylation in the pregnancy and childhood epigenetics (PACE) consortium Doretta Caramaschi Doretta Caramaschi 1 Department of Psychology, Faculty of Health and Life Sciences, University of Exeter, Exeter, UK Find articles by Doretta Caramaschi 1, ✉, # , Sílvia Fernández-Barrés Sílvia Fernández-Barrés 2 Agència de Salut Pública de Barcelona, Barcelona, Spain Find articles by Sílvia Fernández-Barrés 2, # , Emma Casey Emma Casey 3 Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA USA Find articles by Emma Casey 3 , Adrià Cruells Adrià Cruells 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain Find articles by Adrià Cruells 4, 5, 6 , Darina Czamara Darina Czamara 7 Department Genes and Environment, Max Planck Institute of Psychiatry, Munich, Germany Find articles by Darina Czamara 7 , Mohammed El Sharkawy Mohammed El Sharkawy 8 Division of Metabolism and Nutrition, Department of Pediatrics, LMU University Hospital, Dr. von Hauner Children’s Hospital, Munich, Germany Find articles by Mohammed El Sharkawy 8 , Hannah R Elliott Hannah R Elliott 9 Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK 10 MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK Find articles by Hannah R Elliott 9, 10 , Ruby Fore Ruby Fore 11 Division of Chronic Disease Research Across the Lifecourse (CoRAL), Harvard Pilgrim Health Care Institute, Boston, MA USA Find articles by Ruby Fore 11 , Richa Gairola Richa Gairola 12 Department of Epidemiology, Brown University, Providence, RI USA Find articles by Richa Gairola 12 , Olena Gruzieva Olena Gruzieva 13 Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden 14 Centre for Occupational and Environmental Medicine, Region Stockholm, Stockholm, Sweden Find articles by Olena Gruzieva 13, 14 , Anke Huels Anke Huels 3 Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA USA 15 Ganagarosa Department of Environmental Health, Rollins School of Public Health, Emory University, Atlanta, GA USA 16 Department of Biostatistics and Bioinformatics, Rollins School of Public Health, Emory University, Atlanta, GA USA Find articles by Anke Huels 3, 15, 16 , Jari Lahti Jari Lahti 17 Department of Psychology, University of Helsinki, Helsinki, Finland 18 Folkhälsan Research Centre, Helsinki, Finland Find articles by Jari Lahti 17, 18 , Hami Lee Hami Lee 1 Department of Psychology, Faculty of Health and Life Sciences, University of Exeter, Exeter, UK Find articles by Hami Lee 1 , Roberta Magnano San Lio Roberta Magnano San Lio 19 Department of Medical and Surgical Sciences and Advanced Technologies “GF Ingrassia”, University of Catania, Catania, Italy Find articles by Roberta Magnano San Lio 19 , Anni Malmberg Anni Malmberg 17 Department of Psychology, University of Helsinki, Helsinki, Finland Find articles by Anni Malmberg 17 , Toby Mansell Toby Mansell 20 Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Australia 21 Department of Paediatrics, University of Melbourne, Parkville, Australia Find articles by Toby Mansell 20, 21 , Simon K Merid Simon K Merid 22 Department of Clinical Science and Education Södersjukhuset, Karolinska Institutet, Stockholm, Sweden Find articles by Simon K Merid 22 , Boris Novakovic Boris Novakovic 20 Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Australia 21 Department of Paediatrics, University of Melbourne, Parkville, Australia Find articles by Boris Novakovic 20, 21 , Raffael Ott Raffael Ott 23 Institute of Diabetes Research, Helmholtz Munich, German Research Center for Environmental Health, Munich, Germany 24 Forschergruppe Diabetes e.V. at Helmholtz Zentrum München, Munich, Germany 25 School of Medicine, Forschergruppe Diabetes at Klinikum rechts der Isar, Technical University Munich, Munich, Germany Find articles by Raffael Ott 23, 24, 25 , Dolors Pelegrí Dolors Pelegrí 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain Find articles by Dolors Pelegrí 4, 5, 6 , Faisal I Rezwan Faisal I Rezwan 26 Department of Computer Science, Aberystwyth University, Aberystwyth, UK 27 Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, UK Find articles by Faisal I Rezwan 26, 27 , Sheryl L Rifas-Shiman Sheryl L Rifas-Shiman 11 Division of Chronic Disease Research Across the Lifecourse (CoRAL), Harvard Pilgrim Health Care Institute, Boston, MA USA 28 Department of Population Medicine, Harvard Medical School, Boston, MA USA Find articles by Sheryl L Rifas-Shiman 11, 28 , Andreas Weiss Andreas Weiss 23 Institute of Diabetes Research, Helmholtz Munich, German Research Center for Environmental Health, Munich, Germany Find articles by Andreas Weiss 23 , Blandine de Lauzon-Guillain Blandine de Lauzon-Guillain 29 Université Paris Cité and Université Sorbonne Paris Nord, INSERM, INRAE, Center for Research in Epidemiology and Statistics (CRESS), Paris, France Find articles by Blandine de Lauzon-Guillain 29 , Liesbeth Duijts Liesbeth Duijts 30 The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 31 Department of Pediatrics, Division of Respiratory Medicine and Allergology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 32 Department of Neonatal and Pediatric Intensive Care, Division of Neonatology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands Find articles by Liesbeth Duijts 30, 31, 32 , Veit Grote Veit Grote 8 Division of Metabolism and Nutrition, Department of Pediatrics, LMU University Hospital, Dr. von Hauner Children’s Hospital, Munich, Germany Find articles by Veit Grote 8 , John W Holloway John W Holloway 27 Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, UK 33 National Institute for Health Research Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, UK Find articles by John W Holloway 27, 33 , Nastassja Koen Nastassja Koen 34 SAMRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry & Neuroscience Institute, University of Cape Town, Cape Town, South Africa Find articles by Nastassja Koen 34 , Caroline L Relton Caroline L Relton 9 Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK 10 MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK 35 London School of Hygiene and Tropical Medicine, University of London, London, UK Find articles by Caroline L Relton 9, 10, 35 , Dan J Stein Dan J Stein 34 SAMRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry & Neuroscience Institute, University of Cape Town, Cape Town, South Africa Find articles by Dan J Stein 34 , Heather J Zar Heather J Zar 36 SAMRC Unit on Child & Adolescent Health, Department of Paediatrics, University of Cape Town, Cape Town, South Africa Find articles by Heather J Zar 36 , Joseph M Braun Joseph M Braun 12 Department of Epidemiology, Brown University, Providence, RI USA Find articles by Joseph M Braun 12 , Kim M Cecil Kim M Cecil 37 Department of Pediatrics, Department of Radiology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH USA 38 University of Cincinnati College of Medicine, Cincinnati, OH USA Find articles by Kim M Cecil 37, 38 , Marie-France Hivert Marie-France Hivert 11 Division of Chronic Disease Research Across the Lifecourse (CoRAL), Harvard Pilgrim Health Care Institute, Boston, MA USA 28 Department of Population Medicine, Harvard Medical School, Boston, MA USA 39 Diabetes Unit, Massachusetts General Hospital, Boston, MA USA Find articles by Marie-France Hivert 11, 28, 39 , Sandra Hummel Sandra Hummel 23 Institute of Diabetes Research, Helmholtz Munich, German Research Center for Environmental Health, Munich, Germany 24 Forschergruppe Diabetes e.V. at Helmholtz Zentrum München, Munich, Germany 25 School of Medicine, Forschergruppe Diabetes at Klinikum rechts der Isar, Technical University Munich, Munich, Germany Find articles by Sandra Hummel 23, 24, 25 , Vincent W V Jaddoe Vincent W V Jaddoe 30 The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 40 Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands Find articles by Vincent W V Jaddoe 30, 40 , Μarianna Karachaliou Μarianna Karachaliou 30 The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 40 Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 41 Clinic of Preventive and Social Medicine, Medical School, University of Crete, Herakleion, Greece Find articles by Μarianna Karachaliou 30, 40, 41 , Wilfried Karmaus Wilfried Karmaus 42 Division of Epidemiology, Biostatistics and Environmental Health, School of Public Health, University of Memphis, Memphis, TN USA Find articles by Wilfried Karmaus 42 , Manolis Kogevinas Manolis Kogevinas 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain Find articles by Manolis Kogevinas 4, 5, 6 , Berthold Koletzko Berthold Koletzko 8 Division of Metabolism and Nutrition, Department of Pediatrics, LMU University Hospital, Dr. von Hauner Children’s Hospital, Munich, Germany Find articles by Berthold Koletzko 8 , Inger Kull Inger Kull 22 Department of Clinical Science and Education Södersjukhuset, Karolinska Institutet, Stockholm, Sweden 43 Sachsska Children’s Hospital, Stockholm, Sweden Find articles by Inger Kull 22, 43 , Erik Melén Erik Melén 22 Department of Clinical Science and Education Södersjukhuset, Karolinska Institutet, Stockholm, Sweden 43 Sachsska Children’s Hospital, Stockholm, Sweden Find articles by Erik Melén 22, 43 , Emily Oken Emily Oken 11 Division of Chronic Disease Research Across the Lifecourse (CoRAL), Harvard Pilgrim Health Care Institute, Boston, MA USA 28 Department of Population Medicine, Harvard Medical School, Boston, MA USA Find articles by Emily Oken 11, 28 , Katri Räikkönen Katri Räikkönen 17 Department of Psychology, University of Helsinki, Helsinki, Finland Find articles by Katri Räikkönen 17 , Richard Saffery Richard Saffery 20 Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Australia 21 Department of Paediatrics, University of Melbourne, Parkville, Australia Find articles by Richard Saffery 20, 21 , Martine Vrijheid Martine Vrijheid 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain Find articles by Martine Vrijheid 4, 5, 6 , John Wright John Wright 44 Bradford Institute for Health Research, Bradford Royal Infirmary, Bradford, UK Find articles by John Wright 44 , Kimberly Yolton Kimberly Yolton 38 University of Cincinnati College of Medicine, Cincinnati, OH USA 45 Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH USA Find articles by Kimberly Yolton 38, 45 , Barbara Heude Barbara Heude 29 Université Paris Cité and Université Sorbonne Paris Nord, INSERM, INRAE, Center for Research in Epidemiology and Statistics (CRESS), Paris, France Find articles by Barbara Heude 29 , Janine F Felix Janine F Felix 30 The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 40 Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands Find articles by Janine F Felix 30, 40, # , Mariona Bustamante Mariona Bustamante 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain Find articles by Mariona Bustamante 4, 5, 6, ✉, # Author information Article notes Copyright and License information 1 Department of Psychology, Faculty of Health and Life Sciences, University of Exeter, Exeter, UK 2 Agència de Salut Pública de Barcelona, Barcelona, Spain 3 Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA USA 4 ISGlobal, Barcelona, Spain 5 Universitat Pompeu Fabra (UPF), Barcelona, Spain 6 CIBER Epidemiología y Salud Pública, Madrid, Spain 7 Department Genes and Environment, Max Planck Institute of Psychiatry, Munich, Germany 8 Division of Metabolism and Nutrition, Department of Pediatrics, LMU University Hospital, Dr. von Hauner Children’s Hospital, Munich, Germany 9 Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK 10 MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK 11 Division of Chronic Disease Research Across the Lifecourse (CoRAL), Harvard Pilgrim Health Care Institute, Boston, MA USA 12 Department of Epidemiology, Brown University, Providence, RI USA 13 Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden 14 Centre for Occupational and Environmental Medicine, Region Stockholm, Stockholm, Sweden 15 Ganagarosa Department of Environmental Health, Rollins School of Public Health, Emory University, Atlanta, GA USA 16 Department of Biostatistics and Bioinformatics, Rollins School of Public Health, Emory University, Atlanta, GA USA 17 Department of Psychology, University of Helsinki, Helsinki, Finland 18 Folkhälsan Research Centre, Helsinki, Finland 19 Department of Medical and Surgical Sciences and Advanced Technologies “GF Ingrassia”, University of Catania, Catania, Italy 20 Murdoch Children’s Research Institute, Royal Children’s Hospital, Parkville, Australia 21 Department of Paediatrics, University of Melbourne, Parkville, Australia 22 Department of Clinical Science and Education Södersjukhuset, Karolinska Institutet, Stockholm, Sweden 23 Institute of Diabetes Research, Helmholtz Munich, German Research Center for Environmental Health, Munich, Germany 24 Forschergruppe Diabetes e.V. at Helmholtz Zentrum München, Munich, Germany 25 School of Medicine, Forschergruppe Diabetes at Klinikum rechts der Isar, Technical University Munich, Munich, Germany 26 Department of Computer Science, Aberystwyth University, Aberystwyth, UK 27 Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, UK 28 Department of Population Medicine, Harvard Medical School, Boston, MA USA 29 Université Paris Cité and Université Sorbonne Paris Nord, INSERM, INRAE, Center for Research in Epidemiology and Statistics (CRESS), Paris, France 30 The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 31 Department of Pediatrics, Division of Respiratory Medicine and Allergology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 32 Department of Neonatal and Pediatric Intensive Care, Division of Neonatology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 33 National Institute for Health Research Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, UK 34 SAMRC Unit on Risk & Resilience in Mental Disorders, Department of Psychiatry & Neuroscience Institute, University of Cape Town, Cape Town, South Africa 35 London School of Hygiene and Tropical Medicine, University of London, London, UK 36 SAMRC Unit on Child & Adolescent Health, Department of Paediatrics, University of Cape Town, Cape Town, South Africa 37 Department of Pediatrics, Department of Radiology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH USA 38 University of Cincinnati College of Medicine, Cincinnati, OH USA 39 Diabetes Unit, Massachusetts General Hospital, Boston, MA USA 40 Department of Pediatrics, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands 41 Clinic of Preventive and Social Medicine, Medical School, University of Crete, Herakleion, Greece 42 Division of Epidemiology, Biostatistics and Environmental Health, School of Public Health, University of Memphis, Memphis, TN USA 43 Sachsska Children’s Hospital, Stockholm, Sweden 44 Bradford Institute for Health Research, Bradford Royal Infirmary, Bradford, UK 45 Department of Pediatrics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH USA ✉ Corresponding author. # Contributed equally. Received 2025 Oct 6; Accepted 2025 Dec 15; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13085660  PMID: 41764488 Abstract Background Breastfeeding is associated with short- and long-term beneficial effects on child health, including greater cognitive development, and enhanced immune programming. However, the underlying biological mechanisms are only partially understood, with epigenetics emerging as a potential contributor. In this study, we aimed to investigate whether breastfeeding practices are associated with differential DNA methylation (DNAm) in childhood blood. Results We conducted meta-analyses of epigenome-wide association studies (meta-EWASs) in 3421 children from eleven international population-based birth cohorts from the Pregnancy And Childhood Epigenetics (PACE) Consortium. Breastfeeding was assessed as “ever” being breastfed vs. “never”, and duration of any and exclusive breastfeeding. DNAm was measured in childhood blood (ages 5–12 years) using the Illumina 450 K or EPIC arrays, with cord blood at birth used as negative outcome control. At False Discovery Rate (FDR) < 5%, positive associations at six cytosine-phosphate-guanine (CpG) sites were identified in childhood blood: four with duration of exclusive breastfeeding, and three with duration of exclusive breastfeeding of more than three months compared to never. The annotated genes ( ALAD , FNBP4 , and CHFR ) are related to developmental and immune processes. None of these CpG sites were FDR-significant in cord blood prior to breastfeeding. Conclusions Breastfeeding was associated with differential DNAm in childhood blood at a limited number of CpG sites. Future studies in diverse populations are needed to examine the robustness of these associations, the sources of heterogeneity, and the generalizability of the findings. Supplementary Information The online version contains supplementary material available at 10.1186/s13148-025-02042-4. Keywords: Breastfeeding, Exclusive breastfeeding, Epigenetics, DNA methylation, Epigenome-wide association study, Meta-analysis, Birth cohort, Children Background Breastfeeding has substantial short- and long-term beneficial effects on child health, for instance by lowering the risk of respiratory problems, reducing the risk of metabolic diseases, improving cognitive development, and driving immune programming [ 1 – 6 ]. Human milk is preferred compared to formula milk because of its nutritional content, the presence of immunoglobulins, and its antioxidant and anti-inflammatory properties [ 7 ]. The World Health Organization (WHO) recommends exclusive breastfeeding until 6 months, and then continuing alongside complementary foods until at least 2 years of age) [ 8 , 9 ]. Breastfeeding may influence infant development through epigenetic processes. The epigenome, defined as the set of chemical modifications of the DNA or related proteins or RNAs that regulate gene expression, can affect developmental pathways, consistent with the early-life programming of health and disease [ 10 , 11 ]. One of the most studied epigenetic processes in humans is DNA methylation (DNAm), the addition of a methyl group onto a cytosine in a cytosine-phosphate-guanine dinucleotide (CpG) [ 12 ]. Few studies have investigated the association between breastfeeding and DNAm in humans. Two candidate gene studies conducted in blood and buccal epithelial cells found associations with DNAm of the Leptin ( LEP) gene, which encodes a hormone that controls appetite behavior [ 13 , 14 ]. More recent epigenome-wide association studies (EWAS) showed a potential association of breastfeeding with DNAm in early life, but were limited as they were single cohort studies often with relatively small sample sizes, low statistical power and providing few consistent results, with almost no cross-cohort validation [ 15 – 23 ] . We conducted the largest study to date using an EWAS meta-analysis (meta-EWAS) approach aimed at identifying robust associations of breastfeeding with child blood DNAm. Methods Study participants The study overall included data from fourteen birth cohorts from the Pregnancy And Childhood Epigenetics (PACE) Consortium [ 24 ]. Blood DNAm was measured during childhood in eleven of these cohorts and at birth (in cord blood) in six of the cohorts (three cohorts both in childhood and at birth). Results from the analyses at birth were used as negative outcome controls (i.e. it is not expected that breastfeeding practices affect DNAm at birth). For details of the participating cohorts see Table 1 and Supplemental Methods (Supplementary Material 1 ). Table 1. Characteristics of each cohort included in the meta-EWASs of breastfeeding practices Child blood DNAm Any breastfeeding Exclusive breatfeeding Ever vs. Never Duration (months) Duration categories Duration (months) Duration categories Cohort Ancestry Mean age (years) N Never Ever N Median [IQR] N Never ≤4 m >4–12 m ≥12 m N Median [IQR] N Never ≤3 m >3 m ALSPAC European 7.45 767 81 686 767 5.00 [1.00, 9.00] 767 81 294 276 116 766 2.00 [0.00, 3.00] 766 81 593 92 BAMSE.E European 8.30 – – – 300 8.75 [7.00, 11.25] – – – – – 302 5.62 [4.00, 6.00] – – – – BAMSE.M European 8.34 – – – 196 8.38 [6.00, 11.00] – – – – – 197 5.75 [4.00, 6.00] – – – – CHOP European 11.12 373 106 267 – – 342 106 143 60 33 – – – – – GENR European 9.77 339 36 303 339 5.00 [2.50, 8.50] 288 36 116 104 32 – – – – – GLAKU European 12.31 72 26 46 72 9.00 [0.00, 13.62] – – – – – 87 4.00 [0.00, 4.75] 87 26 12 49 HELIX European 8.33 788 128 660 788 4.93 [0.99, 10.84] 788 125 250 278 135 – – – – – HOME Diverse 12.31 157 28 129 157 3.25 [0.50, 12.00] 157 28 56 33 40 157 0.03 [0.03, 0.69] 157 29 110 18 IOW European 10.00 83 18 65 83 1.87 [0.47, 7.00] – – – – – 83 1.40 [0.35, 3.38] 83 20 39 24 POGO.GDM European 5.28 78 10 68 78 5.00 [2.00, 9.00] – – – – – -– – – – – Viva European 7.78 268 25 243 268 8.50 [3.00, 12.00] – – – – – 259 2.79 [0.47, 4.50] – – – – Total N 2925 458 2467 3048 – 2342 376 859 751 356 1851 – 1093 156 754 183 Cord blood DNAm (negative outcome control) Any breastfeeding Exclusive breatfeeding Ever vs. Never Duration (months) Duration categories Duration (months) Duration categories Cohort Ancestry Mean age (years) N Never Ever Median [IQR] N Never ≤4 m >4–12 m ≥12 m N Median [IQR] N Never ≤3 m >3 m ALSPAC European 0 710 76 634 710 5.00 [1.00, 9.00] 710 76 274 251 109 709 2.00 [0.00, 3.00] 709 76 553 80 BIS European 0 745 13 732 745 4.60 [1.38, 8.28] – – – – – 561 1.15 [0.23, 5.06] – – – – DCHS African and admixed 0 250 45 205 250 6.00 [1.39, 12.00] 250 45 62 57 86 250 1.45 [0.46, 3.68] 250 53 125 72 GENR European 0 979 99 880 979 5.00 [1.50, 8.50] 844 99 366 294 85 – – – – – – HOME Diverse 0 261 45 216 261 5.00 [0.75, 12.00] 261 45 82 64 70 261 0.03 [0.03, 0.69] 261 46 183 32 INMA European 0 376 28 348 376 5.26 [2.92, 8.97] 373 28 110 174 61 371 2.07 [0.07-4.36] 371 82 143 146 Total N 3321 306 3015 3321 - 2438 293 894 840 411 2152 – 1591 257 1004 330 Open in a new tab DNAm: DNA methylation, m: months, meta-EWASs: meta-analyses of epigenome-wide association studies With a dash, if the cohort does not participate in that particular analysis because the low number of samples in some group Note that some of the cohorts have lower N in the variable “duration categories” compared to the others due to how the questionnaire was formulated INMA is part of the HELIX cohort Breastfeeding data Data on breastfeeding were collected through questionnaires administered to the mothers. Any breastfeeding was defined as the practice where an infant received breast milk, including milk expressed or donated (human milk bank), regardless of the introduction of formula feed and/or solid foods. Exclusive breastfeeding was defined as the practice where an infant only received breast milk, including milk expressed or donated, but no other food or drink (including water), except for oral rehydration therapy (e.g. due to complications at birth), drops and syrups (e.g. vitamins, minerals and medicines) based on the WHO definitions ( https://www.emro.who.int/nutrition/breastfeeding/index.html ). The following breastfeeding variables were created: (i) any breastfeeding (0: never breastfed, 1: ever breastfed); (ii) any breastfeeding duration in months (to harmonize across cohorts, children breastfed for > 12 months were assigned the value of 12, and children never breastfed were assigned the value of 0); (iii) any breastfeeding duration in four categories (0: not breastfed, 1: ≤4 months, 2: >4 months and < 12 months, 3: ≥12 months); (iv) exclusive breastfeeding duration in months; and (v) exclusive breastfeeding in three categories (0: not breastfed, 1: ≤3 months, 2: >3 months). Any breastfeeding and exclusive breastfeeding categories were defined according to WHO recommendations, average maternity leave across countries, frequency and type of questions asked in each cohort, and reasonable sample size in each category. Blood DNA methylation data DNA was extracted from whole blood or buffy coat and processed with the Illumina EPIC or the 450 K methylation arrays. Data quality control was performed according to each cohort’s preferred procedures and included sample and probe quality control, normalization, and correction for technical batch effects. For cohort-specific details, see Supplementary Material 1 . Extreme DNAm values outside the range of (25th percentile − 3 × interquartile range (IQR)) to (75th percentile + 3 × IQR) were removed. DNAm values were expressed as beta values ranging from 0 (completely unmethylated) to 1 (completely methylated). Six blood cell type proportions (natural killers, monocytes, granulocytes, CD8T cells, CD4T cells, and B cells) were estimated from childhood DNAm data using the Reinius reference panel and the Houseman algorithm [ 25 , 26 ]. For cord blood DNAm, the Gervin and Salas reference panel was used, which includes the same six cell types plus nucleated red blood cells [ 27 ]. Epigenome-wide association analyses Each cohort estimated the association between breastfeeding practices and blood DNAm using an adjusted robust linear regression model for each CpG site, with DNAm as the outcome according to a predefined analysis plan and R code. Covariates included in the models were: age at blood sampling (in years, only for the childhood EWAS), sex (male, female), maternal age at delivery (years), maternal education (primary, secondary, and university), parity (nulliparous, multiparous), pre-pregnancy or early pregnancy maternal body mass index (BMI) (kg/m 2 ), birth weight (grams), gestational age at delivery (weeks), type of delivery (vaginal, caesarean), sustained maternal smoking during pregnancy (non-sustained smoking, sustained smoking), and estimated blood cell type proportions at time of blood sampling. Optional covariates based on specific cohort characteristics and data availability were child’s ancestry (from genome-wide genetic principal components (PCs) or from questionnaire data), technical covariates or selection factors. Meta-analyses Prior to the meta-analyses, we conducted stringent quality control of cohort-specific EWAS results. Probe filtering was done to remove control probes, probes to detect single nucleotide polymorphisms (SNPs), non-CpG probes, probes in sex chromosomes, probes that give non-consistent measures across arrays (EPIC versus 450 k) in blood [ 28 ], and problematic CpG probes [ 29 ]. Quality control of the results from each cohort consisted of checking the effect sizes, standard errors (SE), p-values (p), number of significant CpG sites and calculation of the lambda inflation factors. Fixed-effects inverse variance-weighted meta-analyses were conducted independently by two researchers (AC and HL), using the EASIER R package ( https://github.com/isglobal-brge/EASIER ) [ 30 ] and the metafor R package [ 31 ]. Random-effects meta-analyses were also run for significant probes in the fixed effect meta-analyses. Given that the EPIC array was only available in a subset of the cohorts (four out of eleven cohorts), which in addition had relatively small sample sizes (< 160 participants), we only meta-analysed the CpG sites present on both EPIC and 450 K arrays. CpG sites present in only one cohort were filtered out from the meta-analyses. A main meta-analysis was run for each breastfeeding definition including all children from all available cohorts. We used False Discovery Rate (FDR) method at 5% to control for multiple-testing [ 32 ]. Effect size is reported as the difference in DNAm between breastfed and never breastfed infants (for categorical variables) or by month of any or exclusive breastfeeding (for continuous variables). Sensitivity analyses were performed to examine heterogeneity across cohorts and to reduce the chance of false positives by re-running the models: (i) restricting to European ancestry from all cohorts; (ii) restricting to cohorts with a sample size > 100; and (iii) restricting to European ancestry from cohorts with a sample size > 100. Leave-one-out analyses were conducted for CpG sites that were FDR-significant in the main model. Cohort-specific and meta-analysis results were summarized in tables and plots using ggplot2 and forestplot R packages [ 33 ] ( https://cran.r-project.org/web/packages/forestplot/index.html ). All analyses were conducted in R environment version 4.2.1 (2022-06-23). In silico analyses CpG sites were annotated to genes using the IlluminaHumanMethylation450kanno.ilmn12.hg19 R package. DNAm quantitative trait loci (meQTLs) were retrieved from the Genetics of DNAm consortium (GoDMC) database ( http://mqtldb.godmc.org.uk/index ) [ 34 ], and cis expression quantitative trait DNAm (eQTMs) from the Human Early Life Exposome (HELIX) database ( https://helixomics.isglobal.org/ ) [ 35 ]. Previous associations with exposures or traits were searched in the EWAS catalog ( https://www.ewascatalog.org/ ) [ 36 ]. We also conducted enrichment analyses of the suggestive CpG sites ( p < 1E-05) associated with exclusive breastfeeding variables using the EWAS Toolkit ( https://ngdc.cncb.ac.cn/ewas/toolkit ). We tested enrichment for KEGG pathways and traits/exposures from the EWAS Atlas. Finally, a look-up of previous associations from studies of breastfeeding practices with N > 100 [ 17 , 19 , 21 ], and a candidate-gene look-up across the whole LEP gene were also carried out due to its role in appetite regulation and previous evidence linking it with breastfeeding. Results Participants’ characteristics The meta-EWAS of breastfeeding practices included childhood blood DNAm from eleven cohorts, for a total N of 3421 participants. The study sample is summarised in Table 1 and in Supplemental Tables S1A-B (Supplementary Material 2 ). Nine cohorts participated in the meta-analyses of any breastfeeding ( N = 2467 ever breastfed and N = 458 never breastfed children). Any breastfeeding duration data were available for 3048 children across ten cohorts, with median durations ranging from 1.87 to 9.00 months. Exclusive breastfeeding data were available for 1851 children across seven cohorts, with median durations ranging from 0.03 to 5.75 months. Children from ten out of eleven cohorts were of European ancestry. Three cohorts had < 100 participants. EWAS of any breastfeeding and childhood blood DNA methylation We did not identify CpG sites at which childhood blood DNAm was associated with any breastfeeding (ever vs. never, duration or categories). When restricting the analysis to children of European ancestry, duration of any breastfeeding was inversely associated with DNAm at cg07954414 (at the KIAA0922 gene) (effect= -0.0001, standard error (SE) = 0.0000, p FDR = 0.04) (Supplemental Table S2 (Supplementary Material 3 )). This CpG site showed a similar association in the main model, close to FDR-significance (effect= -0.0001, SE = 0.0000, p FDR = 0.12). No FDR-significant associations were found when restricting to cohorts with > 100 participants. The cohort-specific EWAS and the meta-EWAS models showed good performance, with limited genomic inflation (by cohort: Supplemental Tables S3A-C (Supplementary Material 4 ); meta-EWASs: Supplemental Tables S4A-C (Supplementary Material 5 ) and Supplemental Figures S1-S3 (Supplementary Material 6 )). Top CpG sites ( p < 1E-05) are shown in Supplemental Tables S5A-D, S6A-D and S7A-B (Supplementary Materials 7 , 8 and 9 , respectively). EWAS of exclusive breastfeeding and childhood blood DNA methylation Continuous duration of exclusive breastfeeding was positively associated with DNAm at four CpG sites: cg01257194 ( ALAD ), cg20053493 ( FNBP4 ), cg04942655 (intergenic), and cg20702204 ( CHFR ) (Table 2 ). Cg01257194 ( ALAD ), cg20053493 ( FNBP4 ), and cg04942655 (intergenic) remained FDR-significant in cohorts with > 100 participants, and cg01257194 ( ALAD ) in children of European ancestry. There was high heterogeneity across the cohorts, with values of I 2 > 0.6 for all CpG sites, except for cg20053493 ( FNBP4 ) (I 2 < 0.17), and effect sizes were relatively small (Fig. 1 ). The leave-one-out analysis indicated that some of the associations in the main model were driven by the HOME study, which is the most ethnically diverse sample of participants (Supplemental Table S8A (Supplementary Material 10 )). Three other CpG sites were associated with exclusive breastfeeding duration in the sensitivity models (Supplemental Table S2 (Supplementary Material 3 )). When restricting to European ancestry, longer exclusive breastfeeding duration was associated with higher DNAm at cg02352945 ( ACVR2B ) (effect = 0.0004, SE = 0.0001, p FDR = 0.008), and at cg02592586 (intergenic) (effect = 0.0029, SE = 0.0006, p FDR = 0.035). The effect size for both of these CpG sites was similar in the main model (cg02352945, ACVR2B , effect = 0.0003, SE = 0.0001, p FDR = 0.054; cg02592586, intergenic, effect = 0.0026, SE = 0.0005, p FDR = 0.081). In the cohorts with > 100 participants, exclusive breastfeeding duration was associated with higher DNAm levels at cg06061442 ( PLB1 ) (effect = 0.0017, SE = 0.0003, p FDR = 0.022). In the main model the effect size was similar (effect = 0.0013, SE = 0.0003, p FDR =0.190), with heterogeneity decreasing from I 2 = 0.8 to I 2 = 0 when restricting to cohorts with > 100 participants. Table 2. Association of breastfeeding and DNAm at FDR-significant CpGs identified in the main analysis Main analysis Annotation CpG Breastfeeding Comparison/ Units N Effect size SE p-value FDR p-value I 2 Chromosome Position (hg37) Gene Gene relative position SNPs probe (MAF) GoDMC SNPs EWAS catalog cg01257194 exclusive BF duration months 1852 0.0017 0.0003 9.48E-09 3.93E-03 0.67 chr9 116,161,247 ALAD 5’UTR – – CRP↑, alcohol, sex, age↑ cg04942655 exclusive BF duration months 1852 0.0004 0.0001 3.69E-08 5.09E-03 0.65 chr17 80,303,572 intergenic – rs9889704 (0.31) rs537613896, others age↓ cg20702204 exclusive BF duration months 1852 0.0008 0.0002 3.27E-07 3.39E-02 0.75 chr12 133,430,077 CHFR body – – age*↑ cg20053493 exclusive BF duration months 1852 0.0005 0.0001 2.25E-08 4.66E-03 0.17 chr11 47,776,175 FNBP4 body – rs7935528, others age↓ cg20053493 exclusive BF duration categories ≤3 m vs. never 754 vs. 156 0.0015 0.0005 3.80E-03 7.42E-01 0.19 >3 m vs. never 183 vs. 156 0.0036 0.0006 1.11E-09 4.27E-04 0.00 cg27663031 exclusive BF duration categories ≤3 m vs. never 754 vs. 156 0.0042 0.0011 1.07E-04 4.41E-01 0.51 chr5 105,791,135 intergenic – – – age↓ >3 m vs. never 183 vs. 156 0.0080 0.0015 1.08E-07 2.07E-02 0.62 cg00315563 exclusive BF duration categories ≤3 m vs. never 754 vs. 156 0.0170 0.0038 9.25E-06 2.60E-01 0.00 chr3 182,243,039 intergenic – – rs2700866, others age↓ >3 m vs. never 183 vs. 156 0.0202 0.0038 1.61E-07 2.07E-02 0.10 Open in a new tab BF: breastfeeding, COPD: chronic obstructive pulmonary disease, CRP: C-reactive protein, EWAS: epingenome-wide association study, FDR: false discovery rate, m: months, MAF: minor allele frequency, SE: standard error, SNP: single nucleotide polymorphism *All associations with age are based on the fixed effect model of the paper by Mulder et al., except for this one which is based on the random effect model Associations in bold are FDR-significant Fig. 1. Open in a new tab Forest plots representing the association of exclusive breastfeeding duration (months) with FDR-significant CpG sites across cohorts ordered by age: A cg01257194 at the 5’UTR of ALAD gene, B cg20053493 at the FNBP4 gene body, C cg04942655 intergenic, and D cg20702204 at the CHFR gene body. Both fixed and random effects are shown, as well as heterogeneity (I 2 ). Mean age at DNA methylation assessment in the different cohorts was: ALSPAC (7.45 years), Viva (7.78 years), BAMSE.E (8.30 years), BAMSE.M (8.34 years), IOW (10.00 years), HOME (12.31 years), and GLAKU (12.31 years) Exclusive breastfeeding duration longer than three months compared to never being breastfed was also associated with three CpG sites: cg20053493 ( FNBP4 ), cg27663031 (intergenic), and cg00315563 (intergenic) (Table 2 ). Cg00315563 (intergenic) and cg20053493 ( FNBP4 ) were also FDR-significant in sensitivity models when restricting to European ancestry and to cohorts with > 100 participants. Cg20053493 ( FNBP4 ) was also associated in the meta-EWAS of exclusive breastfeeding duration in months. Cg27663031 (intergenic) presented high heterogeneity (I 2 > 0.6), although the forest plot showed effects in the same direction across cohorts (Fig. 2 ; Supplemental Table S8 B (Supplementary Material 10 )). Additionally, one CpG site was FDR-significant only in cohorts with > 100 participants (> 3 months vs. never: cg22941178, LOC284837 , effect = 0.0152, SE = 0.0029, p FDR = 0.024) (Supplemental Table S2 (Supplementary Material 3 )). The association at cg22941178 ( LOC284837 ) was similar in the main model, though not FDR-significant (effect = 0.0129, SE = 0.0027, p FDR = 0.141). Fig. 2. Open in a new tab Forest plots representing the association of exclusive breastfeeding categories (never, ≤3 months, >3 months) with FDR-significant CpG sites across cohorts: A cg20053493 at the FNBP4 gene body, B cg27663031 intergenic, and C cg00315563 intergenic. Both fixed and random effects are shown, as well as heterogeneity (I 2 ). Mean age at DNA methylation assessment in the different cohorts was: ALSPAC (7.45 years), IOW (10.00 years), HOME (12.31 years), and GLAKU (12.31 years). IOW did not provide results for cg00315563 The lambda inflation factors and EWAS results for exclusive breastfeeding are shown in Supplemental Tables S3D-E, S4D-E, S9A-D, S10A-C (Supplementary Materials 4 , 11 and 12 ), and Supplemental Figures S4, S5A-B (Supplementary Material 6 ). Negative outcome control analysis Six cohorts participated in the negative outcome control analyses focused on cord blood DNAm in relation to breastfeeding (Table 1 and Supplemental Tables S1C-D (Supplementary Material 2 )). Four of these cohorts also participated in the meta-EWASs of breastfeeding and childhood blood DNAm, but with different sample sizes. The number of ever breastfed children was 3015 vs. 306 never breastfed. Data on the duration of any and exclusive breastfeeding were available for 3321 and 2152 children, respectively. Two cohorts were from non-European ancestry and all had > 100 participants. None of the CpG sites identified in the main meta-EWAS of childhood blood DNAm were among the top CpG sites ( p < 0.05) in cord blood. A summary of the full EWAS results is shown in Supplemental Tables S3 (by cohort: Supplementary Material 4 ) and Supplemental Tables S4 (meta-EWASs: Supplementary Material 5 ), volcano plots in Supplemental Figures S1 -S5 (Supplementary Material 6 ), and top CpG sites in Supplemental Tables S5E-F, S6E-F, S7C-D and S9E-F, S10D-E (Supplementary Materials 7 , 8 , 11 and 12 ), for any and exclusive breastfeeding, respectively. In the main model for cord blood, three CpG sites were FDR-significant for any breastfeeding and nine for exclusive breastfeeding duration. When restricting to European ancestry, three CpG sites were FDR-significant, including one associated with any breastfeeding, one with duration of any breastfeeding, and one with exclusive breastfeeding for > 3 months. None of these CpG sites were FDR-significantly associated with breastfeeding in the meta-EWAS of childhood blood (Supplemental Table S11 (Supplementary Material 13 )). Functional characterisation and comparison with previous studies Three of the six CpG sites identified in the main model were located in the 5’UTR promoter and gene body of known genes (Table 2 ). None of the six CpG sites are cis -eQTMs, whereas three of them had mQTLs. Previous associations of these CpG sites with other traits or exposures included childhood age, alcohol, biological sex, and inflammation markers. Functional enrichment analyses of suggestively associated CpG sites ( p < 1E-05) for exclusive breastfeeding definitions are presented in Supplemental Table S12 (Supplementary Material 14 ). Pathways enrichment emerging from these analyses included Ras signalling pathway, smoking, air pollution, and paracetamol exposure, as well as several diseases, including maternal depression and hypertensive disorders. A look up of the literature showed that out of nine CpG sites previously associated with breastfeeding [ 17 , 19 , 21 , 37 ], one (cg11414913 – intergenic and previously described in the ALSPAC cohort) [ 37 ] was replicated in our study ( p < 0.05 and same direction of effect) (Supplemental Table S13 (Supplementary Material 15 )). We also found that four out of the sixteen CpG sites annotated to the LEP gene were inversely associated with any breastfeeding at nominal significance (effect range: -0.001 to -0.006, SE range: 0.0004 to 0.0039, all p < 0.05). Discussion In this meta-EWASs of breastfeeding practices involving eleven cohorts from the PACE Consortium, we identified six CpG sites at which DNAm was associated with exclusive breastfeeding. Five of them remained significant in at least one of the sensitivity analyses, and none were found in the negative outcome control meta-EWASs (cord blood). Substantial heterogeneity across cohorts was observed and the effects were generally of small magnitude, ranging from 0.36% to 2.02% of average DNAm differences between breastfeeding categories, and from 0.04% to 0.17% for each month of exclusive breastfeeding. Although small effect sizes are challenging to interpret biologically, they have been repeatedly reported in EWAS research, likely reflecting cell-specific effects that may nonetheless have a substantial impact on gene expression [ 38 ]. For example, studies in blood have reported differences of approximately 0.5% in DNAm between smokers and non-smokers [ 39 ], and of 0.07% per one-point increase in Mediterranean diet score [ 40 ]. Longer exclusive breastfeeding was associated with increased DNAm at cg01257194 (at the 5’UTR of the ALAD gene), cg04942655 (intergenic), cg20702204 (at the CHFR gene body) and cg20053493 (at the FNBP4 gene body). ALAD encodes an enzyme that catalyses the second step in the porphyrin and heme biosynthetic pathway, CHFR is known to regulate cell cycle entry into mitosis, and FNBP4 is involved in the regulation of cytoskeletal dynamics during cell division, migration and vesicle formation. Cg20053493 was also positively related to exclusive breastfeeding for ≥ 3 months, while two other intergenic CpG sites (cg00315563 and cg27663031) presented higher DNAm in the group of children exclusively breastfed for more than three months. Of the six identified CpG sites, two are particularly noteworthy: cg01257194, previously related to higher levels of CRP and with alcohol consumption [ 41 , 42 ], and cg20702204 at the CHFR gene body. CHFR gene variants have been reported for weight, height, BMI [ 43 ], cognitive abilities [ 44 ], and behavioural traits [ 45 ], suggesting that DNAm at this gene could be involved in developmental effects of breastfeeding. However, further studies need to confirm this and elucidate a potential mechanism, particularly as the association between breastfeeding and higher DNAm at cg01257194 seems to go in the opposite direction to the one expected for the known protective effects of breastfeeding on inflammation [ 46 ]. For the remaining CpG sites, it was not possible to estimate associations with gene or protein expression, making it more challenging to hypothesize their functional consequences. The observation that all FDR-significant CpG sites were previously reported to show differences in DNAm levels with age from birth to adolescence is intriguing and could potentially have a biological explanation [ 47 ]. For instance, both breastfeeding and aging are dynamic processes that can influence DNAm patterns through various mechanisms, such as hormonal changes or maturation of the immune system (e.g. through changes in blood cellular composition) [ 7 , 47 ]. However, the observed overlap could be also attributed to residual confounding or other unaccounted factors in the meta-EWASs, despite adjusting for child age at the moment of blood cell DNAm measurements in all our models. We conducted functional enrichment analyses for suggestive CpG sites associated with exclusive breastfeeding. The annotation of CpG sites associated with the continuous exclusive breastfeeding variable showed enrichment in genes related to the Ras signalling pathway, which regulates cell growth, differentiation, and survival in response to external signals such as growth factors or cytokines. Enrichment was also observed for smoking, air pollution, and paracetamol exposure, as well as for several maternal conditions such as maternal depression and maternal hypertensive disorders during pregnancy. This could suggest potential residual confounding, but these analyses should be interpreted with caution, as they are based on CpG sites that were identified with p < 1E-05 threshold (classified as ‘suggestive’ only). This study has several strengths. First, it is the largest investigation to date, incorporating data from eleven cohorts within the PACE Consortium for the primary meta-EWASs, and thus increasing the statistical power and robustness compared with previous studies. Second, we examined various definitions of both any and exclusive breastfeeding, allowing for a comprehensive investigation. Third, we used data from prospective studies, thereby reducing recall bias on breastfeeding practices. Fourth, we included a negative outcome control study. The aim of the negative outcome control analysis was to filter out potential non-causal associations; however, we acknowledge that the only partial overlap of participants between the main and negative control analyses constrained this approach. Finally, we included several sensitivity analyses, thereby strengthening the robustness of the results. The study also has some limitations. First, there were differences in exposure and outcome assessment across cohorts. Questionnaires on breastfeeding were administered at different postnatal visits in each cohort, and they did not account for potential use of donated human milk. Additionally, blood DNAm was measured at different ages across cohorts, including later childhood. Studies assessing blood DNAm closer to the time of breastfeeding initiation might be more likely to detect associations if effects on the epigenome are not sustained. Differences in both exposure definitions and age at sampling across cohorts could have contributed to the relatively higher heterogeneity observed in associations at some of the CpG sites. Second, the sample size of the study for certain categories of breastfeeding in some of the individual cohorts was still relatively small, thus not allowing us to test sub-categories for either any or exclusive breastfeeding. Breastfeeding categories were defined as a compromise between WHO recommendations, maternity leave, and available sample sizes. Third, while the statistical models were adjusted for several confounding variables, the possibility of residual confounding remains (e.g. by diet/food components, maternal lifestyle or socioeconomic factors). Fourth, we measured DNAm in blood cells, for reasons of accessibility. As DNAm is tissue-specific, this may not reflect associations with DNAm in other organs and tissues that may be relevant for breastfeeding-associated health outcomes. In addition, only DNAm was evaluated, while other epigenetic or molecular mechanisms, such as mitochondrial DNA content [ 48 ], could also contribute. Finally, although we included several cohorts spanning different continents, our sample was of predominantly European genetic ancestry, which may limit the extent to which our findings generalize to the full spectrum of human genetic and environmental diversity. Conclusions This meta-EWAS identified differential DNAm in childhood blood cells in relation to having been exclusively breastfed. Effect sizes observed were of small magnitude, and the results showed some degree of heterogeneity across cohorts. Consequently, the findings should be interpreted with caution. More diverse samples at younger ages, combined with further detailed examination of confounding structures and causal studies, are needed to fully understand potential underlying biological mechanisms. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (210KB, docx) Supplementary Material 2 (38.5KB, xlsx) Supplementary Material 3 (13.3KB, xlsx) Supplementary Material 4 (31KB, xlsx) Supplementary Material 5 (22.8KB, xlsx) Supplementary Material 6 (388.7KB, docx) Supplementary Material 7 (45KB, xlsx) Supplementary Material 8 (40.6KB, xlsx) Supplementary Material 9 (39.8KB, xlsx) Supplementary Material 10 (16.8KB, xlsx) Supplementary Material 11 (43KB, xlsx) Supplementary Material 12 (49.4KB, xlsx) Supplementary Material 13 (64.6KB, xlsx) Supplementary Material 14 (11.4KB, xlsx) Supplementary Material 15 (21.5KB, xlsx) Acknowledgements We would like to thank all the families of the cohorts for their generous contribution. Abbreviations ALL analysis conducted in all participants from all cohorts ALL100 analysis conducted in all participants from cohorts with a sample size > 100 BMI body mass index CpG cytosine-phosphate-guanine dinucleotide DNAm DNA methylation EUR analysis conducted in only children of European ancestry from all cohorts EUR100 analysis conducted in only children of European ancestry from cohorts with a sample size > 100 EWAS epigenome-wide association study FDR false discovery rate IQR interquartile range Meta-EWAS meta-analysis of epigenome-wide association studies PC principal component SE standard error SNP single nucleotide polymorphism TSS transcription start site WHO World Health Organization 3’UTR 3’ untranslated region 5’UTR 5’ untranslated region Author contributions BH, DCa, JFF, MB and SFB designed the study. BLG, BK, BH, BN, CLR, DJS, EM, EO, HJZ, IK, JFF, JL, JMB, JW, JWH, KMC, KR, KY, LD, MB, MFH, MK, MKa, MV, NK, RS, SH, SFB, TM, VG, VWVJ, and WK are the PIs of the cohorts or obtained data for the study. AC, AH, AM, AW, BN, DC, DCa, DP, EC, FIR, HL, HRE, JL, MES, OG, RF, RG, RMS, RO, SFB, SKM, SLR, and TM conducted the analyses in the individual cohorts. AC, HL and MB conducted the meta-analyses. DCa, JFF, MB, and SFB interpreted results and wrote the original draft, and all others contributed to the later drafts, and reviewed and approved the final version. Funding This work was supported by the European Union’s Horizon 2020 research and innovation program - ATHLETE project [grant number 874583], LongITools project [grant number 874739], HELIX project [grant number 308333], and LifeCycle project [grant number 733206], and by the European Joint Programming Initiative “A Healthy Diet for a Healthy Life” (JPI HDHL) [NutriPROGRAM project, ZonMw the Netherlands no.529051022 and and Instituto de Salud Carlos III no. AC18/00006 and PREcisE project ZonMw the Netherlands no.529051023] . Details of funding for each study and for contributions of individual studies can be found in Supplemental Methods, Supplementary Material 1 . No funders listed here or in Supplemental Material 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. Data availability The full genome-wide DNAm meta-analysis summary statistics presented in this manuscript are available at ZENODO (10.5281/zenodo.15624186). Cohort-level data may be available by direct contact to the authors. Participant-level data are not openly available due to informed consent coverage. Access to these data is managed at each institution’s according to their policies. Declarations Ethics approval and consent to participate Research was conducted in accordance with the Declaration of Helsinki. All participating studies received approval from their respective institutional ethics committees, and all participants provided written informed consent. Consent for publication Not applicable. Competing interests JMB has served as an expert witness on behalf of plaintiffs involved in litigation related to PFAS-contaminated drinking water. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Doretta Caramaschi and Sílvia Fernández-Barrés contributed equally to this work. Janine F. Felix and Mariona Bustamante contributed equally to this work. Contributor Information Doretta Caramaschi, Email: [email protected]. Mariona Bustamante, Email: [email protected]. References 1. Smithers LG, Kramer MS, Lynch JW. Effects of breastfeeding on obesity and intelligence: causal insights from different study designs. JAMA Pediatr. 2015. 10.1001/jamapediatrics.2015.0175. [ DOI ] [ PubMed ] [ Google Scholar ] 2. Camacho-Morales A, Caba M, García-Juárez M, Caba-Flores MD, Viveros-Contreras R, Martínez-Valenzuela C. Breastfeeding contributes to physiological immune programming in the newborn. Front Pediatr. 2021. 10.3389/fped.2021.744104. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Horta BL, de Lima NP. Breastfeeding and type 2 diabetes: systematic review and Meta-Analysis. Curr Diab Rep. 2019. 10.1007/s11892-019-1121-x. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Horta BL, De Loret C, Victora CG. Breastfeeding and intelligence: A systematic review and meta-analysis. Acta Paediatr Int J Paediatr. 2015. 10.1111/apa.13139. [ DOI ] [ PubMed ] [ Google Scholar ] 5. Horta BL, Loret De Mola C, Victora CG. Long-term consequences of breastfeeding on cholesterol, obesity, systolic blood pressure and type 2 diabetes: A systematic review and meta-analysis. Acta Paediatr Int J Paediatr. 2015. 10.1111/apa.13133. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Alotiby AA. The role of breastfeeding as a protective factor against the development of the immune-mediated diseases: A systematic review. Front Pediatr. 2023. 10.3389/fped.2023.1086999. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Christian P, Smith ER, Lee SE, Vargas AJ, Bremer AA, Raiten DJ. The need to study human milk as a biological system. Am J Clin Nutr. 2021. 10.1093/ajcn/nqab075. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. WHO. Global Strategy for Infant and Young Child Feeding. Fifthy-fourth world Heal. Assem. 2003. 9. Kramer M, Kakuma R. Optimal duration of exclusive breastfeeding (Review). Cochrane database syst rev. Cochrane Database Syst Rev. 2012. 10.1002/14651858.CD003517.pub2. www.cochranelibrary.com . 10. Canani RB, Di Costanzo M, Leone L, Bedogni G, Brambilla P, Cianfarani S, et al. Epigenetic mechanisms elicited by nutrition in early life. Nutr Res Rev. 2011. 10.1017/S0954422411000102. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Verduci E, Banderali G, Barberi S, Radaelli G, Lops A, Betti F, et al. Epigenetic effects of human breast milk. Nutrients. 2014. 10.3390/nu6041711. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Jones PA, Takai D. The role of DNA methylation in mammalian epigenetics. Science. 2001. 10.1126/science.1063852 [ DOI ] [ PubMed ] [ Google Scholar ] 13. Obermann-Borst SA, Eilers PHC, Tobi EW, De Jong FH, Slagboom PE, Heijmans BT, et al. Duration of breastfeeding and gender are associated with methylation of the LEPTIN gene in very young children. Pediatr Res. 2013. 10.1038/pr.2013.95. [ DOI ] [ PubMed ] [ Google Scholar ] 14. Pauwels S, Symons L, Vanautgaerden EL, Ghosh M, Duca RC, Bekaert B, et al. The influence of the duration of breastfeeding on the infant’s metabolic epigenome. Nutrients. 2019. 10.3390/nu11061408. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Naumova OY, Odintsova VV, Arincina IA, Rychkov SY, Muhamedrahimov RJ, Shneider YV, et al. A study of the association between breastfeeding and DNA methylation in peripheral blood cells of infants. Russ J Genet [Internet]. 2019;55:749–55. 10.1134/S1022795419060103. [ Google Scholar ] 16. Rossnerova A, Tulupova E, Tabashidze N, Schmuczerova J, Dostal M, Rossner P, et al. Factors affecting the 27K DNA methylation pattern in asthmatic and healthy children from locations with various environments. Mutat Res - Fundam Mol Mech Mutagen. 2013. 10.1016/j.mrfmmm.2013.02.003. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Hartwig FP, De Mola CL, Davies NM, Victora CG, Relton CL. Breastfeeding effects on DNA methylation in the offspring: A systematic literature review. PLoS ONE. 2017. 10.1371/journal.pone.0173070. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Sherwood WB, Bion V, Lockett GA, Ziyab AH, Soto-Ramírez N, Mukherjee N, et al. Duration of breastfeeding is associated with leptin (LEP) DNA methylation profiles and BMI in 10-year-old children. Clin Epigenetics. 2019. 10.1186/s13148. 019 0727 9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Walker-Short E, Buckner T, Vigers T, Carry P, Vanderlinden LA, Dong F, et al. Epigenome-wide association study of infant feeding and DNA methylation in infancy and childhood in a population at increased risk for type 1 diabetes. Nutrients. 2021. 10.3390/nu13114057. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Mallisetty Y, Mukherjee N, Jiang Y, Chen S, Ewart S, Hasan Arshad S, et al. Epigenome-wide association of infant feeding and changes in Dna methylation from birth to 10 years. Nutrients. 2021. 10.3390/nu13010099. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Odintsova VV, Hagenbeek FA, Suderman M, Caramaschi D, van Beijsterveldt CEM, Kallsen NA, et al. DNA methylation signatures of breastfeeding in buccal cells collected in mid-childhood. Nutrients. 2019. 10.3390/nu11112804. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Hartwig FP, Smith GD, Simpkin AJ, Victora CG, Relton CL, Caramaschi D. Association between breastfeeding and Dna methylation over the life course: findings from the Avon longitudinal study of parents and children (alspac). Nutrients. 2020. 10.3390/nu12113309. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Briollais L, Rustand D, Allard C, Wu Y, Xu J, Rajan SG, et al. DNA methylation mediates the association between breastfeeding and early-life growth trajectories. Clin Epigenetics Ger. 2021;13:231. 10.1186/s13148. 021 01209 z. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Felix JF, Joubert BR, Baccarelli AA, Sharp GC, Almqvist C, Annesi-Maesano I et al. Cohort profile: pregnancy and childhood epigenetics (PACE) consortium. Int J Epidemiol. 2018. [ DOI ] [ PMC free article ] [ PubMed ] 25. 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. Ting AH, editor. PLoS One [Internet]. 2012;7:e41361. 10.1371/journal.pone.0041361 [ DOI ] [ PMC free article ] [ PubMed ] 26. 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 Bioinf [Internet]. 2012;13:86. 10.1186/1471. 2105 13 86. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Gervin K, Salas LA, Bakulski KM, van Zelm MC, Koestler DC, Wiencke JK, et al. Systematic evaluation and validation of reference and library selection methods for Deconvolution of cord blood DNA methylation data. Clin Epigenetics. 2019;11:125. 10.1186/s13148-019-0717-y. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Solomon O, MacIsaac J, Quach H, Tindula G, Kobor MS, Huen K, et al. Comparison of DNA methylation measured by illumina 450K and EPIC beadchips in blood of newborns and 14-year-old children. Epigenetics. 2018. 10.1080/15592294.2018.1497386. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Zhou W, Laird PW, Shen H. Comprehensive characterization, annotation and innovative use of infinium DNA methylation BeadChip probes. Nucleic Acids Res. 2017. 10.1093/nar/gkw967. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Magi R, Morris AP. GWAMA: software for genome-wide association meta-analysis. BMC Bioinformatics. 2010;11:288. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Viechtbauer W. Conducting Meta-analysis in R with the metafor package. J Stat Softw. 2010. 32. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B. WileyRoyal Statistical Society; 1995. pp. 289–300. 10.2307/2346101 33. Wickham H. Journeal Stat Softw. 2017. 10.1007/978. 0 387 98141 3. ggplot2: Elegant Graphics for Data Analysis. 34. Min JL, Hemani G, Hannon E, Dekkers KF, Castillo-Fernandez J, Luijk R, et al. Genomic and phenotypic insights from an atlas of genetic effects on DNA methylation. Nat Genet. 2021. 10.1038/s41588. 021 00923 x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. 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. 10.7554/eLife.65310. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Battram T, Yousefi P, Crawford G, Prince C, Babei MS, Sharp G, et al. The EWAS catalog: a database of epigenome-wide association studies. OSF Prepr OSF Preprints. 2021;4. 10.31219/OSF.IO/837WN. [ DOI ] [ PMC free article ] [ PubMed ] 37. Hartwig FP, Smith GD, Simpkin AJ, Victora CG, Relton CL, Caramaschi D. Association between breastfeeding and DNA methylation over the life course: findings from the Avon longitudinal study of parents and children (ALSPAC). BioRxiv [Internet]. 2019;800722. 10.1101/800722. [ DOI ] [ PMC free article ] [ PubMed ] 38. Breton CV, Marsit CJ, Faustman E, Nadeau K, Goodrich JM, Dolinoy DC, et al. Small-Magnitude effect sizes in epigenetic end points are important in children’s environmental health studies: the children’s environmental health and disease prevention research center’s epigenetics working group. Environ Health Perspect. 2017;125:511–26. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Hoang TT, Lee Y, McCartney DL, Kersten ETG, Page CM, Hulls PM et al. Comprehensive evaluation of smoking exposures and their interactions on DNA methylation. eBioMedicine. 2024. 10.1016/j.ebiom.2023.104956 [ DOI ] [ PMC free article ] [ PubMed ] 40. 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. [ DOI ] [ PMC free article ] [ PubMed ] 41. Hillary RF, Ng HK, McCartney DL, Elliott HR, Walker RM, Campbell A, et al. Blood-based epigenome-wide analyses of chronic low-grade inflammation across diverse population cohorts. Cell Genomics Authors. 2024;4:100544. 10.1016/j.xgen.2024.100544. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Dugué PA, Wilson R, Lehne B, Jayasekara H, Wang X, Jung CH, et al. Alcohol consumption is associated with widespread changes in blood DNA methylation: analysis of cross-sectional and longitudinal data. Addict Biol. 2021. 10.1111/adb.12855. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Sakaue S, Kanai M, Tanigawa Y, Karjalainen J, Kurki M, Koshiba S, et al. A cross-population atlas of genetic associations for 220 human phenotypes. Nat Genet. 2021. 10.1038/s41588. 021 00931 x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. Davies G, Lam M, Harris SE, Trampush JW, Luciano M, Hill WD, et al. Study of 300,486 individuals identifies 148 independent genetic loci influencing general cognitive function. Nat Commun. 2018. 10.1038/s41467. 018 04362 x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 45. Karlsson Linnér R, Mallard TT, Barr PB, Sanchez-Roige S, Madole JW, Driver MN, et al. Multivariate analysis of 1.5 million people identifies genetic associations with traits related to self-regulation and addiction. Nat Neurosci. 2021. 10.1038/s41593. 021 00908 3. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 46. McDade TW, Metzger MW, Chyu L, Duncan GJ, Garfield C, Adam EK. Long-term effects of birth weight and breastfeeding duration on inflammation in early adulthood. Proceedings Biol Sci. England; 2014;281:20133116. [ DOI ] [ PMC free article ] [ PubMed ] 47. Mulder RH, Neumann A, Cecil CAM, 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. 10.1093/hmg/ddaa280. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 48. Cosemans C, Nawrot TS, Janssen BG, Vriens A, Smeets K, Baeyens W et al. Breastfeeding predicts blood mitochondrial DNA content in adolescents. Sci Rep. 2020. [ DOI ] [ PMC free article ] [ PubMed ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Citations Wickham H. Journeal Stat Softw. 2017. 10.1007/978. 0 387 98141 3. ggplot2: Elegant Graphics for Data Analysis. Supplementary Materials Supplementary Material 1 (210KB, docx) Supplementary Material 2 (38.5KB, xlsx) Supplementary Material 3 (13.3KB, xlsx) Supplementary Material 4 (31KB, xlsx) Supplementary Material 5 (22.8KB, xlsx) Supplementary Material 6 (388.7KB, docx) Supplementary Material 7 (45KB, xlsx) Supplementary Material 8 (40.6KB, xlsx) Supplementary Material 9 (39.8KB, xlsx) Supplementary Material 10 (16.8KB, xlsx) Supplementary Material 11 (43KB, xlsx) Supplementary Material 12 (49.4KB, xlsx) Supplementary Material 13 (64.6KB, xlsx) Supplementary Material 14 (11.4KB, xlsx) Supplementary Material 15 (21.5KB, xlsx) Data Availability Statement The full genome-wide DNAm meta-analysis summary statistics presented in this manuscript are available at ZENODO (10.5281/zenodo.15624186). Cohort-level data may be available by direct contact to the authors. Participant-level data are not openly available due to informed consent coverage. Access to these data is managed at each institution’s according to their policies. Articles from Clinical Epigenetics are provided here courtesy of BMC ACTIONS View on publisher site PDF (1.5 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top