Age-, sex- and disease subtype–related foetal growth differentials in childhood acute myeloid leukaemia risk: A Childhood Leukemia International Consortium analysis

Age-, sexand disease subtypefetal growth differentials in childhood acute myeloid leukemia risk: results from 22 studies participating in a Childhood Leukemia International Consortium analysis Maria A. Karalexi1*a, PhD, Nick Dessypris1*a, PhD, Xiaomei Ma2, PhD, Logan G. Spector3, PhD, Erin Marcotte3, PhD, Jacqueline Clavel4,5, PhD, Maria S. Pombo-de-Oliveira6a, PhD, Julia E. Heck7, PhD, Eve Roman8, PhD, Beth A. Mueller9,10, DrPH, Johnni Hansen11, PhD, Anssi Auvinen12, PhD, Pei-Chen Lee13, PhD, Joachim Schüz14a, PhD, Corrado Magnani15, PhD, Ana M. Mora16, PhD, John D. Dockerty17, PhD, Michael E. Scheurer18, PhD, Rong Wang2, PhD, Audrey Bonaventure4, PhD, Eleanor Kane8, PhD, David R. Doody9, MS, NARECHEM-ST group19, FRECCLE group20, Friederike Erdmann14,21, PhD, Alice Y. Kang22, PhD, Catherine Metayer22, PhD, Elizabeth Milne23a, PhD, Eleni Th. Petridou1,24a, PhD Affiliations: 1Department of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, Athens, Greece 2Department of Chronic Disease Epidemiology, Yale School of Public Health, Cancer Prevention and Control, Yale Comprehensive Cancer Center, Yale School of Medicine, CT, USA 3Division of Epidemiology & Clinical Research, Department of Pediatrics, University of Minnesota, Minneapolis, MN, USA 4CRESS, UMR-S1153, INSERM, Paris-Descartes University, Villejuif, France 5National Registry of Childhood Cancers, APHP, Hôpital Paul-Brousse, CHU de Nancy, France 6Pediatric Hematology-Oncology Program Instituto Nacional de Cancer, Rio de Janeiro, Brazil This is the accepted manuscript of the article, which has been published in European Journal of Cancer. 2020, 130, 1-11. https://doi.org/10.1016/j.ejca.2020.01.018 7Department of Epidemiology, School of Public Health, University of California, Los Angeles, CA, USA 8Epidemiology and Cancer Statistics Group, Department of Health Sciences, University of York, Heslington, York, United Kingdom 9Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA 10Department of Epidemiology, School of Public Health, University of Washington, Seattle, WA, USA 11Danish Cancer Society Research Center, Copenhagen, Denmark 12Faculty of Social Sciences, University of Tampere, Tampere, Finland 13Department of Health Care Management, National Taipei University of Nursing and Health Sciences, Taipei 14International Agency for Research on Cancer (IARC), Section of Environment and Radiation, Lyon, France 15Cancer Epidemiology Unit, Department of Translational Medicine, CPO Piedmont and University of Eastern Piedmont, Novara, Italy 16Central American Institute for Studies on Toxic Substances (IRET), Universidad Nacional, Heredia, Costa Rica 17Department of Preventative and Social Medicine, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand 18Baylor College of Medicine, Department of Pediatrics Texas Children’s Cancer Center, TX, USA 19NARECHEM-ST group, Greece: Margarita Baka, Maria Moschovi, Sophia Polychronopoulou, Maria Kourti, Emmanuel Hatzipantelis, Eftichia Stiakaki, Helen Dana, Maria Kantzanou, Marianna Tzanoudaki, Theodora Anastasiou, Maria Grenzelia, Eleni Gavriilaki, Ioanna Sakellari, Achilles Anagnostopoulos, Vassiliki Kitra, Anna Paisiou, Evdoxia Bouka 20FRECCLE, Finnish Register-Based Case-Control Study of Childhood Leukemia group: Atte Nikkilä, Olli Lohi 21Danish Cancer Society Research Center, Childhood Cancer Research Group, Copenhagen, Denmark 22School of Public Health, University of California, Berkeley, CA, USA 23Telethon Institute for Child Health Research, Centre for Child Health Research, University of Western Australia, WA, Australia 24Clinical Epidemiology Unit, Department of Medicine, Karolinska Institute, Stockholm, Sweden *Equally contributed aCore Writing Group Correspondence to: Eleni Th. Petridou MD, MPH, PhD Professor of Epidemiology and Preventive Medicine Department of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, 75 Mikras Asias Str, Athens Greece 11527 Email: epetr[email protected]a.gr; Tel +30 210-7462187, Fax +30 210-7462105 ABSTRACT Background: Evidence for an association of fetal growth with acute myeloid leukemia (AML) is inconclusive. AML is a rare childhood cancer, relatively more frequent in girls, with distinct features in infancy. In the context of the Childhood Leukemia International Consortium (CLIC), we examined an a priori hypothesis that the association may vary by age, sex and disease subtype using data comprising 22 studies and a total of 3564 AML cases. Methods: Pooled estimates by age, sex and overall for harmonized fetal growth measures in association with AML risk were calculated using the INTERGROWTH 21st project for 17 studies contributing individual-level data; thereafter, meta-analyses were conducted with effect estimates provided ad hoc due to administrative constraints by five more studies. Subanalyses by AML subtype were also performed. Findings: A nearly 50% greater risk was observed among large for gestational age (LGA) infant boys <1 year [odds ratio (OR): 1.49, 95% confidence intervals (CI): 1.03-2.14], reduced to 34% among boys <2 years (OR: 1.34, 95% CI: 1.05-1.71) and 25% for all age boys (OR: 1.25, 95% CI: 1.06-1.46). The association became stronger among boys with M0/M1subtype (OR: 1.80, 95% CI: 1.15-2.83). The association among boys with large birth length for gestational age was 1.38 (95% CI: 1.00-1.92). No index of decelerated fetal growth was associated with AML risk. Interpretation: Greater fetal growth was associated with AML, especially in infant boys and those with minimally differentiated myeloid leukemia. Further cytogenetic research would help inform the underlying mechanisms. Keywords: fetal growth; birthweight; birthweight for gestational age; birth length; weightfor-length ratio; acute myeloid leukemia; subtypes; childhood; sex; meta-analysis Research in context Evidence before this study Fetal growth reflects a complex array of underlying mechanisms, including genetic and epigenetic factors, environmental exposures, maternal pathology and nutritional status. Evidence for an association of fetal growth with acute myeloid leukemia (AML) is inconclusive. AML is a rare childhood cancer with distinct features in infancy and is relatively more frequent in girls as contrasted with other childhood cancers. Added value of this study The pooled analysis and meta-analysis of the largest international dataset (N=3564 AML cases and 9584 controls) comprising 22 studies participating in the Childhood Leukemia International Consortium analysis provide evidence for a positive association of excess fetal growth with childhood AML. The association was robust among large for gestational age newborns, specifically infant boys and those with minimally differentiated myeloid cells subtypes. The impact of accelerated fetal growth on AML risk was attenuated among older ages. A positive association of large birth length for gestational age was also found, confined also to boys, whereas the positive association of the large weight-for-height ratio adjusted for gestational age did not reach statistical significance. Neither low birthweight nor smaller for gestational age were associated with AML risk. Implications of all the available evidence Our results showing accelerated fetal growth age-, sexand disease subtype-related differentials in AML risk, if replicated in future cytogenetic research could further refine our understanding of the biological pathways through which fetal growth environment may be implicated in the risk of childhood AML and its specific subtypes. INTRODUCTION Fetal growth is one of the most commonly studied perinatal risk factors of childhood cancer (1-5). Measures of fetal growth reflect a complex array of underlying mechanisms including genetic and epigenetic factors, environmental exposures, maternal pathology and nutritional status (6). Numerous publications have examined the potential association between high birthweight (HBW), as a gross indicator of fetal growth and childhood acute lymphoblastic leukemia (ALL); however, their results remain inconclusive for acute myeloid leukemia (AML) (1, 4, 7, 8). AML is a rare childhood cancer with distinct features in infancy and relatively more frequent in girls as contrasted with other childhood cancers (9, 10). A U-shape association of AML risk has been reported with birthweight (7); yet, fewer studies have explored the potential relationship of more robust fetal growth measures, such as birthweight adjusted for gestational age, birth length, weight-for-length ratio and proportion of optimal birthweight (POBW) (11, 12). A recent pooled study from the Childhood Leukemia International Consortium (CLIC) (4) and a German study (13), which was also included in the CLIC pooled analysis showed that fetal growth rate, rather than birthweight per se may be more strongly associated with childhood leukemia, especially ALL. Biological mechanisms underlying the possible effect of HBW on AML risk remain indecisive and may involve growth factors and epigenetics with chromatin modifiers (14). Birthweight is, partially, determined by the intrauterine environment and has been linked to cord blood levels of insulin-like growth factors (IGF) I and II, as well as sex steroid hormones (14, 15). Any association between low birthweight and AML is speculated to be due to fetal programming and genomic stability determined by epigenetic pathways, which affect the growth hormone-IGF axis through hyperinsulinemia, according to the “thrifty phenotype” hypothesis (4, 11, 16, 17). Given the rarity of the disease, its distinct age and sex characteristics and the inconclusive results of published literature, we aimed to assess whether measures of fetal growth are related to the risk of AML using data from questionnaire-based case-control (QCC) and registry-based case-control (RCC) studies in an international Childhood Leukemia International Consortium (CLIC) analysis. Specifically, we focused on the potential associations between AML and available fetal growth markers, namely birthweight, birthweight for gestational age, as well as birth length and weight-for-length ratio for gestational age by age, sex and overall, including subgroup analyses by disease subtype. METHODS Study population Fourteen QCC and eight RCC studies provided data for this collaborative CLIC analysis. Registration process and data collection in each study reported elsewhere (18), are summarized in Supplementary Table 1. Information for the RCC studies conducted in Denmark, Finland, Taiwan, and five States of the U.S. (California, Minnesota, New Yorkexcluding New York City, Texas and Washington State) was derived from linkage of nationwide or statewide population-based cancer registries with administrative registries (Supplementary Table 1). Due to administrative constraints, the CLIC Californian RCC study, as well as the non-CLIC RCC studies from Minnesota, New York, Texas and Taiwan contributed only adjusted summary effect estimates for the meta-analyses. Pooled estimates were derived from the remaining 17 studies, which provided individual-level data, namely the RCC CLIC studies from Denmark, Finland and Washington State, as well as the QCC CLIC studies conducted in Brazil, Costa Rica, France (ADELE, ELECTRE, ESCALE and ESTELLE), Germany, Greece, Italy, New Zealand, UK, and the U.S. [Children’s Oncology Group (COG)-AE24, COG-E14 and Texas]. All 22 studies used the same variables (birth characteristics, potential confounders and disease-related information) and the same statistical analysis program. Cases and controls (0-14 years) from the QCC COG-AE24 study born in California, Washington, Minnesota, New York and Texas (40 cases and 84 controls) were excluded due to overlap with the remaining U.S. studies. Children with Down syndrome were also excluded given the particular biological mechanisms of AML leukemogenesis in these patients (19). A total of 3564 AML cases and 9584 controls from the 22 participating studies (19 CLIC studies and 3 non-CLIC studies) were included in this analysis. Data collection and harmonization Information on socio-demographic and birth characteristics of cases and controls including fetal growth measures was harmonized across the studies. Whenever controls were frequencymatched to cases on age and sex (Brazil, Costa Rica, Denmark, France, Germany, New Zealand, Texas QCC study, Taiwan and the U.S. California, Minnesota, New York, Texas and Washington RCC studies), a maximum of three controls were randomly selected from the respective study databases. Data on primary exposures were provided by birth certificates, medical birth records or maternal self-reports, depending on the study. Birthweight for gestational age was examined using the 10th and 90th percentiles of the International Fetal and Newborn Growth Consortium for the 21st Century (INTERGROWTH-21st) birthweight standard (20). Three categories for birthweight for gestational age were used: Small for Gestational Age (SGA) for neonates weighing below the 10th percentile of the reference population, Appropriate for Gestational Age (AGA) for neonates weighing between the 10th and 90th percentile, and Large for Gestational Age (LGA) for neonates weighing above the 90th percentile. The same process was used to categorize the other fetal growth markers, namely birth length for gestational age and weight-for-length adjusted for gestational age ratio. Applying this procedure to our data resulted in approximately 20% of newborn categorized as LGA and 6-7% as SGA. We therefore label them as accelerated fetal growth (AFG) and low fetal growth (LFG) instead. Seeing that INTERGROWTH resulted in such a large number of children in the two extreme categories, for sensitivity analyses, we also calculated the joint growth distribution based on the pooled set of our controls and then applied to the study-specific cases. Statistical analysis The overall analysis model included the matching factors and potential confounders, selected a priori based on the existing literature: child’s age at diagnosis (<1, 1-4, 5-9, 10-14 years), sex (male, female), maternal age at birth (<25, 25-29, 30-34, ≥35years), birth order (1st, 2nd, ≥3rd), index child’s ethnicity (Caucasian, other) (18), plurality (yes, no), study of origin and prematurity (gestational age <37 weeks: yes, no) whenever appropriate to be introduced as an independent variable. The primary exposures of interest were birthweight, birthweight for gestational age, birth length and weight-for-length ratio adjusted for gestational age and they were alternatively introduced into the overall analysis model. Additionally, in order to provide comparable estimates with previous studies, which showed a U-shape association (6), the standard birthweight categories [<2500, 2500-3999 (reference), ≥4000 grams] were used to examine the impact of birthweight alone. The proportion of missing data for each variable per study is presented in Supplementary Table 2. Given the different biological characteristics of AML during infancy and the first two years of age, as well as the increasing incidence of the disease in girls as contrasted with other childhood cancers (10, 21), we initially assessed the associations of birthweight and birthweight for gestational age with AML risk separately for each gender (boys, girls) and for the age groups <1 year and 1-14 years or alternatively <2 years and 2-14 years. A formal test for interaction of birthweight or birthweight for gestational age with age or sex was thereafter applied in the pooled dataset as to explore this a priori hypothesis. In addition, we explored the potential effect modification of sex on the associations above by calculating the relative compared to that of females (49, 50). Hence, the consistent male-specific associations of AFG with AML in our study might be due to differential sex-hormonal interactions, namely higher concentrations of circulating androgens synthesized by the testes and sex-related differences in growth rate before differentiation of the fetal gonads (51, 52), which may lead among others to a higher mean weight of boys at birth as a result of the IGF axis activation, despite the fact that males as a rule are born one week earlier than girls (43, 53). The significant effect modification of sex on the association of AFG with AML risk in the older age group of children (2-14 years) in combination with the small gradual increase in the incidence of the disease in girls (annual percent change: +1.0%) as contrasted with the male preponderance characterizing other childhood cancers (9, 10) provide some evidence for the reliability of our results; nevertheless, the sex-related differentials of our study merit further consideration given the borderline significant interactions of sex with AFG in the total age group (0-14 years). To this end, further research is needed given that sex-specific associations of fetal growth with risk of other cancers, including ALL and central nervous system tumors, have also been described (3, 4). Strengths and limitations Main strengths of the present study include the sound methodological approach including the availability of the largest set of harmonized individual study data for this rare form of childhood cancer -especially infant AMLcontributed by 22 studies around the globe, which were pooled and meta-analyzed as appropriate in comprehensively-adjusted models. Indeed, low power was a substantial limitation of previous studies (54). Despite the proportion of missing values, we assessed several fetal growth markers beyond the gross marker of birthweight, overall and within informative sub-groups. In particular, we performed analyses using alternative fetal growth markers, such as birthweight, birth length and weight-for-length ratio adjusted for gestational age. Moreover, intergrowth standardized curves, based on a population-based, multiethnic, multi-country and sex-specific prospective study, were used (20, 55). Furthermore, results from population-based record linkage studies were materially the same as those springing from QCC studies and they were thus presented jointly. Finally, we performed stratified analyses by sex and AML morphological FAB-subtype, given the different endometrial environments and levels of growth factors by sex of embryos, as well as by age given the differential biological profile of infant AML as contrasted to the disease among older age groups of children. Regarding limitations in the assessment of exposures of interest and despite maternal selfreports of birthweight being considered reliable (56) different methods were used to report/record gestational age depending on the study; diverse were also the diagnostic periods across studies. Yet, the between-study heterogeneity was minimal and any inaccuracies in reporting between cases and the comparison groups are expected to be non-differential, since gestational age is not widely considered as a risk factor for childhood leukemia (57). Several CLIC-QCC studies are nationwide or region-wide easing concerns of control selection. It is true, however, that most studies provide partial or no information on cytogenetic recurrent aberrations according to the International Classification of Diseases for Oncology (ICD)-O-3 coding, especially for the KMT2A/MLL rearrangement status that could have allowed to explore a possible association with fetal growth. Additionally, there have been inherent limitations in application of the POBW formula (12) in this international study beyond the high proportion of missing data mainly on maternal height. Moreover, possible confounding factors, such as maternal smoking, diabetes and weight gain during pregnancy, were not included in the models due to high proportions of missing data leaving room for residual confounding. Lastly, there is no gold standard in defining AFG when comparing across populations and the use of the INTERGROWTH 21st standard international distribution did not only capture the top 10% of babies according to the traditional categorization of LGA, but the approximately top 20%, while the low fetal growth category encompassed roughly 6-7% of babies instead of 10%. There is, however, no reason to assume that increased AML risk was confined to only the top 10% so that AML risk is increased within the top 20% is merely an observation of our study. Future studies in larger samples, i.e. in countries with birth registries, should explore in more detail the dose-response function, if any. Besides that, alternative analyses employing the empirical 10% and 90% distribution within our controls had no impact on the results of the main analyses. Conclusion This is the largest study to-date to explore the association between fetal growth and childhood AML risk using robust markers and a priori designed age and sex sub-analyses. Our results are in line with those of previous studies showing a positive association with indices of accelerated fetal growth, such as HBW. The findings further specify, however, that the association is confined to boys known have a more accelerated fetal growth compared to girls, especially in infancy and with undifferentiated (M0) or with minimal maturation (M1) myeloid leukemia. By contrast, there seems to be no support for an association with decelerated fetal growth. Sub-group analyses on cytogenetics could further refine our understanding of the mechanisms through which accelerated fetal growth may increase childhood leukemia risk. Acknowledgments Disclaimer for use where authors from external institutions are involved: The authors alone are responsible for the views expressed in this article and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated. Specific acknowledgements by study: -Brazilian Collaborative Study Group of Infant Leukemia that notified cases for the epidemiological studies. -COG-AE24, research investigators: Julie Ross, Erin Marcotte (University of Minnesota), clinical investigators at the Children’s Oncology Group (COG) principal and affiliate member institutions. -Costa Rica CRCLS staff, especially CRCLS participants and their families; Noemy Gomez for her support during the data cleaning process. -Germany GCCR study: Drs Peter Kaatsch, Rolf Meinert, Uwe Kaletsch and Jörg Michaelis for their work in the original study -Nationwide Registry for Childhood Hematological Malignancies and Solid Tumors (NARECHEM-ST): Collaborative Study Group and complimentary sources for the nationwide collection of cases and Panagiota Bouka, field coordinator -Veronique Luzon (IARC) for data management of the CLIC Data Coordination Center Declaration of interests-Funding support NARECHEM-ST was partially supported by the Hellenic Society for Social Pediatrics and Health Promotion. The Brazilian study has been supported by CNPq research scholarships (#301594/2015-5) and FAPERJ (#E026/102.337/2013). The Danish study was supported by the US National Institutes of Health (R21CA175959, R03ES021643). The Taiwanese study was supported by Alex’s Lemonade Stand Foundation [grant number 17-01882] and P.C.L. is supported in part by the Taiwan Ministry of Science and Technology (MOST 107-2314-B227-009-MY3); Taipei City Hospital (grant No. 10801-62-008); The Data Coordination Center at the IARC is supported by a grant from Children with Cancer UK for which we are grateful. The German case-control study was funded by the German Federal Ministry of the Environment, Nature Conservation and Nuclear Safety. The COG-AE24 study was supported by the Children’s Oncology Group and the National Institutes of Health Grants R01 CA79940, U10 CA13539, U10 CA98543, U10CA180886 and the Children’s Cancer Research Fund, Minneapolis, MN. Regarding the French studies, funding support was as follows: ADELE: INSERM, the French Ministère de l’Environnement, the Association pour la Recherche contre le Cancer, the Fondation de France, the Fondation Jeanne Liot, the Fondation Weisbrem-Berenson, the Ligue Contre le Cancer du Val de Marne and the Ligue Nationale Contre le Cancer; ELECTRE : INSERM, the French Ministère de l’Environnement, , the Association pour la Recherche sur le Cancer (ARC), the Fondation de France, the Fondation pour la Recherche Médicale, Institut Electricité Santé; ESCALE : Fondation de France, ARC, AFSSAPS, Cent pour Sang la Vie, Inserm, AFSSET, ANR (Grant id : ANR10-COHO-0009), INCA, Cancéropôle Ile de France; ESTELLE: INCa, Ligue Nationale contre le Cancer, association Enfants et Santé, ANSES, the Agence Nationale de Sécurité Sanitaire de l’alimentation, de l’Environnement et du Travail (PNREST Anses, Cancer TMOI AVIESAN, 2013/1/248), INCa-DHOS, Cancéropôle Ile de France, ANR (Grant id : ANR-10COHO-0009). The cancer registry data in Washington are supported by the National Cancer Institute #HHSN261201300012I with additional support from the Fred Hutchinson Cancer Research Center, and the Centers for Disease Control and Prevention's National Program of Central Cancer Registries. The funding source of the Costa Rica study was the Universidad Nacional, Costa Rica and the Research Department of the Swedish International Development Cooperation Agency (Sida/SAREC).The Italian SETIL study was financially supported by research grants received by AIRC (Italian Association on Research on Cancer), MIUR (Ministry for Instruction, University and Research, PRIN Program), Ministry of Health (Ricerca Sanitaria Finalizzata Program), Ministry of Labour and Welfare, Associazione Neuroblastoma, Piemonte Region (Ricerca Sanitaria Finalizzata Regione Piemonte Program), Liguria Region, Comitato per la vita “Daniele Chianelli”-Associazione per la Ricerca e la Cura delle Leucemie, Linfomi e Tumori di Adulti e Bambini, (Perugia). REFERENCES 1. Crump C, Sundquist J, Sieh W, Winkleby MA, Sundquist K. Perinatal risk factors for acute myeloid leukemia. Eur J Epidemiol. 2015;30(12):1277-85. 2. Burton GJ, Fowden AL, Thornburg KL. Placental Origins of Chronic Disease. Physiol Rev. 2016;96(4):1509-65. 3. Georgakis MK, Kalogirou EI, Liaskas A, Karalexi MA, Papathoma P, Ladopoulou K, et al. Anthropometrics at birth and risk of a primary central nervous system tumour: A systematic review and meta-analysis. Eur J Cancer. 2017;75:117-31. 4. Milne E, Greenop KR, Metayer C, Schuz J, Petridou E, Pombo-de-Oliveira MS, et al. Fetal growth and childhood acute lymphoblastic leukemia: findings from the childhood leukemia international consortium. Int J Cancer. 2013;133(12):2968-79. 5. Laurvick CL, Milne E, Blair E, de Klerk N, Charles AK, Bower C. Fetal growth and the risk of childhood non-CNS solid tumours in Western Australia. Br J Cancer. 2008;99(1):179-81. 6. Caughey RW, Michels KB. Birth weight and childhood leukemia: a meta-analysis and review of the current evidence. Int J Cancer. 2009;124(11):2658-70. 7. Hjalgrim LL, Rostgaard K, Hjalgrim H, Westergaard T, Thomassen H, Forestier E, et al. Birth weight and risk for childhood leukemia in Denmark, Sweden, Norway, and Iceland. J Natl Cancer Inst. 2004;96(20):1549-56. 8. McLaughlin CC, Baptiste MS, Schymura MJ, Nasca PC, Zdeb MS. Birth weight, maternal weight and childhood leukaemia. Br J Cancer. 2006;94(11):1738-44. 9. Barrington-Trimis JL, Cockburn M, Metayer C, Gauderman WJ, Wiemels J, McKean-Cowdin R. Trends in childhood leukemia incidence over two decades from 1992 to 2013. Int J Cancer. 2017;140(5):1000-8. 10. Giddings BM, Whitehead TP, Metayer C, Miller MD. Childhood leukemia incidence in California: High and rising in the Hispanic population. Cancer. 2016;122(18):2867-75. 11. Sprehe MR, Barahmani N, Cao Y, Wang T, Forman MR, Bondy M, et al. Comparison of birth weight corrected for gestational age and birth weight alone in prediction of development of childhood leukemia and central nervous system tumors. Pediatr Blood Cancer. 2010;54(2):242-9. 12. Blair EM, Liu Y, de Klerk NH, Lawrence DM. Optimal fetal growth for the Caucasian singleton and assessment of appropriateness of fetal growth: an analysis of a total population perinatal database. BMC Pediatr. 2005;5(1):13. 13. Schuz J, Forman MR. Birthweight by gestational age and childhood cancer. Cancer Causes Control. 2007;18(6):655-63. 14. Petridou E, Dessypris N, Spanos E, Mantzoros C, Skalkidou A, Kalmanti M, et al. Insulin-like growth factor-I and binding protein-3 in relation to childhood leukaemia. Int J Cancer. 1999;80(4):494-6. 15. Oksuzyan S, Crespi CM, Cockburn M, Mezei G, Kheifets L. Birth weight and other perinatal characteristics and childhood leukemia in California. Cancer Epidemiol. 2012;36(6):e359-65. 16. Fong CY, Morison J, Dawson MA. Epigenetics in the hematologic malignancies. Haematologica. 2014;99(12):1772-83. 17. Greenblatt SM, Nimer SD. Chromatin modifiers and the promise of epigenetic therapy in acute leukemia. Leukemia. 2014;28(7):1396-406. 18. Panagopoulou P, Skalkidou A, Marcotte E, Erdmann F, Ma X, Heck JE, et al. Parental age and the risk of childhood acute myeloid leukemia: results from the Childhood Leukemia International Consortium. Cancer Epidemiol. 2019;59:158-65. 19. Tomizawa D, Kolb EA. Down syndrome and AML: where do we go from here? Blood. 2017;129(25):3274-5. 20. Villar J, Cheikh Ismail L, Victora CG, Ohuma EO, Bertino E, Altman DG, et al. International standards for newborn weight, length, and head circumference by gestational age and sex: the Newborn Cross-Sectional Study of the INTERGROWTH-21st Project. Lancet. 2014;384(9946):857-68. 21. Creutzig U, Zimmermann M, Reinhardt D, Rasche M, von Neuhoff C, Alpermann T, et al. Changes in cytogenetics and molecular genetics in acute myeloid leukemia from childhood to adult age groups. Cancer. 2016;122(24):3821-30. 22. Knol MJ, VanderWeele TJ. Recommendations for presenting analyses of effect modification and interaction. Int J Epidemiol. 2012;41(2):514-20. 23. DerSimonian R, Laird N. Meta-analysis in clinical trials revisited. Contemp Clin Trials. 2015;45(Pt A):139-45. 24. Walter RB, Othus M, Burnett AK, Lowenberg B, Kantarjian HM, Ossenkoppele GJ, et al. Significance of FAB subclassification of "acute myeloid leukemia, NOS" in the 2008 WHO classification: analysis of 5848 newly diagnosed patients. Blood. 2013;121(13):242431. 25. Bjorge T, Sorensen HT, Grotmol T, Engeland A, Stephansson O, Gissler M, et al. Fetal growth and childhood cancer: a population-based study. Pediatrics. 2013;132(5):e126575. 26. Jimenez-Hernandez E, Fajardo-Gutierrez A, Nunez-Enriquez JC, Martin-Trejo JA, Espinoza-Hernandez LE, Flores-Lujano J, et al. A greater birthweight increases the risk of acute leukemias in Mexican children-experience from the Mexican Interinstitutional Group for the Identification of the Causes of Childhood Leukemia (MIGICCL). Cancer Med. 2018;7(4):1528-36. 27. Hjalgrim LL, Westergaard T, Rostgaard K, Schmiegelow K, Melbye M, Hjalgrim H, et al. Birth weight as a risk factor for childhood leukemia: a meta-analysis of 18 epidemiologic studies. Am J Epidemiol. 2003;158(8):724-35. 28. O'Neill KA, Bunch KJ, Vincent TJ, Spector LG, Moorman AV, Murphy MF. Immunophenotype and cytogenetic characteristics in the relationship between birth weight and childhood leukemia. Pediatr Blood Cancer. 2012;58(1):7-11. 29. Boulet SL, Alexander GR, Salihu HM, Pass M. Macrosomic births in the united states: determinants, outcomes, and proposed grades of risk. Am J Obstet Gynecol. 2003;188(5):1372-8. 30. Snowden JM, Mission JF, Marshall NE, Quigley B, Main E, Gilbert WM, et al. The Impact of maternal obesity and race/ethnicity on perinatal outcomes: Independent and joint effects. Obesity (Silver Spring). 2016;24(7):1590-8. 31. Ahlsson F, Akerud H, Schijven D, Olivier J, Sundstrom-Poromaa I. Gene Expression in Placentas From Nondiabetic Women Giving Birth to Large for Gestational Age Infants. Reprod Sci. 2015;22(10):1281-8. 32. Kasprzak A, Kwasniewski W, Adamek A, Gozdzicka-Jozefiak A. Insulin-like growth factor (IGF) axis in cancerogenesis. Mutat Res Rev Mutat Res. 2017;772:78-104. 33. Vatten LJ, Nilsen ST, Odegard RA, Romundstad PR, Austgulen R. Insulin-like growth factor I and leptin in umbilical cord plasma and infant birth size at term. Pediatrics. 2002;109(6):1131-5. 34. Tisi DK, Liu XJ, Wykes LJ, Skinner CD, Koski KG. Insulin-like growth factor II and binding proteins 1 and 3 from second trimester human amniotic fluid are associated with infant birth weight. J Nutr. 2005;135(7):1667-72. 35. Wu HK, Weksberg R, Minden MD, Squire JA. Loss of imprinting of human insulinlike growth factor II gene, IGF2, in acute myeloid leukemia. Biochem Biophys Res Commun. 1997;231(2):466-72. 36. Wiznitzer A, Reece EA, Homko C, Furman B, Mazor M, Levy J. Insulin-like growth factors, their binding proteins, and fetal macrosomia in offspring of nondiabetic pregnant women. Am J Perinatol. 1998;15(1):23-8. 37. Olausson H, Lof M, Brismar K, Forsum E, Sohlstrom A. Maternal serum concentrations of insulin-like growth factor (IGF)-I and IGF binding protein-1 before and during pregnancy in relation to maternal body weight and composition and infant birth weight. Br J Nutr. 2010;104(6):842-8. 38. Kosik P, Skorvaga M, Durdik M, Jakl L, Nikitina E, Markova E, et al. Low numbers of pre-leukemic fusion genes are frequently present in umbilical cord blood without affecting DNA damage response. Oncotarget. 2017;8(22):35824-34. 39. Kosik P, Skorvaga M, Belyaev I. Incidence of preleukemic fusion genes in healthy subjects. Neoplasma. 2016;63(5):659-72. 40. Ross JA, Potter JD, Reaman GH, Pendergrass TW, Robison LL. Maternal exposure to potential inhibitors of DNA topoisomerase II and infant leukemia (United States): a report from the Children's Cancer Group. Cancer Causes Control. 1996;7(6):581-90. 41. Spector LG, Xie Y, Robison LL, Heerema NA, Hilden JM, Lange B, et al. Maternal diet and infant leukemia: the DNA topoisomerase II inhibitor hypothesis: a report from the children's oncology group. Cancer Epidemiol Biomarkers Prev. 2005;14(3):651-5. 42. Thakali KM, Faske JB, Ishwar A, Alfaro MP, Cleves MA, Badger TM, et al. Maternal obesity and gestational weight gain are modestly associated with umbilical cord DNA methylation. Placenta. 2017;57:194-203. [32] Table 3. Age and sex-specific meta-analysis-derived Odds Ratios (OR) and 95% Confidence Intervals (95% CI)*on the association of birthweight and birthweight for gestational age with childhood (0-14 years) acute myeloid leukemia (AML) Variables Total Males Females Total Males Females OR (95% CI)** OR (95% CI) OR (95% CI) OR (95% CI)** OR (95% CI) OR (95% CI) Age<2 years Age: 2-14 years Birthweight (grams) <2500 0.99 (0.69-1.42) 1.06 (0.61-1.85) 0.92 (0.55-1.52) 0.96 (0.73-1.26) 0.80 (0.51-1.25) 1.15 (0.79-1.67) 2500-3999 reference reference reference reference reference reference ≥4000 1.20 (0.97-1.66) 1.34 (1.01-1.79) 1.16 (0.74-1.82) 1.13 (0.91-1.41) 1.17 (0.96-1.44) 1.03 (0.78-1.35) Birthweight for gestational agea SGA 0.87 (0.63-1.19) 1.04 (0.67-1.62) 0.72 (0.45-1.15) 1.14 (0.92-1.41) 0.94 (0.53-1.68)b 1.19 (0.88-1.61) AGA reference reference reference reference reference reference LGA 1.24 (1.01-1.53) 1.34 (1.05-1.71) 1.32 (0.80-2.17)b 1.15 (1.01-1.31) 1.22 (1.02-1.46) 1.10 (0.90-1.34) Age<1 year Age: 1-14 years Birthweight (grams) <2500 1.26 (0.77-2.06) 1.88 (0.88-4.03) 0.78 (0.32-1.88) 0.93 (0.73-1.18) 0.75 (0.48-1.17) 1.12 (0.81-1.56) 2500-3999 reference reference reference reference reference reference ≥4000 1.36 (0.83-2.24) 1.21 (0.71-2.07) 1.22 (0.57-2.61) 1.16 (1.01-1.34) 1.22 (1.02-1.46) 1.08 (0.77-1.50) Birthweight for gestational agea SGA 0.91 (0.99-1.39) 1.25 (0.65-2.42) 0.77 (0.41-1.45) 1.08 (0.89-1.31) 0.89 (0.51-1.54)b 1.17 (0.88-1.54) AGA reference reference reference reference reference reference LGA 1.21 (0.95-1.55) 1.49 (1.03-2.14) 1.01 (0.66-1.58) 1.20 (1.03-1.41) 1.23 (1.00-1.51) 1.19 (0.96-1.46) In bold: statistically significant associations; *Meta-analysis comprising pooled analysis-derived estimates from the studies providing primary data along with the provided adjusted estimates; **Adjusted for index child’s age, sex, ethnicity, maternal age at birth, plurality, birth order, prematurity and study of origin; a Intergrowth Curve: IC; SGA: Small for gestational age (<10th of IC), AGA: Appropriate for gestational age (10th-90th of IC), LGA: Large for gestational age (>90th of IC); b Statistically significant heterogeneity: LGAfemales; <2y: I2=68.7%, p=0.04; SGAmales; 2-14yrs: I2=59.8%, p=0.08; SGA males; 1-14yrs: I2=65.1%, p=0.06 [33] [34] Table 4. Meta-analysis derived, subtypeand sexspecific effect estimates [Odds Ratios (OR) and 95% Confidence Interval (95% CI)] on the association of birthweight and birthweight for gestational age with childhood (0-14 years) acute myeloid leukemia (AML) risk Total Males Females OR (95% CI)* OR (95%CI) OR (95% CI) M0-M1 cases (N=210) versus controls (N=2244) Birthweight (grams) <2500 0.83 (0.41-1.68) 0.87 (0.34-2.22) 0.74 (0.25-2.23) 2500-3999 reference reference reference ≥4000 1.32 (0.86-2.05) 1.28 (0.75-2.19) 1.40 (0.66-2.99) Birthweight for gestational age** SGA 1.18 (0.66-2.10) 1.93 (0.95-3.92) 0.55 (0.19-1.61) AGA reference reference reference LGA 1.55 (1.09-2.20) 1.85 (1.18-2.91) 1.22 (0.69-2.17) M2 cases (N=265) versus controls (N=2345) Birthweight (grams) <2500 1.03 (0.37-2.81) 0.99 (0.36-2.72)a 0.50 (0.14-1.82)a 2500-3999 reference reference reference ≥4000 1.17 (0.79-1.77) 1.16 (0.69-1.94) 1.17 (0.59-2.34) Birthweight for gestational age SGA 1.11 (0.66-1.88) 0.71 (0.22-2.29) 2.18 (0.38-12.54)b AGA reference reference reference LGA 1.04 (0.74-1.46) 0.96 (0.61-1.52) 1.23 (0.62-2.44) M3 cases (N=155) versus controls (N=2340) Birthweight (grams) [35] <2500 1.47 (0.63-3.42) 2.09 (0.50-8.78)a 2.14 (0.77-5.95) 2500-3999 reference reference reference ≥4000 0.85 (0.42-1.74) 0.47 (0.14-1.56)a 1.90 (0.92-3.93) Birthweight for gestational age SGA 1.14 (0.56-2.29) 0.95 (0.32-2.78) 1.52 (0.58-4.02) AGA reference reference reference LGA 1.10 (0.71-1.69) 0.59 (0.25-1.44) 1.76 (0.99-3.13) M4-M5 cases (N=600) versus controls (N=2915) Birthweight (grams) <2500 0.99 (0.51-1.91) 0.60 (0.29-1.27) 1.14 (0.67-1.96) 2500-3999 reference reference reference ≥4000 1.00 (0.74-1.34) 1.23 (0.75-2.01) 0.80 (0.48-1.33) Birthweight for gestational age SGA 1.04 (0.72-1.52) 1.03 (0.57-1.84) 1.05 (0.64-1.72) AGA reference reference reference LGA 1.18 (0.95-1.50) 1.30 (0.96-1.77) 1.12 (0.82-1.54) M6-M7 cases (N=223) versus controls (N=2542) Birthweight (grams) <2500 0.90 (0.45-1.80) 1.21 (0.34-4.27) 0.70 (0.25-1.94) 2500-3999 reference reference reference ≥4000 1.08 (0.68-1.72) 0.95 (0.50-1.80) 1.22 (0.57-2.62) Birthweight for gestational age SGA 1.06 (0.56-1.99) 1.08 (0.46-2.52) 0.86 (0.31-2.35) AGA reference reference reference LGA 1.12 (0.79-1.61) 0.87 (0.51-1.48) 1.45 (0.88-2.39) [36] In bold statistically significant associations; *Adjusted for index child’s age, sex, ethnicity, maternal age at birth, plurality, birth order, prematurity and study of origin; ** Intergrowth Curve: IC; SGA: Small for gestational age (<10th of IC), AGA: Appropriate for gestational age (10th-90th of IC), LGA: Large for gestational age (>90th of IC); a No meta-analysis performed: estimates derived from the studies providing primary data; b Statistically significant heterogeneity: I2=76.2%, p=0.04 Supplementary Table 1. Studies participating in the Childhood Leukemia International Consortium analysis of fetal growth markers and risk of childhood (0-14 years) acute myeloid leukemia (AML) Study location, Acronym Recruitment Cases source Controls source/recruitment type Cases, N Controls, N Registry-based case-control studies Denmark 1968-2016 National Cancer Registry Central Population Register/ Electronic linkage 249 747 Finland 1989-2011 National Cancer Registry National Central Population Register/ Electronic linkage 127 381 Taiwan 2004-2014 National Cancer Registry Taiwan Maternal and Child Health Database 112 336 US, California State, CCLRP1 1988-2011 Statewide Cancer Registry Linked State birth-hospital discharge records 846 2976 US, Minnesota, State US, New York, State US, Texas, State 1988-2004 1985-2001 1990-1998 National Cancer Registry National Cancer Registry National Cancer Registry State birth certificates State birth certificates State birth certificates 489 1467 US, Washington State 1974-2014 Cancer Registry (regional 1974-1993; statewide 1994-onwards) Linked State birth-hospital discharge records 177 531 Questionnaire-based case-control studies Brazil 1998-2015 Hospitals Hospitals/Interviews 151 453 Costa Rica, CRCLS 2001-2003 Nationwide Cancer Registry Nationwide Birth registry/Interviews 39 117 France, ADELE France, ELECTRE France,ESCALE France, ESTELLE 1994-1999 1995-1998 2003-2005 2010-2012 Hospitals of Lille, Lyon, Nancy & Paris National Blood Malignancy Registry National Blood Malignancy Registry National Blood Malignancy Registry Same hospitals as cases National Random digit dialing National Random digit dialing National Random digit dialing 291 873 Germany, GCCR 1991-1994 Childhood Cancer Registry (nationwide) German Registries of residents (regional with national coverage)/ Interviews 98 294 Greece, NARECHEM-ST2 1996-2015 Nationwide Clinical Cancer Registry Hospitals/Interviews 138 138 Italy, SETIL 1998-2001 Nationwide Cancer Registry National health system rosters/Interviews 69 97 New Zealand, NZCCS 1989-1994 National Cancer Registry, Children’s Cancer Registry; Hospital Admission/Discharge system Nationwide Birth Registry/ Interviews 22 66 UK, UKCCS 1991-1997 Nationwide General Practitioners’ Registry Nationwide General Practitioners’ registry/ Interviews 235 465 US, Texas State 1997-2015 Hospitals Hospital/Interviews 8 24 US, COG3-AE244 1996-2006 COG institutions; US and Canada Random digit dialing/Birth registries 132 240 US,COG-E14 1989-1993 CCG clinical trials Random digit dialing/Interviews 401 469 1 Effect estimates were provided for the meta-analyses by the registry-based case-control studies of Taiwan, US, California State and US, Minnesota/New York/Texas States; 2 Nationwide Registry for Childhood Hematological Malignancies and Solid Tumors; 3Children's Oncology Group; 4: Overlapping cases with the remaining US studies have been excluded Supplementary Table 2. Proportion of missing values (overall % for cases and controls) of the study variables by participating study Study location, Name Birthweight Birthweight for GA* Birth length for GA Weight-forlength for GA POBW** AML FAB Child’s age Child’s sex Child’s ethnicity Birth plurality Preterm birth Birth order Maternal age Registry-based case-control studies Denmark 9.2 9.3 9.9 9.9 3.6 84.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Finland 16.7 16.7 17.1 17.1 85.8 9.5 0.0 0.0 0.0 16.5 16.7 16.9 0.2 Taiwan 0.0 0.0 100.0 100.0 100.0 47.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 US, California State CCLRP 0.0 0.0 100.0 100.0 100.0 41.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 US, Minnesota/ New York/Texas States 8.8 10.4 100.0 100.0 100.0 100.0 0.0 0.0 0.0 0.1 1.9 0.1 0.0 US, Washington State 0.3 11.4 100.0 100.0 77.4 100.0 0.0 0.0 1.4 0.1 11.4 1.6 0.0 Questionnaire-based case-control studies Brazil 4.8 70.4 75.7 75.8 100.0 5.3 0.0 0.0 0.0 0.0 67.4 25.7 1.0 Costa Rica, CRCLS 11.5 100.0 100.0 100.0 100.0 100.0 0.0 0.0 0.0 100.0 0.6 0.6 0.6 France, ADELE/ ELECTRE/ESCAL E/ESTELLE 0.7 4.8 100.0 100.0 100.0 7.9 0.2 0.0 0.8 0.1 4.2 0.3 0.3 Germany, GCCR 1.0 2.3 100.0 100.0 100.0 100.0 0.0 0.0 0.0 0.0 1.5 0.0 0.3 Greece, NARECHEM-ST* 1.1 63.0 71.0 71.0 68.5 1.5 0.0 0.0 0.0 0.0 62.7 0.0 0.0 Italy, SETIL 0.0 38.0 100.0 100.0 100.0 0.0 0.0 0.0 0.0 0.0 38.0 0.0 0.6 New Zealand, NZCCS 1.1 2.3 100.0 100.0 100.0 4.6 0.0 0.0 0.0 0.0 1.1 0.0 0.0 UK, UKCCS 1.7 1.9 100.0 100.0 100.0 9.8 0.0 0.0 0.4 0.4 0.6 0.4 0.9 US, Texas State 3.1 15.3 40.6 40.6 100.0 100.0 0.0 0.0 0.0 100.0 15.6 100.0 31.3 US, COG*-AE24 0.0 0.0 1.6 1.6 0.0 0.0 0.0 0.0 0.3 0.0 0.0 0.0 0.5 US, COG-E14 0.3 0.3 100.0 100.0 100.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 * Abbreviations: GA, Gestational age; NARECHEM-ST, Nationwide Registry for Childhood Hematological Malignancies and Solid Tumors; COG, Children's Oncology Group POBW, **Proportion of optimal birthweight: approximate estimation based on the % of missing values on maternal height