{"paper_id":"342d7564-a6a0-4108-8210-00030d97d6ff","body_text":"Cohort Profile: The Mendelian Randomization in Pregnancy (MR-PREG) collaboration - \nImproving evidence for prevention and treatment of adverse pregnancy and perinatal \noutcomes \n \nMcBride, Nancy* \nClayton, Gemma L* \nGoncalves Soares, Ana* \nYang, Qian \nBond, Tom A \nTaylor, Amy \nChatzigeorgiou, Charikleia \nAiton, Elisabeth \nWest, Jane \nMagnus, Maria C \nLawlor, Deborah A** \nBorges, Maria Carolina** \n \n*Joint first author \n**Joint last author \n \n \n \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\nAbstract \nPurpose \nAdverse pregnancy and perinatal outcomes (APPOs), including pre-term birth, pre-eclampsia, \nand gestational diabetes, can result in maternal and neonatal morbidity and mortality, \nparental anxiety, and increased health care costs. Better understanding of causes of APPOs is \nessential to inform lifestyle and pharmaceutical intervention s for their prevention and \nmanagement. Given the difficult y of undertaking  randomised control trials in pregnant \nwomen, triangulating evidence from across different methods with d ifferent sources of bias \ncould improve our understanding of the causes of APPOs.  The purpose of the Mendelian \nRandomization in Pregnancy (MR-PREG) collaboration is to support triangulation of evidence \nfrom genetic (e.g. , Mendelian randomization [MR]) and non -genetic (e.g., partner negative \ncontrols) methods to explore causal effects of maternal exposures on a comprehensive set of \nAPPOs \nParticipants  \nThe MR-PREG collaboration includes individual participant data from three birth cohorts (two \nfrom the UK and one from Norway) and UK Biobank, and summary data fro m FinnGen and \npublicly available genome wide association studies (GWAS). We have harmonised data across \nstudies so that currently includes exposures on up to 34 APPOs in up to 678,001 women.   \nFindings to date  \nThe main aims of MR-PREG are to improve the evidence base for 1) prevention, by advancing \nour understanding of maternal modifiable causes of APPOs, 2) better understanding of the \neffect of pre-existing conditions on APPOs, and 3) treatment, by advancing our knowledge of \nthe efficacy and safety of existing medications that women may require for pre -existing \nconditions, and identifying and testing the efficacy and safety of novel medications, and those \nthat might be repurposed to effectively and safely treat APPOs. To date, our published \nresearch mainly addresses aims 1 and 3; some examples include triangulation of evidence \nfrom MR, conventional multivariable regression and a paternal negative control, showing that \nhigher maternal body mass index increases the  risk of many APPOs, and identification of \nmaternal circulating metabolites and proteins that may influence birthweight. \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nFuture Plans \nOur future priorities include increasing diversity in the MR -PREG collaboration by expanding \nparticipant representation from non-European ancestries. We are also integrating molecular \ndata, such as circulating protein levels and placental transcriptomics, to better understand \nthe molecular mechanisms underlying APPOs. Additionally, we are using exome and whole -\ngenome sequencing to identify novel causal genes for APPOs and advance our knowledge on \ncandidate targets for APPOs. \nStrengths and limitations  \n• We have curated data for 34 APPOs and harmonized data across multiple studies to \nsupport large-scale investigations of causes of APPOs.  \n• The scope and type of the data supports triangulation of evidence from a range of \ngenetic and non-genetic methods, with different unrelated sources of bias, to identify \ncauses of APPOs. \n• Over the coming year we will enhance the data to enable identification of molecular \nmechanism underlying APPOs. \n• MR-PREG has limited power to detect causal effects on rarer APPOs , such as \ncongenital anomalies and low Apgar scores at 5 minutes , particularly when using \ngenetic methods . Future work will include larger samples and  rare genetic variant \ndata. \n• Participants are mostly of European ancestry ; future efforts will be focussed on \ndiversifying the ancestry representation in the data. \n  \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nIntroduction   \nA substantial proportion of pregnant women experience adverse pregnancy or perinatal \noutcomes (APPOs), such as miscarriage, stillbirth, gestational diabetes mellitus (GDM), \ngestational hypertension ( GH), pre-eclampsia (PE), pre -term birth (PTB),  or having a small \n(SGA) or large for gestational age (LGA ) baby . Globally, one in six recognised pregnancies \nresults in miscarriage  1, one in six women develops GDM  2, one in ten wome n has \nhypertension during pregnancy 3, and one in ten babies is born preterm   4. Some APPOs are \ncauses of severe morbidity and mortality for mothers and their babies  5. For illustration, \nhypertensive disorders of pregnancy (HDP), including GH and PE,  account for approximately \n14% of all maternal deaths  worldwide 6. In addition, APPOs are associated with long-term \nadverse physical and mental health outcomes, and high costs to families, healthcare systems \nand society 7 8. In the UK, the short-term costs of miscarriage alone are estimated to be around \n£471 million per year due to costs to health services, families and loss of productivity 1.  \nThe marked variation in the prevalence of some APPOs across regions and over time, although \ninfluenced by different screening and diagnostic practices, indicates that they are preventable \n2. Better understanding the underlying causes of APPOs is crucial for guiding effective \ninterventions to prevent them. This knowledge would also enable antenatal care services to \nprovide women and couples with accurate advice on the most plausible risk factor s for \nreducing APPOs, potentially reducing the often conflicting advice given to women about \nexposures that could impact their and their babies' health during pregnancy. Furthermore, \nAPPOs are often related, meaning that preventing one may help reduce the burden of others. \nFor example, both GDM and HDP can influence fetal growth, leading to the delivery of LGA \nand SGA bab ies, respectively 9 10. Therefore, preventing GDM and HDP is a way of avoiding \nfetal over/under growth and the related complications.  \nThere is also an urgent need to better understand the effects of pharmaceutical treatments \nduring pregnancy, both for preventing and managing APPOs and for treating pre -existing \nconditions that are increasingly common among women of reproductive age, such as \nautoimmune, mental health, thyroid, reproductive, and cardiometabolic disorders 11-14. An \nincreasing number of women start pregnancy on medication, and the use of medication \namong pregnant women has also been rising over the past few decades 15. Despite that, \npregnant women are rarely included in pre-licensure randomized controlled trials (RCTs) due \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nto the concerns about potential teratogenic effects of drugs and the potential impact of drug-\ndosing on the physiological changes of pregnancy 16 17. As a result, evidence on the benefits \nand risks of drugs for mother and baby is poor, investment in new drug development for \nAPPOs is scarce, and very few medications are explicitly licensed for use during pregnancy.  \nThis leaves mothers and their babies exposed to unknown risks of drugs, compels pregnant \nwomen and their doctors to undertreat medical conditions, and forces them to make difficult \ndecisions about which medications to prescribe/maintain and at which dose. It also means \nthat APPOs, such as GDM and GH, are managed less well than equivalent conditions outside \nof pregnancy (i.e., type 2 diabetes and hypertension) 16 17.  \nRCTs are the gold standard method for testing the impact of lifestyle and pharmaceutical \ninterventions to prevent or treat APPOs. For several years, national bodies have highlighted \nthe need for clinical trials of medications in pregnant women, largely to no avail 16 18 . \nAcknowledging this, the UK 2021 Report of the Commission on Human Medicines Expert \nWorking Group on Optimising Data On Medicines Used During Pregnancy recommended i) \nthe use of routine healthcare data in research, and ii) the use of observational research to \ninvestigate safety of medicines used in pregnancy and during breastfeeding should be \ncommissioned and a system of ongoing surveillance should be established and funded 18.  \nIn the absence of well -powered, well -conducted RCTs, we need to make the be st use of \nobservational data to improve the current evidence base on causes of APPOs and medication \nefficacy and safety during pregnancy. More than 20 years ago, the use of genetic variants to \ninfer causal effects of exposures − a method known as Mendelian randomization (MR) − was \nfirst proposed 19. MR leverages the random allocation of genetic variants at conception to \ninvestigate the effects of modifiable risk factors on health outcomes. This method mitigates \nconfounding by socioeconomic, behavioural, or health -related factors that frequently affect  \ntraditional observational study analyses ( Figure 1 ) 19-21. Since then, its application has \nexpanded significantly, including for assessing the effects of a few maternal risk factors on a \nsmall range of APPOs 22 23 24. \n \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nFigure 1 Summary of the Mendelian randomisation assumptions to estimate the presence of \nan effect of maternal risk factors on adverse pregnancy and perinatal outcomes \n \nAssumptions \n- The maternal genetic variant(s) are statistically robustly associated with the maternal \nrisk factor during pregnancy [relevance assumption, IV1]  \n- There is no confounding of the maternal genetic variant(s) and adverse pregnancy and \nperinatal outcomes (APPOs) (i.e., population level confounders such as population \nstructure, assortative mating, and intergenerational effects)  25 [independence \nassumption, IV2]  \n- The maternal genetic variant is not associated with APPOs other than through its \nassociation with the maternal risk factor [exclusion restriction criteria, IV3]  \n \n \nIn addition, MR is increasingly used for assessing the effects of medications outside of \npregnancy (aka drug target MR) . The validity of drug target MR  is supported by p roof-of-\nconcept studies comparing MR findings to RCT results for established medications , such as \nantihypertensives and statins 26. Drug targets validated by genetic evidence have higher rates \nof success and MR is increasingly used to prioritise (or de -prioritise) new drug targets to be \ntested in RCTs 27-29.  \nAs with all methods, MR is limited by violation of its assumptions. Triangulation of evidence \nacknowledges, and exploits the fact that all methods have sources of bias 30. It involves \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nintegrating multiple lines of evidence using one or more different approaches (e.g., different \nanalytical methods, different data -sources, or different study designs) with different and \nunrelated key sources of bias. I f results are consistent, despite the different sources of bias, \nthis increases the credibility of that being the correct causal effect (whether it suggests a \nprotective, detrimental or no effect), as it would be unlikely for different biases to produce \ndifferent results.  Where there is disagr eement across the  different methods, prior \nspecification of key sources of bias, the direction of these biases, and statistical efficiency for \neach can help inform whether the inconsistency is explained by the different biases or power \nfor each approach , which can inform what further sensitivity analyses and/or other \napproaches are needed to obtain a valid causal estimate. \nThe Mendelian Randomization in Pregnancy (MR -PREG) collaboration was established to \nimprove causal understanding of the effects of maternal lifestyle and health factors on APPOs, \nas well as efficacy and safety of medication use in pregnancy . In this paper, we  provide the \naims of the MR -PREG collaboration, describe the characteristics of the contributing studies, \nour methods for data generation and exploring causal effects, as well as summarising findings \nto date and discussing ongoing and future work. \nAim 1:  Using triangulation of evidence to improve knowledge on the impact of  maternal \nlifestyle factors on APPOs  \nThe MR-PREG collaboration uses triangulation of genetic methods (e.g., MR, co-localisation, \nrare variant , and joint rare and common variant analyses)  and non -genetic methods (e.g., \nconventional multivariable regression, and negative paternal controls) to explore the effect \nof a range of maternal modifiable lifestyle factors, such as  smoking, alcohol intake, coffee \nconsumption, adiposity, sleeping habits, and physical activity, on the risk of multiple APPOs \n(Figure 2) (Supplementary Table 1) 30.  \nAim 2:  Better understanding the effect of maternal predisposition to  conditions, such  as \nautoimmune, cardiovascular, hormonal, musculoskeletal, and mental health conditions, on \nAPPOs  \nAn increasing number of women begin pregnancy with one or more pre -existing conditions \nthat may elevate the risk of multiple APPOs  12. A systematic understanding of the overall \nimpact of these conditions on APPO risk is  important for guiding clinical decision -making in \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nthe management and treatment of such conditions during pregnancy. Within the MR -PREG \ncollaboration, we employ MR to investigate how genetic predisposition to common \nconditions in women of reproductive age  - such as autoimmune, cardiovascular, hormonal, \nmusculoskeletal, and mental health conditions - affects APPO risk (Figure 2).  \nAim 3: Leveraging genomics to explore drug efficacy and safety of medications in pregnancy \nThe MR -PREG collaboration aims to improve the evidence on the benefits and risks  of \nmedications during pregnancy using drug target MR. We investigate potential effects of \nmedications for mothers and their babies by investigating the efficacy and safety in pregnancy \nof medications used to treat pre-existing conditions, discovering new candidate drug targets \nto prevent/treat APPOs, and identifying opportunities to repurpose existing drugs, designed \nto treat conditions unrelated to pregnancy, to prevent/treat APPOs (Figure 2). \nFigure 2 MR-PREG collaboration aims  \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nCohort description  \nAdverse pregnancy and perinatal outcomes (APPOs) \nA key focus of the MR -PREG collaboration is on assessing the causal effect of exposures on \nmultiple outcomes simultaneously, so that we have a more complete picture of the potential \nbeneficial, detrimental, or n ull effect of an exposure  on APPOs . This  ‘outcome-wide \nepidemiology’ approach 31 is informative so that women, their partners and healthcare \nproviders have a more balanced view of the potential adverse and beneficial effects of risk \nfactors than what might be obtained from studies that focus on a single or small number of \noutcomes.  \nWe have included 34 binary/categorical outcomes that can occur in pregnancy, delivery or \nthe first year after birth, and 2 underlying continuous traits in MR-PREG. Table 1 shows the \nAPPOs and underlying traits available in MR-PREG together with the number of women who \nhave phenotypic and genetic data on each of these and the number of cases for the outcomes. \nCohort-specific APPOs  definitions and exclusions are provided in supplementary material \n(Supplementary Table s 2A and 2B), and as well as  sample size fo r binary /categorical \noutcomes (Supplementary Table 3A and 3B for maternal phenotype and genetic data, \nrespectively, and Supplementary Table 3C for offspring genetic data), and mean and standard \ndeviations for continuous traits (Supplementary Table 3D). Supplementary Tables 4A and 4B \ndescribe any deviations from the collaboration definition for each study.  \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\nTable 1. Adverse pregnancy and perinatal outcomes (APPOs) and underlying traits, total \navailable sample and respective number of cases \nType Outcomes/Traits Total Cases N (%) \nPregnancy Hypertensive disorders of pregnancy 541,768 32,549 (6.0%) \n Gestational hypertension 527,932 20,777 (3.9%) \n Pre-eclampsia 678,001 19,566 (2.9%) \n Gestational diabetes 934,566 26,346 (2.8%) \n Depression 147,601 22,785 (15.4%) \n Anaemia 86,167 4,105 (4.8%) \n Pregnancy loss 298,868 94,152 (31.5%) \n Miscarriage 486,217 89,086 (18.3%) \n Sporadic miscarriage 241,159 55,888 (23.2%) \n Recurrent miscarriage 321,756 6,233 (1.9%) \n Stillbirth 207,670 6,331 (3.0%) \n Hyperemesis 478,466 4,235 (0.9%) \n Severity of nausea and vomiting 82,978  \n   Nausea only  7,166 (8.6%) \n Nausea and vomiting  30,133 (36.3%) \n   On medication  31,210 (37.6%) \nDelivery Induction of labour 95,239 14,495 (15.2%) \n Premature rupture of membranes 310,621 21,839 (7.0%) \n Caesarean section 233,969 33,909 (14.5%) \n Emergency C-section 90,438 9,165 (10.1%) \n Elective C-section 87,283 6,004 (6.9%) \n Very pre-term birth 85,274 1,214 (1.4%) \n Pre-term birth 518,849 29,987 (5.8%) \n Post-term birth 425,673 27,841 (6.5%) \n Low birth weight 275,214 18,617 (6.8%) \n High birth weight 266,835 6,679 (2.5%) \n Small for gestational age 89,033 6,882 (7.7%) \n Large for gestational age 96,305 10,468 (10.9%) \n Low Apgar score at 1 minute 86,668 4,950 (5.7%) \n Low Apgar score at 5 minutes 78,739 876 (1.1%) \nPostnatal NICU admission 77,285 6,996 (9.1%) \n Congenital anomalies 74,701 3,569 (4.8%) \n Congenital heart disease 74,701 612 (0.8%) \n Breastfeeding initiation 84,668 66,408 (78.4%) \n Breastfeeding established 81,258 58,035 (71.4%) \n Breastfeeding sustained 65,189 44,401 (68.1%) \n Breastfeeding duration 82,429  \n   1 to 3  6,372 (7.7%) \n   4 to 6  49,123 (59.6%) \n   >6  3,419 (4.1%) \nUnderlying traits Birth weight* 289,846 - \n Gestational age* 226,810 - \n* continuous traits\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n11 \n \nData sources \nCurrently, the MR-PREG collaboration includes data from four core studies, which comprise \nthree prospective birth cohorts [the Avon Longitudinal Study of Parents and Children \n(ALSPAC), Born in Bradford (BiB), and the Norwegian Mother, Father and Child Cohort Study \n(MoBa)], and a biobank [UK Biobank (UKB)]. Data from additional sources are also used, such \nas from publicly available biobanks (FinnGen) and genome-wide association studies ( GWAS) \nmeta-analyses, described below. \nCore studies \nA brief description of each of the core studies is presented below  and the participant \ncharacteristics are summarised in Table 2 . Supplementary Figures 1-4 show the flow of \nparticipants from recruitment into their cohorts to inclusion in our analyses. The APPOs each \nstudy contributed to MR-PREG are summarised in Supplementary Tables 2A and 2B. Details \non genotyping in each study are presented in supplementary material.  \n  \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n12 \n \nTable 2. Maternal participant characteristics in the participating core studies#  \nCharacteristics ALSPAC BiB MoBa UK Biobank \nTotal N 13,895 9,407 93,512 224,423 \nAge, mean (SD) 28.0 (5.0) 27.4 (5.6) 30.1 (4.7) 25.3 (4.6)* \nMissing 369 (2.7%) 1,169 (12.4%) 458 (0.5%) 45,402 (20.2%) \nEthnicity (White), N (%) 11,600 (97.4%) 3,430 (36.4%) 88,989 (95.1%) 211,677 (94.6%) \nMissing 1,982 (14.3%) 1,025 (10.1%) 2,021 (2.1%) 706 (0.3%) \nEducation (university \nlevel), N (%) 1,544 (12.9%) 1,432 (11.8%) 50,047 (53.5%) 65,334 (29.5%)** \nMissing 1,890 (13.6%) 1,186 (12.6%) 11,064 (11.8%) 2,714 (1.2%) \nIndex of multiple \ndeprivation, (N%)        \n1st quintile (most \ndeprived) \n1,628 (13.9%) \n \n5,232 (55.6%) \n NA \n44,857 (20.0%)§ \n \n5th quintile (least \ndeprived) 3,045 (26.0%) 158 (1.7%) NA 44,805 (20.0%) \nMissing 2,197 (15.8%) 1,171 (12.4%)  384 (0.2%) \nParity (nulliparous), N(%) 5,619 (44. 7%) 3,768 (40.0%) 42,135 (45.1%) 179,385 (83.6%) \nMissing 1,333 (9.6%) 345 (3.6%) 399 (0.4%) 9,762 (4.3%) \nSmoking during \npregnancy (ever), N (%) 3,301 (25.9%) 1,408 (14.9%) 5,939 (6.4%) NA \nMissing 1,124 (8.1%) 1,185 (12.6%) 16,824 (18.0%)  \nAlcohol intake during \npregnancy, N (%) 8,036 (64.0%) 1,707 (18.1%) 2,472 (2.6%) NA \nMissing 1,340 (9.6%) 2,114 (22.9%) 27,648 (29.6%)  \nBody mass index, mean \n(SD) 22.9 (3.8) 26.1 (5.7) 24.1 (4.3) 27.1 (5.1)** \nMissing 2,711 (19.5%) 1,562 (16.6%) 11,561 (12.3%) 888 (0.4%) \n# Participants included if they had data on at least one APPO  \n* age at first birth; ** information assessed at recruitment and not at pregnancy; § this corresponds \nto deprivation index and quintiles were constructed internally (not based on national comparisons) \nAbbreviations: ALSPAC, The Avon Longitudinal Study of Parents and Children; BiB, Born in Bradford; \nMoBa, The Norwegian Mother, Father and Child cohort study; SD, standard deviation; NA, information \nnot available in this study. This includes all available data at the outcome level. \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n13 \n \nALSPAC: The Avon Longitudinal Study of Parents and Children \nThe Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective birth cohort \nthat started recruiting pregnant women resident in the former county of Avon (centred \naround the city of Bristol), with delivery dates between April 1991 and December 1992. A \ntotal of 14,541 women (ALSPAC-G0) were enrolled during pregnancy ( 14,676 fetuses ) and \ngave birth to 14,062 live children (ALSPAC-G1) 32 33. Women responded to four questionnaires \nduring pregnancy (average 8, 12, 18, and 32 weeks gestation) and two postpartum (average \n8 weeks and 8 months). Biological samples were collected during pregnancy (blood and urine) \nand birth  (cord blood and placenta) . Maternal anthropometrics was based on self -report \ncollected at 12 weeks gestation and  children had anthropometrics measured at birth . \nObstetric records were also linked to the participants. Genetic data is available for mothers, \npartners and child ren (details on genotyping array and imputation  are presented in \nsupplementary material). Children of the children  (ALSPAC-G2) are also being assessed and \nfollowed-up, but this data has not yet been included in MR-PREG. The study website contains \ndetails of all the data available through a fully searchable data dictionary and variable search \ntool: http://www.bristol.ac.uk/alspac/researchers/our-data/. Ethical approval for the study \nwas obtained from the ALSPAC Ethics and Law Committee and the Local Research  Ethics \nCommittees. Full details of the ALSPAC consent procedures are available on the study website \n(http://www.bristol.ac.uk/alspac/researchers/research-\nethics/http://www.bristol.ac.uk/alspac/researchers/research-ethics/). \nBiB: Born in Bradford \nBorn in Bradford (BiB)  is a prospective birth cohort that recruited women with expected \ndelivery dates between March 2007 and December 2010 . Most women were recruited at \ntheir oral glucose tolerance test (OGTT) at approximately 26–28 weeks’ gestation, which was \noffered to all women booked for delivery at Bradford Royal Infirmary, except those with \nknown diabetes , during the recruitment period . In BiB, most of the obstetric population \nconsists of women of White British or Pakistani origin (together accounting for 81%, with the \nremaining women being of other ancestries). A total of 12,453 women (13,776 pregnancies) \nwere enrolled during pregnancy who gave birth to 13,858 live children.  Women had \nanthropometrics measured  at recruitment  and responded to questionnaires  during \npregnancy and postpartum to provide information about breastfeeding > 6 months. Biological \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n14 \n \nsamples were collected during pregnancy (blood and urine) and birth (cord blood) and women \nconsented to routine primary and secondary care data linkage . Full details of the study \nmethodology were reported previously 34. Genetic data is available for mothers and children \n(details on genotyping array and imputation presented in supplementary material ). Ethical \napproval for the study was granted by the Bradford National Health Service Research Ethics \nCommittee (ref 06/Q1202/48).  The study website provides further cohort details and an \noverview of available data (https://borninbradford.nhs.uk/). \nMoBa: The Norwegian Mother, Father and Child Cohort Study \nThe Norwegian Mother, Father and Child Cohort Study (MoBa) is a prospective birth cohort \nthat recruited pregnant women from all over Norway from 1999 -2008. The cohort includes \napproximately 95,200 mothers, 75,200 fathers , and  114,500 children. Mothers and their \npartners responded to questionnaires during pregnancy (15, 22, and 30 weeks gestation) and \npostpartum (6 months). Blood samples were obtained from both parents during pregnancy \nand from mothers and children (cord  blood) at birth. Maternal anthropometrics were \ncollected from a questionnaire at 15 weeks gestation.  Data was linked to the Medical Birth \nRegistry (MBRN), which is a national health registry containing information about all births in \nNorway. The current study is based on version 12 of the quality -assured data files released \nfor research in 2019. Genetic data is available for mothers, children and partners ( details on \ngenotyping array and imputation presented in supplementary material ). The establishment \nof MoBa and initial data collection was based on a license from the Norwegian Data \nProtection Agency and approval from The Regional Committees for Medical and Health \nResearch Ethics. The MoBa cohort is currently regulated by the Norwegian Health Registry \nAct. Ethical approval for our study was obtained from The Regional Committees for Medical \nand Health Research Ethics (ref 2018/1256).   \nUKB: UK Biobank \nUK Biobank (UKB) is an adult cohort that retrospectively collected relevant data on APPOs. All \npeople in the UK National Health Service (NHS) registry aged between 40-69 years and living \nwithin an approximately 25-mile radius from one of the 22 study centres were invited to \nparticipate in UKB between 2006-2010 35 36. A total of 500,000 adults (5.5% of the ~9.2 million \ninvited) were recruited into the study (54.4% females). Information was assessed at baseline \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n15 \n \nvia a  self-completed questionnaire , physical measures  (including anthropometrics ), and \ncollection of non-fasting blood, urine, and saliva . Participants have been followed up  by \nlinkage to  electronic health records and a  subset of participants responded to  online \nquestionnaires. Hospital Episode Statistics (HES) from 1997 for England, 1998 for Wales and \n1981 for Scotland  are available 37. HES data also contain maternity -related admissions  for \nEngland and Wales. It is important to note that not every participant has a hospital inpatient \nrecord, as not all have been admitted to hospital within the period covered. Genetic data from \nUKB participants is available ( details on genotyping array and imputation presented in \nsupplementary material). Ethical approval for UKB was obtained from the  Northwest Multi-\nCentre Research Ethics Committee (MREC), and MR-PREG collaboration studies are linked to \nUKB application number 23938.  The UKB showcase website contains details of the data \navailable: https://biobank.ndph.ox.ac.uk/showcase/. \nAdditional data sources \nTo increase statistical power for MR analyses, the MR -PREG collaboration also uses genetic \nassociation data for APPOs from publicly available datasets – i.e., FinnGen and several publicly \navailable GWAS meta-analyses – as described below. \nFinnGen  \nFinnGen is the nationwide network of Finnish biobanks, which are linked to national \nelectronic registries that provide information on prescriptions and diseases (ICD9-10 codes). \nFinnGen includes data from 500,348 individuals (282,064 females and 218,284 males) [12th \ndata release (R1 2)]. Besides clinical endpoints, which also include data on some APPOs, \ngenetic data is also available. The APPOs and number of cases and controls for which FinnGen \ncontributed are detailed in Supplementary Table 3B. The Coordinating Ethics Committee of \nthe Helsinki and Uusimaa Hospital District has approved the FinnGen consortium (Nr \nHUS/990/2017). More information about FinnGen can be found on the website \nhttps://www.finngen.fi/en. The metadata from FinnGen used by the MR -PREG collaboration \nis publicly available at https://www.finngen.fi/en/access_results.  \nGWAS meta-analyses \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n16 \n \nAt the time of writing, the MR-PREG collaboration has harmonised and quality-controlled data \nfrom GWAS meta-analyses for GDM [5,485 cases and 347,856 controls from the GENetics of \nDiabetes in Pregnancy Consortium (GenDIP ) consortium 38]; PE [9,515 cases and 157,719 \ncontrols from the  International Pregnancy Genetics (InterPregGen ) consortium  39]; \ngestational duration-related traits [18,797 PTB cases and 260,246 controls, 15,972 post-term \nbirth cases and 115,307 controls, and 195,555 individuals with gestation al age at delivery \nfrom the Early Growth Genetics (EGG) consortium  40]; post-natal depression [17,339 cases \nand 53,426 controls from the Psychiatric Genomics consortium (PGC) 41]. Details are provided \nin Supplementary Table 5, i ncluding the  number of  participants, outcome definition  and \nsource of data (e.g., electronic health records, maternal report, research data collection), and \navailability of maternal and/or fetal genetic effects. \n \nGenetic association data for adverse pregnancy and perinatal outcomes  \nFor studies with access to individual-level data (ALSPAC, BiB, MoBa, and UKB), we conducted \nGWAS analyses to generate genetic association data for APPOs, enabling two -sample MR \nanalyses across the aims of the MR-PREG collaboration. The procedures used for conducting \nGWAS and quality control in each study are described in detail in the Supplementary text. \nWe excluded genetic variants with low imputation accuracy (INFO score < 0.4) and/or low \nminor allele frequency (MAF < 0.01). In addition, we excluded studies that overlap with public \nGWAS metanalyses or that contributed with less than 50 cases for a given APPO and applied \ngenomic control to each study. We then pooled study -specific genetic association data on \nAPPOs using inverse variance weighted fixed -effects meta-analyses implemented in METAL \n(v. 2020 -05-05) 42 and estimated the Cochrane’s Q statistics to explore between -study \nheterogeneity. Separate meta -analyses were conducted for maternal and offspring genetic \neffects. We used GWASInspector to evaluate the quality of GWAS summary data from each \nGWAS and for the meta-analyses 43. \nConditional estimates \nThe MR-PREG collaboration is mostly interested in the causal effects of maternal exposures  \non APPOs. Due to the correlation between maternal and offspring genetics, accounting for \noffspring genetic effects is crucial to reliably interpret MR findings testing maternal exposure \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n17 \n \neffects (Box 1, Figure 3a and Figure 3b). Genetically instrumented maternal exposure effects \n(not biased due to exclusion restriction violation  via fetal effects) and genetically \ninstrumented fetal effects (not confounded by maternal genotype) can be estimated. To \naccount for mutually adjusted maternal and fetal genetic effects we used a weighted linear \nmodel (WLM) implemented using DONUTS (Decomposing nature and nurture using GWAS \nsummary statistics)  software 44 to estimate mutually adjusted maternal and fetal genetic \neffects at each SNP. The WLM calculates conditional genetic effects (i.e. , the mutually \nadjusted coefficients for maternal, offspring and paternal genotype, fitted jointly in the same \nmodel) as linear combinations of the marginal genetic effects (i.e. , the coefficients for \nmaternal, offspring and paternal genotype, estimated separately in potentially overlapping \nsamples). In MoBa, because we also had genotype data on a large sample of fathers, we did \nthe same to account for paternal genetic effects. Details can be found in Supplementary text. \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n18 \n \nBox 1 Approaches to check the Mendelian randomisation assumptions ( Figure 1), specific to \nmaternal risk factors and adverse pregnancy and perinatal outcomes (APPOs) . These are \nfurther sensitivity methods specific to this scenario, beyond standard MR sensitivity methods \nsuch as weighted median and MR-Egger etc. \n \n \n \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n19 \n \nFigure 3a Illustration of how the exclusion restriction assumption may be violated \n \n \nFigure 3b Illustration of how collider bias might be induced in MR analysis \n \nAdjusting for offspring genetic variants only (which are a collider variable) (Figure 3b) means \nthat a false association between maternal and paternal genetic variants can be created \n(shown as a dashed line). Such associations can also arise via assortative mating. This will only \ncause bias  if the  outcome is influenced by paternal genetic variants, independently of \nmaternal and offspring genetic variants . However, by additionally adjusting for paternal \ngenetic variants, the pathway highlighted in black dashed arrows can be closed.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n20 \n \nFindings to date \n \nTo date there have been 8 peer-reviewed publications 23 45-52 and 1 preprint 53 resulting from \nthe MR-PREG collaboration. We briefly describe a selection of these here to give readers an \nindication of the different ways in which MR-PREG can be used to explore effects of exposures \non APPOs. \nMost publications to date align with Aim 1, enhancing our understanding of how maternal \nlifestyle factors influence the risk of APPOs, thereby informing interventions aimed at their \nprevention. For example, by triangulating evidence from MR, conventional multivariable \nregression and paternal negative control, we have shown that higher maternal BMI increases \nthe risk of GH, PE, GDM, pre -labour membrane rupture, induction of labour, Caesarean \nsection, LGA, high birthweight, low Apgar score, and admission to a  neonatal intensive care \nunit, and reduced risk of breastfeeding . We also found evidence of higher maternal BMI \nreducing the odds of SGA and having no effect on perinatal depression 49. Triangulating \nevidence from MR and conventional multivariable regression, we found evidence that \ninsomnia may increase the risk of perinatal depression but does not appear to influence most \nother APPOs that we were able to explore 51. In a study focussed on fetal growth trajectories, \nassessed by repeat ultrasound scan measures in two cohorts, we triangulated evidence from \nMR, conventional multivariable regression and paternal negative control study and found \nevidence of a consistent linear dose -response association of maternal smokin g with fetal \ngrowth from early in the second trimester onwards. No major growth deficit was found in \nwomen who quit smoking early in pregnancy 54.  \nIn relation to Aim 3, we have published preliminary work identifying novel molecular targets \n(i.e., proteins and metabolites) causally related to APPOs. In one study, we used MR to \ninvestigate the effect of 1,139 maternal and fetal circulating proteins on offspring \nbirthweight. We found evidence that higher maternal levels of PCSK1  potentially increase \nbirthweight whilst higher maternal levels of LGALS4 potentially decrease birthweight. \nConversely, higher fetal levels of PCSK1 potentially decrease birthweig ht and of LGALS4 \npotentially increase birthweight. Higher fetal LEPR increased birthweight. Results support \nmaternal and fetal protein effects on birthweight, implicating roles for glucose metabolism, \nenergy balance, and vascular function  53. We have also identified maternal circulating \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n21 \n \nmetabolites beyond glucose that may influence birthweight, such as amino acids like \nglutamine  46. In further studies, we have used conventional multivariable regression and MR \nto explore the potential effect of > 1,000 maternal circulating mass spectrometry metabolites \non the risk of offspring congenital heart disease. We found that pregnancy amino ac id \nmetabolism, androgenic steroid lipids, and levels of succinylcarnitine could be important \ncontributing factors for congenital heart disease (CHD) 50.  \nStrengths and limitations  \n \nThe collaboration has created curated data for 34 APPOs and harmonised data across several \nstudies to support large-scale investigations of causes of APPOs. In addition, the generation \nof genetic association data enables well-powered genetic studies, including  MR studies, to \nimprove causal knowledge of many key targets for lifestyle and pharmaceutical interventions \naimed at preventing APPOs. Data on equivalent exposures in partners and on a wide range of \nplausible confounders, enables triangulation across genetic and non-genetic approaches, and \nalso appropriate sensitivity analyses. In relation to those sensitivity analyses the collaboration \nhas also derived conditional estimates so that maternal genetic variants can be used as an \ninstrument to test the effect of maternal exposures without biases related to offspring \ngenetic effects. \nDespite the size and scale of the data, we are still underpowered  to detect causal effects on \nrare APPOs, such as congenital anomalies and low Apgar scor es. Furthermore, participants \nare of predominantly European ancestry, except for BiB. Enhancing the ancestry diversity in \nthe data contributing to the collaboration is a key priority as outlined below in ‘Collaborations \nand future plans’. Finally, despite our best efforts to derive accurate and standardised APPO \ndefinitions, there is inevitably a considerable degree of misclassification, especially where \nsome studies derived their data only from self-reported information or only from hospital \nelectronic health records. \nCollaborations and future plans \n \nWe are currently focussing on systematically assessing the effects of predisposition to \nautoimmune, mental health, thyroid, reproductive, and cardiometabolic disorders on APPOs \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n22 \n \n(Aim 2 ), discovering new candidate drug targets for APPOs ( Aim 3 ), and investigating the \nefficacy, safety and potential of repurposing existing drugs during pregnancy (Aim 3).  \nA future priority is to increase diversity in the genetic background of MR -PREG collaboration \nparticipants, given the current effort is predominantly focussed on participants of European \nancestry. This will enable us to improve the internal validity of the MR -PREG studies (e.g., \nselecting better genetic instruments through using trans-ancestral data to improve statistical \nfine mapping ) but also external validity (e.g., testing transportability of effects across \nancestries). Additionally, we are incorporating more molecular data into MR-PREG studies to \nbetter understand the molecular mechanisms underlying APPOs , for example , placental \ntranscriptomics and proteomics.  \nFinally, to enhance our work identifying molecular targets for the prevention and treatment \nof APPOs ( Aim 3), we are utilizing exome and whole -genome sequencing data to pinpoint \ncausal genes linked to these adverse outcomes.  \nWe are advancing collaborations with additional large-scale studies including genetic and \nAPPO data, with a focus on rarer outcomes such as congenital anomalies and participants of \nnon-European ancestry to increase statistical power for key safety outcomes and ancestry \ndiversity. Genetic association data for APPOs generated by the collaboration can only be used \nfor research that is covered by data agreements with current contributing studies.  \nData availability  \nThe ALSPAC access policy that describes the proposal process in detail including any costs \nassociated with conducting research at ALSPAC, which may be updated from time to time \nand is available at: \nhttps://www.bristol.ac.uk/medialibrary/sites/alspac/documents/researchers/dataaccess/AL\nSPAC_Access_Policy.pd \nData is available upon request from Born in Bradford: \nhttps://borninbradford.nhs.uk/research/how-to-access-data/  \nData from MoBa are available from the Norwegian Institute of Public Health after \napplication to the MoBa Scientific Management Group (see its \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n23 \n \nwebsite https://www.fhi.no/en/op/data-access-from-health-registries-health-studies-and-\nbiobanks/data-access/applying-for-access-to-data/ for details). \nResearchers can apply for access to the UK Biobank data via the Access Management \nSystem (AMS) (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access). \nAuthor contributions \nThe MR PREG collaboration was conceptualised by DAL and MCB. GC, NM, AGS, MCB, QY, TB, \nCC, EA, AT undertook data curation and generated GWAS pipelines for the individual cohorts. \nQY and MCB ran the meta-analysis. TB and MCB generated the WLM models. GC, NM, MCB, \nAGS and DAL, wrote, and all authors edited and approved the manuscript.  \nAcknowledgements \nThe authors would like to thank Professor Dave Evans, Dr Marwa Al-Arab, Dr Alba Fernandez-\nSanles, Dr Alice Carter , Dr Helena Urquijo, Jevvy Huang, and Peiyuan Huang  for their \nassistance with statistical analysis, GWAS, and technical support. \nFunding declaration \nAll cohort specific funding information is detailed in the Supplementary text.  \nThis publication is the work of the authors , who will serve as guarantors for the contents of \nthis paper. \nDAL, MCB, GC, AGS, TB, QY, AT, HC, EA, PH, and NM work in a unit supported by the University \nof Bristol and UK Medical Research Council (MC_UU_00032/5)  and the British Heart \nFoundation (AA/18/ 1/34219). AGS is supported by the European Union’s Horizon 2020 \nresearch and innovation programme (grant agreement No 874739, LongITools). AGS, GC and \nDAL are supported by the European Union’s Horizon Europe Research and Innovation \nProgramme (grant agreement No 101137146, STAGE, via UKRI grant number 10099041). EA \nis supported by a Wellcome Trust PhD studentship ( 228276/Z/23/Z). MCM is supported by \nthe Research Council of Norway through its Centres of Excellence funding scheme (project No \n262700) and the European Research Council under the European Union’s Horizon 2020 \nresearch and innovation program (ERC Starting Grant, INFERTILITY grant agreement No \n947684).  \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n24 \n \nReferences \n1. Quenby S, Gallos ID, Dhillon -Smith RK, et al. Miscarriage matters: the epidemiological, \nphysical, psychological, and economic costs of early pregnancy loss. The Lancet  \n2021;397(10285):1658-67. doi: 10.1016/S0140-6736(21)00682-6 \n2. Wang H, Li N, Chivese T, et al. IDF Diabetes Atlas: Estimation of Global and Regional \nGestational Diabetes Mellitus Prevalence for 2021 by International Association of \nDiabetes in Pregnancy Study Group's Criteria. Diabetes Res Clin Pract  \n2022;183:109050. doi: 10.1016/j.diabres.2021.109050 [published Online First: \n20211206] \n3. Abalos E, Cuesta C, Grosso AL, et al. Global and regional estimates of preeclampsia and \neclampsia: a systematic review. Eur J Obstet Gynecol Reprod Biol 2013;170(1):1-7. doi: \n10.1016/j.ejogrb.2013.05.005 [published Online First: 20130607] \n4. Ohuma EO, Moller AB, Bradley E, et al. National, regional, and global estimates of preterm \nbirth in 2020, with trends from 2010: a systematic analysis. Lancet \n2023;402(10409):1261-71. doi: 10.1016/s0140-6736(23)00878-4 \n5. Souza JP, Day LT, Rezende -Gomes AC, et al. A global analysis of the determinants of \nmaternal health and transitions in maternal mortality. The Lancet Global Health  \n2024;12(2):e306-e16. doi: 10.1016/S2214-109X(23)00468-0 \n6. Geller SE, Koch AR, Garland CE, et al. A global view of severe maternal morbidity: moving \nbeyond maternal mortality. Reproductive Health 2018;15(1):98. doi: 10.1186/s12978-\n018-0527-2 \n7. Catov JM, Margerison -Zilko C. Pregnancy as a window to future health: short -term costs \nand consequences. Am J Obstet Gynecol  2016;215(4):406-7. doi: \n10.1016/j.ajog.2016.06.060 \n8. Rich-Edwards JW, Fraser A, Lawlor DA, et al. Pregnancy Characteristics and Women's Future \nCardiovascular Health: An Underused Opportunity to Improve Women's Health? \nEpidemiologic Reviews 2014;36(1):57-70. doi: 10.1093/epirev/mxt006 \n9. Brito Nunes C, Borges MC, Freathy RM, et al. Understanding the Genetic Landscape of \nGestational Diabetes: Insights into the Causes and Consequences of Elevated Glucose \nLevels in Pregnancy. Metabolites 2024;14(9) doi: 10.3390/metabo14090508 \n[published Online First: 20240920] \n10. Wu P, Green M, Myers JE. Hypertensive disorders of pregnancy. BMJ 2023;381:e071653. \ndoi: 10.1136/bmj-2022-071653 \n11. Angum F, Khan T, Kaler J, et al. The Prevalence of Autoimmune Disorders in Women: A \nNarrative Review. Cureus 2020;12(5):e8094. doi: 10.7759/cureus.8094 [published \nOnline First: 20200513] \n12. Singh M, Wambua S, Lee SI, et al. Autoimmune diseases and adverse pregnancy outcomes: \nan umbrella review. BMC Medicine 2024;22(1):94. doi: 10.1186/s12916-024-03309-y \n13. Molenaar NM, Bais B, Lambregtse -van den Berg MP, et al. The international prevalence \nof antidepressant use before, during, and after pregnancy: A systematic review and \nmeta-analysis of timing, type of prescriptions and geographical variability. J Affect \nDisord 2020;264:82-89. doi: 10.1016/j.jad.2019.12.014 [published Online First: \n20191209] \n14. Conrad N, Molenberghs G, Verbeke G, et al. Trends in cardiovascular disease incidence \namong 22 million people in the UK over 20 years: population based study. Bmj \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n25 \n \n2024;385:e078523. doi: 10.1136/bmj -2023-078523 [published Online First: \n20240626] \n15. Subramanian A, Azcoaga -Lorenzo A, Anand A, et al. Polypharmacy during pregnancy \nand associated risk factors: a retrospective analysis of 577 medication exposures \namong 1.5 million pregnancies in the UK, 2000 -2019. BMC Medicine 2023;21(1):21. \ndoi: 10.1186/s12916-022-02722-5 \n16. Sciences AoM. Understanding pregnancy: Accelerating the development of new therapies \nfor pregnancy specific conditions. In: Foundation BHPatC, ed., 2023. \n17. National Academies Sciences E, Medicine Consensus Study Report;. Advancing Clinical \nResearch with Pregnant and Lactating Populations: Overcoming Real and Perceived \nLiability Risks, 2024. \n18. UK Commission on Human Medicines. Report of the Commission on Human Medicines \nExpert Working Group on Optimising Data on Medicines used During Pregnancy, 2021. \n19. Smith GD, Ebrahim S. 'Mendelian randomization': can genetic epidemiology contribute to \nunderstanding environmental determinants of disease? Int J Epidemiol 2003;32(1):1-\n22. doi: 10.1093/ije/dyg070 \n20. Lawlor DA, Harbord RM, Sterne JA, et al. Mendelian randomization: using genes as \ninstruments for making causal inferences in epidemiology. Stat Med 2008;27(8):1133-\n63. doi: 10.1002/sim.3034 \n21. Zheng J, Baird D, Borges MC, et al. Recent Developments in Mendelian Randomization \nStudies. Curr Epidemiol Rep  2017;4(4):330-45. doi: 10.1007/s40471 -017-0128-6 \n[published Online First: 20171122] \n22. Yan M, Chen Z, Tang J, et al. Association between gestational diabetes mellitus and \noffspring health: a two -sample mendelian randomization study. BMC Pregnancy and \nChildbirth 2025;25(1):321. doi: 10.1186/s12884-025-07423-4 \n23. Taylor K, Wootton RE, Yang Q, et al. The effect of maternal BMI, smoking and alcohol on \ncongenital heart diseases: a Mendelian randomisation study. BMC Medicine  \n2023;21(1):35. doi: 10.1186/s12916-023-02731-y \n24. Tyrrell J, Richmond RC, Palmer TM, et al. Genetic Evidence for Causal Relationships \nBetween Maternal Obesity-Related Traits and Birth Weight. (1538-3598 (Electronic)) \n25. Morris TT, Davies NM, Hemani G, et al. Population phenomena inflate genetic associations \nof complex social traits. Science Advances  2020;6(16):eaay0328. doi: \n10.1126/sciadv.aay0328 \n26. Schmidt AF, Finan C, Gordillo -Marañón M, et al. Genetic drug target validation using \nMendelian randomisation. Nature Communications  2020;11(1):3255. doi: \n10.1038/s41467-020-16969-0 \n27. Holmes MV, Richardson TG, Ference BA, et al. Integrating genomics with biomarkers and \ntherapeutic targets to invigorate cardiovascular drug development. Nat Rev Cardiol  \n2021;18(6):435-53. doi: 10.1038/s41569 -020-00493-1 [published Online First: \n20210311] \n28. Zheng J, Haberland V, Baird D, et al. Phenome -wide Mendelian randomization mapping \nthe influence of the plasma proteome on complex diseases. Nature Genetics  \n2020;52(10):1122-31. doi: 10.1038/s41588-020-0682-6 \n29. Atia A, Aboeldahab H, Wageeh A, et al. Safety and Efficacy of Proprotein Convertase \nSubtilisin-Kexin Type 9 Inhibitors After Acute Coronary Syndrome: A Meta -analysis of \nRandomized Controlled Trials. American Journal of Cardiovascular Drugs  \n2024;24(1):83-102. doi: 10.1007/s40256-023-00621-5 \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n26 \n \n30. Lawlor DA, Tilling K, Davey Smith G. Triangulation in aetiological epidemiology. Int J \nEpidemiol 2016;45(6):1866-86. doi: 10.1093/ije/dyw314 \n31. VanderWeele TJ. Outcome -wide Epidemiology. Epidemiology 2017;28(3):399-402. doi: \n10.1097/ede.0000000000000641 \n32. Boyd A, Golding J, Macleod J, et al. Cohort Profile: the ‘children of the 90s’ – the index \noffspring of the Avon Longitudinal Study of Parents and Children. Int J Epidemiol  \n2013;42 doi: 10.1093/ije/dys064 \n33. Fraser A, Macdonald-Wallis C, Tilling K, et al. Cohort profile: The Avon Longitudinal Study \nof Parents and Children: ALSPAC mothers cohort. Int J Epidemiol  2013;42 doi: \n10.1093/ije/dys066 \n34. Wright J, Small N, Raynor P, et al. Cohort Profile: the Born in Bradford multi-ethnic family \ncohort study. Int J Epidemiol  2013;42(4):978-91. doi: 10.1093/ije/dys112 [published \nOnline First: 20121012] \n35. Littlejohns TJ, Sudlow C, Allen NE, et al. UK Biobank: opportunities for cardiovascular \nresearch. Eur Heart J 2019;40(14):1158-66. doi: 10.1093/eurheartj/ehx254 \n36. Sudlow C, Gallacher J, Allen N, et al. UK biobank: an open access resource for identifying \nthe causes of a wide range of complex diseases of middle and old age. PLoS Med  \n2015;12(3):e1001779. doi: 10.1371/journal.pmed.1001779 [published Online First: \n20150331] \n37. UK Biobank. UK Biobank hospital inpatient data, 2023. \n38. Pervjakova N, Moen GA -O, Borges MC, et al. Multi -ancestry genome -wide association \nstudy of gestational diabetes mellitus highlights genetic links with type 2 diabetes. \n(1460-2083 (Electronic)) \n39. Steinthorsdottir V, McGinnis R, Williams NO, et al. Genetic predisposition to hypertension \nis associated with preeclampsia in European and Central Asian women. Nat Commun \n2020;11(1):5976. doi: 10.1038/s41467 -020-19733-6 [published Online First: \n20201125] \n40. Solé -Navais P, Flatley C, Steinthorsdottir V, et al. Genetic effects on the timing of \nparturition and links to fetal birth weight. Nat Genet  2023;55(4):559-67. doi: \n10.1038/s41588-023-01343-9 [published Online First: 20230403] \n41. Guintivano J, Byrne EM, Kiewa J, et al. Meta -Analyses of Genome -Wide Association \nStudies for Postpartum Depression. Am J Psychiatry  2023;180(12):884-95. doi: \n10.1176/appi.ajp.20230053 [published Online First: 20231018] \n42. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta -analysis of genomewide \nassociation scans. Bioinformatics 2010;26(17):2190-1. doi: \n10.1093/bioinformatics/btq340 [published Online First: 20100708] \n43. Ani A, van der Most PJ, Snieder H, et al. GWASinspector: comprehensive quality control \nof genome -wide association study results. Bioinformatics 2021;37(1):129-30. doi: \n10.1093/bioinformatics/btaa1084 \n44. Wu Y, Zhong X, Lin Y, et al. Estimating genetic nurture with summary statistics of \nmultigenerational genome -wide association studies. Proceedings of the National \nAcademy of Sciences 2021;118(25):e2023184118. doi: doi:10.1073/pnas.2023184118 \n45. Yang Q, Magnus MC, Kilpi F, et al. Evaluating causal associations of chronotype with \npregnancy and perinatal outcomes and its interactions with insomnia and sleep \nduration: a Mendelian randomization study. medRxiv 2023 doi: \n10.1101/2023.06.02.23290898 [published Online First: 20230605] \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint \n\n27 \n \n46. Zhao J, Stewart ID, Baird D, et al. Causal effects of maternal circulating amino acids on \noffspring birthweight: a Mendelian randomisation study. (2352-3964 (Electronic)) \n47. Barry CA-O, Lawlor DA, Shapland CA-O, et al. Using Mendelian Randomisation to Prioritise \nCandidate Maternal Metabolic Traits Influencing Offspring Birthweight. LID - \n10.3390/metabo12060537 [doi] LID - 537. (2218-1989 (Print)) \n48. Yang Q, Magnus MC, Kilpi F, et al. Investigating causal relations between sleep duration \nand risks of adverse pregnancy and perinatal outcomes: linear and nonlinear \nMendelian randomization analyses. BMC Medicine  2022;20(1):295. doi: \n10.1186/s12916-022-02494-y \n49. Borges MC, Clayton GL, Freathy RM, et al. Integrating multiple lines of evidence to assess \nthe effects of maternal BMI on pregnancy and perinatal outcomes. BMC Medicine  \n2024;22(1):32. doi: 10.1186/s12916-023-03167-0 \n50. Taylor K, McBride N, Zhao J, et al. The Relationship of Maternal Gestational Mass \nSpectrometry-Derived Metabolites with Offspring Congenital Heart Disease: Results \nfrom Multivariable and Mendelian Randomization Analyses. J Cardiovasc Dev Dis  \n2022;9(8) doi: 10.3390/jcdd9080237 [published Online First: 20220727] \n51. Yang Q, Borges MC, Sanderson E, et al. Associations between insomnia and pregnancy and \nperinatal outcomes: Evidence from mendelian randomization and multivariable \nregression analyses. PLOS Medicine  2022;19(9):e1004090. doi: \n10.1371/journal.pmed.1004090 \n52. Yang Q, Sanderson E, Tilling K, et al. Exploring and mitigating potential bias when genetic \ninstrumental variables are associated with multiple non-exposure traits in Mendelian \nrandomization. Eur J Epidemiol 2022;37(7):683-700. doi: 10.1007/s10654-022-00874-\n5 [published Online First: 20220527] \n53. McBride N, Fernández -Sanlés A, Al Arab M, et al. Effects of the maternal and fetal \nproteome on birth weight: a Mendelian randomization analysis. medRxiv 2025 doi: \n10.1101/2023.10.20.23297135 [published Online First: 20250120] \n54. Brand JS, Gaillard R, West J, et al. Associations of maternal quitting, reducing, and \ncontinuing smoking during pregnancy with longitudinal fetal growth: Findings from \nMendelian randomization and parental negative control studies. PLOS Medicine  \n2019;16(11):e1002972. doi: 10.1371/journal.pmed.1002972 \n \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted March 23, 2025. ; https://doi.org/10.1101/2025.03.22.25324447doi: medRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}