ÌÇÐÄvlog

Key Points

QuestionÌý What are the genetic risk factors associated with preeclampsia and hypertensive disorders of pregnancy?

FindingsÌý In this genome-wide association study, 13 novel preeclampsia- or hypertensive pregnancy–associated genetic loci were discovered. Seven loci are located near genes previously associated with blood pressure traits, and several harbor genes involved in the development of placenta, remodeling of uterine spiral arteries, and kidney function.

MeaningÌý The findings further strengthen the known association between cardiovascular health and preeclampsia and provide new targets for future research of preeclampsia pathophysiology, including genes involved in placental development and kidney function.

ImportanceÌý A genetic contribution to preeclampsia susceptibility has been established but is still incompletely understood.

ObjectiveÌý To disentangle the underlying genetic architecture of preeclampsia and preeclampsia or other maternal hypertension during pregnancy with a genome-wide association study (GWAS) of hypertensive disorders of pregnancy.

Design, Setting, and ParticipantsÌý This GWAS included meta-analyses in maternal preeclampsia and a combination phenotype encompassing maternal preeclampsia and preeclampsia or other maternal hypertensive disorders. Two overlapping phenotype groups were selected for examination, namely, preeclampsia and preeclampsia or other maternal hypertension during pregnancy. Data from the Finnish Genetics of Pre-eclampsia Consortium (FINNPEC, 1990-2011), Finnish FinnGen project (1964-2019), Estonian Biobank (1997-2019), and the previously published InterPregGen consortium GWAS were combined. Individuals with preeclampsia or other maternal hypertension during pregnancy and control individuals were selected from the cohorts based on relevant International Classification of Diseases codes. Data were analyzed from July 2020 to February 2023.

ExposuresÌý The association of a genome-wide set of genetic variants and clinical risk factors was analyzed for the 2 phenotypes.

ResultsÌý A total of 16 743 women with prior preeclampsia and 15 200 with preeclampsia or other maternal hypertension during pregnancy were obtained from FINNPEC, FinnGen, Estonian Biobank, and the InterPregGen consortium study (respective mean [SD] ages at diagnosis: 30.3 [5.5], 28.7 [5.6], 29.7 [7.0], and 28 [not available] years). The analysis found 19 genome-wide significant associations, 13 of which were novel. Seven of the novel loci harbor genes previously associated with blood pressure traits (NPPA, NPR3, PLCE1, TNS2, FURIN, RGL3, and PREX1). In line with this, the 2 study phenotypes showed genetic correlation with blood pressure traits. In addition, novel risk loci were identified in the proximity of genes involved in the development of placenta (PGR, TRPC6, ACTN4, and PZP), remodeling of uterine spiral arteries (NPPA, NPPB, NPR3, and ACTN4), kidney function (PLCE1, TNS2, ACTN4, and TRPC6), and maintenance of proteostasis in pregnancy serum (PZP).

Conclusions and RelevanceÌý The findings indicate that genes related to blood pressure traits are associated with preeclampsia, but many of these genes have additional pleiotropic effects on cardiometabolic, endothelial, and placental function. Furthermore, several of the associated loci have no known connection with cardiovascular disease but instead harbor genes contributing to maintenance of successful pregnancy, with dysfunctions leading to preeclampsialike symptoms.

Preeclampsia is a vascular pregnancy disorder that affects 3% to 5% of all pregnancies.^1,2 The disorder develops only in the presence of a placenta, and especially early-onset preeclampsia is often accompanied by defects in placental development and function.^3,4 Subcellular material and molecules released by the placenta, such as antiangiogenic factors, are thought to evoke the systemic endothelial dysfunction manifested as maternal preeclampsia symptoms, including hypertension and proteinuria.^5-7 However, preeclampsia is both phenotypically and etiologically heterogeneous. The disorder often develops without any evident placental malfunction, and the predisposition to preeclampsia is likely affected by multiple underlying cardiometabolic factors that modify the response to the pregnancy-induced stress.⁸

A genetic contribution to preeclampsia susceptibility has been established,^9,10 but the actual risk loci remain mostly unknown. The genome-wide association study (GWAS) of preeclampsia by Steinthorsdottir et al¹¹ identified 5 risk loci, which have previously been connected to hypertension. Accordingly, epidemiological evidence shows that prior cardiovascular disease inflated the risk of preeclampsia, and individuals with previous preeclampsia were at increased risk of developing cardiovascular disease later in life.^12-15

Preeclampsia is likely to consist of several subtypes with differing etiologies.⁸ The overlap in the genetic risk factors between preeclampsia and related disorders could be subtype specific, with some subtypes sharing more features with hypertensive diseases and others being more closely linked to disorders of placental development. Therefore, we have selected 2 phenotype groups for examination: preeclampsia and preeclampsia or other maternal hypertension during pregnancy.

The aim of our study was to identify genetic risk factors associated with preeclampsia and hypertensive disorders of pregnancy in a maternal genome-wide meta-analysis comprising samples from the closely related populations of Finland and Estonia. For the meta-analysis of the strict preeclampsia phenotype, we supplemented these data with a previously published GWAS on preeclampsia.¹¹ In addition, we performed GWAS in smaller paternal and fetal sample sets from Finland to identify risk loci of preeclampsia conveyed via the fetus.

For the maternal meta-analyses, genome-wide genotyped and imputed samples from the Finnish Genetics of Pre-eclampsia Consortium (FINNPEC), FinnGen, and Estonian Biobank were used. In addition, we included summary statistics from the earlier meta-analysis study of preeclampsia by Steinthorsdottir et al.¹¹ For the fetal and paternal genome-wide association analysis, samples from the FINNPEC cohort were available. All required ethical approvals were obtained from international, national, and regional ethics committees as described in detail in the eAppendix in Supplement 1.

The analyses were performed with 2 phenotypes: (1) preeclampsia, eclampsia, or preeclampsia superimposed on chronic hypertension and (2) preeclampsia or other maternal hypertensive disorders. We also examined a phenotype of preeclampsia or fetal growth restriction (as mother’s diagnosis), used as proxy for small for gestational age (FinnGen and Estonian Biobank) or diagnoses of placental insufficiency or an infant being small for gestational age (FINNPEC; results were comparatively minor and are presented in the eAppendix in Supplement 1). In the FINNPEC cohort, preeclampsia was defined according to the American College of Obstetricians and Gynecologists 2002 criteria as hypertension (systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg) and proteinuria (≥0.3 g/24 hours, 0.3 g/L, or two ≥1+ dipstick readings) occurring after 20 weeks of gestation. Gestational hypertension was defined as hypertension occurring after 20 weeks of gestation in the absence of proteinuria. Birth weights below −2.0 SD units according to the Finnish standards¹⁶ were classified as small for gestational age. Placental insufficiency was defined as umbilical artery resistance index or pulsatility index of +2 or more SD. In FinnGen and Estonian Biobank, the phenotypes were based on International Classification of Diseases codes (Revisions 8, 9, and 10, where available) as detailed in eTable 1 in Supplement 1. All parous women not fulfilling the case inclusion criteria were included in the control group. In addition to FINNPEC, FinnGen, and Estonian Biobank, we included the summary statistics from the previous largest genome-wide association meta-analysis for preeclampsia conducted by Steinthorsdottir et al¹¹ to the meta-analysis of the preeclampsia phenotype. Further details of the phenotype definitions are presented in the eAppendix in Supplement 1. The sample sizes available in each phenotype are presented in Table 1. The study design is visualized in Figure 1.

Basic characteristics of all cohorts are described in eTable 2 in Supplement 1. Detailed description is provided in eTable 3 in Supplement 1 for FINNPEC, as this case-control cohort contains more extensive clinical information compared to the other cohorts. We also report survival analysis between preeclampsia and all other disease end points available in FinnGen. The specifics of this workflow are described in detail in the eAppendix in Supplement 1. The analyses of placental insufficiency or infants who were small for gestational age phenotype yielded limited findings, and the results are discussed in the eAppendix, eTables 11 and 12, and eFigures 6-9 in Supplement 1.

Genotyping was performed using Illumina and Affymetrix arrays (Illumina and Thermo Fisher Scientific) in FINNPEC, FinnGen, and Estonian Biobank. Only high-imputation quality markers (imputation information score >0.7) were used in the analysis. Exact genotyping workflows are detailed in eAppendix in Supplement 1.

In brief, the association analyses were performed for all genotyped variants using generalized mixed model as implemented in SAIGE versions 0.39, 0.39.1, and 0.38 (Lee Lab for Statistical Genetics and Data Science¹⁷) for FINNPEC, FinnGen, and Estonian Biobank, respectively. Summary statistics obtained from these analyses in each cohort were then used to perform inverse variance–weighted meta-analysis with METAL software.¹⁸ Detailed associations and the meta-analysis workflow are presented in the eAppendix in Supplement 1.

To identify plausible candidate genes in each locus, we prioritized genes according to multiple layers of evidence: (1) statistical significance of variants in the GWAS meta-analysis, (2) functional evidence associated with the discovered variants, and (3) careful examination of literature regarding the identified variants and suspected causal genes. The detailed annotation workflow is described in the eAppendix in Supplement 1.

To further examine the correlation of our findings with other disorders, we first conducted a phenome-wide association study (PheWAS) analysis for the lead variants using all 2861 phenotypes provided in FinnGen Data Freeze 6. Second, we used Linkage Disequilibrium Score Regression software^19,20 to evaluate genetic correlation of the studied phenotypes with other traits. Third, we calculated polygenic risk scores (PRS) for preeclampsia and preeclampsia or other maternal hypertensive disorder phenotypes and analyzed the results in the FINNPEC cohort to examine overall genetic contribution uncovered in our GWAS analysis. The methodology of these 3 analyses is explained in detail in the eAppendix in Supplement 1. Data were analyzed from July 2020 to February 2023. The genome-wide association data generated in this study have been deposited in the NHGRI-EBI GWAS Catalogue database.

A total of 16 743 women with prior preeclampsia and 15 200 with preeclampsia or other maternal hypertension during pregnancy were obtained from FINNPEC, FinnGen, Estonian Biobank, and the InterPregGen consortium study. The mean (SD) age in each cohort was 30.3 [5.5], 28.7 [5.6], 29.7 [7.0], and 28 [not available] years, respectively.

Altogether, we identified 9 and 13 genome-wide significant loci for the preeclampsia and the preeclampsia or other maternal hypertensive disorder phenotypes, respectively (Table 2, Figures 2, and eFigures 1-5 in Supplement 1). Four preeclampsia loci in the meta-analysis and 9 preeclampsia or other maternal hypertension loci in the meta-analysis were not significantly associated with preeclampsia in the earlier maternal GWAS.

Preliminary meta-analysis results showed evidence of slight inflation in the meta-analysis test statistics (linkage disequilibrium score intercepts of 1.0291 [SE, 0.0076] and 1.0325 [SE, 0.0079] for preeclampsia and preeclampsia or other maternal hypertension during pregnancy, respectively) and were corrected by rerunning the meta-analysis with genomic correction. Most of the directions of the associations were concordant between the cohorts with the exception of 4 loci on 3q26, 12q24, 19p13, and 20q13 that showed discordant association of direction in the smallest cohort (FINNPEC) compared to the others (Table 2). We did not observe genetic heterogeneity except for the lead variant rs167479 at locus 19p13. Credible sets produced with SuSie version 0.11.92 are listed in eTable 4 in Supplement 2.

In the survival analysis conducted in FinnGen between preeclampsia and all other end points, preeclampsia was associated with an increase in risk of 146 disease end points (eTable 5 in Supplement 1). These included hypertensive diseases, such as gestational hypertension and cardiovascular disease, but also glomerular diseases and glomerulonephritis. Furthermore, increased risk was shown for placental abruption and numerous complications related to induced labor, such as failed induction of labor and infections of genitourinary tract in pregnancy.

In the genetic correlation analysis between our study phenotypes and 894 previously published phenotypes, the preeclampsia phenotypes were correlated most strongly with the phenotypes related to blood pressure or various cardiovascular disease (eTable 6 in Supplement 2). For both phenotypes, the correlations with blood pressure medication and high blood pressure were above 0.59 and 0.56, respectively, and correlations with coronary artery disease were above 0.4. In addition, several measures of body fat were associated with our phenotypes. These findings are also mirrored by the results of the PheWAS analysis conducted for all the lead variants in FinnGen Data Freeze 6 (eTable 7 in Supplement 2). In this test, 11 of 22 lead variants were significantly associated with essential hypertension or cardiovascular disease and 7 with other phenotypes, whereas for 4 variants, no significant associations were found. PheWAS also detected associations to immunology- and autoimmunity-related end points for many lead variants. For instance, various arthritis and rheumatic end points were associated with 4 loci (6p21, 12q13, 12q24, and 16q12).

In the preeclampsia meta-analysis, we replicated all 5 loci reported by Steinthorsdottir et al¹¹ and identified 4 novel genome-wide significant loci of maternal preeclampsia (Figure 2A). Plausible candidate genes were identified for all the novel loci, except for the association at 9q22, where the candidate gene remained elusive for the preeclampsia phenotype. Novel association peaks were detected at locus 13q12 close to FLT1 and at 11q22, where the intergenic variant rs3018700 lies near genes PGR and TRPC6. We also found a genome-wide significant association within the human leukocyte antigen (HLA) region, which is known for its high density of genes involved in immune regulation and recognition. The association lies within an 89 kbp copy number variations region previously associated with preeclampsia²¹ and is located within the major susceptibility locus for psoriasis, PSORS1. In addition, MAGMA gene-based analysis yielded significant results for ‘MTHFR and CLCN6 genes (Figure 3A).

The preeclampsia or other maternal hypertension during pregnancy meta-analysis revealed 9 novel loci not identified by Steinthorsdottir et al¹¹ or in the preeclampsia phenotype studied here (Figure 2B). Four of the 5 associations identified in the GWAS of preeclampsia by Steinthorsdottir et al¹¹ were replicated, with only the locus in 12q24 remaining slightly below genome-wide significance. Association in chromosome 1 locus 1p36 was identified by both the meta-analysis and the MAGMA analysis (Figure 3B), which implicated the MTHFR, CLCN6, and NPPA genes. Other discovered loci harbor genes NPR3, PLCE1, PZP, TNS2, FURIN, FES, RGL3, ACTN4, and PREX1. The variants near PZP on 12p13 and ACTN4 on 19q13 did not appear to contain any genome-wide significant variants directly contributing to hypertensive diseases or genes associated with hypertensive disorders in close proximity, unlike the other discovered loci for preeclampsia or other maternal hypertension during pregnancy. Two additional loci residing in 11q13.1 and 16q23.1 were detected in MAGMA analysis (Figure 3B).

Our analyses in the FINNPEC cohort comparing individuals in the top 10% preeclampsia PRS compared to those in the bottom 90% showed an odds ratio (OR) of 2.21 (95% CI, 1.58-3.10; P < .001) for preeclampsia and 1.89 (95% CI, 1.42-2.52; P < .001) for preeclampsia with severe symptoms. When using the top and bottom percentiles of PRS for preeclampsia or other maternal hypertension during pregnancy, similar associations were shown for the preeclampsia or other maternal hypertension phenotype (OR, 3.66; 95% CI, 2.36-5.69; P < .001). The Nagelkerke R² values calculated for models, including risk factors of age, parity, body mass index, and first trimester blood pressure, showed improvement when PRS information was added, as described in the eAppendix and eTables 8-10 in Supplement 1.

The GWASes of paternal and fetal preeclampsia and preeclampsia or other maternal hypertension during pregnancy did not yield any genome-wide significant associations. Further details can be found in eFigure 5 in Supplement 1.

In this GWAS, we identified multiple novel risk loci for the preeclampsia only and preeclampsia or other maternal hypertensive disorder phenotypes. The discovered loci harbor genes affecting endothelial dysfunction, placental development, and immunology. Six of the 9 loci in the preeclampsia phenotype and 11 of 13 in the preeclampsia or other maternal hypertension loci reached genome-wide significant association in previous GWASes of cardiovascular diseases. This observation, along with our genetic correlation and PheWAS results, imply that the established blood pressure loci are associated with predisposition to hypertension during pregnancy, plausibly via the same mechanisms. The findings of our study, and the results of the survival analysis in particular, support the concept of pregnancy as a window to future cardiovascular health: the increased maternal genetic susceptibility to cardiovascular disease might become evident for the first time during pregnancy.

Systemic endothelial dysfunction characterized by impaired vasodilation, endothelial injury, and reduction in vascular integrity is central to the pathophysiology of preeclampsia.^22-24 The best known biomarker of preeclampsia, sFlt-1, is a soluble antiangiogenic protein that reduces the availability of the proangiogenic proteins VEGF and PlGF to endothelial cells, thus impairing the maintenance of vascular integrity and cellular viability.^25,26 The earlier largest GWAS¹¹ found an association between maternal preeclampsia and the locus in the fetal genome.²⁷ Our study provides evidence of the relevance of the FLT1 gene in the genomes of women with preeclampsia. In addition, NPPA on 1p36, FES and FURIN on 15q26, ACTN4 on 19q13, and PREX1 on 20q13.13 encode for proteins that are involved in regulating endothelial permeability and leukocyte transmigration.^28-32 These findings provide further support to the idea that preeclampsia liability might be modified by alterations in the integrity of the endothelium.

Abnormal leakage of protein to urine is another key feature of preeclampsia. Intriguingly, several of the genes proximal to the associating lead single-nucleotide variants in our study have been associated with kidney disease. Mutations in PLCE1, TNS2, ACTN4, and TRPC6 are associated with nephrotic syndrome characterized by proteinuria.^33-38 Products of these genes have important roles in podocyte function and integrity of the glomerular filtration barrier.^39-42 The mechanisms of action of these genes in causing kidney damage are likely to be variable. It is plausible that such genetic predisposition and putative kidney injuries sustained during preeclampsia may contribute to later glomerular diseases, as suggested by our survival analyses and other literature.⁴³

Examples of putative pleiotropic candidate genes of the associating loci in our study include the natriuretic peptide genes NPPA and NPPB on 1p36 and their clearance receptor NPR3 on 5p13. Natriuretic peptide hormones regulate blood pressure and kidney function, among their several other effects.⁴⁴ In addition, mice that lack the expression of atrial natriuretic peptide develop gestational hypertension and proteinuria and, similar to preeclampsia, exhibit impairment in trophoblast invasion and uterine spiral artery remodeling.^45,46 Changes in the function or expression level of these genes with several effects in the key mechanisms of preeclampsia could contribute to the multiorgan dysfunction that is characteristic of this pregnancy-related disorder.

Many of the genes next to the lead single-nucleotide variants of our study are involved in placental development and function, which are often compromised in preeclampsia. Both genes closest to the lead variant in the locus 11q22, PGR and TRPC6, are known to affect placental functions and maintenance of pregnancy.^47,48 PGR has been suggested to contribute to balanced hormonal signaling during pregnancy and subsequently aid the immune and endothelial cells in the cytotrophoblast invasion.⁴⁸ Also, the 19q13-located ACTN4 is known for regulating the trophoblast proliferation and differentiation during early pregnancy.⁴⁹ Defects in these processes are well documented, especially in early-onset preeclampsia.^3,4 Another pregnancy-related gene, PZP (9q13), is a protease inhibitor that prevents the activity of all 4 classes of proteases and stabilizes misfolded proteins, which have been shown to accumulate in preeclampsia.^50-54 Additional relevant literature regarding these putative candidate genes is reviewed in more detail in the eAppendix in Supplement 1.

Immunological factors have been shown to contribute to the pathophysiology of preeclampsia.⁵⁵ Examples of associating genes putatively modulating the immunological response to pregnancy include PZP (in the preeclampsia or other maternal hypertension phenotype), likely modulating T helper cell response and PSORS1C2 (within the PSORS1 locus in the HLA region). Tissue compatibility, autoimmunity, regulation of inflammation, and cardiovascular diseases are processes with known susceptibility loci within the HLA and pathophysiological relevance in preeclampsia. Our discovered association largely reflects these functions, as discussed in more detail in the eAppendix in Supplement 1. In line with these observations, PheWAS analysis found several associations with immunology-related disorders (although not psoriasis), such as type 1 diabetes, spondylopathies, rheumatoid arthritis, and, to a lesser degree, hypertension. Immunological etiology is typically associated with severe or early-onset preeclampsia.⁵⁶ Due to the high gene density, extreme polymorphism, and complex haplotype structure, identifying likely candidate genes, especially in the HLA region, is challenging. Until now, immunology-related findings have been absent in preeclampsia GWAS results.

When comparing the preeclampsia and preeclampsia or other maternal hypertension phenotypes explored in this study, we notice that the effect sizes of the uncovered loci were largely similar between them. The most likely candidate genes appear to be related to hypertensive diseases, implying that the genetic risk factors of preeclampsia are shared with other hypertensive disorders as well as risk of hypertension later in life. Furthermore, incorporation of PRS to already known risk factors of preeclampsia may yield improvement in prediction of this disorder.

The main strength of this study is the use of the 3 well-characterized cohorts FINNPEC, FinnGen, and Estonian Biobank, originating from 2 closely related populations of Finland and Estonia. Such homogenous populations might facilitate the discovery of rare variants with larger effects and characterization of the genetic basis of complex diseases such as preeclampsia.^57,58 The associations previously reported by Steinthorsdottir et al¹¹ are similar in effect directions to those now discovered in the preeclampsia phenotype in the Finnish and Estonian cohorts, suggesting that similar genetic background may contribute to preeclampsia in other populations as well.

This study has limitations. Generalizability of the results may be limited in more diverse populations. Lack of external validation is also a limitation, although the consistency in effect sizes between the large study cohorts does add to the reliability of the findings. The register-based approach used in this study provides limited phenotype information and can be seen as a limiting factor, although the quality of the Finnish Care Register for Health Care has previously been shown to be excellent.⁵⁹ In support of the robustness of our approach, we replicated the findings of the largest previously published GWAS meta-analysis, both with our preeclampsia and preeclampsia or other maternal hypertension phenotypes and all the study cohorts involved provided uniform support for our findings with little evidence of heterogeneity. Uncovering the paternal and fetal associations may require considerably larger sample sizes than available in the current study. Our PRS analysis has 2 important limitations. First, as the FINNPEC cohort was collected from university hospitals, the included individuals may represent those with a more severe form of the disease. Second, due to its being a case-control cohort, FINNPEC has larger proportion of individuals with preeclampsia compared to other cohorts. Together these factors may lead to inflated estimates of the predictive ability of PRS reported in this work.

Our study uncovered 13 novel loci with genome-wide significant association with preeclampsia or other hypertensive disorders of pregnancy. We found that cardiovascular disease–related genes were associated with preeclampsia, as previously suggested, but many of those genes have pleiotropic effects on cardiometabolic, endothelial, and placental function. In addition, we provide further evidence for an association of several loci not previously associated with cardiovascular disease but containing genes with apparent importance in the maintenance of pregnancy, with dysfunctions leading to preeclampsialike symptoms. Although further functional studies are required in the future, these results offer valuable insights into the genetic architecture and biology behind preeclampsia as well as into the connection between preeclampsia and other maternal hypertensive disorders.

Accepted for Publication: April 14, 2023.

Published Online: June 7, 2023. doi:10.1001/jamacardio.2023.1312

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2023 Tyrmi JS et al. JAMA Cardiology.

Corresponding Author: Jaakko S. Tyrmi, PhD, Center for Child, Adolescent and Maternal Health Research, Faculty of Medicine and Health Technology, Tampere University, Kalevantie 4, 33100 Tampere, Finland (jaakko.tyrmi@tuni.fi).

Author Contributions: Dr Tyrmi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Tyrmi, Kaartokallio, Kettunen, Kivinen, Kere, Laivuori.

Acquisition, analysis, or interpretation of data: Tyrmi, Kaartokallio, Lokki, Jääskeläinen, Kortelainen, Ruotsalainen, Karjalainen, Ripatti, Kivioja, Laisk, Pouta, Kivinen, Kajantie, Heinonen, Kere, Laivuori.

Drafting of the manuscript: Tyrmi, Kaartokallio, Lokki, Kortelainen, Ruotsalainen, Kivinen.

Critical revision of the manuscript for important intellectual content: Kaartokallio, Lokki, Jääskeläinen, Karjalainen, Ripatti, Kivioja, Laisk, Kettunen, Pouta, Kivinen, Kajantie, Heinonen, Kere, Laivuori.

Statistical analysis: Tyrmi, Kaartokallio, Ruotsalainen, Karjalainen, Ripatti, Kivioja, Laisk, Kettunen.

Obtained funding: Jääskeläinen, Kettunen, Kivinen, Kajantie, Heinonen, Laivuori.

Administrative, technical, or material support: Lokki, Jääskeläinen, Kortelainen, Kettunen, Heinonen.

Supervision: Ripatti, Heinonen, Kere, Laivuori.

Conflict of Interest Disclosures: Dr Kajantie reported grants from Academy of Finland, European Commission (RECAP Preterm 733280), Foundation for Pediatric Research, and Signe and Ane Gyllenberg Foundation during the conduct of the study as well as grants from Yrjö Jahnsson Foundation, Foundation for Cardiovascular Research, Finska Läkaresällskapet, and Novo Nordisk Foundation outside the submitted work. Dr Laivuori reported grants from Academy of Finland, ERA PerMed Joint Transnational Call 2020, Academy of Finland (344695), Jane and Aatos Erkko Foundation, Päivikki and Sakari Sohlberg Foundation, Finska Läkaresällskapet, and the Finnish Medical Foundation; competitive state research financing of the expert responsibility area of the Helsinki University Hospital; and research funds from the University of Helsinki during the conduct of the study. Drs Jääskeläinen and Laivuori have received honoraria from Orion Corporation. No other disclosures were reported.

Funding/Support: Dr Kaartokallio is supported by the Jane and Aatos Erkko Foundation during the conduct of the study. Dr Tyrmi is funded by EraPerMed JTC 2020, Academy of Finland (344695). The FINNPEC study was supported by the Jane and Aatos Erkko Foundation, Juho Vainio Foundation, Päivikki and Sakari Sohlberg Foundation, Academy of Finland, research funds of the University of Helsinki, a government special state subsidy for the health sciences for the Hospital District of Helsinki and Uusimaa, Finska Läkaresellskapet, Liv och Hälsa Foundation, NovoNordisk Foundation, Finnish Foundation for Pediatric Research, Emil Aaltonen Foundation, Sigrid Juselius Foundation, Signe and Ane Gyllenberg Foundation, Finnish Foundation for Laboratory Medicine, the Finnish Medical Foundation, and the Finnish Foundation for Cardiovascular Research. The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by the National Cancer Institute; National Human Genome Research Institute; National Heart, Lung, and Blood Institute; National Institute on Drug Abuse; National Institute of Mental Health; and National Institute of Neurological Disorders and Stroke.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: A complete list of the members of the FINNPEC Study Group, FinnGen, and the Estonian Biobank Research Team appears in Supplement 3.

Data Sharing Statement: See Supplement 4.

Additional Contributions: We thank the members and assisting personnel of the FINNPEC study; the InterPregGen consortium members for their assistance in accessing summary data from the study by Steinthorsdottir et al; the Estonian Biobank participants and staff; and the participants and investigators of FinnGen study. The FinnGen project is funded by 2 grants from Business Finland (HUS 4685/31/2016 and UH 4386/31/2016) and the following industry partners: AbbVie, AstraZeneca UK, Biogen MA, Bristol Myers Squibb (and Celgene Corporation & Celgene International II Sàrl), Genentech, Merck Sharp & Dohme, Pfizer, GlaxoSmithKline Intellectual Property Development, Sanofi, Maze Therapeutics, Janssen Biotech, Novartis, and Boehringer Ingelheim. The following biobanks are acknowledged for delivering biobank samples to FinnGen: Auria Biobank (), THL Biobank (), Helsinki Biobank (), Biobank Borealis of Northern Finland (), Finnish Clinical Biobank Tampere (), Biobank of Eastern Finland (), Central Finland Biobank (), Finnish Red Cross Blood Service Biobank () and Terveystalo Biobank (). All Finnish Biobanks are members of BBMRI.fi infrastructure (). Finnish Biobank Cooperative (FINBB; ) is the coordinator of BBMRI-ERIC operations in Finland. The Finnish biobank data can be accessed through the Fingenious services () managed by FINBB.