{"paper_id":"6132c7db-5095-4a3c-a45f-c10c2d19897e","body_text":"Li et al. BMC Medicine          (2023) 21:195  \nhttps://doi.org/10.1186/s12916-023-02881-z\nRESEARCH ARTICLE Open Access\n© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which \npermits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the \noriginal author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or \nother third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line \nto the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory \nregulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this \nlicence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco \nmmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.\nBMC Medicine\nThe effects of coagulation factors on the risk \nof endometriosis: a Mendelian randomization \nstudy\nYan Li1†, Hongyan Liu1†, Shuting Ye1, Bumei Zhang1, Xiaopei Li1, Jiapei Yuan2, Yongrui Du1, Jianmei Wang1* and \nYang Yang1,3*   \nAbstract \nBackground Endometriosis is recognized as a complex gynecological disorder that can cause severe pain and \ninfertility, affecting 6–10% of all reproductive-aged women. Endometriosis is a condition in which endometrial tissue, \nwhich normally lines the inside of the uterus, deposits in other tissues. The etiology and pathogenesis of endometrio-\nsis remain ambiguous. Despite debates, it is generally agreed that endometriosis is a chronic inflammatory disease, \nand patients with endometriosis appear to be in a hypercoagulable state. The coagulation system plays important \nroles in hemostasis and inflammatory responses. Therefore, the purpose of this study is to use publicly available GWAS \nsummary statistics to examine the causal relationship between coagulation factors and the risk of endometriosis.\nMethods To investigate the causal relationship between coagulation factors and the risk of endometriosis, a two-\nsample Mendelian randomization (MR) analytic framework was used. A series of quality control procedures were \nfollowed in order to select eligible instrumental variables that were strongly associated with the exposures (vWF, \nADAMTS13, aPTT, FVIII, FXI, FVII, FX, ETP , PAI-1, protein C, and plasmin). Two independent cohorts of European ancestry \nwith endometriosis GWAS summary statistics were used: UK Biobank (4354 cases and 217,500 controls) and FinnGen \n(8288 cases and 68,969 controls). We conducted MR analyses separately in the UK Biobank and FinnGen, followed by \na meta-analysis. The Cochran’s Q test, MR-Egger intercept test, and leave-one-out sensitivity analyses were used to \nassess the heterogeneities, horizontal pleiotropy, and stabilities of SNPs in endometriosis.\nResults Our two-sample MR analysis of 11 coagulation factors in the UK Biobank suggested a reliable causal effect of \ngenetically predicted plasma ADAMTS13 level on decreased endometriosis risk. A negative causal effect of ADAMTS13 \nand a positive causal effect of vWF on endometriosis were observed in the FinnGen. In the meta-analysis, the causal \nassociations remained significant with a strong effect size. The MR analyses also identified potential causal effects of \nADAMTS13 and vWF on different sub-phenotypes of endometrioses.\nConclusions Our MR analysis based on GWAS data from large-scale population studies demonstrated the causal \nassociations between ADAMTS13/vWF and the risk of endometriosis. These findings suggest that these coagulation \n†Yan Li and Hongyan Liu contributed equally to this work.\n*Correspondence:\nJianmei Wang\nwangjianmei@tmu.edu.cn\nYang Yang\nyy@tmu.edu.cn\nFull list of author information is available at the end of the article\n\nPage 2 of 13Li et al. BMC Medicine          (2023) 21:195 \nfactors are involved in the development of endometriosis and may represent potential therapeutic targets for the \nmanagement of this complex disease.\nKeywords Two-sample Mendelian randomization, Endometriosis, Coagulation, GWAS, ADAMTS13\nBackground\nEndometriosis is defined as the deposit and growth of \nendometrial tissue that normally lines the inside of the \nuterus outside the uterine cavity [1]. Women who have \nendometriosis are more likely to experience dysmen -\norrhea, pelvic pain, and even infertility or difficulty \nconceiving. Endometriosis is a common and complex \ndisorder that affects up to 6–10% of all reproductive-aged \nwomen [2]. Although many factors, including hormones, \ninflammation, genetic factors, epigenetic factors, and \nenvironmental factors, are thought to contribute to the \ndevelopment of endometriosis, the etiology and patho -\ngenesis of endometriosis have not been completely elu -\ncidated [3, 4].\nAmong the hypotheses that have been proposed to \nexplain the pathogenesis of endometriosis, retrograde \nmenstruation, also known as Sampson’s theory, is the \nmost widely accepted [4, 5]. According to the model of \nretrograde menstruation, endometrial tissues are shed \nthrough the fallopian tubes into the pelvic cavity during \nmenstruation, resulting in the formation of ectopic endo-\nmetriotic lesions on the peritoneal tissue or pelvic organs. \nThe ectopic debris could be cleared by the immune sys -\ntem in healthy women, whereas the refluxed endometrial \nfragments might evade the immune surveillance system \nin endometriosis patients [6–8]. Defective immune sur -\nveillance is thought to play a role in the implantation and \ngrowth of ectopic endometrial tissue [6]. Endometrio -\nsis is also considered as a chronic inflammatory disease, \nowing to the presence of ectopic endometrial fragments, \nwhich cause an increase in proinflammatory factors and \nchemotactic cytokines [9–11]. Furthermore, angiogenesis \nis required to replenish the supply of nutrients and oxy -\ngen for the growth and survival of endometriotic lesions \n[12, 13]. Coagulation cascades have been implicated in \nboth inflammatory responses and angiogenesis [12, 14–\n16]. Several epidemiological observational studies have \nfound that patients with endometriosis are hypercoagu -\nlable and hyperfibrinolytic [17, 18]. Plasma fibrinogen, \nd-dimer, and plasminogen activator inhibitor levels are \nhigher in women with endometriosis when compared to \nhealthy controls while thrombin time and activated par -\ntial thromboplastin time decrease [19]. Adenomyosis, a \ncondition characterized by endometrial tissue growth \nwithin the uterine musculature, shares numerous com -\nmon symptoms with endometriosis, including pelvic pain \nand heavy menstrual bleeding [20]. Harmsen et al. have \nreported the increased levels of von Willebrand factor in \nectopic endometrium of adenomyosis patients which are \nassociated with the role of angiogenesis in adenomyosis \n[21]. Although several observational studies have been \nconducted to explore the relationship between coagula -\ntion cascades and endometriosis, the causal associations \nbetween coagulation factors and endometriosis remain \nunclear. The presence of residual confounding and poten-\ntial reverse causality issues in conventional observational \nstudies poses significant challenges in accurately measur-\ning the causal effect of specific coagulation factor on the \nrisk of endometriosis. Residual confounding occurs as a \nresult of inadequate adjustment for confounding varia -\nbles, as measuring a confounder may not fully character -\nize it. In addition, the association between the exposure \nand outcome may occur due to reverse causality, a phe -\nnomenon in which the outcome precedes and causes the \nexposure, rather than the exposure causing the outcome.\nAs an emerging method, Mendelian randomization \n(MR) is a novel statistical method that examines the \ncausal relationship between the exposure and outcome \nby using genetic variants as instrumental variables for \nthe exposure of interest [22, 23]. Because genetic variants \nare randomly allocated during gamete formation and \nconception, MR analysis could reduce confounding bias \nand reverse causality [23]. A two-sample MR analysis was \ncarried out in this study to investigate the causal effects \nof coagulation factors on endometriosis. There were \n11 coagulation factors incorporated as the exposures, \nincluding vWF (von Willebrand factor), ADAMTS13 (A \ndisintegrin and metalloproteinase with thrombospondin \nmotifs 13), aPTT (activated partial thromboplastin time), \nFVIII (factor VIII), FXI (factor XI), FVII (factor VII), FX \n(factor X), ETP (endogenous thrombin potential), PAI-1 \n(plasminogen activator inhibitor-1), protein C, and plas -\nmin. We leveraged summary-level GWAS data from two \nindependent large-scale cohorts of European ancestry, \nincluding the UK Biobank and FinnGen cohorts, to esti -\nmate a putative causal association of a specific coagula -\ntion factor with the risk of endometriosis.\nMethods\nStudy design\nThree critical assumptions must be met in the MR anal -\nysis. The first assumption is that the genetic variables \nshould be significantly related to the exposure, the sec -\nond assumption is that genetic variants extracted as \n\nPage 3 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \ninstrumental variables for the exposure are not related \nto other confounding factors, and the third assump -\ntion is that genetic variants influence the outcome solely \nthrough their effects on the exposure (i.e., no horizontal \npleiotropic effect) [24]. Figure 1 depicts the overall design \nof this study. We began by selecting 11 coagulation fac -\ntors based on publicly available GWAS data. Based on the \nGWAS summary statistics, we selected instrumental var -\niables for each coagulation factor. Then, using summary-\nlevel GWAS data of endometriosis from two independent \ncohorts, including the UK Biobank and FinnGen, we con-\nducted two-sample MR analyses separately to estimate \nthe causal effects of coagulation factors on endometrio -\nsis. To confirm the potential causal effects of coagulation \nfactors, we further meta-analyzed endometriosis GWAS \nsummary statistics from the UK Biobank and FinnGen. \nFinally, MR analyses were also performed to estimate the \ncausal associations of coagulation factors with the risk \nof various sub-phenotypes of endometrioses, including \nendometriosis of the intestine, ovary, pelvic peritoneum, \nfallopian tube, uterus, rectovaginal septum, and vagina.\nEndometriosis GWAS summary statistics\nTo obtain a reliable conclusion of the causal relation -\nships between coagulation factors and the risk of \nendometriosis, we have conducted a systematic analy -\nsis of endometriosis GWAS summary-level data col -\nlected from two large-scale cohorts, including the UK \nBiobank and FinnGen. The GWAS summary statistics \nfor endometriosis among individuals of European \nancestry in the UK Biobank were procured from the \nPan-UK Biobank website (https:// pan. ukbb. broad insti  \ntute. org/) via a phenotype description search for “endo -\nmetriosis” [25]. Correspondingly, the FinnGen cohort’s \nendometriosis GWAS summary statistics were accessi -\nble via the R package TwoSampleMR (v 0.5.6) [26] using \nthe GWAS ID “finn-b-N14_ENDOMETRIOSIS” as \ndocumented in the IEU OpenGWAS database (https://  \ngwas. mrcieu. ac. uk/) [27]. In the UK Biobank, the \ndiagnosis of endometriosis was defined by N80 in the \nInternational Classification of Diseases, 10th Revision \n(ICD-10). The GWASs for endometriosis from the UK \nBiobank of European ancestry were conducted on 4354 \ncases and 217,500 female controls. In FinnGen, endo -\nmetriosis is defined by N80 in ICD-10, 617 in ICD-9, \nand 6253 in ICD-8. The GWAS summary statistics for \nendometriosis from FinnGen included 8288 cases and \n68,969 controls. In addition, we also curated summary-\nlevel GWAS data from the FinnGen cohort for various \nsub-phenotypes of endometrioses, including endome -\ntriosis of the uterus (2372 cases, 68,969 controls), endo -\nmetriosis of the ovary (3231 cases, 68,969 controls), \nendometriosis of the fallopian tube (116 cases, 68,969 \ncontrols), endometriosis of the pelvic peritoneum (2953 \ncases, 68,969 controls), endometriosis of the rectovagi -\nnal septum and vagina (1360 cases, 68,969 controls), \nand endometriosis of the intestine (177 cases, 68,969 \ncontrols).\nFig. 1 Overall design of the MR analysis framework in this study. A flow chart depicts how the MR analysis was conducted step by step in this study\n\nPage 4 of 13Li et al. BMC Medicine          (2023) 21:195 \nGenetic instrumental variable selection\nWe used instrumental variables to investigate the causal \nassociations between coagulation factors and endome -\ntriosis. We searched for GWASs of coagulation factors in \nEuropean populations to curate genetic variants associ -\nated with coagulation factors. vWF, ADAMTS13, aPTT, \nFVII, FXI, FVII, FX, ETP , PAI-1, protein C, and plas -\nmin were chosen as the examined coagulation factors \nwith available genome-wide significant SNPs [28–36] \n(Additional file  1: Table  S1). Then, for each coagulation \nfactor, we went through a stringent quality control pro -\ncedure to select eligible instrumental variables for each \ncoagulation factor. First, we selected SNPs associated \nwith specific coagulation factors at genome-wide signifi -\ncance (P < 5e − 7) as candidate instrumental variables for \nfurther MR analysis. Second, to ensure the instrumental \nvariables for each exposure phenotype are independent, \nwe used the linkage disequilibrium (LD)-based clumping \nto remove SNPs in strong LD (r 2 threshold = 0.1, window \nsize = 10  Mb). The clumping step was carried out based \non the European reference panel of the 1000 Genomes \nProject, which was used to estimate LD between SNPs.\nFor SNPs that were not present in the endometriosis \nGWAS data, we used the LDlink tools to search for the \nmost correlated proxy SNPs using the 1000 Genomes \nof European population data (r 2 > 0.8) [37]. We also dis -\ncarded SNPs with non-concordant alleles and palin -\ndromic SNPs with ambiguous strands that could not be \ncorrected when harmonizing the exposure data and out -\ncome data. These stringently filtered SNPs were used as \nthe instrumental variables for subsequent MR analyses. \nTo determine whether there was a weak instrumental \nvariable bias, we calculated F-statistics to quantify the \nstrength of instrumental variables, where F-statistics \nlarger than 10 indicates a low possibility of weak instru -\nmental variable bias [38, 39] (Additional file 1: Table S1). \nAll the instrumental variable selection and quality con -\ntrol steps are performed using the R package TwoSam -\npleMR (v 0.5.6) [26].\nStatistical power calculation\nWe sought to assess the statistical power of our MR anal-\nyses through the use of an online web tool specialized for \nbinary outcomes (https:// sb452. shiny apps. io/ power) [40]. \nThe assessment of statistical power for MR analyses was \nbased on several parameters, including the total sample \nsize, the significance level of 0.05, the proportion of vari -\nance (R2) in the exposure explained by instrumental vari -\nables, and the ratio of cases to controls.\nMendelian randomization estimates\nWe combined the summary statistics (β coefficients \nand standard errors) to estimate the causal associations \nbetween 11 coagulation factors and endometriosis sepa -\nrately using different MR methods. The MR analyses \nwere first performed separately in the UK Biobank and \nFinnGen cohorts. Three MR methods based on differ -\nent assumptions were applied: inverse variance weighting \n(IVW), weighted mean (WM), and MR-Egger regres -\nsion. The IVW method was utilized as the main statis -\ntical model. There are fixed effects and random effects \nIVW methods available. We first calculated the causal \nestimates using the fixed effects IVW methods by meta-\nanalyzing Wald ratio estimates for each instrumental var-\niable. If significant heterogeneity (P < 0.05) is observed, \nthe random effects IVW method is added. In addition, we \nalso conducted MR analyses based on the meta-analyzed \nsummary statistics which are combined from the UK \nBiobank and FinnGen using the METAL tool [41].\nCausal estimates from MR analyses can only be inter -\npreted reliably if the three critical assumptions are met. \nHeterogeneity in causal estimates among instrumental \nvariables indicates a potential violation of the assump -\ntions of MR analysis [42]. The Cochran’s Q test was used \nto examine the heterogeneity in causal estimates, and \nwe used both the causal estimates of fixed effects IVW \nmethod and MR-Egger regression to detect heterogene -\nity. The heterogeneities were quantified using Cochran’s \nQ statistics and a P-value smaller than 0.05 was consid -\nered significant heterogeneity. To assess the potential \npleiotropic effects of instrumental variables, the MR-\nEgger regression was used. The directional horizontal \npleiotropy in the causal estimates may be indicated by the \nintercept term in MR-Egger regression. Additionally, we \nperformed a leave-one-out analysis where we excluded \neach SNP in turn and then ran MR analysis on the \nremaining SNPs in order to detect potentially outlying \ninstrumental variables [26]. The Steiger test of direction -\nality is also conducted to assess the causal relationship \nbetween the exposure and outcome. All MR analyses \nwere performed using the R package TwoSampleMR (v \n0.5.6) [26].\nResults\nSelection of instrumental variables\nWe systematically curated genome-wide significant \nSNPs associated with 11 coagulation factors (vWF, \nADAMTS13, aPTT, FVIII, FXI, FVII, FX, ETP , PAI-1, \nprotein C, and plasmin) from different GWAS results \nthrough literature searching to examine the potential \ncausal effects of these coagulation factors on the risk of \nendometriosis [28–36] (Additional file  1: Table S1). These \ncoagulation factors could be categorized into five groups, \nincluding platelet adhesion (vWF and ADAMTS13), \nintrinsic pathway (FXI, aPTT, and FVIII), extrinsic path -\nway (FVII), common pathways (ETP and FX), and fibrin \n\nPage 5 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \nclot dissociation (PAI-1, protein C, and plasmin). We \nfirst kept the SNPs that were significantly associated with \neach exposure phenotype in the corresponding GWAS \nstudy (P < 5e − 7). Then, we used LD-based clumping to \nobtain the LD-independent SNPs for the exposure (r 2 \nthreshold = 0.1, window size = 10  Mb). It is critical that \nthe effect of an SNP on the exposure and the effect of that \non the outcome are both attributed to the same allele. \nIn the harmonizing process, ambiguous SNPs with non-\nconcordant alleles and palindromic SNPs with ambigu -\nous strands that cannot be corrected were discarded. \nTherefore, the number of SNPs chosen as instrumental \nvariables for the exposure in subsequent two-sample \nMR analyses would eventually be equal to or less than \nthat listed in Additional file  1: Table  S1. To assess the \nstrength of each instrumental variable, we calculated the \nF-statistics for each instrument-exposure association. \nIn our study, the F-statistics were much greater than 10, \nindicating that those SNPs were strong instrumental vari-\nables (Additional file 1: Table S1). Moreover, we have cal-\nculated the statistical power for every exposure in each \ncohort. Notably, the results indicated that the statistical \npower ranged from 80% to 100% for all coagulation fac -\ntors, thereby affirming the robustness of our subsequent \nMR analyses (Additional file 1: Table S1).\nCausal effects of coagulation factors on endometriosis\nBased on the GWAS summary statistics for endome -\ntriosis in the UK Biobank of European ancestry, which \nincluded 4354 cases and 217,500 controls, we performed \nMR analyses to estimate the causal effects of 11 coagu -\nlation factors on the risk of endometriosis. The MR \nestimates from different methods were shown in Addi -\ntional file 1: Table S2. The findings demonstrated that the \ngenetically predicted plasma ADAMTS13 level is caus -\nally associated with a decreased risk of endometriosis \n(IVW: OR = 0.37, 95%CI: 0.22–0.61, P = 1.25e − 4; WM: \nOR = 0.41, 95%CI: 0.23–0.72, P = 2.05e − 3) (Fig.  2A, \nAdditional file  1: Table  S2, Additional file  2: Fig. S1). \nNotably, after accounting for multiple comparisons \nacross 11 coagulation factors, the negative causal effects \nof plasma ADAMTS13 level on endometriosis remained \nsignificant (IVW: P adjusted = 1.38e − 3). Furthermore, we \ndiscovered a mild negative causal relationship between \ngenetically predicted FXI levels and endometriosis \n(IVW: OR = 0.94, 95%CI: 0.89–0.98, P = 7.08e − 3; WM: \nOR = 0.95, 95%CI: 0.89–1.00, P = 0.059) (Fig.  2A, Addi -\ntional file  1: Table  S2, Additional file  2: Fig. S1). How -\never, other coagulation factors (vWF, aPTT, FVIII, FVII, \nFX, ETP , PAI-1, protein C, and plasmin) had no signifi -\ncant causal effect on endometriosis (Fig.  2A, Additional \nfile 1: Table S2, Additional file  2: Fig. S1). Heterogeneity \ntests revealed heterogeneity in endometriosis for three \ncoagulation factors, vWF (IVW: Cochran’s Q = 20.35, Phet-\nerogeneity = 0.041), aPTT (IVW: Cochran’s Q = 16.57, Phet-\nerogeneity = 0.011), and FVIII (IVW: Cochran’s Q = 11.80, \nPheterogeneity = 0.003) (Additional file  1: Table  S2). Addi -\ntional MR analyses using the random effects IVW \nmethod yielded causal effect estimates that were con -\nsistent with those estimated using the fixed effects IVW \nmethod (Additional file  1: Table  S2). In the MR-Egger \nintercept test, we detected no significant evidence of \nhorizontal pleiotropy (P pleiotropy > 0.05) (Additional file 1: \nTable  S2). Further leave-one-out analyses were carried \nout to ascertain potential outliers in the instrumen -\ntal variable estimation of ADATMS13 and FXI causal \neffects on the risk of endometriosis (Additional file  1: \nTable S3, Additional file  2: Fig. S2). Through the Steiger \ntest of directionality, the results corroborated the nega -\ntive causal effects of ADAMTS13 and FXI on the risk of \nendometriosis (Additional file  1: Table S2). As a result of \nthe MR analyses in the UK Biobank cohort, we were able \nto draw a robust conclusion that the genetically predicted \nplasma ADAMTS13 levels are causally associated with \na decreased risk of endometriosis, and the association \nbetween FXI and the decreased risk of endometriosis is \nlikely to be causal.\nAs a replication analysis, we performed MR analyses \nbased on the GWAS summary statistics for endometriosis \nin FinnGen (8288 cases and 68,969 controls). The findings \nhighlighted that the negative causal effects of the geneti -\ncally predicted plasma ADAMTS13 level on the risk of \nendometriosis remained significant with a large effect \nsize (IVW: OR = 0.46, 95%CI: 0.30–0.71, P = 5.31e − 4; \nWM: OR = 0.53, 95%CI: 0.33–0.85, P = 0.009), which \nwas consistent with the findings from the UK Biobank \n(Fig. 2B, Additional file 1: Table S4, Additional file 2: Fig. \nS3). After multiple test correction, the causal association \nestimated using fixed effects IVW method remained sig -\nnificant (IVW: P adjusted = 5.8e − 3). Despite the presence \nof heterogeneity in the causal estimates for ADAMTS13 \non endometriosis in FinnGen (IVW: Cochran’s Q = 11.91, \nPheterogeneity = 0.003), the causal effects estimated using the \nrandom effects IVW method remained borderline signifi-\ncantly with a strong effect size (IVW: OR = 0.46, 95%CI: \n0.16–0.1.34, P = 0.056) (Additional file  1: Table  S4). The \nresults also showed that the genetically predicted plasma \nvWF level was positively causally associated with the risk \nof endometriosis (IVW: OR = 1.28, 95%CI: 1.06–1.53, \nP = 0.009; WM: OR = 1.33, 95%CI: 1.08–1.62, P = 0.006), \nalthough the effect may not remain significant after \nadjusting for multiple comparisons (Fig.  2B, Additional \nfile 1: Table S4, Additional file 2: Fig. S3). Conversely, the \nsignificant negative causal relationship between FXI and \nendometriosis observed in the UK Biobank was not rep -\nlicated in FinnGen (Fig.  2B, Additional file  1: Table  S3, \n\nPage 6 of 13Li et al. BMC Medicine          (2023) 21:195 \nFig. 2 Causal estimates of 11 coagulation factors on endometriosis by MR analysis. A Forest plots showing causal estimates of 11 coagulation \nfactors on endometriosis estimated in the UK Biobank of European ancestry. B Forest plots showing causal effects of 11 coagulation factors on \nendometriosis estimated in FinnGen. The odds ratio (OR) was estimated using the fixed effect IVW method. The horizontal bars represent 95% \nconfidence intervals (CI)\n\nPage 7 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \nAdditional file 2: Fig. S3). We observed no obvious hori -\nzontal pleiotropy in the MR-Egger intercept test and \nno potentially influential instrumental variable in the \nleave-one-out analysis for ADAMTS13 and vWF (Addi -\ntional file  1: Table  S4 and S5, Additional file  2: Fig. S4). \nThe directionality of their causal effects was also con -\nfirmed using the Steiger test (Additional file  1: Table S4). \nIn conclusion, our FinnGen cohort results suggest \nthat ADAMTS13 levels are causally associated with a \ndecreased risk of endometriosis, and the positive associa -\ntion observed between vWF and the risk of endometrio -\nsis is likely to be causal.\nWith the purpose of verifying the causal effects of \ncoagulation factors on endometriosis, we meta-ana -\nlyzed the GWAS summary statistics obtained from the \nUK Biobank and FinnGen, thereby enhancing the sam -\nple size and statistical power. Subsequent MR analy -\nses were carried out using the meta-analyzed GWAS \nsummary statistics for endometriosis. The results sup -\nported the strong causal effect of ADAMTS13 on the \ndecreased risk of endometriosis (IVW: OR = 0.42, \n95%CI: 0.30–0.58, P  = 2.85e − 7; WM: OR = 0.44, 95%CI: \n0.30–0.66, P = 5.76e − 5) (Fig.  3, Additional file  1: \nTable S6, Additional file  2: Fig. S5). Notably, heteroge -\nneity in causal estimates of ADAMTS13 was detected \nby the heterogeneity test (IVW: Cochran’s Q  = 13.23, \nPheterogeneity  = 0.004), necessitating use of the random \neffects IVW method to evaluate the causal association. \nThe result from random effects IVW analysis confirmed \nthe strong negative causal link between ADAMTS13 \nand endometriosis (Additional file  1: Table S6). The sig-\nnificant MR result of vWF on the risk of endometriosis \nwas also observed (IVW: OR = 1.26, 95%CI: 1.09–1.46, \nP = 0.002; WM: OR = 1.29, 95%CI: 1.10–1.51, P  = 0.002) \n(Fig.  3, Additional file  1: Table  S6, Additional file  2: \nFig. S5). Moreover, the absence of potentially influen -\ntial instrumental variables was ascertained by leave-\none-out analysis (Additional file  1: Table S7, Additional \nfile  2: Fig. S6), and the Steiger test validated the direc -\ntionality of the causal effects on the risk of endome -\ntriosis (Additional file  1: Table  S6). Summarizing the \nfindings from the meta-analysis, we could conclude that \nthe genetically predicted plasma ADAMTS13 levels \nhave a negative causal effect on the risk of endometrio -\nsis, suggesting that ADAMTS13 serves as a protective \nfactor for endometriosis. Conversely, the genetically \npredicted plasma vWF levels are positively associated \nwith the risk of endometriosis, indicating vWF function \nas a risk factor for the development of endometriosis.\nFig. 3 Causal estimates of 11 coagulation factors on endometriosis in a meta-analysis. Forest plots showing causal estimates of 11 coagulation \nfactors on endometriosis in a meta-analysis of UK Biobank and FinnGen. The odds ratio (OR) was estimated using the fixed effect IVW method. The \nhorizontal bars represent 95% confidence intervals (CI)\n\nPage 8 of 13Li et al. BMC Medicine          (2023) 21:195 \nCausal effects of coagulation factors on different \nsub‑phenotypes of endometrioses\nDepending on the location and growth of ectopic endo -\nmetriotic lesions, endometriosis could be categorized. \nThe precise sub-phenotypes of endometriosis expe -\nrienced by patients may have an impact on both their \nsymptoms as well as their chance of infertility. Endo -\nmetrioses of the intestine, ovary, pelvic peritoneum, \nuterus, fallopian tube, and rectovaginal vaginal regions \nwere among the five sub-phenotypes of endometrioses \ndiagnosed in the FinnGen cohort. The GWAS summary \nstatistics of various sub-phenotypes of endometrioses \nwere also available in the FinnGen cohort. The number \nof patients ranged from 116 in endometriosis of the fal -\nlopian tube to 3231 in endometriosis of the ovary. Some \npatients might have more than one sub-phenotype of \nendometriosis because there was an overlap between dif -\nferent sub-phenotypes.\nWe employed MR analyses to further investigate the \ncausal effects of genetically predicted plasma levels of \nADAMTS13 and vWF on the risk of various sub-phe -\nnotypes of endometrioses. The findings demonstrated \nthat ADAMTS13 is negatively causally associated with \nthe risk of endometriosis of the ovary (IVW: OR = 0.48, \n95%CI: 0.25–0.92, P = 0.028; WM: OR = 0.58, 95%CI: \n0.2–81.20, P = 0.140), endometriosis of the pelvic peri -\ntoneum (IVW: OR = 0.32, 95%CI: 0.16–0.64, P = 0.001; \nWM: OR = 0.40, 95%CI: 0.19–0.85, P = 0.017), and \nendometriosis of the uterus (IVW: OR = 0.45, 95%CI: \n0.21–0.97, P = 0.041; WM: OR = 0.44, 95%CI: 0.20–0.99, \nP = 0.048) (Fig. 4, Additional file 1: Table S8). In addition, \nADAMTS13 had a negative but not statistically signifi -\ncant causal effect on endometriosis of the rectovaginal \nseptum and vagina, and there was no evidence of a causal \neffect of ADAMTS13 on endometriosis of the intestine \n(Fig. 4, Additional file  1: Table S8). As heterogeneity was \ndetected, we conducted a random effects IVW analysis \nto validate the findings (Additional file 1: Table S8). From \nthe random effects IVW analysis, the causal estimates of \nADAMTS13 on endometriosis of the uterus remained \nborderline significant (IVW: OR = 0.45, 95%CI: 0.20-\n–1.02, P = 0.051), while the causal estimates for endome -\ntrioses of the ovary (IVW: OR = 0.48, 95%CI: 0.13–1.79, \nP = 0.274) and pelvic peritoneum (IVW: OR = 0.32, \n95%CI:0.08–1.32, P = 0.116) attenuated towards non-\nsignificance (Additional file  1: Table S8). Meanwhile, the \nsignificant causal estimates of vWF were also observed \nfor endometriosis of the ovary (IVW: OR = 1.34, 95%CI: \nFig. 4 Causal estimates of vWF and ADAMTS13 on different sub-phenotypes of endometrioses. Forest plots depicting causal estimates of vWF and \nADAMTS13 on different sub-phenotypes of endometrioses in FinnGen, including endometriosis of intestine, endometriosis of ovary, endometriosis \nof pelvic peritoneum, endometriosis of uterus, endometriosis of the fallopian tube, and endometriosis of the rectovaginal septum and vagina. The \nodds ratio (OR) was estimated using the fixed effect IVW method. The horizontal bars represent 95% confidence intervals (CI). Significant P values \nare highlighted in red\n\nPage 9 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \n1.02–1.77, P = 0.035; WM: OR = 1.37, 95%CI: 1.03–1.81, \nP = 0.028) and endometriosis of the pelvic peritoneum \n(IVW: OR = 1.48, 95%CI: 1.11–1.97, P = 0.008; WM: \nOR = 1.53, 95%CI: 1.13–2.08, P = 0.006) (Fig.  4, Addi -\ntional file 1: Table S8). In summary, the evidence suggests \nthat ADAMTS13 may have a negative causal relationship \nwith endometriosis of the ovary, pelvic peritoneum, and \nuterus, while vWF may have a positive causal relationship \nwith endometriosis of the ovary and pelvic peritoneum.\nIn addition, we noticed that the ratios of cases to con -\ntrols significantly varied across sub-phenotypes, rang -\ning from 1/594 for endometriosis of the fallopian tube to \n1/21 for endometriosis of the ovary. Such disparity may \nimpede the statistical power of a MR study, prompting \nthe need to evaluate the statistical power. To establish the \nvalidity of the results, we additionally calculated the sta -\ntistical power for the MR analysis in each sub-phenotype \ncohort. The statistical power was merely about 14% and \n20% for sub-phenotypes of the fallopian tube and intes -\ntine, respectively (Additional file  1: Table S8). Therefore, \nwe should draw our conclusions with cautions for these \ntwo sub-phenotypes. In contrast, the statistical power for \nthe other four sub-phenotypes, including endometriosis \nof the uterus, ovary, pelvic peritoneum, and rectovagi -\nnal septum and vagina ranged between 80% and 1, thus \naffirming the robustness of the MR results of these sub-\nphenotypes (Additional file 1: Table S8).\nIn addition, the condition of endometriosis of the \nuterus, also referred to as adenomyosis, has been cat -\negorized as a separate disease, despite its classification \nas a form of endometriosis in ICD-10. Several stud -\nies have suggested that endometriosis and adenomyo -\nsis share similar pathophysiology, specifically related to \nsomatic epithelial mutations and epigenetic abnormali -\nties. In order to determine the potential effects of incor -\nporating endometriosis of the uterus in our MR study, \nwe employed LDSC to examine the genetic correlations \nbetween adenomyosis and other sub-phenotypes [43, 44]. \nThe results indicate strong genetic correlations, ranging \nfrom 0.67 to 0.93, indicating a shared genetic architec -\nture and pathophysiological mechanisms between aden -\nomyosis and endometriosis (Additional file  1: Table S9). \nThese findings suggest that the inclusion of adenomyosis \nis unlikely to significantly impact the causal estimation of \ncoagulation factors on the risk of endometriosis.\nDiscussion\nUtilizing summary statistics from two large-scale \nGWASs of European ancestry including UK Biobank \nand FinnGen, we investigated the causal effects of 11 \ncoagulation factors on the risk of endometriosis, employ -\ning a unified MR framework to analyze GWAS data. \nOur results indicate that genetically predicted plasma \nADAMTS13 levels were inversely associated with endo -\nmetriosis, while genetically predicted plasma vWF levels \ndemonstrated a positive causal association with endo -\nmetriosis, as confirmed in the meta-analysis combining \nthe cohorts. Furthermore, MR analyses also revealed the \ncausal associations in different sub-phenotypes of endo -\nmetrioses that are categorized by ectopic location. These \nfindings have significant implications for the develop -\nment of endometriosis prevention strategies and treat -\nment methods. For example, the findings underscore \nthe significance of monitoring the ADAMTS13 plasma \nlevels in individuals diagnosed with endometriosis. Fur -\nthermore, the results also provide a potential therapeutic \napproach that entails regulating the ADAMTS13 plasma \nlevel, thereby enabling the management and prevention \nof endometriosis progression and recurrence.\nAlthough several factors involved in the development \nof endometriosis have been uncovered, the precise etiol -\nogy and pathogenesis of endometriosis remain obscure, \nand its treatment remains controversial [3, 4]. A thor -\nough understanding of endometriosis is required for \nthe development of effective preventative and treatment \nstrategies. Sampson proposed the retrograde menstrua -\ntion theory, which states that menstrual blood contain -\ning endometrial cells retrograde through fallopian tubes \ninto the pelvic cavity instead of out of the body, leading \nto the formation of ectopic endometriotic lesions [45]. \nAlthough Sampson’s theory is the most widely accepted, \nseveral alternative hypotheses have been put forth, such \nas the theories of stem cell origin and altered immunity \n[46, 47]. Endometriosis is considered as a consequence \nof a complex interplay of genetic, anatomical, environ -\nmental, and immunologic factors [1–3]. Despite contra -\ndicting accounts regarding the origin of endometriosis, \nit is generally accepted that endometriosis is associated \nwith a local inflammatory response, and that vasculariza -\ntion at the site of endometriotic invasion plays a crucial \nrole in the development of the lesions [48]. Notably, the \ncoagulation system has been acknowledged as playing \ncritical roles in modulating both inflammatory responses \nand angiogenesis [12, 14–16]. Recently, Li et  al. have \nreported that the fibrinogen alpha chain could promote \nthe migration and invasion of endometrial cells and \npromote angiogenesis in endometriosis [49–52]. Heavy \nmenstrual bleeding (HMB) is a prevalent clinical symp -\ntom of endometriosis. Studies have raised the possibil -\nity of an imbalance in coagulation factors playing a role \nin HMB in patients with endometriosis. Research has \nnoted that women with endometriosis exhibit a hyperco -\nagulable status characterized by elevated levels of specific \ncoagulation factors, such as fibrinogen and vWF [17–19, \n53, 54]. These elevated factors may contribute to HMB \nby promoting the formation of blood clots. As such, an \n\nPage 10 of 13Li et al. BMC Medicine          (2023) 21:195 \nimbalanced coagulation system may represent a plausi -\nble etiologic mechanism behind HMB in endometriosis. \nDespite the growing interest regarding the involvement \nof coagulation factors in the pathogenesis of endometrio-\nsis, the causal roles of these factors in the development of \nendometriosis remain uncertain.\nThis is the first study to investigate the causal relation -\nships between coagulation factors and the risk of endo -\nmetriosis utilizing MR analyses on large-scale population \ncohorts, which provided unconfounded causal estimates. \nThe findings highlighted that the plasma ADAMTS13 \nlevels have a negative causal effect on endometriosis, \nwhereas the plasma vWF levels have a positive causal \neffect on endometriosis. In other words, ADAMTS13 is \nfound to have a protective effect associated with endo -\nmetriosis, while vWF is characterized as a risk factor \nfor the development of the condition. The multimeric \nglycoprotein vWF is stored in the Weibel-Palade bodies \nand α-granules of platelets, awaiting release upon stim -\nulation. Its primary function involves the formation of \na bridge between surface receptors on platelets and the \nendothelium, allowing for platelet recruitment follow -\ning an injury [55]. ADAMTS13 is a multidomain metal -\nloprotease that is predominantly synthesized in the liver \nby hepatic stellate cells, and its primary role is to regu -\nlate thrombogenesis by cleaving hyperactive ultra-large \nmultimers of vWFs into less active, smaller fragments \n[56]. Given the vWF-cleaving function of ADAMTS13, \nthe biological functions of ADAMTS13 and vWF are \nclosely related. The thrombotic thrombocytopenic pur -\npua (TTP) that arises in people with severe ADAMTS13 \ndeficiency has highlighted the relevance of ADAMTS13 \nfunction [57, 58]. ADAMTS13 deficiency may lead to the \naccumulation of vWF multimers, which causes intravas -\ncular platelet aggregation and microthrombosis, resulting \nin TTP . Aside from the well-established role in hemosta-\nsis, the balance between ADAMTS13 and vWF has been \nlinked to a variety of diseases, such as systemic inflam -\nmation, pancreatitis, and multiple sclerosis [59–61]. The \nbiosynthesis and secretion of ADAMTS13 from vascu -\nlar endothelial cells have raised the interests in the role \nof ADAMTS13 in angiogenesis [62–64]. The balance \nbetween ADAMTS13 and vWF is crucial for control -\nling angiogenesis, as demonstrated by numerous studies \n[63]. In addition, Xiao et al. have recently demonstrated \nthe proteolytically active ADAMTS13 is expressed in the \nhuman placental tissues and has a role in trophoblast \ncell proliferation, migration, invasion, and tube forma -\ntion [65]. Overall, the balance between ADAMTS13 and \nvWF not only regulates hemostasis, but also exerts a role \nin inflammation modulation, regulating angiogenesis, \nand tissue remodeling. Our findings of this MR study \nconfirmed the causal roles of ADAMTS13 and vWF on \nendometriosis. Although the UK Biobank and FinnGen \ncohorts were utilized, there remains a need for independ-\nent validation of these causal relationships. Furthermore, \ngiven the potential pathophysiology of endometriosis, a \nmore comprehensive understanding of the molecular \nmechanisms and action of these coagulation factors in \nendometriosis pathogenesis requires additional experi -\nmental validation.\nThere are several strengths in this study. First, because \nit is based on the fact that genetic variants are randomly \nallocated during gamete formation and conception, the \nresults of MR analysis are less susceptible to confounding \nbias and reverse causality [23]. Second, we employed sep-\narate samples for the exposures (coagulation factors) and \nthe outcome (endometriosis) data to ensure two-sam -\nple MR analyses, which avoid inflating the bias of weak \ninstrumental variables. Third, we incorporated two inde -\npendent large-scale cohorts for MR analyses, followed by \na meta-analysis, so that a sufficiently enough sample size \nof the outcome could assure the generalizability of causal \nassociations. In addition, the consistent causal effect \nestimates of ADAMTS13 on endometriosis among the \nUK Biobank, FinnGen, and the meta-analysis alleviated \nconcerns on false-positive results. Fourth, we employed \nmultiple supplementary analyses, such as heterogeneity, \npleiotropy, and leave-one-out sensitivity analyses, to ver -\nify the viability of the assumptions regarding the instru -\nmental variables.\nNonetheless, several limitations also need to be \nacknowledged. First, the number of instrumental vari -\nables for each coagulation factor, as outlined in Addi -\ntional file  1: Table  S1 ranged from three to thirteen. \nFurthermore, some instrumental variables will be dis -\ncarded during MR analyses when harmonizing the \nexposure and outcome data. These limitations suggest \nthat the final MR estimates may be subject to influ -\nence from the limited number of instrumental vari -\nables. Nevertheless, the statistical power calculations \nfor each coagulation factor within each cohort indicate \nthat adequate power was achieved, with estimated sta -\ntistical power ranging from 80% to 1. Therefore, despite \nthe potential limitations, the results presented in this \nstudy remain sufficiently powered to draw robust con -\nclusions. Second, only genome-wide significant SNPs \nfor different coagulation factors were available in the \nexposure GWAS data, preventing us from perform -\ning bi-directional MR analyses. Third, in the context of \nendometriosis, a female-specific condition, it is note -\nworthy that existing GWASs examining diverse coagu -\nlation factors have been conducted on a sex-combined \nbias. As two-sample MR necessitates consistency in the \nunderlying population for both sample sets, it is impor -\ntant to consider potential discrepancies with regard \n\nPage 11 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \nto genetic estimates of coagulation factors in females \nversus males, which may introduce bias into our MR \nfindings. Fourth, because this study was limited to peo -\nple of European ancestry, the findings may not be gen -\neralizable to other populations. More studies into the \ncausal associations between coagulation factors and \nendometriosis in other populations are needed.\nConclusions\nTo the best of our knowledge, this is the first MR study \nto examine the causal associations between coagula -\ntion factors and the risk of endometriosis in the Euro -\npean population. The findings convincingly support the \ncausal associations between ADAMTS13/vWF and the \nrisk of endometriosis. This study contributes to a bet -\nter understanding of the involvement of coagulation \ncascades in the development of endometriosis. These \nfindings may have important implications for endome -\ntriosis prevention and treatment strategies.\nAbbreviations\nCI  Confidence interval\nGWAS  Genome-wide association study\nIVW  Inverse variance weighting\nLD  Linkage disequilibrium\nMR  Mendelian randomization\nOR  Odds ratio\nRCTs  Randomized controlled trials\nSNP  Single nucleotide polymorphism\nWM  Weighted mean\nSupplementary Information\nThe online version contains supplementary material available at https:// doi. \norg/ 10. 1186/ s12916- 023- 02881-z.\nAdditional file 1: Table S1. Selected instrumental variables for coagula-\ntion factors in this study. Table S2. Summary statistics of the causal esti-\nmates of coagulation factors on endometriosis in UK Biobank. Table S3. \nThe results of leave-one-out analyses for endometriosis in UK Biobank. \nTable S4. Summary statistics of the causal estimates of coagulation fac-\ntors on endometriosis in FinnGen. Table S5. The results of leave-one-out \nanalyses for endometriosis in FinnGen. Table S6. Summary statistics of \nthe causal estimates of coagulation factors on endometriosis in the meta-\nanalysis. Table S7. The results of leave-one-out analyses for endometriosis \nin the meta-analysis. Table S8. Summary statistics of the causal estimates \nof vWF and ADAMTS13 on different sub-phenotypes of endometrioses. \nTable S9. Genetic correlations between endometriosis of uterus (adeno-\nmyosis) and other sub-phenotypes.\nAdditional file 2: Fig. S1. Scatter plots for MR analyses of the causal effect \nof 11 coagulation factors on endometriosis in UK Biobank. Fig. S2. Plots of \nleave-one-out analyses for the causal associations in UK Biobank. Fig. S3. \nScatter plots for MR analyses of the causal effect of 11 coagulation factors \non endometriosis in FinnGen. Fig. S4. Plots of leave-one-out analyses for \nthe causal associations in FinnGen. Fig. S5. Scatter plots for MR analyses of \nthe causal effect of ADAMTS13 and vWF on endometriosis in a meta-anal-\nysis. Fig. S6. Plots of leave-one-out analyses for the causal associations in \nthe meta-analysis.\nAcknowledgements\nWe would like to thank the UK Biobank consortia, FinnGen consortia, GWAS \nCatalog, and Neale Lab for sharing the GWAS data. We would like to thank the \nanonymous reviewers for their constructive comments.\nAuthors’ contributions\nYY, JW, and YD conceived and designed the study. YY supervised the study \nand data analysis. YL and HL performed the data analysis with help from SY, \nBZ, and XL. YY, JY, YL, and HL wrote the manuscript. All authors revised and \napproved the final manuscript.\nFunding\nThis study was funded by the Natural Science Foundation of China (Grant No. \n32100534 and No. 32200514), Talent Excellence Program from Tianjin Medical \nUniversity (to YY).\nAvailability of data and materials\nThe GWAS summary statistics for coagulation factors are available in the \nGWAS Catalog or the published article and its supplementary files. The GWAS \nsummary statistics for endometriosis are available on the Neale lab Pan-UK \nBiobank website (https:// pan. ukbb. broad insti tute. org/) for the UKBB cohort \nand the IEU GWAS database (https:// gwas. mrcieu. ac. uk/) for FinnGen [25, 27].\nDeclarations\nEthics approval and consent to participate\nThe analyses were based on publicly available data that have been approved \nby relevant review boards. The UK Biobank was approved by the Research \nEthics Committee (REC reference: 21/NW/0157). The FinnGen was approved \nby the Coordinating Ethics Committee of the Hospital District of Helsinki and \nUusimaa (HUS/990/2017).\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare that they have no competing interests.\nAuthor details\n1 Department of Family Planning, The Second Hospital of Tianjin Medical \nUniversity, The Province and Ministry Co-Sponsored Collaborative Innova-\ntion Center for Medical Epigenetics, Tianjin Key Laboratory of Inflammation \nBiology, School of Basic Medical Sciences, Tianjin Medical University, Tian-\njin 300070, China. 2 State Key Laboratory of Experimental Hematology, National \nClinical Research Center for Blood Diseases, Haihe Laboratory of Cell Ecosys-\ntems, Institute of Hematology and Blood Diseases Hospital, Chinese Academy \nof Medical Sciences and Peking Union Medical College, Tianjin 300020, China. \n3 Department of Bioinformatics, School of Basic Medical Sciences, Tianjin Medi-\ncal University, Tianjin 300070, China. \nReceived: 20 January 2023   Accepted: 26 April 2023\nReferences\n 1. Vercellini P , Vigano P , Somigliana E, Fedele L. Endometriosis: pathogenesis \nand treatment. Nat Rev Endocrinol. 2014;10(5):261–75.\n 2. Burney RO. The genetics and biochemistry of endometriosis. Curr Opin \nObstet Gynecol. 2013;25(4):280–6.\n 3. Wang Y, Nicholes K, Shih IM. The origin and pathogenesis of endometrio-\nsis. Annu Rev Pathol. 2020;15:71–95.\n 4. Saunders PTK, Horne AW. Endometriosis: etiology, pathobiology, and \ntherapeutic prospects. Cell. 2021;184(11):2807–24.\n 5. Vercellini P , Abbiati A, Vigano P , Somigliana ED, Daguati R, Meroni F, Cro-\nsignani PG. Asymmetry in distribution of diaphragmatic endometriotic \nlesions: evidence in favour of the menstrual reflux theory. Hum Reprod. \n2007;22(9):2359–67.\n\nPage 12 of 13Li et al. BMC Medicine          (2023) 21:195 \n 6. Christodoulakos G, Augoulea A, Lambrinoudaki I, Sioulas V, Creatsas G. \nPathogenesis of endometriosis: the role of defective ‘immunosurveillance.’ \nEur J Contracept Reprod Health Care. 2007;12(3):194–202.\n 7. Du Y, Liu X, Guo SW. Platelets impair natural killer cell reactivity and \nfunction in endometriosis through multiple mechanisms. Hum Reprod. \n2017;32(4):794–810.\n 8. Yu JJ, Sun HT, Zhang ZF, Shi RX, Liu LB, Shang WQ, Wei CY, Chang KK, Shao \nJ, Wang MY, et al. IL15 promotes growth and invasion of endometrial \nstromal cells and inhibits killing activity of NK cells in endometriosis. \nReproduction. 2016;152(2):151–60.\n 9. Taylor HS, Kotlyar AM, Flores VA. Endometriosis is a chronic sys-\ntemic disease: clinical challenges and novel innovations. Lancet. \n2021;397(10276):839–52.\n 10. Othman Eel D, Hornung D, Salem HT, Khalifa EA, El-Metwally TH, Al-Hendy \nA. Serum cytokines as biomarkers for nonsurgical prediction of endome-\ntriosis. Eur J Obstet Gynecol Reprod Biol. 2008;137(2):240–6.\n 11. Tseng JF, Ryan IP , Milam TD, Murai JT, Schriock ED, Landers DV, Taylor RN. \nInterleukin-6 secretion in vitro is up-regulated in ectopic and eutopic \nendometrial stromal cells from women with endometriosis. J Clin Endo-\ncrinol Metab. 1996;81(3):1118–22.\n 12. May K, Becker CM. Endometriosis and angiogenesis. Minerva Ginecol. \n2008;60(3):245–54.\n 13. Taylor RN, Yu J, Torres PB, Schickedanz AC, Park JK, Mueller MD, Sidell N. \nMechanistic and therapeutic implications of angiogenesis in endometrio-\nsis. Reprod Sci. 2009;16(2):140–6.\n 14. Levi M, van der Poll T. Inflammation and coagulation. Crit Care Med. \n2010;38(2 Suppl):S26-34.\n 15. Hess R, Wujak L, Hesse C, Sewald K, Jonigk D, Warnecke G, Fieguth \nHG, de Maat S, Maas C, Bonella F, et al. Coagulation factor XII regu-\nlates inflammatory responses in human lungs. Thromb Haemost. \n2017;117(10):1896–907.\n 16. Belting M, Ahamed J, Ruf W. Signaling of the tissue factor coagulation \npathway in angiogenesis and cancer. Arterioscler Thromb Vasc Biol. \n2005;25(8):1545–50.\n 17. Vigano P , Ottolina J, Sarais V, Rebonato G, Somigliana E, Candiani \nM. Coagulation status in women with endometriosis. Reprod Sci. \n2018;25(4):559–65.\n 18. Ottolina J, Bartiromo L, Dolci C, Salmeri N, Schimberni M, Villanacci R, \nVigano P , Candiani M. Assessment of coagulation parameters in women \naffected by endometriosis: validation study and systematic review of the \nliterature. Diagnostics (Basel). 2020;10(8):567.\n 19. Lin Q, Ding SJ, Zhu TH, Li TT, Huang XF, Zhang XM. Role and clinical signif-\nicance of coagulation and inflammatory factors in moderate and severe \novarian endometriosis. Zhonghua Fu Chan Ke Za Zhi. 2018;53(3):167–71.\n 20. Bulun SE, Yildiz S, Adli M, Chakravarti D, Parker JB, Milad M, Yang L, \nChaudhari A, Tsai S, Wei JJ et al. Endometriosis and adenomyosis: shared \npathophysiology. Fertil Steril. 2023;119(5):746–50.\n 21. Harmsen MJ, Wong CFC, Mijatovic V, Griffioen AW, Groenman F, Hehen-\nkamp WJK, Huirne JAF. Role of angiogenesis in adenomyosis-associated \nabnormal uterine bleeding and subfertility: a systematic review. Hum \nReprod Update. 2019;25(5):647–71.\n 22. Grover S, Del Greco MF, Stein CM, Ziegler A. Mendelian randomization. \nMethods Mol Biol. 2017;1666:581–628.\n 23. Verduijn M, Siegerink B, Jager KJ, Zoccali C, Dekker FW. Mendelian ran-\ndomization: use of genetics to enable causal inference in observational \nstudies. Nephrol Dial Transplant. 2010;25(5):1394–8.\n 24. Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisa-\ntion studies: a guide, glossary, and checklist for clinicians. BMJ. 2018;362: \nk601.\n 25. Pan-UK Biobank. Pan-ancestry genetic analysis of the UK Biobank. 2020. \nhttps:// pan. ukbb. broad insti tute. org.\n 26. Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D, Laurin \nC, Burgess S, Bowden J, Langdon R, et al. The MR-Base platform sup-\nports systematic causal inference across the human phenome. Elife. \n2018;7:e34408.\n 27. Elsworth B, Lyon M, Alexander T, Liu Y, Matthews P , Hallett J, Bates P , \nPalmer T, Haberland V, Smith GD, et al. The MRC IEU OpenGWAS data \ninfrastructure. 2020.\n 28. Sabater-Lleal M, Huffman JE, de Vries PS, Marten J, Mastrangelo MA, Song \nC, Pankratz N, Ward-Caviness CK, Yanek LR, Trompet S, et al. Genome-\nwide association transethnic meta-analyses identifies novel associations \nregulating coagulation factor VIII and von Willebrand factor plasma levels. \nCirculation. 2019;139(5):620–35.\n 29. Ma Q, Jacobi PM, Emmer BT, Kretz CA, Ozel AB, McGee B, Kimchi-Sarfaty \nC, Ginsburg D, Li JZ, Desch KC. Genetic variants in ADAMTS13 as well as \nsmoking are major determinants of plasma ADAMTS13 levels. Blood Adv. \n2017;1(15):1037–46.\n 30. de Vries PS, Sabater-Lleal M, Huffman JE, Marten J, Song C, Pankratz N, \nBartz TM, de Haan HG, Delgado GE, Eicher JD, et al. A genome-wide \nassociation study identifies new loci for factor VII and implicates factor VII \nin ischemic stroke etiology. Blood. 2019;133(9):967–77.\n 31. Sun BB, Maranville JC, Peters JE, Stacey D, Staley JR, Blackshaw J, Burgess \nS, Jiang T, Paige E, Surendran P , et al. Genomic atlas of the human plasma \nproteome. Nature. 2018;558(7708):73–9.\n 32. Rocanin-Arjo A, Cohen W, Carcaillon L, Frere C, Saut N, Letenneur L, \nAlhenc-Gelas M, Dupuy AM, Bertrand M, Alessi MC, et al. A meta-analysis \nof genome-wide association studies identifies ORM1 as a novel gene \ncontrolling thrombin generation potential. Blood. 2014;123(5):777–85.\n 33. Tang W, Basu S, Kong X, Pankow JS, Aleksic N, Tan A, Cushman M, Boer-\nwinkle E, Folsom AR. Genome-wide association study identifies novel loci \nfor plasma levels of protein C: the ARIC study. Blood. 2010;116(23):5032–6.\n 34. Tang W, Schwienbacher C, Lopez LM, Ben-Shlomo Y, Oudot-Mellakh T, \nJohnson AD, Samani NJ, Basu S, Gogele M, Davies G, et al. Genetic asso-\nciations for activated partial thromboplastin time and prothrombin time, \ntheir gene expression profiles, and risk of coronary artery disease. Am J \nHum Genet. 2012;91(1):152–62.\n 35. Suhre K, Arnold M, Bhagwat AM, Cotton RJ, Engelke R, Raffler J, Sarwath \nH, Thareja G, Wahl A, DeLisle RK, et al. Connecting genetic risk to disease \nend points through the human blood plasma proteome. Nat Commun. \n2017;8:14357.\n 36. Huang J, Sabater-Lleal M, Asselbergs FW, Tregouet D, Shin SY, Ding J, \nBaumert J, Oudot-Mellakh T, Folkersen L, Johnson AD, et al. Genome-\nwide association study for circulating levels of PAI-1 provides novel \ninsights into its regulation. Blood. 2012;120(24):4873–81.\n 37. Machiela MJ, Chanock SJ. LDlink: a web-based application for exploring \npopulation-specific haplotype structure and linking correlated alleles of \npossible functional variants. Bioinformatics. 2015;31(21):3555–7.\n 38. Li B, Martin EB. An approximation to the F distribution using the chi-\nsquare distribution. Comput Stat Data Anal. 2002;40(1):21–6.\n 39. Georgakis MK, Gill D, Rannikmae K, Traylor M, Anderson CD, Lee JM, \nKamatani Y, Hopewell JC, Worrall BB, Bernhagen J, et al. Genetically \ndetermined levels of circulating cytokines and risk of stroke. Circulation. \n2019;139(2):256–68.\n 40. Burgess S. Sample size and power calculations in Mendelian randomi-\nzation with a single instrumental variable and a binary outcome. Int J \nEpidemiol. 2014;43(3):922–9.\n 41. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of \ngenomewide association scans. Bioinformatics. 2010;26(17):2190–1.\n 42. Bowden J, Del Greco MF, Minelli C, Davey Smith G, Sheehan N, Thompson \nJ. A framework for the investigation of pleiotropy in two-sample sum-\nmary data Mendelian randomization. Stat Med. 2017;36(11):1783–802.\n 43. Bulik-Sullivan BK, Loh PR, Finucane HK, Ripke S, Yang J, Schizophrenia \nWorking Group of the Psychiatric Genomics C, Patterson N, Daly MJ, \nPrice AL, Neale BM. LD Score regression distinguishes confounding \nfrom polygenicity in genome-wide association studies. Nat Genet. \n2015;47(3):291–5.\n 44. Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, \nReproGen C, Psychiatric Genomics C, Genetic Consortium for Anorexia \nNervosa of the Wellcome Trust Case Control C, Dunca L, et al. An atlas \nof genetic correlations across human diseases and traits. Nat Genet. \n2015;47(11):1236–41.\n 45. Tsamantioti ES, Mahdy H. Endometriosis. In: StatPearls. edn. Treasure \nIsland (FL); 2022.\n 46. Maruyama T. A revised stem cell theory for the pathogenesis of endome-\ntriosis. J Pers Med. 2022;12(2):216.\n 47. Shigesi N, Kvaskoff M, Kirtley S, Feng Q, Fang H, Knight JC, Missmer \nSA, Rahmioglu N, Zondervan KT, Becker CM. The association between \nendometriosis and autoimmune diseases: a systematic review and meta-\nanalysis. Hum Reprod Update. 2019;25(4):486–503.\n 48. Koninckx PR, Fernandes R, Ussia A, Schindler L, Wattiez A, Al-Suwaidi \nS, Amro B, Al-Maamari B, Hakim Z, Tahlak M. Pathogenesis based \n\nPage 13 of 13\nLi et al. BMC Medicine          (2023) 21:195 \n \n•\n \nfast, convenient online submission\n •\n  \nthorough peer review by experienced researchers in your field\n• \n \nrapid publication on acceptance\n• \n \nsupport for research data, including large and complex data types\n•\n  \ngold Open Access which fosters wider collaboration and increased citations \n \nmaximum visibility for your research: over 100M website views per year •\n  At BMC, research is always in progress.\nLearn more biomedcentral.com/submissions\nReady to submit y our researc hReady to submit y our researc h  ?  Choose BMC and benefit fr om: ?  Choose BMC and benefit fr om: \ndiagnosis and treatment of endometriosis. Front Endocrinol (Lausanne). \n2021;12:745548.\n 49. Li H, Cai E, Cheng H, Ye X, Ma R, Zhu H, Chang X. FGA controls VEGFA \nsecretion to promote angiogenesis by activating the VEGFR2-FAK signal-\nling pathway. Front Endocrinol (Lausanne). 2022;13:791860.\n 50. Zhao Y, Liu YN, Li Y, Tian L, Ye X, Cui H, Chang XH. Identification of bio-\nmarkers for endometriosis using clinical proteomics. Chin Med J (Engl). \n2015;128(4):520–7.\n 51. Chen Y, Li H, Cheng HY, Rui-Qiong M, Ye X, Cui H, Hong-Lan Z, Chang XH. \nFibrinogen alpha chain is up-regulated and affects the pathogenesis of \nendometriosis. Reprod Biomed Online. 2019;39(6):893–904.\n 52. Li H, Ma RQ, Cheng HY, Ye X, Zhu HL, Chang XH. Fibrinogen alpha chain \npromotes the migration and invasion of human endometrial stromal cells \nin endometriosis through focal adhesion kinase/protein kinase B/matrix \nmetallopeptidase 2  pathway†. Biol Reprod. 2020;103(4):779–90.\n 53. Rizzello F, Ralli E, Romanelli C, Coccia ME. Severe recurrent endometrio-\nmas in a young woman with congenital von Willebrand disease. Gynecol \nEndocrinol. 2019;35(12):1040–2.\n 54. Yang B, Gu N, Shi S, Zhang C, Chen L, Ouyang J, Lin Y, Sun F, Xu H. Immu-\nnoreactivity of plasminogen activator inhibitor 1 and its correlation with \nenomyosis. Reprod Sci. 2021;28(8):2378–86.\n 55. Bryckaert M, Rosa JP , Denis CV, Lenting PJ. Of von Willebrand factor and \nplatelets. Cell Mol Life Sci. 2015;72(2):307–26.\n 56. Crawley JT, de Groot R, Xiang Y, Luken BM, Lane DA. Unraveling the scis-\nsile bond: how ADAMTS13 recognizes and cleaves von Willebrand factor. \nBlood. 2011;118(12):3212–21.\n 57. Warwicker P , Goodship JA, Goodship TH. von Willebrand factor-cleaving \nprotease in thrombotic thrombocytopenic purpura and the hemolytic-\nuremic syndrome. N Engl J Med. 1999;340(17):1368–9.\n 58. Tsai HM, Lian EC. Antibodies to von Willebrand factor-cleaving pro-\ntease in acute thrombotic thrombocytopenic purpura. N Engl J Med. \n1998;339(22):1585–94.\n 59. Reiter RA, Varadi K, Turecek PL, Jilma B, Knobl P . Changes in ADAMTS13 \n(von-Willebrand-factor-cleaving protease) activity after induced release \nof von Willebrand factor during acute systemic inflammation. Thromb \nHaemost. 2005;93(3):554–8.\n 60. Morioka C, Uemura M, Matsuyama T, Matsumoto M, Kato S, Ishikawa M, \nIshizashi H, Fujimoto M, Sawai M, Yoshida M, et al. Plasma ADAMTS13 \nactivity parallels the APACHE II score, reflecting an early prognostic indi-\ncator for patients with severe acute pancreatitis. Scand J Gastroenterol. \n2008;43(11):1387–96.\n 61. Lu K, Liu L, Xu X, Zhao F, Deng J, Tang X, Wang X, Zhao BQ, Zhang X, \nZhao Y. ADAMTS13 ameliorates inflammatory responses in experimental \nautoimmune encephalomyelitis. J Neuroinflammation. 2020;17(1):67.\n 62. Randi AM. Angiogenesis and the ADAMTS13-VWF balance. Blood. \n2017;130(1):1–2.\n 63. Xu H, Cao Y, Yang X, Cai P , Kang L, Zhu X, Luo H, Lu L, Wei L, Bai X, et al. \nADAMTS13 controls vascular remodeling by modifying VWF reactivity \nduring stroke recovery. Blood. 2017;130(1):11–22.\n 64. Shang D, Zheng XW, Niiya M, Zheng XL. Apical sorting of ADAMTS13 in \nvascular endothelial cells and Madin-Darby canine kidney cells depends \non the CUB domains and their association with lipid rafts. Blood. \n2006;108(7):2207–15.\n 65. Xiao J, Feng Y, Li X, Li W, Fan L, Liu J, Zeng X, Chen K, Chen X, Zhou X, \net al. Expression of ADAMTS13 in normal and abnormal placentae and \nits potential role in angiogenesis and placenta development. Arterioscler \nThromb Vasc Biol. 2017;37(9):1748–56.\nPublisher’s Note\nSpringer Nature remains neutral with regard to jurisdictional claims in pub-\nlished maps and institutional affiliations.","source_license":"CC0","license_restricted":false}