Further insights into the roleof the annexin A5 M2 haplotypeas recurrent pregnancy loss factor,assessing timing of miscarriageand partner risk

Frank T€uttelmann, M.D.,a Petar Ivanov, M.D., Ph.D.,b,c Charlotte Dietzel,a Anna Sofroniou,a

Tsvetomira M. Tsvyatkovska,b Regina S. Komsa-Penkova, Ph.D.,c Arseni Markoff, Ph.D.,d

Peter Wieacker, M.D.,a and Nadja Bogdanova, M.D.a

a Institute of Human Genetics, ZMBE and IZKF, University Clinic Muenster and University of Muenster, Muenster, Germany;b Clinical Institute for Reproductive Medicine and c Department of Chemistry and Biochemistry and Physics and Biophysics,Medical University, Pleven, Bulgaria; and d Institute ofMedical Biochemistry, ZMBE and IZKF, University ClinicMuenster andUniversity of Muenster, Muenster, Germany

Objective: To study the influence ofM2/ANXA5 for recurrent pregnancy loss (RPL), according to the timing of miscarriages and assessthe male partner risk.Design: Genetic association study.Setting: Academic research center.Patient(s): Female patients from two academic centers in Germany and Bulgaria with two or more unexplained miscarriages wereselected for this study. Male partners were available for a part of the German sample. Population controls were recruited from healthyindividuals of respective populations.Intervention(s): None.Main OutcomeMeasure(s): Incidence of M2 carriage and odds ratios were calculated between patient and control groups, and RPL riskwas evaluated.Result(s): TheM2 haplotype inANXA5was associated with greater overall RPL risk in German and in Bulgarian women, and a trend ofhigher prevalence was seen for male partners of German RPL patients. The highest relative risk of M2 carriage was observed in women ofboth populations with ‘‘early’’ fetal losses between the 10th and 15th gestational weeks, which was significant in the meta-analysis.

Use your smartphone

Conclusion(s): M2 carriage seems to have an RPL risk role mostly for early abortions, gesta-tional weeks 10–15. In the first phase of pregnancy this correlates with vascular remodelingto accomplish the transition from high- to low-resistance blood vessels. (Fertil Steril�2013;100:1321–5. �2013 by American Society for Reproductive Medicine.)Key Words: M2/ANXA5, ANXA5, annexin A5, recurrent pregnancy loss, risk factor

Discuss: You can discuss this article with its authors and with other ASRM members at http://fertstertforum.com/tuttelmannf-annexina5-m2-haplotype-recurrent-pregnancy-loss/

to scan this QR codeand connect to thediscussion forum forthis article now.*

* Download a free QR code scanner by searching for “QRscanner” in your smartphone’s app store or app marketplace.

2 is a haplotype in the prox- gene, which was originally found to with two or more consecutive miscar-

M imal core promoter region ofthe annexin A5 (ANXA5)

Received May 7, 2013; revised June 11, 2013; accepteF.T. has nothing to disclose. P.I. has nothing to disclo

to disclose. T.M.T. has nothing to disclose. R.S.WO/2012/025616 pending and a patent WO/20nothing to disclose. N.B. has a patentWO/2006/053725 licensed to IHG Pharmaco.

F.T. and P.I. contributed equally to this work.Reprint requests:ArseniMarkoff, Ph.D., InstituteofM

ster, Von Esmarch Strasse 56, 48149 Muenster, G

Fertility and Sterility® Vol. 100, No. 5, November 20Copyright ©2013 American Society for Reproductivehttp://dx.doi.org/10.1016/j.fertnstert.2013.06.046

VOL. 100 NO. 5 / NOVEMBER 2013

be associated with recurrent pregnancyloss (RPL) in a German patient cohort

d June 21, 2013; published online July 27, 2013.se. C.D. has nothing to disclose. A.S. has nothingK.-P. has nothing to disclose. A.M. has a patent06/053725 licensed to IHG Pharmaco. P.W. hasWO/2012/025616 pending and a patent

edical Biochemistry, ZMBEand IZKF,WWUMuen-ermany (E-mail: markoff@uni-muenster.de).

13 0015-0282/$36.00Medicine, Published by Elsevier Inc.

riages (1). The haplotype, present in15% of the general population,conferred a trend of about twofoldelevated RPL risk in this patient group.Later work confirmed the M2/ANXA5association with RPL among Italianwomen and estimated a threefoldhigher risk in patients with three ormore miscarriages (2). The same studycharacterized M2 as miscarriage riskfactor for pregnancies %15 weeks of

1321

ORIGINAL ARTICLE: EARLY PREGNANCY

gestation. Further, this work showed an M2 association withthrombophilia related to placental obstetric complicationssuch as gestational hypertension and preeclampsia (PE). Laterstudies confirmed a risk role for fetal growth restriction (FGR)(3) and pregnancy-related venous thromboembolism (4). Thesame relative risk but a lower population incidence of thehaplotype has been determined for the Japanese populationthat presented independent evidence for M2 as a predisposerto RPL (5).

Despite uncertainties concerning the exact pathologytriggered by M2 carriage, the most favored explanation ofits effects on early pregnancy losses points to a reducedexpression ofANXA5 in placenta. Retrospectively, lower pro-tein levels resulting from promoters carrying M2/ANXA5were documented in a gene reporter analysis (1). In addition,reduced ANXA5 expression has been previously described inthrombophilia-related placental complications of pregnancysuch as obstetric antiphospholipid syndrome (APS) (6), PE(3, 7), and FGR (3, 8) and in nonpregnant women withprevious recurrent RPL (9). In fact, M2 carriage is found tobe remarkably enriched exactly in these patient groups,including obstetric APS (10). Moreover, a murine model ofAnxA5 loss of function was characterized with fetal loss,FGR, and placental thromboses in homo- and heterozygousanimals (11). A study on ANXA5 mRNA expression in RPLplacentas demonstrated M2 allele-specific reduction of tran-script quantities, independent of parental origin (12). Indeed,a very recent pilot work showed comparable associated RPLrisk for paternal M2 carriers as well (13).

ANXA5 is most abundantly enriched at the syncytiotro-phoblast (SCT) apical surface covering the placental villi(14) that is well in line with its extensively studied anticoag-ulant properties, which are important for the maintenance ofhemodynamic balance during pregnancy. Since reducedANXA5 expression should be largely specific for the SCT(1, 6), the development timing of the relevant pathology inthe course of pregnancy is an important question.

With the current study we sought to determine when inthe course of gestation M2/ANXA5 is mostly associatedwith RPL and to confirm the role of paternal carriage.

MATERIALS AND METHODSStudy Populations

The present genetic association study complied with theethical guidelines of the institutions involved. Approval bythe Review Board of the Westfalian Wilhelms-University ofMuenster, Germany, and by the Ethics Commission for Scien-tific Research Activity at the Medical University of Pleven,Bulgaria, was obtained upon anonymization of participatingpatients and control cohorts. Informed consent was obtainedfrom all subjects examined. The criteria of strengthening thereporting of association studies in genetics were observed asfar as applicable.

German patients (n ¼ 243) attended the Institute ofHuman Genetics, Muenster, Germany, for genetic RPL coun-seling between August 1999 and June 2011. Male partners of109 of these women agreed to participate in the study uponsigning informed consent. Bulgarian patients (n ¼ 236)

1322

were recruited between January 2006 and December 2010in the Clinical Institute for Reproductive Medicine, Pleven.All RPL patients had two or more abortions by the timethey were selected for this study. Respective patient cohortswere prescreened for any potential causes of their repetitivemiscarriages as described elsewhere (13, 15). Uterineanomalies, endocrinologic dysfunctions (polycystic ovarysyndrome according to the Rotterdam criteria (16), thyroidaldysfunctions, thyroid autoantibodies) were excluded, aswell as fetal and parental chromosomal disorders (numericalaberrations). Inherited thrombophilia (FVL, PTm) anddeficiencies in antithrombotic factors (protein C, proteinS, factor XII, antithrombin III) were ruled out.

According to fetal development at the time of pregnancyloss (17), patients were stratified into three subgroups: sub-group 1, embryonal losses, 5–10 gestational weeks (GW); sub-group 2, early fetal losses, 10–15 GW; subgroup 3, late fetallosses, >15 GW.

Previously recruited German control groups (1, 13)consisted of a population control sample drafted from thePopGen biobank at the University Clinic Schleswig-HolsteinKiel (n ¼ 533) and a cohort of fertile women (n ¼ 500) fromthe registry of the Institute of Human Genetics, Universityof Muenster. PopGen population controls were from North-west Germany, federal state of Schleswig-Holstein, and con-sisted of healthy subjects identified through officialpopulation registers (18). The sample used in this study con-sisted of about equal numbers of men and women, distributedamong the three age groups (18–30, 30–50, 50–80 years). TheBulgarian control group (n ¼ 200; 102 men and 98 women)was composed of 33 women from the Institute for Reproduc-tive Medicine, Pleven, with a history of normal pregnancies,more than one normal term deliveries of healthy, and normalweight singletons and with no gestational pathology and of167 nonrelated healthy Bulgarian individuals (102 men and65 women) from the registry of National Genetics Laboratory,Sofia. These 167 subjects were recruited from healthy spousesof individuals counseled for genetic disorders upon signingappropriate informed consent.

Genotyping and Statistical Analysis

Genotyping of DNA extracted from peripheral blood followedby amplicon sequencing was as described in reference 1.Genotypes were scored in table format and four-digit codedfor further processing.

All analyses were carried out with Stata/SE software (Sta-taCorp LP, ver. 9.1) and visualized with GraphPad Prismversion 5.00 for Windows (GraphPad Software). Departuresfrom the Hardy-Weinberg equilibrium were assessed usingan MCMC implementation of an exact test, which is part ofthe Genepop package (http://genepop.curtin.edu.au/) (19).Meta-analysis was performed using Stata with the Manteland Haenszel method under a fixed-effects model.

RESULTSThe RPL risk of M2 carriage was assessed in two comparablysized and independently recruited cohorts of German andBulgarian origin using same selection criteria. The population

VOL. 100 NO. 5 / NOVEMBER 2013

TABLE 1

Clinical characteristics of the study groups.a

All patientsSubgroup 1: embryonal

losses, 5–10 GWSubgroup 2: early fetallosses, 10–15 GW

Subgroup 3: late fetallosses >15 GW

German sample n ¼ 243 n ¼ 131 n ¼ 70 n ¼ 42Age 32.7 � 4.9; 33 (20–45) 32.3 � 4.6; 32 (21–43) 33.9 � 4.9; 35 (20–42) 31.7 � 5.7; 32 (22–45)No. of miscarriages 2.8 � 1.0; 3 (2–9) 2.7 � 0.9b; 2 (2–8) 2.8 � 1.1; 3 (2–9) 3.2 � 1.2; 3 (2–7)

Bulgarian sample n ¼ 237 n ¼ 115 n ¼ 75 n ¼ 47Age 32.4 � 5.6c; 33 (19–45) 31.8 � 5.6; 32 (20–45) 32.4 � 5.5; 33 (20–43) 33.7 � 6.0; 35 (19–42)No. of miscarriages 2.2 � 0.7c; 2 (2–6) 2.4 � 0.7d; 2 (2–5) 2.1 � 0.5; 2 (2–6) 2.1 � 0.5; 2 (2–5)

a Data are presented as mean � SD and median (minimum–maximum).b Significantly lower (P<.05) than subgroup 3.c Comparable ages of German and Bulgarian samples (P¼ .53), number of miscarriages significantly higher in German patients (P<.0001).d Significantly higher (P<.05) than both other subgroups.

T€uttelmann. Insights into M2/ANXA5 role for RPL risk. Fertil Steril 2013.

Fertility and Sterility®

control groups were both in Hardy-Weinberg equilibrium(P>.05). The mean age of 32.4 � 4.9 years for the Germanand 32.4� 5.6 years for the Bulgarian patients was compara-ble (P¼ .53; Table 1), while German women had an overallhigher number of losses (2.8 � 1.0 vs. 2.2 � 0.7; P<.0001).

Overall RPL Risk

The M2 carriage rates were comparable between German andBulgarian controls (14.5% vs. 15.4%; P¼ .82) as were theoverall elevated carrier frequencies in RPL women of bothpopulations (23.5% vs. 19.9%; P¼ .38; Table 2). Comparisonof German patients with population controls yielded an over-all odds ratio of 1.7 (95% confidence interval, 1.2–2.5) forM2/ANXA5 carriers to belong to the RPL group. The calcu-lated odds ratio for Bulgarian patients was slightly lower at1.5 (95% confidence interval, 0.9–2.4). The meta-analysis ofboth populations confirmed a highly significant associationofM2/ANXA5 with RPL (combined odds ratio, 1.6; 95% con-fidence interval, 1.2–2.1; P¼ .003; Fig. 1A). When German

TABLE 2

Genotype frequencies of ANXA5 promoter haplotypes in control groups a

Populationcontrols RPL patients

Subgroup 1:losses, 5–

German sampleGenotypes n ¼ 533 n ¼ 243 n ¼ 1

N/N 77.9 (415) 63.4 (154) 70.2 (9N/M1 6.6 (35) 13.2 (32) 8.4 (1M1/M1 0.2 (1) — —

N/M2;M1/M2a 13.5 (72;5) 23.5 (53;4) 21.4 (2M2/M2 0.9 (5) — —

M2 carriage 15.4 (82) 23.5 (57) 21.4 (2Bulgarian sampleGenotypes n ¼ 200 n ¼ 236 n ¼ 1

N/N 77.0 (154) 73.7 (174) 74.5 (8N/M1 8.0 (16) 5.6 (13) 7 (8M1/M1 0.5 (1) 0.9 (2) 0.9 (1N/M2;M1/M2a 14.0 (28) 17.3 (37;4) 15.8 (1M2/M2 0.5 (1) 2.5 (6) 1.8 (2M2 carriage 14.5 (29) 19.9 (47) 17.5 (2

Note: Data are % (n).a Genotype M1/M2 was only observed rarely and combined with N/M2 for analyses. Numbers in p

T€uttelmann. Insights into M2/ANXA5 role for RPL risk. Fertil Steril 2013.

VOL. 100 NO. 5 / NOVEMBER 2013

patients and fertile controls were compared, the odds ratiowas significantly higher at 3.4 (95% confidence interval,2.2–5.3). Relative overall RPL risks of 1.3–1.4 for M2 carrierswere similar between the two study samples. When comparedwith population controls, German male partners carrying M2exhibited a trend towards a similar RPL risk (odds ratio, 1.5;95% confidence interval, 0.8–2.5).

Risk in Patient Subgroups in Relation to Timing ofPregnancy Loss

The clinical stratification of RPL women according to thetiming of miscarriage (about 50%, 30%, and 20% in sub-groups 1–3 of the German and Bulgarian populations) wascomparable in both samples as was the mean age betweenthe subgroups (Table 1). The M2 carriage was highest inwomen with early fetal losses (subgroup 2), 27.1% of Germanand 24.0% of Bulgarian women, and lowest in subgroup 1with embryonal losses, 21.4% and 17.5%, respectively(Table 2). When genotype distributions in each subgroup of

nd RPL patients.

embryonal10 GW

Subgroup 2: early fetallosses, 10–15 GW

Subgroup 3: late fetallosses, >15 GW

31 n ¼ 70 n ¼ 422) 52.9 (37) 59.5 (25)1) 20 (14) 16.7 (7)

— —

6;2) 27.1 (17;2) 23.8 (10)— —

8) 27.1 (19) 23.8 (10)

14 n ¼ 75 n ¼ 475) 70.7 (53) 76.6 (36)) 4 (3) 4.3 (2)) 1.3 (1) —

5;3) 20 (15) 17 (7;1)) 4 (3) 2.1 (1)0) 24.0 (18) 19.1 (9)

arentheses indicate counts of respective genotypes.

1323

FIGURE 1

Forrest plots showingodds ratios forM2carriage in specific RPLgroupscompared with population controls. An enrichment of M2 carriers isevident in total patient samples (A, odds ratio, 1.6; 95% confidenceinterval, 1.2–2.1; P¼.003) as well as in all subgroups (B, embryonallosses; C, early fetal losses; D, late fetal losses). The highest andsignificant increase was found for women with early fetal losses(C, odds ratio, 1.9; 95% confidence interval, 1.3–3.0; P¼.003).T€uttelmann. Insights into M2/ANXA5 role for RPL risk. Fertil Steril 2013.

1324

ORIGINAL ARTICLE: EARLY PREGNANCY

both populations were compared with controls, all showedtrends of RPL association with M2 carriage (Fig. 1B, 1C,1D). The comparison of German RPL women with fertilecontrols yielded significantly higher odds ratios of 3.0 (95%confidence interval, 1.8–5.2), 4.2 (95% confidence interval,2.3–7.7), and 3.5 (95% confidence interval, 1.6–7.6) forsubgroups 1–3, respectively. After meta-analysis of theGerman and Bulgarian samples, the odds ratios for subgroup1 (Fig. 1B) and subgroup 3 (Fig. 1D) were 1.4 (95% confidenceinterval, 0.9–2.0; P¼ .093) and 1.8 (95% confidence interval,1.0–3.1; P¼ .049). However, the highest and most significantodds ratio was found for subgroup 2 (Fig. 1C) at 1.9 (95%confidence interval, 1.3–3.0; P¼0.003), which represents mis-carriages in GWs 10–15, corresponding to an elevated relativerisk of 1.7 (95% confidence interval, 1.2–2.4; P¼ .004).

In four German couples where both partners were M2carriers, all fetal losses recorded were early, 9th to 13thweek of gestation.

No significant associations were found for the otherANXA5 genotypes.

DISCUSSIONNovel genetic markers for RPL have been sought extensivelyover the last years (20), but the current genetic workup of thisdistressing problem for both patients and physicians has notprogressed so far (21). Moreover, the classic inherited throm-bophilia conditions associated with pregnancy loss lead tolate events (>20th week of gestation) that are rare in compar-ison with early miscarriages (22). The haplotype M2 of theANXA5 gene emerged as a marker for proposed earlythrombophilia-related obstetric complications (2), an inde-pendent RPL risk factor in various populations (1, 2, 5) andpatient groups with recurrent miscarriage (23, 24). Thecurrent study confirms the overall RPL risk associated withM2 carriage in the German and in the Bulgarian population,which was significant in the meta-analysis (combined oddsratio, 1.6; 95% confidence interval, 1.2–2.1). The study alsoshows a similar elevated risk trend in a larger male partnersample, as was previously proposed in a pilot work (13). Sincemost previous research on the role of M2/ANXA5 as an RPLpredisposition factor has concentrated on maternal carriageof the haplotype, it is important to note the confirmation ofsimilar risk contributed by paternal carriage, as expectedfrom the nonpreferentially and equally reduced ANXA5mRNA expression in chorionic placenta carrying maternalor paternal M2 alleles (12). Current data do not support domi-nance of either parental allele for RPL, although maternalM2 carriage may be specifically associated with otherthrombophilia-related complications.

Clinically and biologically it is an important questionwhen in the course of a pregnancy-relevant pathology asso-ciated with M2/ANXA5 carriage would mainly develop inthe course of pregnancy. To elucidate the role of M2 on timingof pregnancy losses, only well-characterized patients withdetailed miscarriage history were included. After stratifyingpatients into three subgroups based on fetal development,the strongest association of the M2 haplotype was found forwomen with early fetal losses between GWs 10 and 15 in

VOL. 100 NO. 5 / NOVEMBER 2013

Fertility and Sterility®

both populations and in the meta-analysis (combined oddsratio, 1.9; 95% confidence interval, 1.3–3.0). This result is inagreement with the previously noted risk trend for repeatedmiscarriages%15 GW, which was previously assessed in Ital-ian patients (2).

ANXA5 found abundantly on SCT forms unique two-dimensional crystals upon binding to phosphatydilserinesurfaces, thus being essential for the dynamics of membranerepair in living cells (25). Repair of SCT outer membranes,naturally enriched on phosphatidylserine, should progres-sively activate upon enhanced mechanical stimuli, concomi-tant with the dramatic increase of uteroplacental andintervillous blood flow after the 10th week of gestation (26),when the growing fetus and placenta are exposed to amore normoxic environment (27). Therefore, the finding ofa significant M2/ANXA5 association with fetal losses in10th–15th GW and lower odds ratios for earlier and laterlosses correlates well with the biological hypothesis.

Conclusion

M2/ANXA5 appears as an RPL risk factor in male and femalecarriers with most distinguishable effects between the 10thand 15th week of gestation, which relates to a time in thecourse of pregnancy when vascular remodeling is most activeto accomplish the transition from high- to low-resistanceblood vessels (26).

Treatment of M2 carriers in RPL couples is currently lack-ing, but this study provides relevant pointers to be consideredfor future potential therapies in large patient screens.

Acknowledgments: The authors thank the women and theirpartners who participated in this study and Ursula Antkowiakfor the technical assistance. The continuous support of theIZKF Muenster is gratefully acknowledged.

REFERENCES1. Bogdanova N, Horst J, Chlystun M, Croucher PJ, Nebel A, Bohring A, et al. A

common haplotype of the annexin A5 (ANXA5) gene promoter is associatedwith recurrent pregnancy loss. Hum Mol Genet 2007;16:573–8.

2. Tiscia G, Colaizzo D, Chinni E, Pisanelli D, Sciannam�e N, Favuzzi G, et al.Haplotype M2 in the annexin A5 (ANXA5) gene and the occurrence ofobstetric complications. Thromb Haemost 2009;102:309–13.

3. Chinni E, Tiscia GL, Colaizzo D, Vergura P, Margaglione M, Grandone E.Annexin V expression in human placenta is influenced by the carriershipof the common haplotype M2. Fertil Steril 2009;91:940–2.

4. Grandone E, TisciaG,ColaizzoD,Chinni E, Pisanelli D, BafunnoV, et al. Role oftheM2haplotypewithin the annexinA5gene in the occurrenceof pregnancy-related venous thromboembolism. Am J Obstet Gynecol 2010;203:461.e1–5.

5. Miyamura H, Nishizawa H, Ota S, Suzuki M, Inagaki A, Egusa H, et al.Polymorphisms in the annexin A5 gene promoter in Japanese womenwith recurrent pregnancy loss. Mol Hum Reprod 2011;17:447–52.

6. Rand JH, Wu XX, Guller S, Gil J, Guha A, Scher J, et al. Reduction of annexin-V (placental anticoagulant protein-I) on placental villi of women withantiphospholipid antibodies and recurrent spontaneous abortion. Am J Ob-stet Gynecol 1994;171:1566–72.

7. Shu F, Sugimura M, Kanayama N, Kobayashi H, Kobayashi T, Terao T.Immunohistochemical study of annexin V expression in placentae of pre-eclampsia. Gynecol Obstet Invest 2000;49:17–23.

VOL. 100 NO. 5 / NOVEMBER 2013

8. Sifakis S, Soufla G, Koukoura O, Soulitzis N, Koutroulakis D, Maiz N, et al.Decreased annexin A5 mRNA placental expression in pregnancies compli-cated by fetal growth restriction. Thromb Res 2010;125:326–31.

9. Rand JH, Arslan AA, Wu XX, Wein R, Mulholland J, Shah M, et al. Reductionof circulating annexin A5 levels and resistance to annexin A5 anticoagulantactivity in women with recurrent spontaneous pregnancy losses. Am JObstet Gynecol 2006;194:182–8.

10. Bogdanova N, Baleva M, Kremensky I, Markoff A. The annexin A5 protectiveshield model revisited: inherited carriage of the M2/ANXA5 haplotype inplacenta as a predisposing factor for the development of obstetric antiphos-pholipid antibodies. Lupus 2012;21:796–8.

11. Ueki H, Mizushina T, Laoharatchatathanin T, Terashima R, Nishimura Y,Rieanrakwong D, et al. Loss of maternal annexin A5 increases the likelihoodof placental platelet thrombosis and foetal loss. Sci Rep 2012;2:827.

12. Markoff A, Gerdes S, Feldner S, Bogdanova N, Gerke V, Grandone E.Reduced allele specific annexin A5 mRNA levels in placentas carrying theM2/ANXA5 allele. Placenta 2010;31:937–40.

13. Rogenhofer N, Engels L, Bogdanova N, T€uttelmann F, Markoff A, Thaler CJ.Paternal and maternal carriage of the annexin A5 M2 haplotype are equalrisk factors for recurrent pregnancy loss: a pilot study. Fertil Steril 2012;98:383–8.

14. Krikun G, Lockwood CJ, Wu XX, Zhou XD, Guller S, Calandri C, et al.The expression of the placental anticoagulant protein, annexin V, by villoustrophoblasts: immunolocalization and in vitro regulation. Placenta 1994;15:601–12.

15. Li TC, Makris M, Tomsu M, Tuckerman E, Laird S. Recurrent miscarriage:aetiology, management and prognosis. Hum Reprod Update 2002;8:463–81.

16. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group.Revised 2003 consensus on diagnostic criteria and long-term health risksrelated to polycystic ovary syndrome. Fertil Steril 2004;81:19–25.

17. Heuser C, Dalton J, Macpherson C, Branch DW, Porter TF, Silver RM.Idiopathic recurrent pregnancy loss recurs at similar gestational ages. Am JObstet Gynecol 2010;2034:343.e1–5.

18. Krawczak M, Nikolaus S, von Eberstein H, Croucher PJP, El Mokhtari NE,Schreiber S. PopGen: population-based recruitment of patients and controlsfor the analysis of complex genotype-phenotype relationships. J CommunGenet 2006;9:55–61.

19. Rousset F. genepop’007: a complete re-implementation of the genepopsoftware for Windows and Linux. Mol Ecol Resour 2008;8:103–6.

20. Rull K, Nagirnaja L, Laan M. Genetics of recurrent miscarriage: challenges,current knowledge, future directions. Front Genet 2012;3:34.

21. Branch DW, Gibson M, Silver RM. Clinical practice. Recurrent miscarriage.N Engl J Med 2010;363:1740–7.

22. Roqu�e H, Paidas MJ, Funai EF, Kuczynski E, Lockwood CJ. Maternal throm-bophilias are not associated with early pregnancy loss. Thromb Haemost2004;91:290–5.

23. Rogenhofer N, Engels L, Bogdanova N, T€uttelmann F, Thaler CJ, Markoff A.Independent association of the M2/ANXA5 haplotype with recurrentpregnancy loss (RPL) in PCOS patients. Metabolism 2013. http://dx.doi.org/10.1016/j.metabol.2013.02.005.

24. Rogenhofer N, Engels L, Bogdanova N, T€uttelmann F, Markoff A, Thaler CJ.The haplotypeM2 of the ANXA5 gene is not associated with antitrophoblastantibodies. J Assist Reprod Genet 2013;30:711–6.

25. Bouter A, Gounou C, B�erat R, Tan S, Gallois B, Granier T, et al. Annexin-A5assembled into two-dimensional arrays promotes cell membrane repair.Nat Commun 2011;2:270.

26. Burton GJ, Jauniaux E, Watson AL. Maternal arterial connections tothe placental intervillous space during the first trimester of humanpregnancy: the Boyd collection revisited. Am J Obstet Gynecol 1999;181:718–24.

27. Rodesch F, Simon P, Donner C, Jauniaux E. Oxygen measurements inendometrial and trophoblastic tissues during early pregnancy. ObstetGynecol 1992;80:283–5.

1325