Thrombosis Research 131, Suppl. 1 (2013) S8–S10

Anticoagulant choices in pregnant women with mechanical heart valves:Balancing maternal and fetal risks – the difference the dose makes

Claire McLintock *National Women’s Health, Auckland City Hospital, Auckland, New Zealand

A R T I C L E I N F O A B S T R A C T

Keywords:PregnancyMechanical prosthetic heart valvesWarfarinThrombosisLow molecular weight heparin

Long-term anticoagulation is required in all patients with mechanical prosthetic heart valves to prevent com-plications with valve thrombosis and valve failure or systemic thromboembolism. The prothrombotic envi-ronment of pregnancy further increases the risks of these complications. Anticoagulant choices for pregnantwomen include oral vitamin K antagonists such as warfarin, or heparin – either unfractionated heparin (UFH)or low molecular weight heparin (LMWH). None of the options is without risk for the mother or her baby.Warfarin crosses the placenta and is associated with warfarin embryopathy and fetopathy but is very effec-tive at preventing thromboembolic complications. The dose ofwarfarinmay play a role in the risk of some, butnot all fetal complications. Heparin does not cross the placenta but is less effective at preventing thrombosisand LMWH may be more effective than UFH. The optimal dose and target anti-Xa levels for LMWH have notbeen established. Measurement of trough anti-Xa levels in addition to peak anti-Xa levels may be important.

© 2013 Elsevier Ltd. All rights reserved.

Introduction

Long-term anticoagulation is recommended in all patients withmechanical prosthetic heart valves (MPHV) to prevent valve throm-bosis with its sequelae of valve failure and systemic thromboem-bolism. The prothrombotic changes of pregnancy further enhancethe need for continued anticoagulation. However, the choice ofanticoagulant for pregnant women with MPHV remains a highlycontentious issue. Oral anticoagulants or vitamin K antagonists maybe the most effective agents at preventing valve thrombosis butreadily cross the placenta and are associated with development ofwarfarin-specific embryopathy, fetal ocular and neurological abnor-malities and also late fetal loss and stillbirth. Unfractionated heparin(UFH) and more recently low molecular weight heparin (LMWH) donot cross the placenta and have been demonstrated to be associatedwith better fetal outcomes. A number of reviews have summarisedmaternal and fetal outcomes with various anticoagulant options[1,2].

The decision of which anticoagulant to use will be driven by anumber of factors including maternal preference but also organi-zational and economic factors relating to the availability and costof the anticoagulants themselves, access to laboratory testing andtrained staff to manage decisions relating to anticoagulant dosesand complications. These factors will vary from one country to

* Corresponding author: Dr. Claire McLintock, Clinical Director, ObstetricMedicine, National Women’s Health, Auckland City Hospital, Level 9, SupportBuilding, Park Road, Grafton, Private Bag 92024, Auckland 1142, New Zealand. Fax+64 9 375 4344.E-mail address: claire.mclintock@adhb.govt.nz (C. McLintock).

0049-3848/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.

another and also among different clinical settings within countriesthemselves. The options available to women in tertiary hospitalsin countries such as the US, UK, New Zealand and Europe willbe different from those for women in countries from the Africanor Indian subcontinents. Irrespective of the clinical setting, healthpractitioners and women want to know the pros and cons of theanticoagulant options available to them to inform their choice butalso improve their ability to argue for better resources where theyfeel their options are limited.

While it is recognized that vitamin K antagonists (VKA) aremore effective than UFH or LMWH at preventing thromboemboliccomplications, they are not 100% effective. Recent studies fromtertiary hospitals [3,4] in developing countries report that valvethrombosis occurred in 6.7% of pregnancies where women withMPHV were prescribed VKA; three of 45 pregnancies in an Indiancentre [3] and four of 59 pregnancies in a South African centre[4]. Valve thrombosis resulting in maternal death [3] was reportedin two women who had INR values within the therapeutic rangeat their previous antenatal visit. Two of the four women from theSouth African centre [4] had stopped taking warfarin prior to preg-nancy suggesting that non-compliance may also have played a role.Non-compliance is thought to have been a contributing factor todevelopment of thromboembolism in three of the five women withMPHV from the Auckland cohort [5] who were assigned treatmentwith dose-adjusted twice-daily enoxaparin and low dose aspirin.

Vitamin K antagonist dose and fetal outcome

The impact of the dose of warfarin (or other vitamin K an-tagonist) on the risk of development of fetal complications is

C. McLintock / Thrombosis Research 131, Suppl. 1 (2013) S8–S10 S9

frequently included in discussions about options for anticoagulationfor women in this clinical setting [2,6]. Confounding these discus-sions is a degree of confusion relating to the terminology usedto describe the fetal complications. The focus of this review is toclarify the existing clinical data relating to VKA dose and specificfetal complications of VKA therapy.

The seminal review by Hall and co-workers [7] of maternaland fetal outcomes in 418 pregnancies where mothers had takencoumarin derivatives during pregnancy established that the specificform of warfarin embryopathy or fetal warfarin syndrome character-ized by nasal hypoplasia and/or stippled epiphyses occurred afterexposure during the first trimester, between six and nine weeksgestation, while other features such as central nervous systemabnormalities or ocular abnormalities were more likely to occur fol-lowing exposure during the second and third trimester. The clinicalfeatures of warfarin embryopathy are a phenocopy of the chon-drodysplasia punctata, a group of inherited disorders characterizedby stippled epiphyses. Hall and co-workers hypothesized that theclinical features can be explained by inhibition of vitamin-K de-pendent osteocalcins that have a critical role in calcification duringembryogenesis. Subsequent and current literature frequently failsto make a distinction between classical warfarin embryopathy andthe other fetal effects of the vitamin K antagonists that include CNSabnormalities as well as fetal loss and stillbirth, which may be best,described as warfarin or VKA “fetopathy”. This ongoing confusionleads to erroneous statements relating to the risk of developmentof these complications and hence clinical recommendations.

Vitale and co-workers were the first to suggest that there isa dose-dependent effect of warfarin on fetal complications [8].In their 1999 study of outcome in 58 pregnancies in 43 womenwith MPHV, they reported that fetal loss at <28 weeks gestationoccurred in 12% (4 of 33 pregnancies) of women taking ≤5 mgwarfarin. In contrast 75% of the 25 pregnancies in women taking>5 mg warfarin ended in fetal loss (18 “spontaneous abortions”prior to 28 weeks and one stillbirth after 28 weeks). Evidence ofclassical warfarin embryopathy was noted in two second trimesterlosses to women on >5 mg warfarin. This Italian group [9] wenton to publish a second retrospective study of outcomes in womenwith MPHV. Although not stated explicitly, the cohort of womenin this second study appears to overlap their first cohort with fetaloutcomes from 71 pregnancies in 52 consecutive patients. Differentgestations were used to define spontaneous abortion (<20 weeksgestation) and stillbirth (>20 weeks gestation) and again fetalloss occurred more often in the 33 pregnancies in women taking>5 mg warfarin (78%; stillbirth n= 5, spontaneous abortion n =21)compared to only 2 fetal losses (5.3%) at 17 and 18 weeks gestation

Table 1Summary of recommendations for anticoagulation during pregnancy in pregnant women with mechanical prosthetic heart valves.

European Society of Cardiology [11]All patients OAC are recommended during the 2nd and 3rd trimester until the 36th week

Warfarin dose <5 mg/day * Continue OAC during the first trimester

Warfarin dose <5 mg/day * Discontinuation of OACs between weeks 6 and 12 and replacement by UFH or LMWH may be considered on an individual basis

Warfarin dose >5 mg/day * Discontinuation of OAC between weeks 6 and 12 and replacement by adjusted-dose UFH or LMWH twice daily

Warfarin dose >5 mg/day * Continuation of OAC between weeks 6 and 12 may be considered

Target anti-Xa levels for LMWH 4–6 h post dose 0.8–1.2 IU/mL and for UFH aPTT ≥2× control

American College of Chest Physicians [12]Adjusted-dose twice daily LMWH throughout pregnancy (doses adjusted to achieve manufacturer’s peak anti-Xa 4 h post dose)

Adjusted-dose subcutaneous UFH twice daily throughout pregnancy (doses adjusted aiming for mid-interval aPTT at least twice controlor for anti-Xa heparin level of 0.35–0.70 units/mL)

UFH or LMWH (as above) until the 13th week with substitution by vitamin K antagonists until close to delivery when UFH or LMWH isresumed

Women at high risk of thromboembolism vitamin K antagonists with replacement of I+UFH or LMWH (as above) close to delivery

* Equivalent doses of phenprocoumon 3 mg/day and acenocoumarol 2 mg/day.OAC, oral anticoagulants; LMWH, low molecular weight heparin; UFH, unfractionated heparin.

in 38 pregnancies to women taking >5 mg warfarin. Three casesof warfarin embryopathy were reported: nasal hypoplasia in oneinfant born at term in the low-dose group (2.6%) and in the high-dose group there were two spontaneous abortions, one with nasalhypoplasia and one with cervical spine abnormalities. Therefore,while this pregnancy cohort shows much lower rates of warfarin-related fetopathy at warfarin doses ≤5 mg it does not show adose-dependent relationship for development of classical warfarinembryopathy.

A prospective cohort of 62 pregnancies from Pretoria, SouthAfrica [10], where women with MPHV were treated with warfarin,reported warfarin embryopathy in five of 41 pregnancies wherewomen took warfarin during the first trimester. Again, no dose-dependent relationship for warfarin embryopathy was evidentand it was reported in two pregnancies where women took ≤5mg warfarin and three pregnancies where women took >5 mgwarfarin. This finding is consistent with another pregnancy cohortfrom KwaZulu-Natal, South Africa [4] where warfarin embryopathywas reported in four warfarin-exposed pregnancies (7.1%), two inwomen on doses of 7.5 mg and two in women taking ≤5 mgwarfarin.

A dose-dependent rate of warfarin fetopathy was confirmedin the Pretoria cohort [10] with stillbirth rates (fetal loss after24 weeks gestation) were significantly increased in women takinghigher doses of warfarin during pregnancy: 39% (5/13) with war-farin dose ≥7.5 mg, 14% (3/21) with warfarin dose 5.1–7.5 mg and3.6% (1/28) of pregnancies with warfarin dose ≤5 mg.

In the Auckland cohort, four of 13 women who took predomi-nantly warfarin during pregnancy did not have a surviving infant.One woman required a termination of pregnancy at 17 weeks forfetal death due to intracerebral haemorrhage (warfarin dose 5 mg),there were two stillbirths in women taking 4 mg and 5 mg warfarinand one death at two months of age of an infant with warfarin em-bryopathy whose mother took 6 mg warfarin throughout pregnant[5].

An appreciation of the impact of the dose of VKA on therisk of developing warfarin embryopathy and warfarin fetopathybecomes relevant in providing counselling for women and also indevelopment of clinical practice guidelines.

The European guidelines [11] recommend only oral anticoag-ulants (OAC) in the second and third trimester (Table 1). Theirrecommendations in the first trimester are dependent on the doseof OAC, apparently based on the authors’ perception that the rate ofwarfarin embryopathy is low and also dose related. The recommen-dations do not discuss the risk of fetal loss and stillbirth with OACuse in the second and third trimester.

S10 C. McLintock / Thrombosis Research 131, Suppl. 1 (2013) S8–S10

The most recent guidelines from the American College of ChestPhysicians [12] suggest that taking some form of anticoagulantduring pregnancy is better than taking nothing. They make nostatement regarding the dose of OAC but state that the decisionshould be completely individualized. Options include adjusted dose,twice-daily LMWH or UFH throughout pregnancy or given until13 weeks gestation followed by VKA until close to delivery. Inwomen at very high risk of thromboembolism such as womenwith older style valves in the mitral position or with a historyof thromboembolism they recommend VKA throughout pregnancy,substituted by LMWH or UFH close to delivery. The American HeartAssociation Guidelines [13] are more directive in using OAC in thesecond and third trimester, discussing their relative safety. Theirrecommendations do not include comment regarding the dose ofOAC but suggest that the target INR should be 2.5–3.5.

The Italian group [14] has gone further with their approachusing low dose warfarin in pregnant women with MHPV. From acohort of young women referred for aortic valve replacement, theyselected 17 women who, in pre-operative anticoagulation tests,could maintain an INR in their target range 1.5–2.5 with warfarindoses <5 mg to have a third generation St Judes mechanical deviceplaced. Sixteen of these women went on to have pregnancies wherethey continued to take low dose warfarin to maintain a target INRof 1.5–2.5 with good maternal and fetal outcomes. As was pointedout in an accompanying editorial [15], the safety of this approachis questionable given the lack of data to support the safety ofusing such a low target INR and data from other studies that showthat adverse fetal outcomes are also described with low doses ofwarfarin.

The key driver in recommendation of OAC over LMWH foranticoagulation of women with MPHV is improved maternal safetyand a lower risk of valve thrombosis. Good outcomes are reportedin women who are compliant with twice daily therapeutic doseLMWH with dose adjustment using anti-Xa levels [1]. Treatmentfailures are described [16] but perhaps these could be minimizedif trough as well as peak anti-Xa levels are used to guide dose-adjustments. The trough anti-Xa level is likely to be more importantthan the peak anti-Xa level to indicate that women are maintainingan adequate baseline anticoagulant effect. However, measurementof both peak and trough anti-Xa levels is only advocated by twogroups [5,15]. Efforts should be directed at determining the safetyand efficacy of this approach, as women with MPHV will continueto seek alternatives to VKA for anticoagulation during pregnancy.

Economic factors will be a major determinant of the anticoag-ulant options offered to pregnant women with MPHV. It shouldcome as no surprise that the articles reporting extensive experiencewith warfarin come mainly from countries with lower per capitahealthcare spending and more limited economic resources [3,4].Warfarin and other VKA are cheap and INR testing is inexpensiveand generally readily available. In contrast, the cost of LMWH ismuch higher and access to laboratory testing for anti-Xa levels islimited and also expensive placing this therapeutic option beyondthe means of many low-income countries.

Summary

The debate about the optimal choice of anticoagulant therapy inpregnant women with MPHV seems set to continue. Ultimately, theclinical setting of where pregnancy care is provided will determinethe choices available because of economic realities. To be able tooffer safe and effective treatment with therapeutic LMWH requiresa healthcare budget that has the resources not only to buy the drugitself but that can provide laboratory facilities for regular testingof anti-Xa levels and clinical facilities of a dedicated, multidisci-plinary team to provide close clinical follow-up and to ensure

maternal compliance with the treatment regimen. Where this is notavailable, warfarin or other VKA will be a more appropriate optionfor women. Whatever the clinical setting, women with MPHV havethe right be educated about the risk that pregnancy poses for themand understand the nature of the risks to them and their baby. Weshould also ensure that effective contraception is available to helpwomen avoid unplanned pregnancies and make an active choice tobecome pregnant at the safest possible time.

Conflict of interest statement

Claire McLintock has received honoraria from Sanofi for pre-sentation at clinical meetings and financial support to attendconferences.

References

[1] McLintock C. Anticoagulant therapy in pregnant women with mechanicalprosthetic heart valves: no easy option. Thromb Res 2011;127(Suppl 3:)S56–60.

[2] Castellano JM, Narayan RL, Vaishnava P, Fuster V. Anticoagulation duringpregnancy in patients with a prosthetic heart valve. Nat Rev Cardiol 2012;9(7):415–24.

[3] Suri V, Keepanasseril A, Aggarwal N, Chopra S, Bagga R, Sikka P, et al.Mechanical valve prosthesis and anticoagulation regimens in pregnancy: atertiary centre experience. Eur J Obstet Gynecol Reprod Biol 2011;159(2):320–3.

[4] Mazibuko B, Ramnarain H, Moodley J. An audit of pregnant women withprosthetic heart valves at a tertiary hospital in South Africa: a five-yearexperience. Cardiovasc J Afr 2012;23(4):216–21.

[5] McLintock C, McCowan LM, North RA. Maternal complications and pregnancyoutcome in women with mechanical prosthetic heart valves treated withenoxaparin. BJOG 2009;116(12):1585–92.

[6] Pieper PG, Balci A, Van Dijk AP. Pregnancy in women with prosthetic heartvalves. Neth Heart J 2008;16(12):406–11.

[7] Hall JG, Pauli RM, Wilson KM. Maternal and fetal sequelae of anticoagulationduring pregnancy. Am J Med 1980;68(1):122–40.

[8] Vitale N, De Feo M, De Santo LS, Pollice A, Tedesco N, Cotrufo M. Dose-dependent fetal complications of warfarin in pregnant women with mechani-cal heart valves. JACC 1999;33(6):1637–41.

[9] Cotrufo M, De Feo M, De Santo LS, Romano G, Della Corte A, Renzulli A, et al.Risk of warfarin during pregnancy with mechanical valve prostheses. ObstetGynecol 2002;99(1):35–40.

[10] Soma-Pillay P, Nene Z, Mathivha TM, Macdonald AP. The effect of warfarindosage on maternal and fetal outcomes in pregnant women with prostheticheart valves. Obstet Med 2011;4:24–7.

[11] Regitz-Zagrosek V, Blomstrom Lundqvist C, Borghi C, Cifkova R, Ferreira R,Foidart JM, et al. ESC Guidelines on the management of cardiovascular dis-eases during pregnancy: the Task Force on the Management of CardiovascularDiseases during Pregnancy of the European Society of Cardiology (ESC). EurHeart J 2011;32(24):3147–97.

[12] Bates SM, Greer IA, Middeldorp S, Veenstra DL, Prabulos AM, Vandvik PO.VTE, thrombophilia, antithrombotic therapy, and pregnancy: AntithromboticTherapy and Prevention of Thrombosis, 9th ed: American College of ChestPhysicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141(2Suppl):e691S–736S.

[13] Bonow RO, Carabello BA, Chatterjee K, de Leon AC, Jr., Faxon DP, Freed MD, etal. 2008 focused update incorporated into the ACC/AHA 2006 guidelines for themanagement of patients with valvular heart disease: a report of the AmericanCollege of Cardiology/American Heart Association Task Force on Practice Guide-lines (Writing Committee to revise the 1998 guidelines for the management ofpatients with valvular heart disease). Endorsed by the Society of CardiovascularAnesthesiologists, Society for Cardiovascular Angiography and Interventions,and Society of Thoracic Surgeons. JACC 2008;52(13):e1–142.

[14] De Santo LS, Romano G, Della Corte A, D’Oria V, Nappi G, Giordano S,et al. Mechanical aortic valve replacement in young women planning onpregnancy: maternal and fetal outcomes under low oral anticoagulation,a pilot observational study on a comprehensive pre-operative counselingprotocol. JACC 2012;59(12):1110–5.

[15] Elkayam U, Goland S. The search for a safe and effective anticoagulationregimen in pregnant women with mechanical prosthetic heart valves. JACC2012;59(12):1116–8.

[16] Yinon Y, Siu SC, Warshafsky C, Maxwell C, McLeod A, Colman JM, et al. Useof low molecular weight heparin in pregnant women with mechanical heartvalves. Am J Cardiol 2009;104(9):1259–63.