Portal de Periódicos da CAPES

Communication from the ISTH SSC Subcommittee on Women's Health Issues in Thrombosis and Haemostasis: A Survey on Anticoagulation for Mechanical Heart Valves in Pregnancy

2021; Elsevier BV; Volume: 19; Issue: 3 Linguagem: Inglês

10.1111/jth.15213

1538-7933

Isabelle Malhamé, Maha Othman, Patricia Casais, Rohan D’Souza, Rachel M. Wald, Candice K. Silversides, Mathew Sermer, Nadine Shehata,

Blood Coagulation and Thrombosis Mechanisms

Patients with mechanical heart valves (MHV)s require life-long anticoagulation to prevent thromboembolic complications (TECs). Women with MHVs in pregnancy experience an increased risk of severe maternal morbidity and mortality from both thrombotic and bleeding complications. A large international registry of pregnancies with MHVs revealed that valve thrombosis and maternal mortality occurred in 4.7% and 1.4% of women, respectively, and hemorrhage complicated 23% of pregnancies.1 Of importance, only 58% of women with MHVs were free from serious adverse events during pregnancy compared with 79% of women with bioprosthetic valves and 78% of women with cardiac disease and no prosthetic valves.1 Vitamin K antagonists (VKAs), such as warfarin, are the standard anticoagulation modality for nonpregnant patients with MHVs. However, VKAs readily traverse the placenta and are teratogenic. Indeed, warfarin-associated embryopathy may occur with first-trimester VKA exposure consisting of developmental anomalies affecting bones and cartilage.2, 3 Moreover, administration of VKAs in the second and third trimesters can lead to a warfarin-associated fetopathy, characterized by central nervous system anomalies potentially from microhemorrhages in brain tissue.2, 3 Alternative options to VKAs are low molecular weight heparins (LMWHs). However, LMWHs have been associated with higher frequency of maternal valvular thrombosis and mortality, and their use for anticoagulation of MHVs remains off-label.4 The main options for anticoagulation regimens have been described in pregnancy with the aim of reducing maternal fetal risks include (a) VKAs throughout pregnancy, (b) LMWH throughout pregnancy, (c) LMWH in the first trimester and VKAs in the second and third trimester (sequential treatment), and (d) unfractionated heparin (UFH) throughout pregnancy.5, 6 Outcomes associated with these strategies have predominantly been described by observational studies, which are inherently at increased risk of bias.4 In addition, society guidelines have mostly issued recommendations regarding the choice of anticoagulation modality,5 and practices regarding adjunctive antiplatelet therapy in pregnancy are not well described. The optimal anticoagulation management strategy for pregnant women with MHVs thus has not yet been determined. Assessing current anticoagulation practice is a necessary step for planning of future prospective studies. Thus, we sought determine practice patterns for anticoagulation therapy for pregnant women with MHVs by conducting an international survey. We conducted a cross-sectional study using online questionnaires to be administered and filled out by respondents. The survey was sent to hematologists, cardiologists, obstetric medicine physicians, and obstetricians internationally in collaboration with the Subcommittee on Women's Health Issues in Thrombosis and Haemostasis of the ISTH. Approvals were obtained from the institutional review boards of Mount Sinai Hospital in Toronto, Ontario (IRB#19-0045-E), and the McGill University Health Centre in Montreal, Quebec (IRB#2020-5819), for this survey. An electronic questionnaire was developed according to previously published techniques.7 The final survey instrument included two case scenarios: a patient with a mitral MHV and a patient with an aortic MHV, as these are the two most common MHVs. Partially close-ended questions addressed (a) type of anticoagulation (ie, LMWH, VKA, UFH, dosing of LMWH) on the basis of a previous reviews of anticoagulation regimens for MHV in pregnancy4; (b) low-dose ASA use because low-dose ASA may improve outcomes, yet has been associated with an increased bleeding risk1; (c) peripartum management (including regional anesthesia and mode of delivery) because minimal data on peripartum management are available; and (d) postpartum anticoagulation regimens because increased bleeding risk with early transition to VKAs postpartum has been described.8 To gauge whether additional clinical factors could potentially affect decisions for anticoagulation selection during pregnancy, a partially close-ended question listed several characteristics including an option to select other characteristics. Because a dose-dependent teratogenic effect of VKAs has been suggested,9 we aimed to assess whether anticoagulation choices depended on the equivalent of a warfarin dose of ≤5 mg orally daily.5 To permit for an analysis of potential physician variables that may be associated with decisions regarding anticoagulation, the following physician characteristics were gathered: years of practice, physician specialty, type of practice (primary, secondary, or tertiary health care center), and continent of practice. The survey instrument was pilot tested by investigators, resident physicians, and members of the ISTH for validity and clarity. Modifications to the survey were made based on this feedback. The complete survey instrument can be found in the Supplementary Appendix. Potential participants included all members of the ISTH, the North American Society of Obstetric Medicine, and the International Society of Obstetric Medicine. These societies were selected because they were believed to represent worldwide physicians most likely to be primarily managing anticoagulation for pregnant women with MHVs. In addition, researchers who published studies on the use anticoagulation for MHVs in pregnancy previously identified by the investigative team in a systematic review of the use of anticoagulation in pregnant patients with MHVs were also invited to participate. All potential respondents were asked to participate by email invitation sent out by their respective societies. The following screening question assessed their eligibility to participate: "Do you manage anticoagulants in pregnant patients with mechanical heart valve replacements?" The survey was administered using the REDCap software platform,10 hosted by ISTH. We adopted several strategies to try to maximize the response rates by developing a respondent-friendly questionnaire and making several contacts with each subject.11 The survey was open to respondents from September 24, 2019, to April 20, 2020. The numbers and proportion of physicians selecting response categories was described for each case scenario. We hypothesized that responses could potentially vary according to physicians' training experience, and location of practice. Thus, we assessed the association between physician characteristics (ie, specialty, years of practice, and continent of practice) and the respondents' choice of anticoagulation modality and dose adjustment practices for LMWH using chi-squared test or Fisher's exact test, as appropriate. The proportion missing for these analyses were less than 5%; thus, complete case analysis was used. All statistical analyses were performed using IBM SPSS version 26, 2019. Significant associations were defined if P values were less than .05. A total of 121 potential participants answered the initial survey screening question for eligibility, and 84 (69%) were involved in anticoagulation management of these patients. Respondents were from the specialties of hematology/thrombosis (n = 49, 61%), obstetric medicine (14, 17.5%), cardiology (n = 8, 10%), and maternal fetal medicine (n = 3, 3.8%). They almost exclusively practiced in tertiary referral hospitals (n = 74, 89%), and were in mid-late career, having practiced for >11 years after completion of training (n = 55, 65%). Participants were clinicians from Europe (n = 35, 42%), North America (n = 31, 37%), Asia (n = 8, 10%), South America (n = 5, 6%), Africa (n = 3, 4%), and Australia and New Zealand (n = 2, 2%). Most centers (n = 48, 62%) saw fewer than 5 pregnant patients with MHVs per year, and therefore most respondents (n = 61, 78%) reported managing fewer than 5 cases per year. The most commonly used anticoagulation regimens for mitral MHVs in pregnancy were sequential therapy with therapeutic LMWH at 12 to 14 weeks followed by a VKA through the remaining pregnancy (n = 28, 33%), therapeutic LMWH throughout pregnancy (n = 27, 32%), and VKAs throughout pregnancy if the dose was equivalent to ≤5 mg of warfarin daily (n = 9, 11%) (Figure 1A). No respondents chose UFH as the anticoagulant of choice for women with mitral MHVs. The most commonly used anticoagulation regimens for pregnant women with aortic MHVs were therapeutic LMWH throughout pregnancy (n = 47, 56%), therapeutic LMWH until 12 to 14 weeks followed by a VKA throughout remaining pregnancy (n = 21, 25%), and VKAs throughout pregnancy if the dose was equivalent to ≤5 mg of warfarin daily (n = 7, 8%) (Figure 1B). Aspirin was selected by 31 (37%) participants, predominantly from the first trimester onward (n = 25/31, 80%), among women with either mitral or aortic MHVs receiving LMWH in 14/31 (47%), and women with either mitral or aortic MHVs regardless of anticoagulation modality in 9/31 (30%). Factors considered when choosing the anticoagulation modality are described in Table 1. Four characteristics were selected by the majority (>60%) of respondents, the type of valve, the location of the replacement valve, previous embolic event, and patient preference. Cost was selected by 17% of respondents. Anticoagulation selection for mitral MHVs varied according to specialty, but not according to years or continent of practice (Table S1 of the Supplementary Appendix). Anticoagulation selection for aortic MHVs varied according to specialty and continent of practice (Table S2 of the Supplementary Appendix). Respondents n (%) Seventy-seven (92%) respondents reported LMWH use in pregnancy. Although most respondents decided on the initial LMWH dose based on pregnancy weight (n = 52, 68%), some used prepregnancy weight (n = 22, 29%). Adjustments to LMWH dosage were made according to pregnancy weight, peak anti-Xa (n = 22, 29%), peak and trough anti-Xa (n = 20, 26%), peak anti-Xa only (n = 20, 26%), or pregnancy weight only (n = 11, 14%). Hematology/thrombosis specialists predominantly used peak anti-Xa (n = 15, 31%). Dosage adjustments did not vary according to a respondent's characteristics (Table S3 of the Supplementary Appendix). Fifty (60%) respondents reported VKA use in pregnancy. Most reported aiming for an international normalized ratio (INR) of 2 to 3 for pregnant women with aortic MHVs and 2.5 to 3.5 for mitral MHVs (n = 33, 66%). Some respondents, however, aimed for INR of 2.5 to 3.5 (n = 5, 10%) and INR of 2 to 3, (n = 8, 16%) regardless of valve position. Epidural anesthesia was the preferred mode of peripartum analgesia (n = 35, 44%). The preferred mode of delivery was based on obstetrical indications for most respondents (n = 44, 55%), whereas vaginal and a cesarean delivery was specified for 29 (36%) and 7 (9%) individuals, respectively. VKAs were resumed 24 to 72 hours following delivery in 38 (48%) respondents, and within 24 hours (n = 15, 19%) or 7 days after delivery (n = 15, 19%) for others. Aspirin was either discontinued 5 to 10 days before delivery and it was not restarted postpartum (n = 7, 23%), discontinued 5 to 10 days before delivery and restarted following delivery until at least 6 to 12 weeks postpartum (n = 4, 13%), or discontinued at delivery (9, 30%). Certain respondents, however, did not discontinue aspirin before delivery and continued it until 6 to 12 weeks postpartum (n = 7, 23%). We surveyed physicians' strategies of anticoagulation for women with MHVs during pregnancy. We found that sequential therapy and LMWH throughout pregnancy were the anticoagulation strategies most commonly chosen for women with mitral and aortic MHVs, respectively. Although weight in pregnancy was used to determine the starting dose of LMWH, anti-Xa peak and trough levels were used to guide further dosing adjustments in most instances. Aspirin was often prescribed concomitantly with LMWH and discontinued 5 to 10 days before delivery. Delivery mode was guided by obstetrical indications. We observed considerable variations in practice in the choice of anticoagulation modality for pregnant patients with mitral MHVs. The choice of anticoagulant was made predominantly according to patient-specific factors, whereas published guidelines were only used by a minority of participants. This may reflect the paucity of clinical data to guide the optimal method of anticoagulation for these women.5, 6 Sequential therapy and LMWH alone were reported to be more frequently used than UFH and VKAs among respondents. A systematic review of 2468 pregnancies in 1874 women evaluating maternal, fetal, and neonatal outcomes among women with MHVs described 44 studies of pregnant patients with MHV of which 68% were in the mitral position.4 Data on the use of sequential therapy and LMWH were available for 530 pregnancies (20 studies) and 132 pregnancies (10 studies) respectively; these strategies were associated with 5.8% and 8.7% maternal TECs, respectively.4 Importantly, anticoagulation-related fetal/neonatal adverse events associated with LMWH throughout pregnancy did not occur in any of 103 infants included, and the live birth rate was 92%.4 As such, the strategy of LMWH throughout pregnancy is currently being used and may be an option from a fetal/neonatal standpoint, so long as maternal safety can be optimized. A summary of associated risks for each anticoagulation strategy and current recommendations has been provided in Tables S4 and S5 of the Supplementary Appendix, respectively. Most respondents reported using anti-Xa levels for dose adjustment of LMWH. The monitoring of trough and peak anti-Xa levels has been recommended by the European Society of Cardiology (level IC) (Table S5).5 However, whether anti-Xa level monitoring improves clinical outcomes among women with MHVs remains unclear. Indeed, adverse outcomes have been described among in women within the recommended therapeutic target anti-Xa level peak range of 1.0 to 1.2 IU/mL.12 Moreover, how to adjust dosing as a function of anti-Xa levels is uncertain because achieving target peak anti-Xa levels did not always ensure maintenance of minimum trough levels among pregnant participants with MHVs.13 In a national registry from the United Kingdom, monitoring practices using anti-Xa levels were highly variable.14 The frequency at which anti-Xa levels were measured ranged from none to weekly and the target range for peak and trough levels were heterogeneous, highlighting a lack of standardized approach.14 Observational studies have not demonstrated any clear clinical benefit in monitoring of anti-Xa levels for women using LMWH for the treatment of venous thromboembolism (VTE) in pregnancy.15-18 In addition, considering the lack of calibration curves for the anti-Xa assay in pregnancy and the costs associated with anti-Xa monitoring, the American Society of Hematology has recommended against the use of anti-Xa levels in the management VTE in pregnancy.19 Given differences in thrombotic risks between VTEs and MHVs, recommendations regarding anticoagulation for one condition may not be generalizable to the other. Thus, the use of anti-Xa levels specifically for MHVs in pregnancy needs to be clarified; in particular, whether peak and/or trough levels should be used, optimal anti-Xa levels, and best time intervals for anti-Xa monitoring require further research."5 Low-dose aspirin, usually defined as ranging from 75 to 100 mg daily, was selected as an adjunct antithrombotic therapy for women taking LMWH. In the European Society of Cardiology's Registry of Pregnancy and Cardiac Disease, none of the 13 patients with aspirin in addition to anticoagulation had TECs, whereas 2.5% of patients without aspirin developed mitral valve thrombosis.1 In the nonpregnant setting, a Cochrane review highlighted that antiplatelet agents given in addition to anticoagulation decreased the risk of systemic embolism or death among patients with MHV, although the risk of bleeding events was slightly increased.20 Further guidance on the addition of aspirin as part of the antithrombotic regimen is currently needed. A minority of respondents used UFH in contemporary practice, in contrast to the use of UFH reported in systematic reviews.4 In addition, VKAs were selected less frequently than LMWH-based strategies, contrasting with findings from prior studies most commonly describing the use of VKAs throughout pregnancy.4 Therefore, we report a potential shift in clinical practice, whereby LMWHs appears to be increasingly selected either in conjunction with measurement of anti-Xa levels, at least in high-resource settings. This may reflect that UFH is the preferred heparin in middle- and low-resource settings, where LMWH use and monitoring is often cost prohibitive. As such, among respondents, the selection of UFH in high-income settings may have declined. This international survey captured the anticoagulation regimens practiced by a variety of specialists, from diverse geographical settings. The data highlight important knowledge gaps in regard to the optimal dose adjustment and drug monitoring strategies for LMWH, which will serve to plan for future studies. This study has some limitations. First, the total number and baseline characteristics of all potential respondents is not known. Furthermore, results reflected the practice of the respondents only; practices of nonrespondents could not be assessed. A major limitation is that cardiologists were underrepresented because the survey was not sent to cardiology societies. Furthermore, most respondents had access to anti-Xa level monitoring. Thus, results may not be generalizable to areas of low- and middle-income. Lastly, case scenarios may not reflect actual physician practice.21, 22 To minimize this limitation, we developed scenarios that closely simulated real patients. In addition, we pilot tested the survey for face and content validity. In summary, in an international survey of physicians, mostly from tertiary referral centers in high-resource settings, we described an important shift in clinical practice regarding the use and monitoring of LMWH and the addition of low-dose aspirin in those using LMWH among respondents. Most physicians reported using peak anti-Xa levels to adjust LMWH dose during pregnancy, with an increasing number reporting the additional use of trough anti-Xa levels. In light of these changes in clinical practice, and given the paucity of high-quality evidence to inform clinical decision making, international multicenter trials are required to determine the optimal anticoagulant strategy. We thank Shannon Brooks for managing the dissemination of the survey throughout the conduct of the study. We thank all practitioners who responded to the survey and members of the subcommittee of Women's Health Issues in Thrombosis and Haemostasis of the ISTH for their insightful comments. Dr. Casais has received travel support from Sanofi. Dr. D'Souza has received a Canadian Institutes of Health Research Early Career Award, as well as speaking honoraria and grant funding from Ferring Inc. for topics unrelated to the current publication. Dr. Shehata has received honoraria and an educational grant from Sanofi. Drs. Malhamé, Othman, Wald, Silversides, and Sermer have no conflicts of interest to declare. Isabelle Malhamé contributed to the design, analysis, interpretation of data, and writing of the initial draft; Maha Othman, Patricia Casais, Rohan D'Souza, Rachel M. Wald, Candice K. Silversides, and Mathew Sermer contributed to the design, interpretation of data, and critical revisions of the manuscript; Nadine Shehata supervised the design, analysis, interpretation of data, and critically revised the manuscript. All authors approved of the final version of the manuscript. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.