Case-based discussion on the implications of exogenous estrogens in hemostasis and thrombosis: the hematologist’s view

In the childbearing years, women with congenital disorders of hemostasis may experience greater bleeding complications of the menstrual cycle or delivery, whereas those with congenital thrombophilia may experience greater thrombotic complications of childbirth or of exogenous estrogen use during assisted reproductive technologies. Despite improvements in clinical assessment and disease classification, as well as innovations in therapeutic agents and procedures, the optimal management of women with disorders of hemostasis and thrombosis remains elusive. This constitutes a public health problem. In women with bleeding disorders (WBD), despite refined bleeding assessment tools and diagnostic assays, the burden of bleeding remains high. WBD still experience a delay in diagnosis and a lack of effective therapies to reduce heavy menstrual bleeding (HMB) and prevent postpartum hemorrhage (PPH). Up to 80% of women with von Willebrand disease (VWD), the most common congenital bleeding disorder, have HMB, and ∼50% have depleted iron stores and iron-deficiency anemia, with physical and cognitive dysfunction, anxiety, depression, and poor quality of life. At delivery, despite treatment, women with VWD may experience PPH, resulting in longer hospitalization and greater transfusion requirements. The optimal therapeutic approach, including antifibrinolytic therapy and/or weight-vs-volume–based von Willebrand factor (VWF) dosing, remains unknown. In women with congenital thrombophilia, thrombotic risk may be associated with exogenous hormone therapy, such as during assisted reproductive procedures. However, this risk has been reduced by improvements in oocyte transfer techniques and in vitro fertilization (IVF). This article reviews the persistent burden and management controversies of HMB and delivery in WBD and the role for anticoagulation in oocyte retrieval in women with thrombophilia. Although clinical data are limited and randomized trials are rare, the most current data and treatment guidelines for each of these scenarios are provided, identifying gaps and future approaches to management.

An 18-year-old-woman with severe menorrhagia since starting menses at age 14 develops severe abdominal pain with a 2 g/dL decrease in hemoglobin. Laboratory tests reveal hemoglobin, 8.6 g/dL; normal values for PT, PTT, and platelet count, VWF:RCo, 32 IU/dL; VWF:Ag, 38 IU/dL; VIII:C, 49 IU/dL; PFA closure time with epinephrine, 186 seconds; and PFA closure time with adenosine diphosphate, 178 seconds; with normal multimers. The ISTH BAT bleeding score is 5, based on menorrhagia, dental extraction bleeding, bruising, and epistaxis. The genotype is a missense mutation in the VWF coding sequence. The family history includes heavy menses in her mother, maternal aunt, and maternal grandmother. Her abdomen is diffusely tender to palpation with guarding and rebound tenderness. Ultrasound reveals a 4.4 × 4.0 × 3.6-cm right ovarian mass with cystic and solid components.

HMB is a significant public health problem, affecting up to 15% of all women but up to 80% of WBD, including those with VWD, the most common congenital bleeding disorder. VWD is characterized by mucosal bleeding due to defective or deficient VWF causing defective platelet plug formation and primary hemostasis.¹  Menorrhagia, defined as >80 mL of blood loss per month, a level at which progressive iron loss and iron deficiency occurs, is measured by the validated Pictorial Blood Assessment Score, with >80 mL of blood loss equivalent to a score > 100.² 

Bleeding symptoms, including HMB, vary across the range of VWF levels,³  even in those with the same mutations,⁴  but debate continues about whether those with “low VWF” (ie, VWF 30-50 IU/dL) have “true” disease. At least two thirds of those with “low VWF” have mutations, as in the case patient, whose missense mutation is common in type 1 VWD. Among those patients with VWF < 30 IU/dL, or “true VWD,” nearly all have causal mutations.⁴  Yet, there is little difference in bleeding scores and HMB between these groups, suggesting that the separation is artificial. For example, the presence of HMB in both groups (“low VWF” and “true VWD”) was associated with the highest bleeding scores.⁵  Bleeding score tools determine the type and severity of bleeding symptoms and have been validated to predict the presence of a bleeding disorder. Although highly dependent on age and hemostatic challenges, a bleeding score > 5 in a woman has moderately high sensitivity and specificity for the presence of a bleeding disorder.^1,6  It is recognized that identification of VWD in the childbearing years may be difficult because hormonal therapy and pregnancy increase VWF levels and may mask a diagnosis. Further, VWF levels are influenced by stress, inflammation, ABO blood type, age, and age-related morbidities, including hypertension, diabetes, thyroid disease, cancer, body mass index,^7,8  and proteins contributing to VWF variability (eg, CLEC4M, SNARE, STXBP5) and sialylation.^9,10  The assay for VWF function, VWF:RCo, may itself be problematic because it is based on ristocetin-induced platelet aggregation, which has a high coefficient of variation with potentially falsely high or low results. By using assays not dependent on ristocetin, such as the VWF:GPIbM assay, the coefficient of variation is reduced, and the assay is unaffected by VWF A1, 2M, or 2B mutations^11,12  or “benign” sequence variations (eg, D1472H).¹³  Although no formal guidance exists regarding the use of the VWF:GPIbM assay, it appears helpful in patients with undiagnosed bleeding symptoms and a family bleeding history¹⁴  but not in those with low VWF,^13,14  as in our patient.

The burden of HMB is underscored by monthly blood loss (Table 1), which depletes iron stores, and, in ≥60% of patients, results in severe iron deficiency (ferritin < 15 ng/mL).¹⁵  Iron deficiency is associated with impaired cognitive functioning, mood, physical functioning, and quality of life.^15,16  Although symptoms may be reversible with iron repletion, oral iron therapy may be poorly tolerated, with nausea, vomiting, constipation, and poor compliance. The appropriate dose and frequency of iron replacement are not established, because even small increases in serum iron may increase hepcidin, which limits iron absorption.¹⁷  A recent randomized trial found that iron absorption was higher with alternate-day dosing¹⁸  and may also be better tolerated. Thus, iron studies should be obtained for our patient, even if her hemoglobin is normal. If iron deficiency is confirmed, I would initiate an alternate-day regimen. If she is unresponsive or intolerant, I would initiate IV iron with the third-generation ferric carboxymaltose (INJECTAFER), which is simple and well tolerated, in 2 doses 7 days apart.^17,19  Carboxymaltose (INJECTAFER) is also more effective in fewer doses and results in a greater and more rapid increase in hemoglobin in women with iron-deficiency anemia in pregnancy compared with iron sucrose, the previous standard of care in pregnancy.²⁰ 

In addition to the burden of iron deficiency, quality of life in WBD is poor and associated with a delay in diagnosis and poor access to treatment.²¹  Although a family history of bleeding can shorten the time to diagnosis by 6 years, and care at a hemophilia treatment center (HTC) can increase the likelihood of receiving treatment twofold,²¹  many patients are not diagnosed until years later when they experience surgical bleeding. HTCs are federally supported centers that provide comprehensive care for all patients with congenital bleeding disorders. Their supportive health care staff can enable otherwise reluctant women to discuss their social, psychosocial, and stress experiences.^22,23  Our patient should receive comprehensive care at an HTC, including assessment by quality-of-life and psychosocial tools, counseling, and follow-up with psychosocial and hemostasis professionals, to optimize her quality of life.

With what agents should our patient’s HMB be managed? This is a challenging question because treatment for menorrhagia that is effective and well tolerated is a major unmet health need in WBD (Table 2).²⁴  The currently recommended nonhormonal agent, tranexamic acid (Lysteda), an antifibrinolytic agent dosed at 1300 mg 3 times daily during the first 5 days of the cycle, reduces menstrual loss by 50% but is limited by nasal congestion, headache, and nausea.²⁵  Desmopressin (1-deamino-8-d-arginine vasopressin [DDAVP]), a synthetic vasopressin analog that stimulates VWF release from the vascular endothelium at the site of injury, is limited by hyponatremia, tachyphylaxis, and local infusion reaction, and the intranasal form, Stimate, although convenient, is less potent.²⁴  Combined oral contraceptives (COCs) stimulate VW synthesis but are avoided in 35% of patients because of headaches and hypertension.²⁴  The levonorgestrel intrauterine system, Mirena, stimulates an antifibrinolytic effect by releasing hormone into the endometrial cavity but is limited by weight gain and depression in 20% of patients, and it may be expelled as a result of large clot burden.²⁶  VWF concentrate replaces defective or deficient VWF protein but is limited by invasiveness and cost.²⁴  A survey of hemostasis physicians regarding management of menstrual bleeding in VWD revealed that the most common first-line therapy was COCs in 70%, followed by tranexamic acid (Lysteda) in 30%, and DDAVP in 20%. VWF concentrate was a third-line therapy used only when other agents failed.²⁴  Few data exist on recombinant VWF (rVWF; Vonvendi), although it has the full range of multimers and a longer half-life than plasma-derived VWF (pdVWF; Humate-P). An ongoing randomized superiority cross-over trial, the VWD Minimize Menorrhagia Trial (NCT020606045) is comparing rVWF with tranexamic acid, the currently recommended nonhormonal agent, to reduce menorrhagia in women with VWD. For our patient, the decision regarding treatment of HMB involves discussion of available therapeutics, the pros and cons of each, what has and has not worked, and her past experience and personal preference. I also encourage consultation with our gynecologic colleagues with whom I collaborate in the care of our patients, with follow-up by phone or in person, until there is confirmation of success in reducing HMB and improving quality of life.

The acute symptoms of abdominal pain and acute blood loss are suggestive of hemorrhagic rupture of a corpus luteum cyst. Although women with VWD are at no greater risk for developing an ovarian cyst, they are at greater risk for bleeding than women without bleeding disorders. Acute management would include VWF concentrate immediately, starting at 80 IU/kg, followed by 50 IU/kg 6 to 8 hours later, with supportive IV fluids, red blood cell transfusion for blood loss, and pain management, as well as imaging (magnetic resonance imaging) after factor has been given. Urgent discussion regarding the need for surgical intervention, as well as temporizing with factor, if possible, can reduce major morbidity. Once the diagnosis is confirmed, I would monitor blood loss by hemoglobin and continue VWF concentrate for up to 3 to 5 days or longer for ongoing blood loss. Once our patient is stable and the bleeding and pain have stopped, longer-term prevention with hormonal therapy should be discussed. DDAVP may be another alternative, although its release of VWF from endothelial stores is limited, because VWF stores are depleted after 3 days. Finally, iron replacement should be initiated after resolution of the acute event.

The patient returns 5 years later to reestablish care and is now seeking advice on becoming pregnant. She is worried about bleeding if she stops her combined hormonal contraceptive to become pregnant. She is also worried about bleeding risk during and after delivery.

Primary PPH, defined as >500 mL blood loss within 24 hours of delivery or >1000 mL following cesarean delivery (Table 3), is the leading cause of maternal death worldwide, and it accounts for one third of maternal deaths in the United States.²⁷  Uterine atony is the major cause of PPH,²⁷  and uterotonic agents are the first-line treatment that is effective in reducing PPH.²⁷ 

Among women with VWD, the odds of PPH are 1.5-fold greater than in controls.²⁸  In a Pennsylvania database of >4.5 million inpatient hospital discharges, PPH incidence was 5.5% in VWD, similar to national rates.²⁷  Of women with VWD who developed PPH, 27% were anemic, twofold greater than in those without PPH, and were more likely to require a red blood cell transfusion and longer hospitalization.²⁸  These data suggest that anemia may be a predictor of PPH and that hemoglobin should be monitored in women with VWD and other bleeding disorders, in addition to traditional VWF and other factor levels. Although there were no deaths in women in that study,²⁸  severe anemia (<7 g/dL) has been associated with a nearly twofold odds ratio (OR) for maternal death worldwide.²⁹  Yet, based on an analysis of 11 published trials by the US Preventive Services Task Force,³⁰  routine iron supplementation does not improve these maternal outcomes. Moreover, no data are available on testing or outcomes in WBD.

When uterotonics fail, treatment of women with persistent PPH should incorporate hemostatic agents, including tranexamic acid, an inhibitor of plasmin-mediated fibrinogen and fibrin breakdown, along with transfusion support (blood, plasma products).²⁷  In the WOMAN trial, a large randomized double-blind placebo-controlled trial of >15 000 women from low- to middle-income countries, tranexamic acid was shown to reduce PPH and mortality when given within 3 hours of PPH, with no thrombotic complications.³¹  The causes of PPH included uterine atony (63.7%), placenta previa or placenta accreta (9.4%), surgical trauma or tears (18.4%), other (7.3%), and unknown (1.2%). Whether the findings of this trial are valid for all women with PPH with hypovolemic shock is not known. Controversy exists regarding why there was benefit with tranexamic acid in patients undergoing cesarean section^30,32  but not in those undergoing vaginal delivery when oxytocin was given.²⁹  A recent meta-analysis of 25 trials showed that when tranexamic acid was given preventively before cesarean section and vaginal delivery, it reduced the rate of PPH significantly, with no increase in thrombosis.^33,34  No randomized trial has evaluated tranexamic acid to prevent PPH in WBD. Clearly, future trials are needed to determine the role of tranexamic acid in WBD at delivery.

Hemostatic support with clotting factor is the recommended approach to PPH prevention in women with VWD (Table 4). Based on expert guidance,^27,35-43  a VWF level > 50 IU/dL is recommended before epidural analgesia. Although some advocate a dose of VWF 50 IU/kg at delivery, there is no specific guidance regarding the optimal dose or duration to prevent PPH, except to avoid DDAVP, which may be associated with hyponatremia with fluid losses and fluid replacement at delivery.³⁵  Yet, at least one third of women with VWD develop PPH.³⁸  Even when VWF concentrate is given before and after delivery in women with VWD, VWF levels are lower and blood loss is greater in women with VWD than in controls.³⁹  Although it is known that third trimester VWF levels are inversely related to PPH,³⁸  and lower prepregnancy VWF and higher prepregnancy and third trimester body weight are associated with PPH,⁴⁰  there are no established predictors of PPH.

Why there is increased blood loss at delivery in women with VWD, despite VWF concentrate treatment, remains a conundrum. It may be related to the concomitant presence of a bleeding disorder, in addition to the well-described hormonal-associated decrease in VWF and factor VIII levels at delivery,^38,41  as well as the reduced platelet count and platelet function that occur during pregnancy.⁴²  Several questions remain regarding the optimal management of the woman with VWD at delivery. What are the optimal VWF level and duration of treatment? Are third trimester levels sufficient to determine who should receive factor replacement? However, it is clear that 50 IU/kg is insufficient to prevent PPH. It may be that delivery should be managed as a surgical procedure, and the recommended VWF concentrate dose for major surgery (80 IU/kg) should be used at delivery.⁴⁰  Whether the latter dose is sufficient to prevent PPH is unknown, but 1 small observational study found no difference in PPH rates with the use of 50 IU/kg vs 80 IU/kg at delivery in women with VWD.⁴³ 

The currently recommended VWF concentrate dose of 50 IU/kg at delivery in women with VWD does not take into account the 1.5-fold pregnancy-associated increase in blood volume.⁴⁴  Yet volume-based dosing, rather than weight-based factor dosing, is routinely used in children and obese adults with hemophilia.⁴⁵  Moreover, during pregnancy, the physiologic increases in cardiac output, extracellular fluid, renal blood flow, and glomerular filtration rate accelerate drug clearance.^46,47  In fact, pregnancy-associated increased drug clearance has been the basis for blood volume–based dosing for a number of drugs during pregnancy (eg, antihypertensive agents).⁴⁸  Despite this, blood volume–based VWF dosing has not been recommended at delivery.^40,45  There is a potential risk for thrombosis with a 1.5-fold higher than recommended VWF dose at delivery, although a recent review of >570 patients receiving VWF at doses up to 200 IU/kg found that thrombosis risk was low (0.4%).⁴⁰ 

How should we manage our patient at delivery? A multidisciplinary team, including an obstetrician, hematologist, anesthesiologist, and the patient, should be set up to develop a treatment plan at delivery and, when possible, to encourage enrollment in clinical studies. Because VWF levels peak 4 hours postpartum and then decline rapidly, with return to baseline within 3 weeks,³⁹  clinical vigilance is needed for ≥3 weeks postpartum. I would obtain a late third trimester VWF level, and, given a low baseline VWF and a strong bleeding history, as in our patient, we would give a dose of VWF 80 IU/kg immediately before epidural anesthesia or delivery, whichever is first. I would continue VWF at 50 IU/kg on days 1 and 2 postpartum. If the patient should develop excessive postpartum bleeding despite this approach, I would consider adding tranexamic acid, additional supportive fluid resuscitation, or other interventions per our obstetric colleagues. Because bleeding risk may persist for 4 to 6 weeks postpartum, she would require close monitoring and active follow-up at 4 to 6 weeks; once stable, iron supplementation should be given if needed.

Two years later, the patient returns for her annual appointment at the local HTC and brings her 28-year-old cousin who is heterozygous for factor V Leiden (FVL) mutation, is planning to undergo IVF in several months, and seeks a second opinion regarding the risk of thrombosis with the procedure.

Approximately 170 000 assisted reproduction technology (ART) procedures are performed annually. For ART procedures, which include ovulation-induction therapy and IVF, gonadotropin or gonadotropin-releasing hormones have been used to promote pharmacologic stimulation of ovarian follicles. In the past, gonadotropins were used as part of ART, leading to multiple follicle development in the ovary. The latter caused supraphysiologic levels of estrogens, leading to increased levels of coagulation factors, prothrombin fragment 1+2, and d-dimers, and reduced natural anticoagulants, antithrombin and protein S, and an overall shortened clotting time.⁴⁹  Some women with an excessive response to exogenous gonadotropins developed ovarian hyperstimulation syndrome (OHSS). OHSS may lead to more profound changes in hemostasis, including increases in fibrinogen, factors II, V, VII, VIII, and IX, and activation of the fibrinolytic system⁴⁹ ; it may be fatal in some patients.

Prior to the use of human chorionic gonadotropin (HCG) to stimulate ovulation, thrombosis rarely developed as a complication of ART; however, following its introduction, HCG was shown in clinical trials to be the major factor for the enhanced thrombosis risk associated with ART (Table 5). The subsequent use of goanadotropin-releasing hormone agonists, instead of exogenous HCG, to trigger ovulation greatly eliminated OHSS and reduced VTE.⁴⁹  Compared with the normal population, the absolute risk for VTE in women undergoing ART with no gonadotropin is 10-fold greater, whereas in those undergoing gonadotropin-induced ovulation who develop OHSS, the absolute risk is 100-fold greater.^49-51  In the large REITE Registry, the OR for developing VTE in women undergoing ART without HCG (OR, 4.13; 95% confidence interval [CI], 1.4-12.4) was comparable to the VTE risk in contraceptive users (OR, 2.96; 95% CI, 1.95-4.5) and in the puerperium (OR, 1.96; 95% CI, 1.16-3.3).⁵² 

Among women with thrombophilia, it has been well established that exogenous hormones constitute the major risk for thrombosis. The risk of thrombosis among those undergoing non-HCG ART is not different from the general population.⁵³  Thus, the risk for VTE appears to be related to the hormonal ovarian stimulation approach, and the approach to VTE prevention should be based on the approach to prevention of VTE during pregnancy per the American Society of Hematology 2018 Guidelines for Management of Venous Thromboembolism: Venous Thromboembolism in the Context of Pregnancy.⁵³ 

Low molecular weight heparin is the anticoagulant of choice in pregnancy⁵³  and is also the agent of choice to treat OHSS. The duration of anticoagulation is up to 6 weeks postpartum or to complete ≥3 months of treatment following a VTE and for up to 3 months after symptom resolution of OHSS. In women with less severe ovarian hyperstimulation and risk factors that warrant anticoagulation, anticoagulation should follow the current American Society of Hematology guidelines in the context of pregnancy.⁵³  Specifically, in women with unprovoked VTE, hormone-associated VTE, or higher-risk thrombophilia, including homozygous FVL mutation, homozygous prothrombin gene mutation, or a strong family history and lower risk thrombophilia, low molecular weight heparin anticoagulation prophylaxis is recommended. In women with a history of prior provoked VTE in which the risk factor is resolved or with a low-risk thrombophilia (ie, heterozygous FVL or heterozygous prothrombin gene mutation), no prophylaxis is recommended for ART.

Management of patients undergoing ART is based on the current guidelines for management of venous thrombosis in pregnancy.⁵³  Because our case patient is heterozygous for the FVL mutation, I would recommend against anticoagulation, even with a family history of venous thromboembolism (VTE) or a personal history of VTE associated with a transient risk unrelated to pregnancy or estrogen use. Therefore, I would recommend clinical vigilance only, unless other risk factors exist (eg, obesity, immobilization, or prolonged hospitalization).⁵³  If our patient had a previous VTE associated with pregnancy or estrogen use, anticoagulation would be considered during hormone use with ART.⁵³  However, anticoagulation should be held during the 12-hour period before oocyte retrieval. Anticoagulation can be resumed the same day, as long as the procedure was not complicated by bleeding. If OHSS should occur, the risk of VTE lasts for up to several weeks after OHSS is resolved, so anticoagulation would be continued for up to 3 months after OHSS resolution. If OHSS does not occur, no anticoagulation beyond the procedure is required.

Finally, if our patient becomes pregnant in the future, current pregnancy guidelines suggest clinical vigilance only. The recommendation in patients with heterozygous FVL mutation, as in the cousin, is against anticoagulation during or after pregnancy.⁵³  This applies even when there is a family history of VTE or a personal history of VTE associated with a transient risk factor unrelated to pregnancy or estrogen use. If our patient desires pregnancy but had had a previous VTE with pregnancy or estrogen or ART use, prophylactic anticoagulation is recommended during pregnancy, as well as for ≥6 weeks following delivery.⁵³ 

Margaret V. Ragni, Division of Hematology/Oncology, Department of Medicine, University of Pittsburgh, Hemophilia Center of Western Pennsylvania, 3636 Blvd of the Allies, Pittsburgh, PA 15213-4306; e-mail: ragni@pitt.edu.