Abstract
In patients with mechanical heart valve protheses, warfarin is usually recommended because of its exceptional anticoagulation effects. However, warfarin can cross the placenta, leading to teratogenicity and even catastrophic hemorrhage in the fetus. The present article describes a case of warfarin-associated fetal intracranial hemorrhage. The patient was a woman in her early 30s. At the age of 11 years, she had undergone aortic valve replacement (mechanical) for aortic regurgitation. Since then, she had been taking oral warfarin. During her pregnancy, her prothrombin time–international normalized ratio was maintained between 1.5 and 2.5. At 35 weeks of gestation, fetal ultrasonography revealed an intracranial mass in the left hemisphere. An emergency cesarean section was performed because fetal intracranial hemorrhage was suspected. A male infant was delivered with a 1- 5-, and 10-minute Apgar score of 1, 5, and 7, respectively. Cranial computed tomography revealed multiple hemorrhage sites with newly emerged bleeding spots. In patients with mechanical heart valve protheses, obstetricians face the dilemma of individual-patient differences and the difficulty of intensive monitoring of the coagulation parameters in the fetus. Tailor-made anticoagulation therapy and a more intensive ultrasonic monitoring strategy, even that involving regular magnetic resonance imaging, are necessary in these patients.
Keywords
Introduction
In patients with mechanical heart valve protheses, the anticoagulation strategy should be individualized and thoroughly assessed after balancing the pros and cons. Pregnancy undoubtedly places these women at high risk, even when asymptomatic.1–3 Bearing in mind the risk of life-threatening thromboembolism in these women, warfarin therapy is commonly suggested even during pregnancy because of its excellent anticoagulation effects.4–6 Warfarin is the drug of choice because of its high efficacy as a vitamin K antagonist. In one study of pregnant women with mechanical valves, the incidence of thrombosis was 2.7%, 5.8%, 8.7%, and 11.2% in the warfarin group, warfarin after heparin group, low-molecular-weight heparin group, and unfractionated heparin group, respectively. 7 However, the potential adverse effects of warfarin are non-negligible, with a fetal loss rate of 12% and teratogenesis rate of 0.45% to 29%.2,8 Although rare, warfarin may also cause fetal bleeding, particularly intracranial bleeding, which can have life-threatening outcomes. We herein present a case of warfarin-associated fetal intracranial hemorrhage.
Case report
The reporting of this study conforms to the CARE guidelines. 9 The patient’s details have been deidentified. The patient and her husband (authorized client) provided written consent for all treatments of the patient and newborn. The present study was performed according to the Declaration of Helsinki and approved by the Ethics Committee of West China Second University Hospital (Approval No. 2023Shen077).
The patient was a woman in her early 30 s (gravida 2, para 0). At the age of 11 years, she had undergone aortic valve replacement (mechanical) for aortic regurgitation. Since then, she had been taking oral warfarin at a daily dose of 3.75 mg. Two years previously, she had experienced spontaneous abortion at 10 gestational weeks.
At the current presentation, she was found to be pregnant at 5 gestational weeks as confirmed by ultrasound and underwent regular antenatal care in our hospital. The oral warfarin was discontinued, and subcutaneous unfractionated heparin at 5000 IU twice a day was prescribed until 16 gestational weeks. Oral warfarin was then resumed at a dosage of 3.75 mg/day. The prothrombin time–international normalized ratio (PT/INR) was maintained between 1.5 and 2.5.
At 35 weeks of gestation, routine antepartum fetal ultrasonography revealed an intracranial mass in the left hemisphere (7.5 ×3.2 × 4.5 cm) (Figure 1). The left ventricle was compressed, and the brain midline was shifted to the right. The patient was immediately admitted to the obstetric department.
Ultrasonography revealed a fetal intracranial mass in the left hemisphere.
A cardiologist, neonatologist, and anesthesiologist were soon contacted for a multidisciplinary consultation. The patient’s PT/INR was 2.18 at admission. The cardiologist suggested injection of 20 mg vitamin K before cesarean section (CS). Cardiotocography revealed a suspected sinusoid wave, indicating that the fetal conditions were unfavorable. Because prothrombin complex concentrate (PCC) was not available in the obstetric department, fresh frozen plasma (FFP) was used during CS. A male infant weighing 3250 g was delivered with a 1-, 5-, and 10-minute Apgar score of 1, 5, and 7, respectively. The umbilical arterial blood pH was 7.35. The infant was transferred to the neonatal intensive care unit immediately after intubation and positive-pressure ventilation.
The neonatal hemoglobin level was 57 g/L, PT/INR was 10.01, and activated partial thromboplastin time was 103.6 s. Invasive mechanical ventilation was applied. Blood products including red blood cells, PCC, and FFP were administered, and the coagulation parameters were corrected to within the normal range. However, cranial computed tomography revealed multiple hemorrhage sites with newly emerged bleeding spots in the frontal, parietal, occipital, and temporal lobes of the left hemisphere accompanied by small bleeding sites in the right hemisphere. The midline was shifted (Figure 2). The infant was still unconscious after 2 days of treatments. On the third day after delivery, the newborn was discharged at the parents’ request.
Computed tomography showed that the hemisphere midline had shifted and that a large intracranial mass was present in the cerebrum.
One day after CS, warfarin was prescribed as suggested by the cardiologist. The initial dosage was the same as that during pregnancy (3.75 mg/day). The INR was 1.37, and no low-molecular-weight heparin (LMWH) was added. On the second day after CS, the INR decreased to 1.00. LMWH was administered at 4000 IU every 12 hours. This bridging plan with LMWH continued for 12 days, during which the daily dose of warfarin increased from 3.75 to 5.00 mg. On the 12th day after CS, the INR was in the target range and the LMWH was stopped. The patient was discharged on the 13th day after CS and was followed up 6 weeks after delivery. She recovered well with no development of abnormal vaginal bleeding and with good uterine involution. Her INR was also controlled in the therapeutic range (Figure 3).
Coagulation function tests during pregnancy and after cesarean section.
Discussion
After artificial valve replacement surgery in women, pregnancy-related risks are huge challenges for both the patients and obstetricians. The high mortality rate and low live birth rate in these patients are inevitable considering the increased cardiopulmonary burden and the risks associated with anticoagulation therapy.1,2 A meta-analysis in 2017 summarized maternal and fetal outcomes in women with mechanical heart valves from 46 published studies. The maternal mortality rate ranged from 0.9% to 3.4% according to the difference in the anticoagulation treatment regimen. The live birth rate ranged from 64.5% to 92.0%, and the prevalence of thrombotic events ranged from 2.7% to 11.2%. 7 Studies in recent years have also demonstrated unfavorable pregnancy outcomes. Erba et al. 10 summarized 307 pregnancies in 253 women who had mechanical heart valve prostheses and were taking warfarin in Sudan. The maternal death rate was 4.9%, and only 59.5% of the women had live births. Sillesen et al. 11 collected data from the Denmark national birth registry and reported 155 pregnancies in 79 women. There were two maternal fatalities and four thromboembolic episodes. All the thrombosis events were associated with heparin application. Two intrauterine fetal deaths occurred in the late trimester, and one of them was caused by intracerebral hemorrhage.
Our case involved catastrophic warfarin-associated intracranial bleeding in the fetus. To identify similar cases, we performed a systematic search of three English-language databases (PubMed, Embase, and Web of Science) and four Chinese databases (Chinese Biological Medical Literature database, Chinese National Knowledge Infrastructure database, Chinese Medicine Premier (Wanfang database), and Chinese Journals Full-text Database (VIP)) without language restriction from 1970 to the present. The key words used were “warfarin,” “fetus,” “pregnancy,” and “intracranial hemorrhage” or “intracranial bleeding.” Eleven articles involving 12 cases were identified,12–22 and only 5 live births occurred. Two of the infants had normal nervous system development at 1 year of age (Tables 1 and 2). Two of the newborns developed nervous system defects including polycystic encephalomalacia and cerebral palsy. Among the identified articles, the warfarin dose in most women was less than 5 mg, and the INR was balanced. One study reported a sudden increase in the INR followed by stillbirth. 21 The detection of fetal intracranial hemorrhage occurred from 23 to 35 weeks. Only Fujiwara et al. 19 described a case at 35 gestational weeks, similar to our case. Fetal intracranial bleeding was found before 32 weeks in many cases.
Warfarin-associated fetal intracranial hemorrhage in previously published cases.
CS, cesarean section; GA, gestational age; INR, international normalized ratio; LMWH, low-molecular-weight heparin; MRI, magnetic resonance imaging; NA, not available; PT, prothrombin time; TEE, transesophageal echocardiography; UFH, unfractionated heparin.
Neonatal conditions and prognosis in previously published cases of warfarin-associated fetal intracranial hemorrhage.
APTT, activated partial thromboplastin time; GA, gestational age; Hb, hemoglobin; HC, head circumference; INR, international normalized ratio; MRI, magnetic resonance Imaging; NA, not available; PT, prothrombin time.
As in the previous cases, regular antenatal care and good compliance could not prevent this tragedy in our case. Despite close monitoring of the INR, devastating hemorrhage still occurred. This indicates that a balanced INR and relatively low dose of warfarin do not guarantee fetal safety. Oral administration of warfarin at a dosage of <5 mg/day has been described as relatively safe 23 and associated with a lower frequency of bleeding complications. Nevertheless, we should bear in mind that the efficacy of warfarin is dose-dependent. Maintaining a balanced INR is still a relatively safer choice for decreasing the risk of bleeding in both the mother and fetus. Notably, an increased dosage of warfarin is sometimes needed to maintain the target INR during pregnancy. During this process, intensive monitoring of the INR is needed to determine the appropriate dosage of warfarin. In our literature review, we found that higher dosages of warfarin might lead to fetal hemorrhage at earlier gestational weeks. Additionally, an uncontrolled high INR was associated with sudden uterine demise or severe fetal bleeding events.
Importantly, the emergency situations of pregnant patients with mechanical heart valves involved both maternal- and fetal-associated factors. The balance of bleeding events and thrombosis must be carefully considered for pregnant women. We reviewed the literature regarding the management of mechanical heart valves during pregnancy and found that the major causes of maternal death included valve thrombosis, intracranial bleeding, and disseminated intravascular coagulation secondary to severe postpartum hemorrhage (PPH). In the study by Erba et al., 10 307 patients were on warfarin therapy. Five patients died of valve thrombosis, and one patient died of intracranial bleeding. The thrombosis rate was 7.8%, and the major bleeding rate was 7.2%. Most of the bleeding events occurred at delivery or abortion, with only one patient developing hemorrhagic stroke. 10 The authors concluded that thrombosis leading to death was more frequent than hemorrhage (33.3% vs. 6.7%, respectively). Ranjan et al. 24 evaluated 182 patients with mechanical valves who were undergoing sequential treatment (heparin followed by warfarin), and the thrombosis rate was 11.54% while the PPH rate was 6.04%. One patient died of disseminated intravascular hemorrhage caused by PPH after CS. 24 In the Registry of Pregnancy and Cardiac Disease (ROPAC) study, 212 pregnant women with mechanical heart valves were analyzed, and most of them were undergoing sequential treatment. Ten patients developed valve thrombosis, and two of them died. Three patients underwent valve replacement surgery in the first trimester, and two of them subsequently delivered two live newborns. Two patients underwent valve replacement surgery in the second trimester, and both newborns died. One patient underwent CS in the third trimester, and valve replacement surgery was performed 3 months after CS. 25 The incidence of bleeding events in their study was as high as 23.1%, and the bleeding events mainly occurred around the time of delivery. However, the authors concluded that hemorrhage did not induce other adverse events such as heart failure, maternal mortality, or fetal demise. Data from the United Kingdom Obstetric Surveillance System (UKOSS) collection system showed that the maternal mortality rate in pregnant women with mechanical prosthetic heart valves was 8.6% (5 of 58 patients). Two women died of valve thrombosis, and the other three died of valve dysfunction, cardiac failure, and cerebrovascular accident, respectively. 26 Severe PPH occurred in 29% of the patients. From these studies, it seems that the bleeding events were mainly associated with delivery of the fetus. Although the rate of bleeding events was not low, most cases of hemorrhage could be successfully managed by transfusion and surgery. However, the consequences of valve thrombosis can be devastating. In pregnant patients with mechanical heart valves, valve thrombosis can lead to a certain rate of maternal mortality. Retrospective studies and case reports revealed similar results.11,26–30
Management of such patients requires multidisciplinary teamwork involving a cardiologist, obstetrician, neonatologist, and anesthesiologist. To minimize the possibility of adverse pregnancy outcomes, regular visits to the cardiologist should be emphasized. If emergency CS is deemed necessary, a treatment plan for the correction of coagulation function should be provided by the cardiologist. In our case, suggestions were obtained from the cardiologist. Because of the worsening fetal condition as reflected by the fetal monitor, however, CS was performed shortly after vitamin K injection. Notably, we did not use PCC, the administration of which is the most efficacious strategy to overcome warfarin-induced coagulopathy, because it was not available in the obstetric department. Instead, we used FFP during CS. After this treatment combined with careful placement of homeostatic sutures, our patient did not develop PPH.
As a potent antagonist of vitamin K-dependent proteins, warfarin is the first-line drug for women after artificial valve replacement, even during pregnancy. Some pregnant women, including the patient in our case, are on continuous warfarin therapy. The most obvious disadvantage of this drug is that it crosses the placenta, which can lead to pregnancy loss, fetotoxicity, teratogenicity, and stillbirth.31,32 Warfarin embryopathy has historically focused on fetal abnormalities caused by exposure in the first trimester. With the increasing awareness of severe PPH, however, recent guidelines recommended application of LMWH in the first trimester and after 36 gestational weeks. 5 Except for its teratogenic risk in the first trimester, warfarin is a comparatively safe anticoagulation choice because of its lower thrombosis rate. In our literature review, Kamei et al. 12 reported a case of thrombosis formation during heparin administration. After increasing the dosage of heparin and restarting warfarin for 2 weeks, the thrombus disappeared at the 24th gestational week. This is direct evidence of the superiority of warfarin for anticoagulation.
In our case, the fetal anticoagulation function was not monitored until delivery. However, previous studies have shown that fetal adverse events might be more sensitive to the warfarin dosage.23,32 Through analysis by cordocentesis, Ville et al. 20 reported lower levels of procoagulant factors II, VII, IX, and X in the fetus than in the mother, but the plasma concentration of warfarin was similar. Their findings indicated that the fetus were more susceptible to the effects of warfarin, which might be caused by the poor development of the glucuronide conjugation enzymatic pathway in the fetal liver and the subsequently limited renal elimination of warfarin metabolites. 33 Notably, the elimination half-life of warfarin is very long (mean of 40 hours), which is determined by its plasma protein-binding characteristic. 32 This long half-life increases the risk of hemorrhage. Major dilemmas faced by clinicians are that the maternal INR does not represent the actual anticoagulant effect in the fetus and that fetal coagulation function cannot be monitored.
The prognosis is usually very poor for a fetus with intracranial hemorrhage, as in our case. In our review, only a few newborns survived without cerebral functional disorders. The PT/INR in these live newborns were relatively lower than in our case. The warfarin dosage and maternal INR were comparable. Although the coagulation function of the newborn was corrected in our case, multiple sites of intracranial hemorrhage inevitably led to a poor prognosis.
Conclusion
Anticoagulation therapy with warfarin during pregnancy has bidirectional targets; both the mother and the fetus are influenced. After artificial valve replacement surgery, attention should be paid to the risk of not only maternal thrombosis but also fetal cerebral hemorrhage. Warfarin-associated fetal hemorrhage is usually a fatal event. Obstetricians face the challenges of individual-patient differences and the difficulty of intensive monitoring of the coagulation parameters in the fetus. Management of pregnant patients with mechanical heart valves requires multidisciplinary teamwork. Tailor-made anticoagulation therapy that is adjusted in a timely manner should be implemented after consultation with the cardiologist. A more intensive ultrasonic monitoring strategy for the fetus, even with regular magnetic resonance imaging examinations during pregnancy, should be considered in these patients.
Footnotes
Author contributions
Dan Shan: Project development, data collection, and manuscript writing.
Yurou Ji: Provision of figures and data collection.
Tao Li: Supervision and manuscript writing.
Yayi Hu: Supervision and manuscript writing.
Availability of data and materials
The datasets used in the current study are available from the corresponding author on reasonable request.
Declaration of conflicting interests
The authors declare that they have no competing interests.
Funding
This work was supported by the Sichuan Science and Technology Program (Nos. 2022YFS0043 and 2023YFS0217).
