Mechanical heart valves and pregnancy: Issues surrounding anticoagulation. Experience from two obstetric cardiac centres

Abstract

Background

Pregnant women with mechanical heart valves are at significant risk of obstetric/cardiac complications. This study compares the anticoagulation management in two obstetric cardiac centres.

Methods

Retrospective case-note review from Chelsea and Westminster/Royal Brompton Hospitals (CR) and Erasmus Medical Centre (EMC). Main outcome measure was mechanical heart valve thrombosis.

Results

Nineteen pregnancies from CR and 25 pregnancies from EMC were included. Most women were on low-molecular-weight heparin (LMWH) throughout pregnancy at CR, whereas at EMC most had LMWH in the first trimester and vitamin K antagonists in subsequent trimesters. Peak anti-factor Xa were performed monthly at CR, levels 0.39–1.51 IU/mL (mean 0.82 IU/mL). Anticoagulation management peri-partum was inconsistent. Delivery was mainly by caesarean section at CR (74%) and vaginal delivery at EMC (64%). No maternal deaths and only one mechanical heart valve thrombosis at CR. Two mechanical heart valve thromboses and one maternal death at EMC.

Conclusion

Peri-partum anticoagulation strategies, anticoagulation monitoring and mode of delivery inconsistencies reported.

Keywords

High-risk pregnancy pregnancy complications heart valve diseases mechanical heart valve anticoagulation

Introduction

Women with mechanical heart valves (MHVs) require life-long anticoagulation to avoid valve thrombosis.¹ The number of women with MHV of childbearing age is increasing,² and since pregnancy is a prothrombotic state,³ those who choose to become pregnant are at higher risk of thromboembolic events. Pregnant woman with MHV fall into class III of the Modified World Health Organization (mWHO) classification of maternal cardiovascular risk, identifying them to be at significantly increased risk of maternal mortality and/or severe morbidity.⁴

The United Kingdom Obstetric Surveillance System (UKOSS) reported that women with MHV have a 28% chance of good maternal and fetal outcome.⁵ The main maternal complications reported were maternal death (9%) and serious maternal morbidity, defined as secondary postpartum haemorrhage (PPH, 10%), wound haematoma (10%), thrombosed valve (9%), cerebrovascular accident (7%), intraabdominal bleed (7%), primary PPH (2%) and vaginal haematoma (2%).⁵ Poor fetal outcome, mainly related to admission to the Neonatal Intensive Care Unit, birthweight below the 10th centile, and preterm birth, occurred in nearly half the pregnancies.⁵ The Registry of Pregnancy and Cardiac Disease (ROPAC) reported an uncomplicated pregnancy in 58%.⁶ Maternal mortality was 1.4%, MHV thrombosis occurred in 4.7% and haemorrhagic complications in 23.1%.⁶ Fifty per cent of MHV thrombosis occurred in the first trimester in women being treated with heparin and no MHV thromboses occurred in women treated with vitamin K antagonists (VKA).⁶ Fetal mortality ≥24 weeks was 2.8%.⁶

The type of anticoagulation used during pregnancy was the most important factor in determining the risk of thromboembolic complications.³ One of three anticoagulation regimens are typically used by pregnant women with MHV: VKAs throughout pregnancy, low-molecular-weight heparin (LMWH) throughout pregnancy, or LMWH in the first trimester switching to VKAs in the second and third trimester and back to LMWH from 36 to 37 weeks, in a so-called “sequential treatment”.^7,8 There are no published randomised controlled trials (RCTs) comparing the efficacy of the different anticoagulation regimens in pregnant women with MHV, and current evidence is based on secondary research.^1,3 Overall, it appears that VKAs throughout pregnancy offer the best thromboembolic protection to the mother but higher rates of fetal loss and complications.^2,3,7 LMWH is associated with higher livebirth rates but a higher rate of maternal complications.^1,7

The latest 2018 European Society of Cardiology (ESC) Guidelines⁴ recommend that in women on low-dose VKA, the preferred management is to continue VKAs after preconception counselling (PCC) until 36 weeks. The evidence for the use of low-dose VKAs throughout pregnancy is controversial, as comparison between studies is hindered due to reporting differences.⁴ In women on high-dose VKAs, the preferred management is in-hospital change to LMWH with in-hospital reconversion to VKA for the second and third trimesters until 36 weeks. In both cases, women should be admitted at 36 weeks of gestation for conversion to intravenous (IV) unfractionated heparin (UFH) or LMWH; 36 h prior to delivery women should be on IV UFH and this should be stopped 4–6 h before delivery and restarted 4–6 h after delivery in the absence of active bleeding.

We undertook a review of the maternal and fetal outcomes of pregnant women with MHV seen in two tertiary European centres with different approaches to anticoagulation during pregnancy and delivery. The objective of this review was to compare the management of pregnancy in women with MHV between both centres, document associated outcomes and discuss current guidance.

Material and methods

Retrospective case-note review from two Obstetric Cardiac Medicine Centres: Chelsea and Westminster/Royal Brompton Hospitals (CR), London, United Kingdom and Erasmus Medical Centre (EMC), Rotterdam, the Netherlands.

Women with MHV were identified from each centres’ database. Medical records were reviewed and data analysed using Excel (Microsoft® EXCEL for MAC, version 15.33, 2017 – Microsoft Corp., Redmond, WA, USA). Data extracted included demographics, valve replacement indication, valve location, valve type, PCC, type of anticoagulation before conception and target International Normalised Ratio (INR), labour onset, indications for induction of labour (IOL), mode of delivery, analgesia/anaesthesia, estimated blood loss, gestational age at delivery, birthweight, anticoagulation use at 10 and 20 weeks of gestation, pre-delivery (36–37 weeks) and postpartum, fetal and maternal outcomes.

Results

We identified 13 women (19 pregnancies) at CR and 15 women (25 pregnancies) at EMC. The period reviewed was 1999–2017 for CR and 1993–2017 for EMC. Maternal demographics, indication for valve replacement, valve position and type and pre-pregnancy anticoagulation of women from both centres are shown in Table 1.

Table 1.

Demographic details, indication for valve replacement, valve position and type of women included in the review of Mechanical Heart Valves and Pregnancy: Issues Surrounding Anticoagulation.

Centre	Chelsea and Westminster/Royal Brompton Hospital – CR(13 women, 19 pregnancies)	Erasmus Medical Centre – EMC (15 women, 25 pregnancies)
Dates reviewed	1999–2017	1993–2017
Age	Median 31 (IQR 19–43)	Median 31 (IQR 20–39)
Body mass index (kg/m²)	Median 24 (IQR 18–40)	Median 25 (IQR 21–35)
Parity, n (%)	0–11 (58) ≥1–8 (42)	0–8 (32) ≥1–17 (68)
Race, n (%)	White – 7 (54) South Asian – 2 (15) Black African– 2 (15) Mixed – 1 (8) Unknown – 1 (8)	White – 12 (80) South Asian – 2 (13) Black African – 1 (7)
Smoking status, n (%)	Non-smoker – 16 (84) Smoker – 3 (16)	Non-smoker – 21 (84) Unknown – 4 (16)
Indication for valve replacement, n (%)	Endocarditis – 2 (15) Rheumatic heart disease – 3(24) Congenital – 6 (46) Unknown – 2 (15)	Endocarditis – 2 (13) Rheumatic heart disease – 6(40) Congenital – 6 (40) Unknown – 1 (7)
Valve position, n (%)	Aortic – 5 (38) Mitral – 5 (38) Aortic and Mitral – 3 (24)	Aortic – 5 (33) Mitral – 7 (47) Aortic and Mitral – 2 (13) Tricuspid 1 (7)
Type of valve, n (%)	St. Jude Valve – 12 (92%) Björk-Shiley Valve – 1 (8%)	St Jude Valve – 12 (80%) Bicer Valve – 2 (13%) Unknown (performed abroad) – 1 (7%)

IQR: interquartile range.

All women were on VKAs prior to pregnancy in both centres, and for most women, the target INR was between 2.5 and 3.5. Most women were also on low-dose acetylsalicylic acid (75 mg) during pregnancy. Conversion of VKAs to LMWH throughout pregnancy was the most common management adopted at CR, mainly due to the women's expressed wish to avoid the risk of fetal damage from VKAs. We could not obtain the specific discussion records regarding PCC. At EMC, most women were converted from VKAs to LMWH in the first trimester and then converted back to VKAs in the second and third trimesters, when the teratogenic risk of VKAs decreases. One women at CR and three at EMC stayed on VKA throughout pregnancy. Most women were on LMWH around the time of delivery with the last dose of LMWH being omitted greater than 24 h prior. Specific information regarding anticoagulation management throughout pregnancy, pregnancy and maternal outcomes, mode of delivery, labour onset and anaesthesia in labour in both centres is shown in Tables 2 and 3.

Table 2.

Anticoagulation management throughout pregnancy of women included in the review of Mechanical Heart Valves and Pregnancy: Issues Surrounding Anticoagulation. Data reported per pregnancy, not woman.

Centre	Chelsea and Westminster/Royal Brompton Hospital – CR(13 women, 19 pregnancies) N, %	Erasmus Medical Centre – EMC(15 women, 25 pregnancies) N, %
Anticoagulation regimen
VKA throughout	1 (5)	3 (12)
LMWH throughout	14 (74)	9 (36)
LMWH 1st T/VKA 2nd/3rd T	0	11 (44)
Other/unknown	4 (21)	2 (8)
Anticoagulation at 10/40
VKA	1 (5)	3 (12)
LMWH	14 (74)	20 (80)
Unknown	4 (21)	2 (8)
Anticoagulation at 20/40
VKA	1 (5)	14 (56)
LMWH	14 (74)	9 (36)
Unknown	4 (21)	2 (8)
AC around time of delivery
Converted to UFH prior to delivery	1 (5)	7 (28)
Converted to LMWH prior to delivery/LMWH dose omitted 24 h prior to delivery	14 (74)	16 (64)
Other/unknown	4 (21)	2 (8)

LMWH: low-molecular-weight heparin; UFH: unfractionated heparin; VKA: vitamin K antagonist.

Table 3.

Pregnancy and maternal outcomes, mode of delivery, labour onset and anaesthesia in labour of women included in the review of Mechanical Heart Valves and Pregnancy: Issues Surrounding Anticoagulation. Data reported per pregnancy, not woman.

Centre	Chelsea and Westminster/Royal Brompton Hospital – CR(13 women, 19 pregnancies) N, %	Erasmus Medical Centre – EMC(15 women, 25 pregnancies) N, %
Pregnancy outcomes
Intrauterine death	0 (0)	2 (8) (18 weeks, on VKAs (dose and INR not known); 21 weeks, on 10,000 U Dalteparin twice daily – Anti-Xa levels not known )
Stillbirth	0 (0)	1 (4) (26 weeks, on LMWH – Peak Anti-Xa on admission: 0.56 IU/L)
Livebirth	19 (100)	22 (88)
Preterm birth (PTB)	6 (32) – one spontaneous PTB, five iatrogenic PTB (fetal growth restriction or maternal cardiac decompensation)	7 (28) – two spontaneous PTB, five iatrogenic PTB (fetal growth restriction or maternal cardiac decompensation)
Neonatal death	2 (5) (25 and 32 weeks; intracranial and pulmonary haemorrhage. Both mothers on LMWH; dosing and Anti-Xa levels not known)	1 (4) (36 weeks, intracranial and pulmonary haemorrhage. Mother on VKAs; dosing and INR not known)
Unknown	0 (0)	1 (4)
Maternal outcomes
Maternal death	0 (0)	1 (4) (on LMWH, difficulty in achieving therapeutic levels and poor compliance)
MHV thrombosis	1 (5) (on VKA at 24 weeks, poor compliance before and during pregnancy)	2 (8) (one VKA 2nd T with poor compliance and INR 1.7; one on LMWH, difficult in achieving therapeutic levels and poor compliance)
Heart failure	0 (0)	1 (4) (34 weeks)
Atrial fibrillation	0 (0)	1 (4)
Primary PPH	1 (5)	3 (12)
Secondary PPH	4 (31)	0 (0)
Wound haematoma	4 (31) – two rectus sheath + two subcutaneous	0 (0)
Intrabdominal haematoma	2 (10) – one laparoscopic drainage, one secondary infection and hysterectomy	0 (0)
Mode of delivery
Spontaneous vaginal delivery	2 (10)	15 (60)
Instrumental	3 (16)	1 (4)
CS before labour	6 (32)	5 (20)
CS in labour	8 (42)	2 (8)
Unknown	0 (0)	2 (8)
Onset of Labour
Spontaneous	4 (21)	6 (32)
Induced	5 (26) (artificial rupture of membranes only)	11 (44) (prostaglandins, mechanical balloon, artificial rupture of membranes)
Did not labour	6 (32)	5 (20)
Unknown	4 (21)	3 (12)
Anaesthesia
Regional anaesthetic	12 (63) (8 for CS, 4 for vaginal delivery)	2 (8) (1 for CS, 1 for vaginal delivery)
General anaesthetic	6 (32)	5 (20)
None	1 (5)	10 (4)
Unknown	0 (0)	8 (32)

CS: caesarean section; LMWH: low-molecular-weight heparin; MHV: mechanical heart valve; PPH: postpartum haemorrhage; T: trimester; UFH: unfractionated heparin; VKA: vitamin K antagonist.

There was one maternal death at EMC (4%). This woman presented at 26 weeks of gestation with an ischemic stroke secondary to emboli from a MHV thrombus. She was on LMWH and underwent emergency valve replacement surgery. Anti-Xa levels to confirm therapeutic range unavailable, but this woman was known to have poor compliance with LMWH, and it had been difficult to achieve therapeutic levels in pregnancy. During the postoperative period, she suffered a haemorrhagic stroke, underwent emergency craniotomy but was declared brain dead. An emergency caesarean section (CS) was performed to deliver the baby, who was born alive. There was one (5.3%) valve thrombosis at CR and two (8.0%) at EMC. No women presented to CR with heart failure or dysrhythmias but two (8.0%) needed admission at EMC for these reasons. At CR, eight women (42.1%) had a primary PPH (defined as blood loss greater than 500 ml at vaginal delivery or greater than 1000 ml at CS, as per local guidelines) and one (5.3%) had a secondary PPH.

At EMC, three women (12.0%) had a primary PPH, with no secondary PPH reported. There were no postoperative wound complications in the EMC cohort and at CR, six (31.6%) had a wound or intra-abdominal haematoma (Table 3). All women with secondary PPH/haematoma had a prophylactic dose of LMWH 4–6 h after delivery, and had restarted VKAs with LMWH bridging 24 h after delivery, as per protocol, at the time of diagnosis of secondary PPH. They were all diagnosed between day 2 and 3 after the CS whilst inpatients. The women who required laparoscopic drainage of intrabdominal infection, had this performed at six days postoperatively and, for the woman with secondary infection who required a hysterectomy, this was performed three weeks after delivery.

Most notably, at CR, the mode of delivery was predominantly by CS – 14 (73.7% – five category I CS, two category II CS for fetal distress, one category III CS for deteriorating maternal cardiac function and six category IV) and only 5 (26.3%) women had a vaginal delivery. Most women (63.2%) received regional (neuraxial) analgesia or anaesthesia (RA – four for labour and eight for CS), whilst a third of women undergoing CS had general anaesthesia (GA). Of these, one woman had spinal fusion and so RA was not attempted, one had failed attempts at RA, two women required GA for fetal distress/speed and another two underwent planned GA (for unspecified reasons). At EMC, most women had a vaginal delivery (64.0%) with only 28% undergoing CS (one category I for fetal distress, one category III for severe growth restriction and five category IV). Only 1 out of 16 vaginal deliveries and one out of seven women undergoing CS were known to receive RA. Of note, all the emergency CS at EMC were performed under GA, which likely reflects the fewer number of regional blocks sited in labour in the EMC group.

The mean gestational age at delivery was 36 + 5 weeks of gestation at CR and the mean birthweight was 2554 g ± 714 g. The mean gestational age at delivery at EMC was 36 + 6 weeks of gestation and the mean birthweight was 2551 g ± 674 g.

At EMC, INR and Anti-Xa monitoring occurred mainly in the community and we could only obtain a few readings performed in Hospital. INR levels were only obtained in Hospital for three women at EMC, and of six INR measurements performed, only two were within the therapeutic range (INR 2.5–3.5, 2/6 = 33.3%) and levels ranged from 1.7 to 5.8. Anti-Xa levels were obtained in Hospital, at EMC, for four women and ranged from 0.21 to 1.38 IU/mL. We were unable to obtain the specific gestational ages at which these were taken.

Peak anti-Xa levels were obtained antenatally for some of the women at CR. Most (65%) of these levels were outside the recommended range of 0.8–1.2 IU/mL (predominantly sub-therapeutic) and most women only achieved therapeutic levels in the second trimester – Figure 1.

Figure 1.

Peak (4–6 h after administration of LMWH). Anti-Xa levels in IU/mL according to gestational age of women included in the review of Mechanical Heart Valves and Pregnancy: Issues Surrounding Anticoagulation. Each line represents a single woman. The grey area shows the optimal levels between 0.8 and 1.2 IU/mL. X-axis – gestational age in weeks; Y-axis – anti-Xa levels in IU/mL.

Discussion

Pregnancy puts women with MHV at considerable risk. Optimising their management is critically important to ensure the lowest risk of complications and the highest chance of a live birth. Achieving a balance between these demands is often difficult. This series of cases enables us to review different approaches to the management of women with MHV in two leading European centres. One favours the use of LMWH throughout pregnancy and the other the use of a sequential regimen. In one centre, most deliveries were by CS and, in the other, most by vaginal delivery. It appears that the difference in approach is likely to reflect clinician preference. Each of the centres report, on average, on only one pregnancy per annum, which reflects the rarity of pregnancy in women with MHV and highlights the relative inexperience in managing these women especially in the peripartum period. Nonetheless, these differences allow us to compare the risks and benefits of the alternative approaches.

The most contentious area of the management of a pregnancy in women with MHV is anticoagulation, with several regimens available.^8–10 The risk of valve thrombosis is lower with the use of VKA throughout pregnancy (0–4%) when compared to the use of LMWH throughout pregnancy (4.4–8.7%).⁴ Even when LMWH is used in the first trimester only, the risk of valve thrombosis is still high at 5.8–7.4%.⁴ Although there has not been an RCT comparing VKA and LMWH, the data suggest that VKA are best at preventing valve thrombosis, while LMWH offers the greatest chance of a live birth without malformation.^9–13 In terms of valve thrombosis, three occurred across the whole group, two in women on VKA and one in a woman on LMWH. Compliance was an issue in both cases taking VKA (as demonstrated by subtherapeutic INR in one case), and the one case using LMWH was thought to have an underlying prothrombotic state, predisposing her to thrombosis which culminated in maternal death. These data illustrate the overwhelming importance of compliance.⁹ The numbers of cases is too low to draw conclusions about the better regimen regarding the risk of valve thrombosis. However, the period of greatest risk is well recognised to be during therapy change and this may be best achieved as an inpatient with close monitoring to ensure adequate anticoagulation, but avoiding the need by staying on VKA throughout may the lowest risk.

Women taking VKA are at greatest risk of fetal loss and the alternative approach of switching to LMWH in early pregnancy, when the prothrombotic effect of pregnancy is limited and staying on LMWH throughout pregnancy may be safer. However, of the three women suffering pregnancy loss (miscarriage and stillbirth) in this series, two were on LMWH and only one on VKAs. Of the three women who lost their babies due to intracranial and pulmonary haemorrhage (neonatal death), two were on LMWH and one on VKAs. In this series, poor fetal outcome does not seem to be associated with VKAs in particular, although our numbers are small. The key is defining the optimal management of anticoagulation on LMWH. Again, there are no RCTs to guide us,^4,14 but monitoring both peak and trough levels seems to provide a clearer idea of the anticoagulation status. Being willing to use a three times daily regimen and adding aspirin to LMWH may reduce the risk of valve thrombosis.^4,13,15,16 There were no cases of maternal haemorrhage during pregnancy, and all women were converted to LMWH or UFH for delivery in both centres. This means that differences in pregnancy anticoagulation regimens are unlikely to be apparent in post-partum rates of haemorrhage.

The literature suggests that the overall fetal risk (composite of spontaneous miscarriage, fetal death and congenital anomaly) for women on VKAs throughout pregnancy is 39.2% vs. 13.9% in the LMWH group and 16.4% in the sequential treatment group.¹ The better outcomes from regimens using LMWH reflect the fact that they do not cross the placenta and are not teratogenic.¹⁰ Overall, their use is associated with a 95% live birth rate.¹⁰

The difference in use of anticoagulation regimens is striking. We ascribed this to woman preference, but we could not obtain the PCC discussion for each centre. It seems unlikely that the majority of women in one centre would opt for one approach and a second group opt for a different approach in another centre. It seems more likely that clinician preference was a major factor in the women’s decisions. There has also been a shift in the anticoagulation management of such women over the years with a clear preference at CR in using LMWH throughout pregnancy, whereas at EMC, the sequential treatment remained the preferred option throughout the years. It is clear that we need an RCT of the different anticoagulation regimens in order to be able to give women unambiguous advice.

The analysis of the ROPAC data showed that performing an elective CS conferred no benefit to women and meant that babies were born earlier at a lower birthweight.⁶ The incidence of preterm birth was high for both centres, 31.6% for CR and 28% for EMC, mainly due to iatrogenic preterm birth due to maternal or fetal reasons and is in line with the literature.^5,6 The difference in mode of delivery could be explained by the higher proportion of multiparous women at EMC (68%) than at CR (42%) and, possibly, local policy. Despite mode of delivery differences between our centres, gestation at delivery and birthweight were very similar. Primary PPH rates were also similar, but secondary PPH and wound haematomas occurred more commonly in the centre with higher CS section rates, requiring re-operation in two cases. Overall rates of haemorrhage vary in the literature from 12.8 to 42.8%,^5,6,17–19 with the results of the centre with higher CS rates being higher than this range, if we include the occurrence of haematoma. These data again suggest that vaginal delivery is the better option, despite the greater ability to predict the anticoagulation-free period with CS delivery. Minimising the duration without anticoagulation may be best achieved by converting to UFH, which can be maintained until closer to delivery or need for regional analgesia/anaesthesia, as it has a shorter half-life and can be more easily reversed than LMWH, if necessary.

Anticoagulation monitoring during pregnancy

Women on VKA should have INR monitoring weekly/every two weeks with target INR as pre-pregnancy.⁴ For women on LMWH in the first trimester or throughout pregnancy, strict monitoring of anti-Xa levels is recommended.⁴ Due to increased plasma volume, increased glomerular filtration rate and increased renal clearance of LMWH,^3,8 twice daily dosage adjusted to peak (3–4 h after administration) anti-Xa levels is recommended.^4,8 The target peak levels should be 0.8–1.2 U/l for valves in the aortic position and 1.0–1.2 U/l for valves in the mitral positions, as these are more thrombogenic.⁴ We only obtained anti-Xa levels for nine pregnancies at CR (only in-hospital levels were plotted) but most were outside (mainly below) the recommended range. Additional pre-dosing levels (trough) with levels targeted at ≥0.6 U/l can be considered.⁴ As pregnancy progresses, peak anti-Xa activity decreases due to greater renal clearance and increased volume distribution of LMWH. The trough levels might be more informative as increasing the dose of LMWH (to peak levels) might result in over anticoagulation and more haemorrhagic complications.¹⁵ However, even with intensive monitoring, the rate of complications in women on LMWH is reported to be high.²⁰ Haemorrhagic complications are common with any anticoagulation regimen but less so with VKA compared with LMWH throughout pregnancy or sequential treatment.⁴ As we observed, haemorrhagic complications are more often secondary in nature (wound haematoma, intra-abdominal bleeding, secondary PPH, 33.3%).⁵

Anticoagulation peripartum

There was a notable difference with regards to peripartum management of anticoagulation between the two centres. At CR, the preferred method was to stop LMWH for ≥24 h prior to IOL or CS to allow for safe insertion of RA, as recommended by several UK bodies²¹ and to also reduce the risk of PPH. Although we did not have any peripartum thromboembolic complications with this approach, the risk of valve thrombosis during the anticoagulation-free period is high and should be kept to a minimum.⁴ EMC also used this approach in 64% of the cases. The recommended management for women on VKA is planned delivery, switching to IV UFH (target aPTT ≥2× control) and stopping this 4–6 h prior to delivery or insertion of regional analgesia. VKAs should be restarted 4–6 h after delivery if there is no active bleeding. It is suggested that CS could be considered to minimise the duration of anticoagulation-free time.⁴ However, our higher rates of CS result in similar rates of primary PPH but higher rates of secondary PPH and haematoma, suggesting that vaginal delivery, in most situations, should be the preferred mode of delivery.⁸ The policy of maintaining LMWH until 36 h prior to delivery and then switching to IV UFH may minimise anticoagulation-free time, but this approach requires intensive monitoring and we have experienced several complications due to a failure of interpretation of the aPTT.

Strengths and limitations

Our biggest strength is that this review analysed 44 pregnancies from two large specialist obstetric cardiac centres and gives important insight into our current practice, raising awareness on the management of pregnant women with MHV. Data were collected onto a predefined spreadsheet and cross-checked by two authors at each centre. Our biggest limitation is that this was a retrospective review and a significant amount of data are missing – we only had information on peak anti-Xa levels for nine women at CR and for four at EMC; INR levels were only available for three women at EMC and none at CR; gestational age at delivery and birthweight are missing for one pregnancy at CR and nine pregnancies at EMC. Most of our maternal and fetal complications occurred up to the early years of the year 2000, and we could not obtain specific information on the anticoagulation doses and INR/Anti-Xa levels at the time that the complications occurred. We included pregnancies from 1993 and, at that time, mechanical valves were more thrombogenic and anticoagulation management has also since evolved with use of LMWH being favoured in the last few years.

Conclusion

Women with MHV should be managed in tertiary centres during pregnancy with a dedicated team of maternal medicine specialists, cardiologists, haematologists and obstetric anaesthetists.²² Anti-coagulation management for women with MHV during pregnancy is challenging,²⁰ and the associated maternal and fetal morbidity and mortality is high.^7,20 Although VKAs generally seem to offer the best maternal protection in terms of MHV thrombus, they are associated with significant adverse fetal outcomes; hence, women often prefer anticoagulation with LMWH.^7,22 Our data support these conclusions. We suggest that further research is needed; ideally, a multicentre RCT comparing women anticoagulated with LMWH vs. VKA throughout pregnancy. This could provide invaluable information on the optimal management of women with MHV. However, there could be huge ethical implications in conducting such a trial due to the increased risk of warfarin embryopathy/fetopathy with high-dose VKAs and due to the potential increase in thrombotic complications with LMWH which would make obtaining consent for each of the arms difficult.

The management of anticoagulation around the time of delivery should be targeted at minimising the anticoagulation-free period while minimising the risk of haemorrhage. This may be best achieved with the use of UFH.

The mode of delivery for women with MHV on the basis of our data, and the overall data in women with heart disease, should be vaginal, as it is associated with similar rates of primary PPH and lower rates of secondary PPH and haematoma formation.

Footnotes

Acknowledgements

None.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

This review was undertaken as part of the Joint Cardiac Clinic service evaluation at CR and therefore Ethical review was not deemed necessary. For the EMC, Medisch Etische Toetsings Commissie (METC) approval was not required, as the data were collected retrospectively and are anonymised.

Informed consent

This review was an evaluation of service. Participant informed consent was not required.

Guarantor

FDS.

Contributorship

FDS and MRJ conceived the idea. FDS and AV collected the data from CR, LB and JC collected the data from EMC. FDS and AV analysed all the data. FDS wrote the first draft of the article. All authors gave significant input during all stages of manuscript development and approved the final version of the article.

ORCID iD

Francois Dos Santos

References

Steinberg

Dominguez-Islas

Otto

, et al. Maternal and fetal outcomes of anticoagulation in pregnant women with mechanical heart valves. J Am Coll Cardiol 2017; 69: 2681–2691.

Bhagra

D’Souza

Silversides

CK.

Valvular heart disease and pregnancy part II: management of prosthetic valves. Heart 2017; 103: 244–252.

Alshawabkeh

Economy

Valente

AM.

Anticoagulation during pregnancy: evolving strategies with a focus on mechanical valves. J Am Coll Cardiol 2016; 68: 1804–1813.

Regitz-Zagrosek

Roos-Hesselink

Bauersachs

, et al. 2018 ESC Guidelines for the management of cardiovascular diseases during pregnancy. Eur Heart J 2018; 39: 3165–3241.

Vause

Clarke

Tower

, et al.; on behalf of UKOSS. Pregnancy outcomes in women with mechanical prosthetic heart valves: a prospective descriptive population based study using the United Kingdom Obstetric Surveillance System (UKOSS) data collection system. BJOG 2017; 124: 1411–1419.

van Hagen

Roos-Hesselink

Ruys

, et al. Pregnancy in women with a mechanical heart valve: data of the European society of cardiology registry of pregnancy and cardiac disease (ROPAC). Circulation 2015; 132: 132–142.

D’Souza

Ostro

Shah

, et al. Anticoagulation for pregnant women with mechanical heart valves: a systematic review and meta-analysis. Eur Heart J 2017; 38: 1509–1516.

D’Souza

Silversides

McLintock

Optimal anticoagulation for pregnant women with mechanical heart valves. Semin Thromb Hemost 2016; 42: 798–804.

McLintock

Anticoagulant therapy in pregnant women with mechanical prosthetic heart valves: no easy option. Thromb Res 2011; 127: S56–S60.

10.

McLintock

Anticoagulant options in pregnancy for women with mechanical valves. BJOG 2017; 124: 1421.

11.

Baumgartner

Falk

Bax

, et al. 2017 ESC/EACTS Guidelines for the management of valvular heart disease. Eur Heart J 2017; 38: 2739–2791.

12.

Castellano

Narayan

Vaishnava

, et al. Anticoagulation during pregnancy in patients with a prosthetic heart valve. Nat Rev Cardiol 2012; 9: 415–424.

13.

Elkayam

Anticoagulation therapy for pregnant women with mechanical prosthetic heart valves: how to improve safety?

J Am Coll Cardiol 2017; 69: 2692–2695.

14.

McLintock

Anticoagulant choices in pregnant women with mechanical heart valves: balancing maternal and fetal risks–the difference the dose makes. Thromb Res 2013; 131: S8–S10.

15.

Patel

Green

Patel

, et al. Population pharmacokinetics of enoxaparin during the antenatal period. Circulation 2013; 128: 1462–1469.

16.

Goland

Schwartzenberg

Fan

, et al. Monitoring of anti-Xa in pregnant patients with mechanical prosthetic valves receiving low-molecular-weight heparin: peak or trough levels? J Cardiovasc Pharmacol Ther 2014; 19: 451–456.

17.

Wang

Marnoch

Khurana

, et al. Haemorrhagic complications of peripartum anticoagulation: a retrospective chart review. Obstet Med 2014; 7: 77–83.

18.

Knol

Schultinge

Veeger

, et al. The risk of postpartum haemorrhage in women using high dose of low-molecular-weight heparins during pregnancy. Thromb Res 2012; 130: 334–338.

19.

Irani

Santa-Ines

Elder

, et al. Postpartum anticoagulation in women with mechanical heart valves. Int J Womens Health 2018; 10: 663–670.

20.

Snape

Thachil

Clarke

, et al. Anti-Xa based dose changes during low molecular weight heparin anticoagulation for mechanical prosthetic heart valves during pregnancy. J Obstet Gynaecol 2018; 38: 721–722.

21.

Party

Ireland

AoAoGB

Association

RA.

Regional anaesthesia and patients with abnormalities of coagulation: the Association of Anesthetists of Great Britain & Ireland The Obstetric Anesthetists’ Association Regional Anaesthesia UK. Anaesthesia 2013; 68: 966–972.

22.

Aggarwal

Economy

Valente

AM.

State of the art management of mechanical heart valves during pregnancy. Curr Treat Options Cardiovasc Med 2018; 20: 102.