Introduction

Venous thromboembolism (VTE), which may manifest as pulmonary embolism (PE) or deep vein thrombosis (DVT), complicates 0.5–2.2 per 1000 deliveries, depending on the population studied [18]. During pregnancy, the risk of VTE is increased five to tenfold compared to non-pregnant women of comparable age [1, 9, 10]. The postpartum period poses a higher risk [1, 7, 10] and during this time frame, the daily risk of VTE is increased 15- to 35-fold compared to age-matched non-pregnant women [1, 9]. The daily risk of pregnancy-associated VTE appears greatest during the first 3–6 weeks postpartum [1, 7]. After that, the risk declines rapidly, although a small residual risk increase may persist for 12 weeks after delivery [7, 10, 11]. Although the absolute VTE rates are low, pregnancy-associated VTE is an important cause of maternal morbidity [1214] and mortality [15, 16].

The treatment and prevention of pregnancy-associated VTE is challenging because of the potential for both fetal and maternal complications, as well as the paucity of relevant high quality research. Although evidence-based guideline recommendations for the use of anticoagulants in this patient population have been published [1725], they are based largely upon observational studies and extrapolated from data in non-pregnant patients. The lack of high quality data specific to pregnancy results in a lack of consistency in their recommendations. This chapter reviews the published evidence base and uses that information, as well as published guidelines, to provide practical guidance for the management and prevention of VTE during pregnancy.

Methods

The goal of this chapter is to provide guidance to providers on how best to individualize care to patients with pregnancy-associated VTE, with specific focus on the questions listed in Table 1. Questions were developed by consensus from the authors. To address these questions, current guidelines from the American College of Obstetricians and Gynecologists (ACOG) [17, 18], the Society of Obstetricians and Gynaecologists of Canada (SOGC) [19], the Royal College of Obstetricians and Gynaecologists (RCOG) [20, 21], clinicians from Australia and New Zealand [22], and American College of Chest Physicians (ACCP) [2325] were reviewed and relevant recommendations were extracted (Tables 2A–2D). The literature was reviewed and data from relevant systematic reviews, randomized trials, and observational studies were incorporated. The authors’ consensus interpretation of these studies, in the context of the realities of VTE care, was distilled into the practical recommendations that are presented in this article.

Table 1 Guidance questions to be considered
Table 2A Guideline summary—anticoagulant choice
Table 2B Guideline summary—management of acute venous thromboembolism
Table 2C Guideline summary—prevention of first and recurrent pregnancy-associated VTE
Table 2D Guideline summary—anticoagulant management around the time of delivery

In making recommendations regarding the need for prophylaxis, the panel used a risk threshold of 3 % and greater for antepartum prophylaxis and 3 % and greater for postpartum prophylaxis. For risk factors for which only case control data are available, a relative risk of at least 30-fold antepartum and 60-fold postpartum are required to reach our thresholds, assuming antepartum and postpartum baseline risks of 0.1 and 0.05 %, respectively [26]. The prophylaxis thresholds above were determined by the majority result of an anonymous vote of the authors. It is important to note that there was inconsistency between the authors in their risk threshold for recommending prophylaxis. For antepartum prophylaxis, three chose 3 % or greater, one 5 % or greater, and one 1 % or greater. For postpartum prophylaxis, four selected a threshold of 3 % or greater and one chose 1 % or greater. The variability in risk thresholds is not surprising given the limitations of the available evidence, as well as the competing benefits and drawbacks of prophylaxis. The panel would emphasize that changes in the antepartum threshold to 5 or 1 % and to the postpartum threshold to 1 % would markedly change the recommendations that follow. When making recommendations, the panel also took into account the estimated risks of major bleeding with prophylactic LMWH (antepartum: 0; 95 % CI 0–0.6 % and postpartum: 0.3 %; 95 % CI 0–1.0 %) [26]; the variability in risk estimates reported in the literature, the 95 % confidence intervals around the risk estimates, and the strengths or weaknesses of relevant study methodology in addition to the above threshold limits.

Guidance

  1. 1.

    What are the risks of anticoagulant use during pregnancy?

During pregnancy, the risks posed to the fetus by anticoagulant therapy, as well as maternal efficacy and safety must be considered. Vitamin K antagonists cross the placenta and have the potential to cause teratogenicity as well as pregnancy loss, fetal bleeding, and neurodevelopmental deficits [2734]. Discontinuation of vitamin K antagonists prior to the 6th week of gestation essentially eliminates the risk of warfarin embryopathy [29, 30, 32]. Pregnant women were excluded from participating in clinical trials evaluating the oral direct thrombin and factor Xa inhibitors (e.g. dabigatran, rivaroxaban, apixaban, edoxaban). These agents are likely to cross the placenta and their human reproductive risks are unknown [3538]. Fondaparinux appears to cross the placenta in small quantities [39]. Reports of the successful use of fondaparinux in pregnant woman have been published [3946] but it is important to recognize that many of these involve second trimester or later exposure.

Unfractionated heparin (UFH), low molecular weight heparin (LMWH) and danaparoid (a heparinoid) do not cross the placenta and are safe for the fetus [4755]. Although UFH can be used during pregnancy for both prevention and treatment of thromboembolism, LMWH has a better safety profile than UFH [56, 57] and the incidence of bleeding and other complications (e.g. heparin induced thrombocytopenia [HIT], and heparin-associated osteoporosis) are lower in pregnant women receiving LMWH than with UFH [5869]. LMWHs are eliminated primarily by renal excretion and may accumulate in patients with significant renal dysfunction. In the non-pregnant population, it has been suggested that therapeutic dose LMWH not be used in patients with significant renal impairment (e.g. a glomerular filtration rate (GFR) of less than 30 mL/min), although it is recognized that accumulation in patients with renal impairment may differ between the various LMWHs [70].

As outlined in Table 2A, there is clear consensus amongst the reviewed guideline documents that, in general, LMWH is the preferred anticoagulant for the management and treatment of VTE in pregnancy [1725].

Guidance Statement

  • Physicians should counsel women receiving long-term therapy with vitamin K antagonists and the oral direct-acting anticoagulants about the fetal risks of these medications before pregnancy occurs.

  • LMWH is the drug of choice for treatment and prevention of VTE in pregnancy, except in patients with HIT, a history of HIT, or significant renal dysfunction. UFH is preferred in patients with significant renal dysfunction.

  • For women taking vitamin K antagonists, two options are available to reduce the risk of warfarin embryopathy. The first is to advise women to perform frequent pregnancy tests and substitute LMWH for warfarin once pregnancy is achieved and before 6 weeks gestation. Alternatively, LMWH or UFH can be substituted for vitamin K antagonists before conception is attempted. Although the latter approach minimizes the risks of early miscarriage associated with vitamin K antagonist therapy, it lengthens the duration of exposure to LMWH or UFH and, therefore, is costly and exposes the patient to a greater burden of treatment associated with the use of injectable heparin therapy. Since warfarin embryopathy is unlikely to result from warfarin exposure before 6 weeks, the first option is usually favored by guidelines. Although the management of women who are receiving long-term therapy with oral direct thrombin and factor Xa inhibitors and attempting to conceive remains controversial, it has been suggested that these women should be converted to a coumarin or LMWH before conception is attempted; failing that, the switch should be made as soon as pregnancy is confirmed.

  • In pregnant women with severe cutaneous allergies to UFH or LMWH or with HIT or a history of HIT, danaparoid or fondaparinux (if danaparoid not available) may be used. Appropriate dosage and management of these anticoagulants around the time of delivery should be discussed with a hematologist or thrombosis specialist.

  1. 2.

    What are the risks of anticoagulation in breastfeeding women?

Neither warfarin, the most commonly used vitamin K antagonist in North America and the United Kingdom nor acenocoumarol, which is commonly used in Europe, is detected in breast milk and neither medication induces an anticoagulant effect in the breast-fed infant when nursing mothers consume the drug [7174]. Phenprocoumon, another vitamin K antagonist with a long half-life, is also widely used outside of North America. This agent is more lipophilic than warfarin and acenocoumarol and so can be excreted into breast milk, although since it is highly protein-bound, the amounts detected are small [75, 76]. Small amounts of LMWH have been detected in the breast milk of women receiving this medication [77]; however, given the very low bioavailability of heparin when ingested orally [78], there is unlikely to be any clinically relevant effect on the nursing infant. All of the guidelines that address the issue of anticoagulant use during breastfeeding agree that warfarin, LMWH, and UFH are safe to use in this setting (see Table 2A) [17, 2022]. Phencoumaron should be reserved for women who are unstable on short-acting acenocoumarol in countries where warfarin is not available [76].

According to the manufacturer’s prescribing information, fondaparinux was excreted in the milk of lactating rats [79]. There are no published data on the excretion of fondaparinux into human milk and the effects on the nursing infant are unknown. The manufacturer recommends that caution be used when administering fondaparinux to breastfeeding women [79]. That said, significant absorption by the nursing infant would be unlikely as orally ingested heparins have low bioavailability [78].

There are no clinical data on the effect of maternally ingested oral direct thrombin and factor Xa inhibitors on breastfed infants. The manufacturers of these agents all recommend against using these medications in breastfeeding women [3537].

Guidance Statement

  • UFH, LMWH, warfarin and acenocoumarol are safe for the breast-fed infant when administered to the nursing mother.

  • The oral direct thrombin and factor Xa inhibitors should not be used while breastfeeding.

  1. 3.

    How is venous thromboembolism during pregnancy treated?

The guideline recommendations for management of acute VTE during pregnancy are summarized in Table 2B. There have been no large studies examining the safety and efficacy of outpatient treatment of VTE diagnosed during pregnancy. Data from the non-pregnant population suggest that outpatient DVT treatment is not associated with an increase in mortality, recurrent VTE, or major bleeding [24]. In non-pregnant patients with acute DVT, outpatient treatment is recommended as long as the patient feels well enough to be treated at home (e.g. does not have severe leg symptoms or comorbidity) and has well-maintained living conditions, strong support from family or friends, telephone access, and the ability to quickly return to hospital if conditions deteriorate [24]. The safety of treating PE at home, even in the non-pregnant population, is uncertain. Prediction rules have been developed for identifying non-pregnant patients with acute PE who might be suitable for outpatient treatment because they are at low risk of serious complications [24, 80].

The results of large trials in non-pregnant patients demonstrating that LMWHs are at least as safe and effective as UFH for the acute treatment of VTE [81, 82] and as vitamin K antagonists for the prevention of recurrent VTE [83, 84], as well as data from subsequent observational studies in pregnant women, support the use of LMWH for treatment of VTE in this patient population [60, 61, 85, 86].

There are no large trials examining the optimal dose of anticoagulants for treatment of acute VTE during pregnancy. Some pharmacokinetic studies suggest that increases in GFR and in patient weight (and, hence, LMWH volume of distribution) that occur during pregnancy may lead to lower LMWH levels and that the dose of LMWH should be adjusted over the course of pregnancy to maintain “therapeutic” anti-Xa LMWH levels [87, 88], or according to changes in weight [89]. However, other researchers have demonstrated that few women require dose-adjustment when therapeutic doses of LMWH are used [9094]. Some recommend a twice-daily LMWH dosing schedule during pregnancy to compensate for increases renal clearance of this medication that occur in the second trimester. In non-pregnant patients, once daily LMWH is as safe and effective as twice daily LMWH when used to treat acute VTE [95]. Observational studies in pregnant women with acute VTE have not demonstrated any increase in the risk of recurrence with a once-daily regimen compared with twice-daily schedules [85, 86] and many clinicians use once-daily therapy to simplify administration and enhance compliance.

There are issues with reliability of anti-Xa LMWH tests [96, 97] and these assays are costly. In the absence of robust data demonstrating that there is an optimal “therapeutic anti-Xa LMWH range” and that dose-adjustments increase the safety or efficacy of LMWH therapy, current guidelines do not mandate routine monitoring of LMWH with anti-Xa levels [2123]. Anti-Xa monitoring may be helpful to ensure appropriate anticoagulant effect in patients with renal impairment and in those at the extremes of body weight [21].

Regimens in which the intensity of LMWH is reduced later during the course of therapy to an intermediate dose regimen [98] or 75 % of a full treatment dose [84] have been used successfully in cancer patients. A recent systematic review that identified four studies in which pregnant women with symptomatic VTE were transitioned from full-dose anticoagulation to intermediate-dose LMWH (less than 75 % of a full treatment dose but greater than prophylactic dose) within 6 weeks of VTE diagnosis, reported a low risk of VTE recurrence (one of 152 patients) during intermediate-dose LMWH therapy; however, the number of patients with PE was small (four) and at least one of the included studies enrolled patients with isolated calf vein thrombosis, which could lead to an overestimation of the positive effect [99]. Some guidelines suggest a dose-reduction strategy for pregnant women at risk of anticoagulant-related bleeding and heparin-induced osteoporosis [23] and in those with isolated calf vein thrombosis [22]. That said, a survey of members of the North American Society of Obstetric Medicine and Thrombosis Canada found that only one-quarter of respondents utilized this strategy in their patients [100].

The risk of HIT in pregnant women treated with LMWH alone is low (less than 0.1 %) [59]; it is higher in pregnant women who have received UFH. Several guidelines suggest that routine platelet count monitoring for detection of HIT is not required in pregnant women treated exclusively with LMWH [21, 25].

Intravenous UFH is preferred when rapid reversal of anticoagulation may be required (i.e. in situations in which urgent delivery or surgery may be necessary) and in patients in whom thrombolysis may be considered (e.g. high risk or massive PE) [24, 80]. UFH should be used in preference to LMWH to treat acute VTE in patients with GFR of less than 30 mL/min [80]. When UFH is preferred, it can be given intravenously or subcutaneously every 12 h in doses adjusted to prolong a mid-interval (6 h post-injection) activated partial thromboplastin time (aPTT) into therapeutic range [101], although it is recognized that aPTT monitoring is less reliable in pregnancy [102].

Concerns about the use of thrombolytic therapy during pregnancy center on its maternal effects (major hemorrhage), as well as those on the placenta (i.e. premature labor, placental abruption, fetal demise), as transplacental passage of tissue plasminogen activator and streptokinase is minimal [103]. There have been several reports of successful thrombolysis in pregnancy with no harm to the fetus; however, the number of cases is small and most cases involved streptokinase [104107]. Therefore, there is agreement amongst available guidelines that the use of thrombolytic therapy in pregnancy is best reserved for limb or life-threatening maternal thromboembolism (e.g. PE with refractory cardiorespiratory compromise) [2123, 80].

There is limited experience with inferior vena caval filters during pregnancy and serious complications, including filter fracture, filter migration, failed retrieval of temporary devices, and inferior vena caval perforation, have been reported [108112]. Current guidelines recommend insertion of temporary inferior vena caval filters in pregnant women with acute VTE and contraindications to anticoagulant therapy [18, 22] or recurrent VTE despite therapeutic anticoagulation [17, 2123]. An alternate strategy involving anticoagulant dose-escalation may also be appropriate for managing the latter situation, based on favorable (but limited) data in cancer patients, in which recurrent VTE despite anticoagulant therapy is treated by increasing the dose of LMWH by approximately 25 % or to therapeutic levels in those receiving lower doses [113, 114].

There are conflicting data concerning the long-term effectiveness of graduated compression stockings to prevent post-thrombotic syndrome. On the basis of two positive open label randomized trials in the non-pregnant population [115, 116], several guidelines have suggested that graduated compression stockings be prescribed to reduce the likelihood of developing post-thrombotic syndrome [19, 22, 23]. However, a recent multicenter placebo-controlled trial that enrolled non-pregnant patients reported that these stockings neither prevented this complication nor reduced the risk of recurrent VTE [117]. In addition, although it is thought that graduated stockings may be useful for acute symptom relief, a subgroup analysis of this study suggests that, at least in the non-pregnant population, this may not be the case [118].

There have been no studies assessing optimal duration of anticoagulant therapy for treatment of pregnancy-related VTE. In non-pregnant patients with VTE, evidence supports a minimum treatment duration of 3 months [24]. Given the increased risk of VTE in pregnant women and following delivery, available guidelines suggest that anticoagulants be continued throughout pregnancy and the postpartum period, and for a minimum of 3 months [19, 2125].

Guidance Statement

  • Outpatient treatment of VTE can be considered in patients who are clinically stable and have good cardiorespiratory reserve, no major risk factors for bleeding and good social support with easy access to medical care. Hospitalization is indicated in patients who are hemodynamically unstable or do not have good social support and those who have extensive VTE, or maternal co-morbidities that limit their tolerance of recurrent VTE or increase their risk of major bleeding.

  • LMWH is the preferred anticoagulant for most pregnant women with acute VTE. UFH should be used instead of LMWH in patients with GFR less than 30 mL/min. Intravenous UFH should be considered in patients who may require thrombolysis, surgery or urgent delivery.

  • If LMWH is used for treatment of acute VTE in pregnancy, the same weight-adjusted dosing regimen as in the nonpregnant population should be utilized (Table  3). Routine monitoring of LMWH dosing with anti-Xa LMWH is likely not required.

    Table 3 Accepted LMWH dosing regimens for treatment of pregnancy-related VTE
  • Thrombolytic therapy should be reserved for pregnant women with PE associated with life-threatening cardiorespiratory compromise or limb-threatening DVT.

  • Insertion of a temporary inferior vena caval filter should be considered in pregnant women with acute VTE and a contraindication to anticoagulant therapy.

  • Anticoagulant therapy for treatment of VTE during pregnancy should be continued throughout pregnancy and for at least 6 weeks postpartum for a minimum duration of 3 months.

  1. 4.

    How is pregnancy-associated VTE prevented?

Decisions regarding the use of prophylactic anticoagulation during pregnancy depend on the balance between the estimated risk of VTE and associated reduction in risk with prophylaxis, along with the burdens associated with anticoagulant therapy. The appropriate use of prophylaxis depends on identifying those at sufficiently high risk of VTE to benefit from this intervention. Risk factors to be considered include prior VTE, familial VTE history, the presence of a known thrombophilia, and clinical factors, including cesarean delivery, prolonged antepartum immobilization, increased body mass index (BMI), as well as significant pregnancy complications and medical comorbidities.

Prophylaxis during pregnancy typically involves long-term subcutaneous injections of LMWH. Although prophylactic LMWH is safe for the fetus [27, 29, 30, 4951] and does not appear to appreciably increase the risk of adverse maternal outcomes [58, 59, 6267]; it is expensive, inconvenient and uncomfortable to administer. Depending on local practice, prophylaxis with LMWH may also necessitate a planned delivery to permit epidural analgesia and women may perceive that it creates an undesirable “medicalization” of their pregnancy.

A Cochrane systematic review of thromboprophylaxis in pregnancy and the early postnatal period that examined 16 randomized trials involving 2592 women concluded that the current available information is insufficient to make firm recommendations for prophylaxis [119]. Current clinical guidelines are based on these small trials, additional observational studies and indirect evidence suggesting that LMWH substantially decreases the risk of VTE in a wide variety of clinical settings. As shown in Table 2C, there is incomplete agreement between the guidelines as to which patients should receive thrombosis prophylaxis and only a few guidelines (SOGC [19] and ACCP for postpartum prophylaxis [23]) explicitly provide information about the risk threshold used to determine whether or not patients should receive prophylaxis.

Given the competing potential drawbacks and benefits of prophylaxis, as well as the limitations of the available evidence, the decision to use or not use LMWH is likely to be value and preference sensitive. In addition to holding different attitudes toward the risk of recurrent thrombosis and about the burdens associated with the use of prophylaxis, women are also likely to place varying importance on minimizing medicalization of their pregnancy. All women, therefore, merit an individualized risk–benefit assessment of their need for prophylaxis and the opportunity to share in a decision making process about this intervention that takes into account their values and preferences.

If the decision is made to use antepartum prophylaxis, it should be initiated early in pregnancy as there is evidence of an increased risk of VTE during all three trimesters [120, 121]. Postpartum prophylaxis is less burdensome than antepartum prophylaxis as the duration of prophylaxis is shorter (i.e. 6 weeks) and an oral anticoagulant is available for those uncomfortable with subcutaneous injections (vitamin K antagonists, except for those with protein C or S deficiency who are at risk for developing warfarin-induced skin necrosis) [122124].

The optimal prophylaxis strategy is unknown. Several LMWH dosing regimens have been used for prophylaxis of VTE during pregnancy (Table 4) [59, 62, 118, 125133]. Although all of the studies evaluating these regimens reported low VTE rates, most were cohort studies and, therefore, lacked data from untreated controls. Some investigators have reported failures of prophylactic LMWH; however, it is unclear whether these represent true failures or were due to noncompliance with long-term subcutaneous injections [59, 60, 127, 134, 135]. Different dosing strategies have not been directly compared, although one randomized trial comparing higher doses of LMWH prophylaxis with usual fixed dose prophylaxis is ongoing (Highlow Randomized Controlled Trial: Comparison of Low and Intermediate Dose Low-molecular-weight Heparin to Prevent Recurrent Venous Thromboembolism in Pregnancy; NCT001828697).

Table 4 Suggested LMWH dosing regimens for prophylaxis against pregnancy-related VTE

In hospitalized women, mechanical prophylaxis with elastic stockings and/or intermittent pneumatic compression is an alternative for those with contraindications to anticoagulant prophylaxis [23]; although there is limited evidence that these devices are less effective at prevention of VTE [136].

Duration of anticoagulant prophylaxis after delivery remains controversial. Available guidelines recommend 6 weeks of postpartum prophylaxis in patients with prior VTE and those with some thrombophilias (varies between guidelines) [17, 19, 21, 23]. However, there is minimal evidence to guide duration of prophylaxis in women with other clinical risk factors and recommendations vary. A shorter course of postpartum prophylaxis (until discharge or for one to 2 weeks post discharge) is often suggested for women with transient risk factors [19, 21, 23]. A recent study that used linked primary and secondary care data to assess VTE risk during specific postpartum periods reported that women with pre-eclampsia/eclampsia and acute systemic infection, obesity (body mass index or BMI ≥ 30 kg/m2), and cesarean delivery had elevated VTE risks up to 6 weeks postpartum; while VTE risk was increased only for the first 3 weeks after delivery in those with postpartum hemorrhage or preterm birth [137]. However, the absolute VTE risk during those time frames was low (less than 1 %).

All pregnant women at risk of VTE should be educated about the signs and symptoms of DVT and PE and the need to seek urgent medical attention should they develop. Objective testing is mandatory if symptoms suspicious of DVT or PE occur.

Prevention of recurrent VTE

The most important individual risk factor for pregnancy-associated VTE is a prior history of thrombosis [138]. The absolute risk of recurrent VTE during pregnancy in women not given antepartum prophylaxis remains controversial. In more recent studies, the reported incidence ranged from 2.4 % (95 % CI 0.2–6.9) in a prospective study of 125 pregnant women [139] to approximately 6 % in larger retrospective cohort studies [140, 141]. Differences in study population (later median gestational age at enrollment; inclusion of women with more than one prior episode of VTE in the retrospective studies), as well as failure to independently adjudicate recurrent events in the retrospective studies, may explain the higher risk of recurrence in the latter studies. However, the overall risk of antepartum recurrent VTE in both prospective and retrospective studies was less than 10 % and CI’s around the risk estimates of individual studies are overlapping.

Data regarding prognostic factors for recurrent VTE during pregnancy are inconsistent. Although a subgroup analysis of the prospective cohort study mentioned above found a lower risk of recurrence in women without thrombophilia who had a temporary risk factor (including oral contraceptive therapy or pregnancy) at the time of their prior VTE, than in those with abnormal thrombophilia testing and/or an unprovoked event [138]; in the two subsequent retrospective studies, the presence or absence of a definable thrombophilia did not appear to influence the risk of recurrent pregnancy-associated VTE [140, 141]. Studies in nonpregnant patients have also demonstrated that thrombophilic abnormalities do not play an important role in determining the risk of recurrent VTE, despite being clear risk factors for a first episode of DVT or PE [142]. There was a suggestion in the two retrospective studies that women with a first VTE provoked by oral contraceptives or related to pregnancy might be at higher risk of recurrence in a subsequent pregnancy than those with an unprovoked event or VTE related to a transient non-hormonal risk factor [140, 141]. The latter findings are consistent with those from an observational administrative dataset from California [143].

The above data suggest that pregnant women with a single prior episode of VTE associated with a transient risk factor not related to pregnancy or use of estrogen are likely at lower risk of recurrent antepartum VTE compared to pregnant women with a history of unprovoked, pregnancy or estrogen-related VTE. The ACCP guidelines estimated the risk of recurrent antepartum VTE without prophylaxis to be 2 % in the first group and 8 % in the second group [23]. Current guidelines favor a strategy of antepartum clinical vigilance for those with a single prior episode of VTE associated with a transient risk factor not related to pregnancy or hormone use and antepartum LMWH with a history of unprovoked, pregnancy or estrogen-related VTE [17, 19, 20, 23]. However, as the available data have significant limitations, antepartum clinical vigilance may also acceptable for higher risk patients accepting of the risks of recurrence and for whom the burden of LMWH prophylaxis outweighs potential benefits. Similarly, women with a prior VTE associated with a transient risk factor not related to pregnancy or use of estrogen may benefit from antepartum prophylaxis if they have additional major risk factors for thrombosis. Although supportive data from clinical trials are lacking, postpartum prophylaxis for 6 weeks with prophylactic or intermediate dose LMWH or vitamin K antagonists targeted at INR 2.0–3.0 is generally recommended for all pregnant women with prior VTE not receiving long-term anticoagulants [17, 19, 20, 23].

Prevention of VTE in pregnant women with thrombophilia and no prior VTE

Thrombophilias are laboratory abnormalities associated with an increased risk of thrombosis and can be either inherited or acquired. The majority of studies that have examined the risk of VTE in pregnancy have focused on inherited thrombophilic mutations. Although it has been reported that approximately 50 % of pregnancy-associated VTE are associated with inherited thrombophilia; these abnormalities are very common and collectively are present in at least 15 % of the population [144, 145].

As shown in Table 5, in a systematic review of nine case control studies (n = 2526) that evaluated the association between thrombophilia and pregnancy-associated VTE, the highest risks were associated with homozygosity for factor V Leiden or the prothrombin G20210A variant [146]. Pregnant women with the most common heritable thrombophilias (i.e. heterozygosity for factor V Leiden or the prothrombin G20210A variant) had lower risks. Deficiencies of antithrombin, protein C, and protein S were associated with moderate risk increases. Estimated absolute VTE risks, calculated using the provided odds ratios and a background incidence of VTE during pregnancy of approximately 1/1000 deliveries, suggest a low thrombosis risk (0.5–1.2 % of affected pregnancies) for most of the inherited thrombophilias, except perhaps for homozygous carriers of the factor V Leiden or the prothrombin mutations, where the risk estimate is approximately 4 % (Table 5). However, these findings are limited by the fact that most of the included women would not have had a family history of VTE. A positive family history of VTE increases the risk for VTE two- to four-fold, depending on the number of affected relatives [147, 148] and thrombophilic subjects without a personal or family history of DVT or PE have lower rates of VTE than patients with thrombophilia and a positive family history [149]. Family-based cohort studies not included in the above-mentioned systematic review suggest that the risks of developing a first VTE during pregnancy and the postpartum period are two to four times greater than estimated in thrombophilic women without a positive family history (Table 5) [150161]. However, it should be noted that many of the events occurred during the postpartum period and these risk estimates are very imprecise, particularly for the less common thrombophilias.

Table 5 Risks of pregnancy-related VTE in asymptomatic thrombophilic women

Acquired thrombophilias have been less well studied but repeated antiphospholipid antibody positivity (lupus anticoagulants [non-specific inhibitors], anticardiolipin antibodies, or anti-β2glycoprotein-I antibodies) is associated with an increased risk of VTE [162]. The risk or pregnancy-related VTE in women with antiphospholipid antibodies and no previous history of venous thrombosis is uncertain [163, 164].

There is considerable disagreement between current guidelines about the indication for antepartum thrombosis prophylaxis in pregnant women with deficiencies of antithrombin, protein C, or protein S. The inconsistency in recommendations likely results from the use of different risk thresholds for suggesting prophylaxis, uncertainty in risk estimates in recent studies, as described above, and concerns about VTE risks presented in older studies that suggested that these are high risk thrombophilias [165167]. However, this data is somewhat problematic as these papers have methodologic limitations, including acceptance of non-objectively diagnosed outcome events, failure to clearly specify criteria for the diagnosis of VTE, including recurrent VTE episodes in women who already had had a VTE, and the potential for referral and recall bias, that have the potential to lead to an overestimation of risk.

Prevention of pregnancy-associated VTE in patients with clinical risk factors

Most studies that have assessed clinical and pregnancy-related risk factors for VTE have utilized a case control or cross-sectional design (Table 6) [35, 7, 87, 168170]; although a few recent publications have used large databases to provide population-level absolute and relative risks for VTE [171, 172]. In methodologically stronger studies, most established risk factors have only a modest effect on VTE risk, with few increasing the absolute risk about 1 %. How combinations of independent risk factors might affect overall VTE risk has not been extensively studied and in most cases, it is unclear whether combinations result in additive or multiplicative risks. Further research in this area is required.

Table 6 Clinical risk factors for VTE as determined from case–control or cross-sectional studies

Prevention of pregnancy-associated VTE following cesarean delivery

Several observational studies have assessed the risk of VTE after cesarean delivery. Small prospective studies in which patients underwent screening ultrasounds following cesarean section and were then followed post-discharge for at least 6 weeks reported symptomatic VTE event rates of 0 (95 % CI 0–6.1 %) [173] and 0.5 % (95 % CI 0.1–2.8 %) [174]. The latter is consistent with estimates based on hospital discharge data that antedate the use of thrombosis prophylaxis [1, 175]. Emergency cesarean delivery approximately doubles the risk of VTE [7, 169, 172].

In the Cochrane systematic review mentioned above, four (840 women) of the nine included trials that examined prophylaxis following cesarean delivery compared heparin (UFH or LMWH) with placebo [119]. There was no evidence that using any form of heparin following delivery reduced the risk of maternal VTE (risk ratio [RR] vs no heparin for symptomatic events of 1.30; 95 % CI 0.39–4.27) and the authors concluded there was insufficient evidence on which to base recommendations.

Guidance Statement

Note: Given the uncertainty around optimal prophylactic strategies, all women should be provided with the opportunity to participate in shared decision making regarding this intervention, including a discussion of VTE risks, potential benefits (reduction in VTE risk) and drawbacks (risks of bleeding and localized skin reactions; cost; potential limitation of analgesic options at the time of delivery; anxiety associated with injections) of prophylaxis along with their values and preferences. Physicians and patients (and, perhaps, societies) with a lower threshold for recurrent VTE may choose a more aggressive anticoagulant strategy than recommended, whereas withholding prophylaxis may be appropriate in those who are willing to accept a higher risk of recurrence in order to forgo the drawbacks associated with prophylaxis.

General Comments

  • All pregnant women at risk of VTE should be educated about the signs and symptoms of DVT and PE and the need to seek urgent medical attention should they develop. Objective testing is mandatory if symptoms suspicious of DVT or PE occur.

  • All women should undergo an individualized risk assessment for VTE prior to pregnancy, once pregnancy is achieved and throughout pregnancy as new clinical situations arise.

  • When considering the use of thrombosis prophylaxis during pregnancy and/or the postpartum period, the absolute risk of VTE, the risk reduction with prophylaxis, drawbacks of prophylaxis, and the woman’s values and preferences should all be taken into account. Given the limitations of the available data, clinical vigilance rather than prophylaxis may also be acceptable for patients accepting the VTE risks quoted above and for whom the burden of LMWH prophylaxis outweighs potential benefits.

  • If the decision is made to use antepartum prophylaxis, it should be initiated early in pregnancy.

  • Six weeks of postpartum prophylaxis is recommended in patients with prior VTE and those with some thrombophilias. A shorter course of postpartum prophylaxis (until discharge or for 1–2 weeks post discharge) is suggested for women with transient risk factors.

Prevention of recurrent VTE

  • Pregnant women with prior VTE who are not receiving long-term anticoagulation should receive 6 weeks of postpartum prophylaxis.

  • Antepartum prophylaxis should be considered in pregnant women with prior unprovoked VTE or pregnancy- or estrogen-related VTE not receiving long-term anticoagulation.

Prevention of VTE in women with thrombophilia and no prior VTE

  • Asymptomatic women who are homozygous for the factor V Leiden mutation or prothrombin gene mutation and who have a family history of VTE should receive antepartum and postpartum prophylaxis.

  • Consideration should be given to providing postpartum prophylaxis in asymptomatic women who are heterozygous for the factor V Leiden mutation or prothrombin gene mutation or who have protein C or protein S deficiency and who have a family history of VTE.

  • Given the variability in the VTE risk estimates for asymptomatic women with antithrombin deficiency and a family history of VTE, either a strategy of antepartum and postpartum prophylaxis or postpartum prophylaxis alone is reasonable.

  • Asymptomatic women who are homozygous for the factor V Leiden mutation or prothrombin gene mutation and who have no family history of VTE should receive postpartum prophylaxis.

  • Asymptomatic women with all other thrombophilias who do not have a family history of VTE do not require prophylaxis, in the absence of other risk factors. However, given the variability in the VTE risk estimates for asymptomatic women with antithrombin deficiency and no family history of VTE, consideration could also be given to utilizing postpartum prophylaxis in these patients.

Prevention of VTE in women with clinical risk factors

  • Antepartum prophylaxis should be provided to immobilized (strict bedrest) women with a pre-pregnancy BMI of at least 25 kg/m 2 and to those with a prior history of VTE regardless of their BMI. Consideration should be given to providing prophylaxis during antepartum immobilization (as defined above) to women with a lower body mass index who have other significant comorbidities (e.g. systemic lupus erythematosus, sickle cell disease, heart disease) associated with an increased risk of VTE or a thrombophilia.

  • Consideration should be given to providing postpartum prophylaxis while in hospital to women with a history of antepartum immobilization (as defined above) for at least 7 days and to those who are immobilized postpartum who have a known thrombophilia or significant medical comorbidity

Prevention of VTE in after cesarean delivery

  • Prophylaxis should be provided after cesarean delivery to women with the following risk factors:

    • One or more of prior VTE, a history of antepartum immobilization (strict bedrest for at least 1 week), significant postpartum infection, postpartum hemorrhage of at least 1000 mL requiring re-operation, pre-eclampsia with growth restriction, significant medical co-morbidities (systemic lupus erythematosis, heart disease, or sickle cell disease) or a known thrombophilia.

    • Two or more of (or one or more in the setting of emergency cesarean delivery) of postpartum hemorrhage of at least 1000 mL that does not require re-operation, BMI >30 kg/m 2 , fetal growth restriction, pre-eclampsia, multiple pregnancy and tobacco use during pregnancy (at least 10 cigarettes per day)

  1. 5.

    How is peripartum anticoagulation managed?

The delivery options in women using anticoagulants are best considered by a multidisciplinary team in order to minimize the risks of maternal hemorrhage, epidural hematoma, and VTE around the time of delivery. In a systematic review of 2777 pregnancies in which LMWH was utilized in either therapeutic or prophylactic doses, postpartum hemorrhage of greater than 500 mL occurred in 26 pregnancies (0.94 %; 95 % CI 0.61–1.37 %) and wound hematoma in 17 pregnancies (0.61 %, 95 % CI 0.36–0.98 %) [59]. A more recent systematic review of 18 studies that focused solely on pregnant women (n = 981) receiving treatment for acute VTE during pregnancy reported an incidence of major bleeding during the first 24 h after delivery of 1.90 % (95 % CI 0.80–3.60 %) [60]. Although epidural hematomas in obstetrical patients receiving epidural analgesia/anesthesia are rare, with an estimated incidence of less than 1 in 150,000 [176]; the potential complications are devastating and include permanent neurologic dysfunction.

Depending on local practice, delivery options include spontaneous labor and delivery, induction of labor, and planned cesarean delivery. Induction of labor may help to avoid an unwanted anticoagulant effect during delivery (especially with neuroaxial anesthesia) in women receiving LMWH. Current guideline recommendations for management of anticoagulant therapy around the time of delivery are outlined in Table 2D. Anesthesia and obstetrical guidelines agree that 24 h should pass between the last dose of therapeutic LMWH and insertion of a neuroaxial catheter [17, 1923, 177]. For prophylactic LMWH, catheter insertion should occur no sooner than 10–12 h after the last LMWH dose [17, 1923, 177].

At some centers, women are converted from therapeutic adjusted-dose LMWH to subcutaneous twice daily therapeutic dose UFH in the last month of pregnancy. However, therapeutic doses of subcutaneous UFH may cause a persistent anticoagulant effect. One study reported that six of 11 women receiving subcutaneous UFH during pregnancy had an elevated aPTT at delivery despite discontinuing their injections at the onset of labour to 12 h prior to elective induction [178].

Patients should be instructed to withhold their injections if they believe that they have entered labor spontaneously. In centers with laboratory support that allows for rapid assessment of heparin levels, testing can be considered to guide anesthetic and surgical management; otherwise time since last injection should be used. If anticoagulation precludes regional techniques, alternative analgesic options include intravenous analgesia or general anesthesia for cesarean delivery [2022].

The potential increased risk of wound hematoma after cesarean delivery in patients receiving anticoagulant therapy has led to the suggestion that wound drains and closure techniques that allow easy hematoma drainage be considered in this population [179]. If bleeding occurs that is considered secondary to LMWH rather than an obstetric cause, protamine sulfate may provide partial neutralization [180].

Women diagnosed with proximal DVT or PE within two to 4 weeks of delivery are at very high risk for recurrent VTE with prolonged anticoagulant cessation [181, 182]. A strategy involving planned delivery with transition to intravenous UFH will minimize time off therapeutic anticoagulation [22, 23]. Discontinuation of intravenous UFH four to 6 h prior to the expected time of delivery or epidural insertion with a repeat activated partial thromboplastin time drawn after 4 h to confirm normalization will ensure that there is no residual anticoagulant effect. For the highest risk patients (e.g. VTE within 2 weeks), consideration can be given to insertion of a temporary inferior vena caval filter that can be removed postpartum.

Anticoagulants should be recommenced post-delivery as soon as adequate hemostasis is assured. Guidelines generally recommend resumption of prophylactic LMWH four to 12 h following delivery [17, 19]; and not sooner than 4 h after epidural catheter removal (with a longer delay for bloody or traumatic neuroaxial procedures) [177]. There are no definitive recommendations for resumption of full-dose LMWH following epidural catheter removal; however, it appears safe to do so 24 h of catheter removal (again, with a delay if placement was bloody or traumatic) [19, 22]. The timing of resumption of postpartum vitamin K antagonists for patients who choose this option remains controversial; some guidelines recommend a delay of at least 5 days [21, 22], although this recommendation appears based on the results of a single centre retrospective audit [183]. Once an INR of at least 2.0 is achieved, bridging LMWH can be discontinued.

Guidance Statement

  • All pregnant women receiving anticoagulants should have an individualized delivery plan that addresses obstetrical, anesthetic and thrombotic concerns.

  • All pregnant women should be advised to discontinue anticoagulant therapy upon the onset of spontaneous labor.

  • If there is a planned delivery, therapeutic LMWH should be discontinued at least 24 h prior to the expected time of epidural analgesia or delivery. Prophylactic LMWH should be stopped at least 10–12 h prior to epidural analgesia.

  • For planned deliveries, intravenously administered unfractionated heparin should be stopped at 4–6 h prior to the expected time of epidural analgesia or delivery and the aPTT checked to ensure normalization. For therapeutic doses of unfractionated heparin administered subcutaneously, the last dose should be given no sooner than 12 h and preferably closer to 24 h prior to expected time of epidural analgesia or delivery and the aPTT checked to ensure normalization. Guidelines differ in their requirement for a delay prior to epidural analgesia in patients receiving prophylactic dose unfractionated heparin up to 10,000 units daily; when possible prophylactic unfractionated heparin should be discontinued 8–10 h prior to planned procedures.

  • Prophylactic LMWH may be started/restarted 6–12 h after delivery (no sooner than 4 h after epidural catheter removal), as long as hemostasis is assured and there has not been a bloody or traumatic epidural. For prophylactic unfractionated heparin, the recommended time interval from epidural catheter removal is one to 8 h.

  • Therapeutic LMWH may be started/restarted 24 h after delivery (no sooner than 24 h after epidural catheter removal), as long as hemostasis is assured and there has not been a bloody or traumatic epidural. Attainment of therapeutic levels of intravenous unfractionated heparin should be delayed for the same amount of time.

Conclusion

Women are at increased risk of VTE during pregnancy and the postpartum period. Treatment and prevention of VTE in this patient population is complicated by the need to consider fetal, as well as maternal, wellbeing when making management decisions. Although our knowledge of risk factors for pregnancy-related VTE and the safe and effective use of anticoagulants used to prevent and treat VTE in this population continues to increase, there are still important gaps and high quality research in this area should be a priority. In the interim, all women should be provided with the opportunity to participate in shared decision making regarding their management. To make the best decisions, absolute risks and potential benefits of interventions, guideline recommendations, and patient values and preferences should all be taken into account. Table 7 summarizes these guidance statements.

Table 7 Summary of guidance statements