Key Abbreviations
Activated partial thromboplastin time aPTT
American College of Obstetricians and Gynecologists ACOG
Activated protein C APC
Adenosine diphosphate ADP
Antiphospholipid antibody APA
Antiphospholipid syndrome APS
Computed tomography CT
Computed tomographic pulmonary angiography CTPA
Deep venous thrombosis DVT
Disseminated intravascular coagulation DIC
Enzyme-linked immunosorbent assay ELISA
Factor V Leiden FVL
Heparin-induced thrombocytopenia HIT
Inferior vena cava IVC
International normalized ratio INR
Low-molecular-weight heparin LMWH
Magnetic resonance angiography MRA
Magnetic resonance imaging MRI
Protein Z–dependent protease inhibitor ZPI
Pulmonary embolus PE
Systemic lupus erythematosus SLE
Thrombin-activatable fibrinolysis inhibitor TAFI
Thromboxane A 2 TXA 2
Tissue factor TF
Tissue factor pathway inhibitor TFPI
Type 1 plasminogen activator inhibitor PAI-1
Unfractionated heparin UFH
Urokinase-type plasminogen activator uPA
Venous thromboembolism VTE
Venous ultrasonography VUS
Ventilation-perfusion scan V/Q scan
Background and Historic Notes
Pregnancy, childbirth, and the puerperium pose serious challenges to a woman’s hemostatic system. Whereas implantation, placentation, and uterine spiral artery remodeling lead to the development of the high-volume, high-flow, low-resistance uteroplacental circulation required for human fetal development, they require enhanced hemostatic responsiveness to avoid potentially fatal hemorrhage. The price paid for this essential hemostatic adaptation to human hemochorial placentation is an increased risk of superficial and deep venous thrombosis (DVT) and pulmonary embolus (PE). Acquired or inherited thrombophilias, obesity, advanced maternal age, advanced parity, antepartum hospitalizations, surgery, and infection are major risk factors for DVT and PE in pregnancy and the puerperium. The expeditious identification and prompt treatment of thrombotic events is critical to avoiding death and serious postphlebitic sequelae.
Diagnoses and Definitions
Thrombosis is the obstruction or occlusion of a vessel by a blood clot. Venous thromboembolism (VTE) includes venous thrombosis of the deep venous system of the lower (common) or upper (uncommon) extremity (DVT). Thrombosis or inflammation of the superficial venous system is generally not associated with morbidity, although in some cases it can develop into or be associated with DVT or PE. In fact, 10% to 20% of superficial thrombosis cases in nonpregnant patients are associated with DVT. Pulmonary embolus is the obstruction of the pulmonary artery or one of its branches, arising from a clot from a DVT in approximately 90% of cases. A majority of PE cases are due to deportment of thrombus from the lower extremities; for the purposes of this chapter, pulmonary embolus will refer to VTE of the pulmonary vasculature (rather than air, fat, or amniotic fluid embolism).
Symptoms
Deep Venous Thrombosis
Clinical findings typical of DVT include erythema, warmth, pain, edema, and tenderness localized to the area of the thrombosis. Occasionally, a palpable cord may be present that corresponds to a thrombosed vein. Homans sign is pain and tenderness elicited on compression of the calf muscles by squeezing the muscles or by dorsiflexion of the foot. These are rather nonspecific signs and symptoms that involve a broad differential diagnosis, including cellulitis, ruptured or strained muscle or tendon, trauma, ruptured popliteal (Baker) cyst, cutaneous vasculitis, superficial thrombophlebitis, and lymphedema. The specificity of these manifestations is less than 50%, and among patients with these signs and symptoms, the diagnosis of DVT is confirmed by objective testing in only approximately one third of the group.
Pulmonary Embolism
Tachypnea (>20 breaths/min) and tachycardia (>100 beats/min) are present in 90% of patients with acute PE, but these findings lack specificity and generate a broad differential diagnosis. Presyncope and syncope are rarer symptoms and indicate a massive embolus.
Epidemiology and Incidence
Occurring in approximately in 1 in 1500 pregnancies, VTE is a relatively uncommon disorder but is a leading cause of mortality and serious morbidity in pregnant women. This rate represents a nearly tenfold increase compared with nonpregnant women of comparable childbearing age. According to the most recent U.S. vital statistics, from 2006 through 2010, VTE was a leading cause of maternal mortality that contributed to 9% of pregnancy-related deaths. Classic teaching viewed the postpartum period as the period of maximal thrombotic occurrence. However, management styles of prior eras that included prolonged puerperal bed rest and estrogen to suppress lactation likely inflated this risk. More recent studies have shown that a majority of thromboembolic events happen in the antepartum period. Given its shorter duration, and after adjusting for duration of exposure, the day-to-day relative risk of VTE is about threefold to eightfold higher in the puerperium. New evidence suggests that the risk for a thrombotic event extends out to 12 weeks postpartum, although the absolute increase in risk is quite low after 6 weeks.
Genetics
It is well known that inherited mutations in various components of the coagulation cascade, the so-called inherited thrombophilias, contribute to significant risk for thrombosis, especially in the presence of other risk factors such as pregnancy, surgery (e.g., cesarean delivery), trauma, infection, or immobility. Factor V Leiden is the most common mutation and accounts for over 40% of inherited thrombophilias in most studies. Most of these genetic mutations act in an autosomal-dominant manner, thus one mutation will incur an elevated risk for VTE; individuals with two mutations will have higher risks for thrombotic events than those with one. Patients with a strong family history of thrombotic events who have screened negative for the panel of known thrombophilia mutations likely have an as yet unrecognized gene defect in a specific component of the coagulation cascade. The details of the known inherited thrombophilias are discussed later in this chapter.
Physiology of Hemostasis
Vasoconstriction and Platelet Action
Vasoconstriction and platelet activity play a primary initial role in limiting blood loss following vascular disruption and endothelial damage. Vasoconstriction limits blood flow and also limits the size of thrombus necessary to repair the defect. Platelet adherence to damaged vessels is mediated by the formation of von Willebrand factor (vWF) “bridges” anchored at one end to subendothelial collagen and at the other to the platelet glycoprotein Ib (GP Ib)/factor IX/V receptor. Platelet adhesion stimulates release of α-granules that contain vWF, thrombospondin, platelet factor 4, fibrinogen, β-thromboglobulin, and platelet-derived growth factor as well as dense granules that contain adenosine diphosphate (ADP) and serotonin. These latter molecules, when combined with the release of thromboxane A 2 (TXA 2 ), contribute to further vasoconstriction and platelet activation. In addition, ADP causes a conformational change in the platelet GP IIb/IIIa receptor that promotes aggregation by forming interplatelet fibrinogen, fibronectin, and vitronectin bridges.
Coagulation Cascade
Platelet action alone is insufficient to provide adequate hemostasis in the face of a substantial vascular insult; in this setting, the coagulation cascade—with resultant fibrin plug formation—is required to restore hemostasis. Tissue factor (TF), a cell membrane-bound glycoprotein, is the primary initiator of the coagulation cascade. It is expressed constitutively by epithelial, stromal, and perivascular cells throughout the body and in abundance by endometrial stromal cells and the pregnant uterine decidua. TF is also present in low concentrations in blood, on activated platelets, and in high levels in amniotic fluid, which accounts for the coagulopathy seen in amniotic fluid embolism. It is interesting to note that although intrauterine survival is possible in the absence of platelets or fibrinogen, it is not possible in the absence of TF. Clotting is initiated by the binding of TF to factor VII, the only clotting factor with intrinsic coagulation activity in its zymogenic form ( Fig. 45-1 ).
Following endothelial injury and in the presence of ionized calcium, perivascular cell- or platelet-bound TF comes into contact with factor VII on anionic cell membrane phospholipids. Factor VII has low intrinsic clotting activity but can be autoactivated after binding to TF, or it can be activated by thrombin or activated factors such as IXa, Xa, or XIIa. The TF–activated factor VII (VIIa) complex initiates the elements of the coagulation cascade by activating both factors IX and X. Activated factor IX (IXa) complexes with its cofactor VIIIa to indirectly activate X. Once generated, Xa binds with its cofactor Va to convert prothrombin (factor II) to thrombin (factor IIa). Cofactors V and VIII can each be activated by either thrombin or Xa, and XIIa activates XI on the surface of activated platelets, which provides an alternative route to IX activation. Factor XII can be activated by kallikrein/kininogen as well as by plasmin. The key event of hemostasis occurs when thrombin cleaves fibrinogen to produce fibrin. Fibrin monomers self-polymerize and are cross-linked by thrombin-activated factor XIIIa. Although TF is the initiator of hemostasis, thrombin is the ultimate arbiter of clotting; it not only activates platelets and generates fibrin, it also activates critical clotting factors and cofactors (V, VII, VIII, XI, and XIII). Figure 45-1 provides a diagram of the interaction of the various components of the coagulation cascade.
Anticoagulant System
The risk of thrombosis, the inappropriate and excessive activation of the clotting cascade, is restrained by the anticoagulant system (see Fig. 45-1 ). Evidence shows that the coagulation system “idles” like a car engine to quickly respond to vascular injury, and thus the anticoagulant system performs the critical role of preventing inappropriate acceleration of clotting. Tissue factor pathway inhibitor (TFPI) binds to the prothrombinase complex (factor Xa/TF/factor VIIa) to stop TF-mediated clotting. However, as noted, factor XIa generation can bypass this block. Moreover, in the 10 to 15 seconds before TFPI-mediated prothrombinase inhibition, sufficient quantities of factors Va, VIIIa, IXa, and Xa and thrombin are generated to sustain clotting for some time. As a result, additional physiologic anticoagulant molecules are required to maintain blood fluidity.
Paradoxically, thrombin also plays a pivotal role in the anticoagulant system by binding to thrombomodulin, which causes a conformation change that allows it to activate protein C. The activated protein C (APC) molecule binds to anionic endothelial cell membrane phospholipids on damaged vessels or to the endothelial cell protein-C receptor (EPCR) to inactivate factors Va and VIIIa. Protein S is an important cofactor in this process because it enhances APC activity. Factor Va is also a cofactor in APC-mediated factor VIIIa inactivation.
Factor Xa can also be inhibited by the protein Z–dependent protease inhibitor (ZPI). When ZPI forms a complex with its cofactor, protein Z, its inhibitory activity is enhanced a thousandfold, although ZPI can also inhibit factor XIa independent of protein Z. Deficiencies of protein Z can promote both bleeding and thrombosis, although the latter predominates particularly in the presence of other thrombophilias.
Thrombin activity is modulated by a number of serine protease inhibitors—such as heparin cofactor II, α-2 macroglobulin, and antithrombin—which serve to inactivate thrombin and Xa. The most active inhibitor within this group is antithrombin, which binds to either thrombin or factor Xa and then to heparin or other glycosaminoglycans, augmenting antithrombin’s rate of thrombin inactivation more than a thousandfold. The other two inhibitors work in a similar fashion to inhibit thrombin.
Clot Lysis and Fibrinolysis
Fibrinolysis is a further critical element in preventing overwhelming thrombosis (see Fig. 45-1 ). Tissue-type plasminogen activator (tPA), an endothelial enzyme metabolized by the liver, becomes embedded in fibrin and cleaves plasminogen to generate plasmin, which in turn cleaves fibrin into fibrin degradation products; the latter are indirect measures of fibrinolysis. These fibrin degradation products can also inhibit thrombin action, a favorable effect when production is limited, but a contributor to disseminated intravascular coagulation (DIC) when production is excessive. A second plasminogen activator, urokinase-type plasminogen activator (uPA), is produced by endothelial cells. A series of fibrinolysis inhibitors also prevent hemorrhage from premature clot lysis. The α-2 plasmin inhibitor is bound to the fibrin clot, where it prevents premature fibrinolysis. Platelets and endothelial cells release type 1 plasminogen activator inhibitor (PAI-1), an inactivator of tPA. In pregnancy, the decidua is also a rich source of PAI-1, whereas the placenta produces mostly type 2 (PAI-2). The thrombin-activatable fibrinolysis inhibitor (TAFI) is another fibrinolytic inhibitor that is also activated by the thrombin-thrombomodulin complex. TAFI modifies fibrin and renders it resistant to inactivation by plasmin.
Pathophysiology of Thrombosis in Pregnancy
Characteristic physiologic changes in decidual and systemic hemostatic systems occur in pregnancy in preparation for the hemostatic challenges of implantation, placentation, and childbirth. Decidual TF and PAI-1 expression are greatly increased in response to progesterone, and levels of placental-derived PAI-2, which are negligible prior to pregnancy, increase until term. Pregnancy is associated with systemic changes that enhance hemostatic capability and promote thrombosis. For example, a doubling occurs in circulating concentrations of fibrinogen, and 20% to 1000% increases are seen in factors VII, VIII, IX, X, and XII, all of which peak at term in preparation for delivery. Levels of vWF also increase up to 400% at term. In contrast, levels of prothrombin and factor V remain unchanged, and levels of factor XIII and XI decline modestly. Concomitantly, there is a 40% to 60% decrease in the levels of free protein S, conferring an overall resistance to activated protein C. Further reductions in free protein-S concentrations are caused by stress, cesarean delivery, and infection; this accounts for the high rate of PE following cesarean deliveries, particularly in association with prolonged labor and endomyometritis. Coagulation parameters may normalize as early as 3 weeks postpartum, but they generally return to baseline at 6 to 12 weeks.
The risk of thrombosis in pregnancy is also related to physical changes in the gravid woman. Venous stasis in the lower extremities results from compression of the inferior vena cava (IVC) and pelvic veins by the enlarging uterus . Despite the presence of the sigmoid colon promoting uterine dextrorotation, ultrasound findings indicate lower flow velocities in the left leg veins throughout pregnancy. This would explain why multiple studies have confirmed that the incidence of thrombosis is far greater in the left leg than in the right . Hormone mediated increases in deep vein capacitance secondary to increased circulating levels of estrogen and local production of prostacyclin and nitric oxide also contribute to the increased risk of thrombosis.
Antiphospholipid Syndrome
Overall, antiphospholipid syndrome (APS) is responsible for approximately 14% of thromboembolic events in pregnancy . The diagnosis of APS requires the presence of prior or current vascular thrombosis or characteristic obstetric complications together with at least one laboratory criterion, such as anticardiolipin antibodies (immunoglobulin G [IgG] or IgM greater than 40 GPL [1 GPL unit is 1 µg of IgG antibody] or 40 MPL [1 MPL unit is 1 µg of IgM antibody] or greater than the 99th percentile), anti–β-2 glycoprotein-I (IgG or IgM greater than the 99th percentile), or lupus anticoagulant. Refer to Chapter 46 for details of diagnosis and treatment of APS.
The antiphospholipid antibodies (APAs) are a class of self-recognition immunoglobulins whose epitopes are proteins bound to negatively charged phospholipids. These antibodies must be present on two or more occasions at least 12 weeks apart for diagnosis and are present in 2.2% of the general obstetric population, and most affected patients have uncomplicated pregnancies. Thus providers should use caution when ordering and interpreting tests in the absence of APS-qualifying clinical criteria.
APS has been associated with both venous (DVT, PE) and arterial vascular events (stroke). A meta-analysis of 18 studies has shown elevated risk of DVT, PE, and recurrent VTE among patients with systemic lupus erythematous (SLE) who test positive for APA. Overall, when compared with those SLE patients who do not test positive for either test, those with lupus anticoagulants and anticardiolipin antibodies have a respective sixfold and twofold increased risk of venous thrombosis. These antibodies also pose a risk to patients without SLE. The lifetime prevalence of arterial or venous thrombosis in affected non-SLE patients is approximately 30%, with an event rate of 1% per year . The risks of thromboembolic events are highly dependent on the presence of other predisposing factors that include pregnancy, estrogen exposure, immobility, surgery, and infection. As noted above, APS has also been associated with adverse pregnancy outcome and accounts for 14% of VTE in pregnancy. In fact, the risk of a thrombotic event in pregnancy is 5% even with prophylaxis. All patients who present with VTE in pregnancy or in the postpartum period should have an appropriate APS workup.
Inherited Thrombophilias
The inherited thrombophilias are a heterogeneous group of genetic disorders associated with arterial and venous thrombosis as well as fetal loss. As with APAs, the occurrence of a thromboembolic event is highly dependent on other predisposing factors such as pregnancy, exogenous estrogens, immobility, obesity, surgery, infection, trauma, and the presence of other thrombophilias. However, the most important risk modifier is a personal or family history of venous thrombosis. Table 45-1 presents the prevalence and risk of venous thrombosis among pregnant patients with and without a personal or family history of venous thrombosis. As noted, the thrombophilias are divided into high and low risk based on the overall risk of VTE. The screening for and management of inherited thrombophilias during and around the time of pregnancy has been recently addressed by American College of Obstetricians and Gynecologists (ACOG). All patients who present with VTE in pregnancy or postpartum should be considered for an appropriate workup for inherited thrombophilias.
| RISK | THROMBOPHILIA TYPE | PREVALENCE IN THE EUROPEAN POPULATION | PREVALENCE IN PATIENTS WITH VTE IN PREGNANCY | RR/OR OF VTE IN PREGNANCY (95% CI) | PROBABILITY OF VTE IN PREGNANT PATIENTS WITH PERSONAL OR FAMILY HX | PROBABILITY OF VTE IN PREGNANT PATIENTS WITHOUT PERSONAL OR FAMILY HX | STUDY |
|---|---|---|---|---|---|---|---|
| High risk | FVL homozygous | 0.07% * | <1% * | 25.4 (8.8-66) | ≫10% | 1.5% | |
| Prothrombin gene G20210A mutation homozygous | 0.02% * | <1% * | N/A | ≫10% | 2.8% | ||
| Antithrombin III deficiency | 0.02%-1.1% | 1%-8% | 119 | 11%-40% | 3.0%-7.2% | ||
| Compound heterozygous (FVL/prothrombin G20210A) | 0.17%+ | <1%+ | 84 (19-369) | 4.7% (overall probability of VTE in pregnancy) | |||
| Low risk | FVL heterozygous | 5.3% | 44% | 6.9 (3.3-15.2) | >10% | 0.26% | |
| Prothrombin G20210A mutation heterozygous | 2.9% | 17% | 9.5 (2.1-66.7) | >10% | 0.37%-0.5% | ||
| Protein C deficiency | 0.2%-0.3% | <14% | 13.0 (1.4-123) | NA | 0.8%-1.7% | ||
| Protein S deficiency | 0.03%-0.13% | 12.4% | NA | NA | <1%-6.6% | ||
Recent prospective studies have suggested that lower-risk inherited thrombophilias may have an even weaker association with maternal thrombosis than that reported by the retrospective studies cited in Table 45-1 . For example, a prospective study of 4885 low-risk women screened in the first trimester of pregnancy noted that 134 (2.7%) carried the factor V Leiden mutation, but none had a thromboembolic event during pregnancy or the in puerperium (95% confidence interval [CI], 0% to 2.7%). In two other prospective studies that involved 584 Irish and 4250 British pregnant women again screened for factor V Leiden early in pregnancy, no thrombotic episodes were noted among carriers. Said and associates tested 1707 Australian nulliparous women blindly for factor V Leiden, the prothrombin gene G20210A mutation, and a thrombomodulin polymorphism prior to 22 weeks and reported an expected prevalence of heterozygosity for the factor V Leiden and prothrombin G20210A gene mutations and homozygosity for the thrombomodulin polymorphism of 5.39%, 2.38%, and 3.51%, respectively; again, none of the patients developed VTEs. However, one prospective study of 2480 women tested for activated protein-C resistance/factor V Leiden early in pregnancy observed that affected patients had an eightfold increase in VTE. Thus the true risk of thrombosis in patients with low-risk inherited thrombophilias is probably lower than that suggested by retrospective case-control and cohort studies and is also likely dependent on the presence of concomitant risk factors such as a strong family history, obesity, and surgery.
Risk Factors and Associations
Virchow triad—vascular stasis, hypercoagulability, and vascular trauma—describes the three classic antecedents to thrombosis, and many of the physiologic changes of pregnancy contribute to these criteria. Other pregnancy-specific risk factors for thrombosis include increased parity, postpartum endomyometritis, operative vaginal delivery, and cesarean delivery. The latter is associated with a substantial ninefold increase in VTE risk compared with vaginal delivery. Risk factors not unique to pregnancy include age greater than 35 years, obesity, trauma, immobility, infection, smoking, nephrotic syndrome, hyperviscosity syndromes, cancer, surgery (particularly orthopedic procedures), and a history of DVT or PE ( Table 45-2 ). Admission to the hospital in pregnancy may be associated with a seventeenfold increased risk for VTE compared with a nonhospitalized cohort, with this risk remaining high (sixfold) for the 28 days after admission. In vitro fertilization (IVF) has been shown to increase the risk for VTE in the first trimester in a cross-sectional study in Sweden, although the overall absolute risk is still rather low.
| GENERAL | PREGNANCY RELATED |
|---|---|
| Age >35 years | Increased parity |
| Obesity | Postpartum endomyometritis |
| Trauma | Operative vaginal delivery |
| Immobility | Cesarean delivery |
| Infection | |
| Smoking | |
| Nephrotic syndrome | |
| Hyperviscosity syndromes | |
| Cancer | |
| Surgery, especially orthopedic | |
| Prior deep venous thromboembolism or pulmonary embolism | |
| Hospital admission |
Complications
Thromboembolism is associated with serious complications that include arrhythmia, hypoxia, pulmonary hypertension, heart failure, and postthrombotic syndrome of the extremities. Thromboembolism is a major cause of death worldwide; thus prompt diagnosis and treatment is a priority. When confronted with the signs and symptoms suggestive of VTE, rapid initiation of the workup and treatment is essential to avoid complications. Complications of anticoagulation, such as bleeding or thrombocytopenia, are also a reality and should be avoided.
Considerations in Management of Pregnant Women
Pregnancy and the postpartum period are considered a high-risk period for thromboembolism. Key considerations in the workup and management of pregnant women are the selection of appropriate diagnostic tools and anticoagulation regimens with special concern regarding pregnancy-related changes and fetal exposures.
Diagnosis of Venous Thromboembolism
Deep Venous Thrombosis
Clinical Signs and Symptoms
The clinical findings typical of DVT include erythema, warmth, pain, edema, and tenderness localized to the area of the thrombosis. Occasionally, a palpable cord may correspond to a thrombosed vein. Homans sign is the pain and tenderness elicited on compression of the calf muscles by squeezing the muscles or by dorsiflexion of the foot. These are rather nonspecific signs and symptoms that involve a broad differential diagnosis that includes cellulitis, ruptured or strained muscle or tendon, trauma, ruptured popliteal (Baker) cyst, cutaneous vasculitis, superficial thrombophlebitis, and lymphedema. In fact, as noted above, evidence suggests that the specificities of these manifestations are less than 50%, and among patients with these signs and symptoms, the diagnosis of DVT is confirmed by objective testing in only approximately one third of the group.
Risk-Scoring System
Work by Chan and colleagues demonstrated that subjective assessment by “thrombosis experts” of clinical risk of DVT in symptomatic pregnant patients can categorize patients prior to diagnostic testing into two groups, one low risk (1.5% prevalence, 98.5% negative predictive value) and the other non-low risk (25% prevalence). Three factors—symptoms in the left leg, a leg circumference discrepancy of 2 cm or more, or first-trimester presentation—were highly predictive of DVT. Furthermore, this study and one additional validation study demonstrated that patients who lack all three criteria were at no risk for VTE. Thus determination of prediagnostic study risk can be helpful before proceeding to the selection and interpretation of diagnostic tests.
Imaging
Contrast venography, an invasive technique that involves the injection of dye into a vein distal to the site of suspected thrombosis, followed by radiographic imaging is now rarely used for diagnosis of DVT. The risks of radiation and contrast allergy, as well as its technical difficulty, particularly preclude its use in pregnancy.
The most common diagnostic modality used in the evaluation of patients with suspected DVT is venous ultrasonography (VUS) with or without color Doppler. This modality has virtually replaced the cumbersome and less accurate impedance plethysmography technique. The ultrasound transducer is placed over the common femoral vein beginning at the inguinal ligament and is then sequentially moved to image the greater saphenous vein, the superficial femoral vein, and the popliteal vein to its trifurcation with the deep veins of the calf. Compression VUS involves the application of pressure with the probe to determine whether the vein under investigation is compressible. The most accurate ultrasonic criterion for diagnosing venous thrombosis is noncompressibility of the venous lumen in a transverse plane under gentle probe pressure using duplex and color flow Doppler imaging. The overall sensitivity and specificity of VUS has been reported at 90% to 100% for proximal vein thromboses, but traditionally it has been considered lower for the detection of calf vein thromboses. This is confirmed by a more recent meta-analysis, which demonstrated that in nonpregnant patients, duplex ultrasound has a pooled sensitivity of 96.4% for proximal (knee or thigh) DVT and 75.2% for distal (calf) DVT, and total specificity is 94.3%.
Magnetic resonance imaging (MRI) appears to have equivalent test performance characteristics to VUS. One meta-analysis demonstrated a sensitivity of 91.5% and specificity of 94.8% for diagnosis of DVT in nonpregnant patients with suspected DVT or PE. Similar to VUS, sensitivity is also improved for distal, compared with proximal, thrombi (93.9% vs. 62.1%). The advantage of MRI may be in the ability to detect more centralized DVT, like those in the pelvic, iliac, or femoral veins.
D-Dimer Assays
Laboratory evaluation of serum concentrations of D-dimer, a product of the degradation of fibrin by plasmin, has been increasingly advocated as a helpful test in the diagnosis of DVT in the nonpregnant population. Testing relies on the use of monoclonal antibodies to D-dimer fragments. The most accurate and reliable tests for D-dimer appears to be two rapid enzyme-linked immunosorbent assays (ELISAs; Instant-IA D-dimer [Stago] and Vidas DD [bioMérieux]) and a rapid whole blood assay (SimpliRED D-dimer [Agen Biomedical]). The test is limited by factors that may contribute to false-positive testing, including pregnancy, postpartum and postoperative periods, and superficial thrombophlebitis. In particular, normal pregnancy causes a physiologic increase in D-dimer, with levels that exceed the threshold for normal in 78% and 100% of patients in the second and third trimesters, respectively. The performance of the SimpliRED D-dimer test was evaluated by Chan and colleagues in a prospective cohort of pregnant patients at risk for DVT. In this population, who had a DVT prevalence of 8.7%, sensitivity of D-dimer testing was 100% and specificity was 60%. Utility in particular appeared to be stratified across gestation; false-positive rates were 0%, 24%, and 51% for the first, second, and third trimesters, respectively. The value of D-dimer testing in pregnancy may lie in its ability to rule out disease because this same study showed a negative predictive value of 100% with a 95% confidence interval of 95% to 100%. The value of D-dimer testing in the late second and third trimesters is likely to be lower than in the first half of pregnancy because patients in the late second and third trimester will likely have levels above the threshold. Thus at this time, although D-dimer appears useful as a test to rule out a DVT, its routine use in pregnancy cannot be endorsed.
Workup of Patients with Suspected Deep Venous Thrombosis
The diagnostic paradigm outlined in Figure 45-2 can be used to diagnose DVT in pregnant patients with maximal sensitivity and specificity. Risk may be stratified as in the section above (see “ Risk-Scoring System ”).
Pulmonary Embolus
Clinical Signs and Symptoms
Tachypnea (>20 breaths/min) and tachycardia (>100 beats/min) are present in 90% of patients with acute PE, but these findings lack specificity and generate a broad differential diagnosis. Presyncope and syncope are rarer symptoms and indicate a massive embolus.
Nonspecific Studies
The classic electrocardiography (ECG) changes associated with PE are the S1, Q3, and inverted T3. Other findings include nonspecific ST changes, right bundle branch block, or right axis deviation. These findings are usually associated with cor pulmonale and right-sided heart strain or overload, reflective of more serious cardiopulmonary compromise. About 26% to 32% of patients with massive PE had the above-mentioned ECG changes. These findings are why the Royal College of Obstetricians and Gynaecologists (RCOG) have recommended that an ECG be performed in women who present with signs and symptoms of an acute PE. Arterial blood gases and oxygen saturation have limited value in the assessment of acute PE, particularly in a pregnant population. Measurements of PO 2 are greater than 80 mm Hg in 29% of PE patients younger than 40 years of age. In another study, up to 18% of patients with PE had PO 2 measurements of more than 85 mm Hg.
The chest radiograph may be abnormal in up to 84% of affected patients. The common findings on radiography are pleural effusion, pulmonary infiltrates, atelectasis, and elevated hemidiaphragm. The eponymous findings of pulmonary infarction such as a wedge-shaped infiltrate (Hampton hump) or decreased vascularity (Westermark sign) are rare. Although a normal chest radiograph in the setting of dyspnea, tachypnea, and hypoxemia in a patient without known preexistent pulmonary or cardiovascular disease is suggestive of PE, a chest radiograph cannot be used to confirm the diagnosis. It is useful, however, in the workup for PE as a tool for selecting the correct diagnostic modality (ventilation-perfusion [V/Q] scan vs. computed tomography [CT] scan; see below).
A large PE can create changes consistent with cor pulmonale and right-sided heart strain. Large emboli in the main pulmonary artery and its primary branches can result in acute right-sided heart failure, which is the ultimate cause of death in most patients with PE. Between 30% and 80% of patients with PE display echocardiographic abnormalities in right ventricular size or function. Typical findings include a dilated and hypokinetic right ventricle, tricuspid regurgitation, and absence of preexisting pulmonary arterial or left-sided heart pathology. Transesophageal echocardiography with or without contrast appears to improve the imaging of main or right pulmonary artery emboli and, occasionally, of left pulmonary artery clots. For these reasons, bedside echocardiography can be very useful for the unstable patient or the patient unable to travel for other imaging studies.
Ventilation-Perfusion Scanning.
Perfusion scanning uses intravenously injected radioisotope-labeled albumin macroaggregates that deposit in the pulmonary capillary bed. Ventilation scanning involves the inhalation of radiolabeled aerosols, whose distribution is evaluated by gamma camera. The comparison of these two images allows for interpretation of characteristic patterns that are then used to assign diagnostic probabilities (high, intermediate, or low). More than 90% of high-risk patients with high probability ventilation-perfusion (V/Q) scans have a PE, whereas less than 6% of low-risk patients with low-probability scans have a PE. Given that most young, healthy women have little underlying lung pulmonary pathology, the diagnostic efficacy of V/Q scanning in pregnancy is substantially higher than that in older, nonpregnant patients .
Spiral Computed Tomographic Pulmonary Angiography.
Spiral CT scanning (computed tomographic pulmonary angiography [CTPA]) requires injection of intravenous (IV) contrast while simultaneously imaging the distribution of contrast in the pulmonary vasculature with a CT scanner. Although this can be an effective judge of large, segmental, and central emboli, CT is of limited value with small subsegmental vessels and horizontally oriented vessels in the right middle lobe. One strength of CTPA is its utility in diagnosing nonembolic pulmonary phenomena such as pneumonia or pulmonary edema.
In nonpregnant patients, CTPA is often the modality of choice; however, this is not likely the case in pregnancy. Several studies have shown that CTPA performs less well in pregnant compared with nonpregnant cohorts. Andreou and associates demonstrated in a small study of 32 patients with suspected PE that contrast enhancement of the pulmonary arteries is reduced in pregnant women compared with nonpregnant women, likely due to the increased cardiac output of pregnancy. Another reason may be the more frequent dilution or interruption of contrast by nonopacified blood from the lower vena cava in pregnancy. Comparing CTPA imaging in 40 pregnant patients with that in 40 nonpregnant patients, U-King-Im and colleagues also demonstrated a threefold higher rate of suboptimal studies in the pregnant population (27.5% vs. 7.5%).
Ventilation-Perfusion Scanning Versus Computed Tomographic Pulmonary Angiography.
V/Q scanning and CTPA have been directly compared in pregnant patients by Cahill and colleagues, who retrospectively evaluated a cohort of 304 pregnant or postpartum women evaluated for pulmonary embolism. CTPA was performed in 108 women, and a V/Q scan was done in 196. In women with a normal chest radiograph, CTPA yielded nondiagnostic results 5.4 times more often than V/Q scanning (30.0% vs. 5.4%). As would be expected, in those patients with an abnormal chest radiograph, V/Q scanning was more often nondiagnostic. Ridge and colleagues evaluated these two techniques prospectively in a total of 50 patients. Again, they found that CTPA more often gave a nondiagnostic result (35.7% vs. 2.1%), and V/Q scanning was adequate for diagnosis more frequently (35.7% vs. 4%). For these reasons, as well as considerations of radiation exposure (see below), V/Q scanning is the modality of choice for pregnant patients with suspected PE and a normal chest radiograph .
Magnetic Resonance Angiography.
Magnetic resonance angiography (MRA) may be performed using MRI during IV injection of gadolinium. Faster image-acquisition rates and improved image timing to respiratory and cardiac motion have allowed this technique to be utilized. Preliminary experience by Meaney and colleagues suggested a sensitivity of 100%, specificity of 95%, and positive and negative predictive values of 87% and 100%, respectively, for MRA in 30 patients also assessed by classic pulmonary angiography. Another prospective study that involved 141 patients showed an overall sensitivity of only 77% in comparison to pulmonary angiography, with the sensitivity broken down to 40%, 84%, and 100% for isolated subsegmental, segmental, and central pulmonary emboli, respectively. Because MRA does not involve ionizing radiation, it is an appealing alternative to CT scanning and angiography for pregnancy, and further assessment in large trials are needed to prove its ultimate utility as a primary diagnostic modality. Gadolinium contrast crosses the placenta and enters the amniotic fluid through excretion by the fetal kidneys. A majority of animal studies suggest no teratogenic effects by gadolinium, and it is considered a category C agent by the U.S. Food and Drug Administration (FDA), although it should be used with caution.
D-Dimer Assays.
The D-dimer assay appears to have much lower sensitivity in patients with PE and thus it is not helpful as a test for ruling out disease. Negative D-dimer testing has been observed in patients with confirmed PE. Damodaram and associates report a sensitivity and specificity of 73% and 15% for D-dimer in testing for suspected PE in pregnancy. The difference in D-dimer test performance for PE in pregnancy may be due to the smaller clot burden compared with DVT coupled with the increase in intravascular plasma volume of pregnancy. This unacceptably high false-negative rate indicated no role for D-dimer testing in the workup of pregnant patients with suspected PE .
Lower Extremity Evaluation.
Most PEs (90%) arise from lower extremity DVTs, and among patients with the diagnosis of PE, half will be found to harbor a lower extremity DVT; this includes up to 20% of PE patients without signs or symptoms of lower extremity DVT . In patients who present with signs or symptoms of PE who also have left-sided lower extremity symptoms, it is reasonable to begin with lower extremity VUS to detect DVT because the need for therapeutic anticoagulation is similar. This avoids exposure to the ionizing radiation, as well as the burden of more complicated testing, associated with CTPA and V/Q scanning. Furthermore, in high-risk patients in whom CTPA or V/Q scanning is nondiagnostic or even negative, evaluation of the leg veins for DVT can be helpful to reinforce results. In such cases, however, a negative VUS study is still associated with a 25% risk of PE, which suggests further studies are generally needed.
Workup of Patients with Suspected Pulmonary Embolus
Workup of patients with suspected PE should begin with evaluation of their cardiorespiratory status to determine whether they are critically ill; this would include an ECG and chest radiograph. Patients who are stable hemodynamically with favorable oxygenation status (oxygen saturation >80%) should be evaluated in a carefully ordered fashion, taking into account the possibility of lower extremity DVT and the results of the chest radiograph ( Fig. 45-3 ), according to an algorithm advocated by a panel of experts from the American Thoracic Society, the Society of Thoracic Radiology, and ACOG. If lower extremity symptoms are present, particularly if left sided, VUS can be performed; if results are positive, therapeutic anticoagulation should begin. Chest radiography should definitely be performed when VUS is negative or when no lower extremity signs or symptoms are apparent. Although the chest radiograph is not typically used in the actual diagnosis, it allows for proper selection of the next diagnostic test. V/Q scan is performed if the chest radiograph is normal . A positive (high or moderate probability) V/Q scan should prompt therapeutic anticoagulation. A negative test should exclude the diagnosis. A nondiagnostic result—such as intermediate probability/equivocal test in any patient and a low-probability result in a high-risk patient (one with prior VTE, known thrombophilia, family history of thrombosis in a first-degree relative under 50 years of age, or other clinical risk factors)—should trigger performance of CTPA.
For such patients, and for those with an abnormal chest radiograph, CTPA is the diagnostic modality of choice. Positive tests will prompt initiation of anticoagulation. In the patient with an abnormal chest radiograph who has a negative CTPA, the true etiology of the chest pathology is often seen on CT. Alternatively, if the CTPA is nondiagnostic, further testing is performed, such as MRA or serial lower extremity VUS studies.
In the critically ill patient, anticoagulation should be started—in the absence of contraindications—and a similar protocol is followed ( Fig. 45-4 ). Very unstable patients who cannot be transported safely can be evaluated with a bedside echocardiogram, which may have more test accuracy in this setting.
Radiation Exposure from Diagnostic Procedures
Fetal Exposure
The diagnosis of DVT and PE in pregnant patients poses unique challenges because of concerns in regard to fetal radiation exposure. ACOG advises that exposure to less than 5 rads has not been associated with increases in pregnancy loss or fetal anomalies. Exposure to ionizing radiation doses above 1 rad, however, may create a marginally increased risk of childhood leukemia (from 1/3000 baseline to 1/2000). Table 45-3 outlines the fetal radiation exposure of different radiation modalities. A combination of chest x-ray, V/Q scan, and pulmonary angiography exposes the fetus to less than 0.5 rads . CTPA is associated with only slightly lower radiation levels for the fetus compared with V/Q scanning. Although the concern for possible adverse effects should not prevent a medically important test from being performed, judicious use and selection of tests is advised.
| RADIOLOGIC MODALITY | FETAL RADIATION EXPOSURE (RADS) |
|---|---|
| Chest radiograph | <0.01 |
| Venography | |
| Limited, shielded | <0.05 |
| Full (unilateral), unshielded | 0.31 |
| Pulmonary angiography | |
| Brachial vein | 0.05 |
| Femoral vein | 0.22-0.37 |
| Ventilation-perfusion scan | 0.007-0.031 |
| Ventilation scan | 0.001-0.019 |
| Perfusion scan | 0.006-0.012 |
| Spiral computed tomography | 0.013 |
Maternal Exposure
Radiation exposure to the maternal breast is also an important consideration when choosing tests because the increased glandularity and proliferative state of the pregnant breast presumably make it more radiosensitive. This is an important concern with respect to cumulative radiation exposure and cancer risk. V/Q scanning results in substantially lower maternal breast radiation exposure. CTPA is estimated to expose maternal breast tissue to 150 times more ionizing radiation than V/Q scanning with doses estimated at 2 to 6 rads. This dose can be reduced approximately 50% to 60% with breast shielding, without significant reduction in image quality.
MRI and ultrasonography have not been associated with any adverse fetal effects, and teratogenic effects have not been described after administration of gadolinium contrast media. Concerns about fetal goiter following maternal radiographic contrast exposure suggest that fetal heart rate monitoring may be used to detect the reduced variability that can be seen with fetal hypothyroidism, and neonatal thyroid function should be tested during the first week of life.
Management of Venous Thromboembolism
Prevention
Perioperative Prevention
Evidence is very limited regarding the value of perioperative thromboprophylaxis with cesarean delivery. Perioperative administration of low-dose unfractionated heparin may be appropriate in patients undergoing cesarean delivery with clear risk factors such as obesity, malignancy, immobility, or a high-risk chronic medical disease. As noted, patients with low-risk thrombogenic thrombophilias require postoperative prophylaxis. Nonpharmacologic interventions aimed at preventing VTE include graduated elastic compression stockings and pneumatic compression devices. In pregnancy, a cohort study suggested that the use of graduated compression stockings reduced the prevalence of postpartum VTE from 4.3% to 0.9%. Because these stockings and pneumatic compression devices pose no hemorrhagic risk and do little harm, they should be strongly considered for thromboprophylaxis in all patients with risk factors, such as those patients who are hospitalized or immobilized, including pregnant or postoperative cesarean patients. The left lateral decubitus position in the third trimester may also reduce the risk of VTE.
Preconception Counseling
Preconception counseling is particularly important for patients at high-risk for recurrent VTE in pregnancy and for patients with a recent VTE. Patients on long-term anticoagulation before pregnancy, for example, should be advised of the risks of pregnancy. In particular, patients on warfarin can be counseled that they would be advised to switch from warfarin to a heparin-based regimen upon finding out about pregnancy. In particular, it would be helpful for the woman to know that the risks of warfarin embryopathy are highest between the sixth and twelfth weeks of pregnancy, so vigilance on tracking the last menstrual period is important.
Therapy
Unfractionated Heparin
Unfractionated heparin (UFH) affects anticoagulation by enhancing antithrombin activity, increasing factor Xa inhibitor activity, and inhibiting platelet aggregation. UFH, an FDA category C agent, does not cross the placenta and is not teratogenic. The chief side effects of heparin include hemorrhage, osteoporosis, and thrombocytopenia. The former is more common with treatment coinciding with surgery or liver disease or with concomitant aspirin use. Heparin-associated bone loss is usually reversible and correlates more with therapy that exceeds 15,000 U/day for more than 6 months and can be opposed by supplemental calcium use (1500 mg/day). Heparin-induced thrombocytopenia (HIT) occurs in 3% of patients, and type 1HIT is the most common form; it occurs within days of exposure, is self-limited, and is not associated with a significant risk of hemorrhage or thrombosis. Immunoglobulin-mediated type 2 HIT, on the other hand, is rare and usually occurs 5 to 14 days following initiation of therapy; paradoxically, it increases the risk of thrombosis. Monitoring for HIT should include serial platelet counts every 2 to 3 days from day 4 until day 14 or until heparin is stopped, whichever occurs first. A 50% decline in platelet count from its pretreatment maximum suggests immune-mediated HIT type 2 and should prompt cessation of all heparin exposure, including that in IV flushes. The diagnosis of HIT type 2 can be confirmed by serotonin release assays, heparin-induced platelet aggregation assays, flow cytometry, or solid-phase immunoassays.
Protamine sulfate reverses the effect of intravenously administered UFH . It is given by slow IV infusion of less than 20 mg/min, with no more than 50 mg given over 10 minutes. The total dose of protamine required is calculated based on the residual amount of circulating heparin, with a ratio of 1 mg of protamine sulfate necessary for every 100 U of residual circulating heparin. Residual heparin is calculated by assuming a half-life of 30 to 60 minutes of IV heparin. However, repeated serial administrations of lower doses of protamine, coinciding with serial measurements of the activated partial thromboplastin time (aPTT), are required when the heparin is dosed subcutaneously.
Low-Molecular-Weight Heparin
The low-molecular-weight heparins (LMWHs)—dalteparin, enoxaparin, and tinzaparin—are reliable and safe alternatives to UFH and have fewer side effects. Enzymatic manipulation of standard heparin produces lower molecular weight molecules with equivalent anti–factor Xa but little or no antithrombin effects. The LMWHs have longer half-lives, and a closer correlation exists between anti–factor Xa activity and body weight than subcutaneously administered UFH. We suggest following anti–factor Xa levels in pregnant patients, although this is not universal practice, because the variability in binding, distribution, metabolism, and excretion is far greater. However, this approach may be justified in pregnancy because of an increased rate of subprophylactic levels in pregnancy in patients on standard 40-mg once-daily enoxaparin dosing, which may be due to enhanced renal clearance. The LMWHs (pregnancy category B) do not cross the placenta and do not enter breast milk, and the risk of hemorrhage associated with LMWH is lower. However, regional anesthesia is contraindicated within 18 to 24 hours of therapeutic LMWH administration . Accordingly, we recommend switching to UFH at 36 weeks or earlier if preterm delivery is expected. Protamine is not as effective in fully reversing the anti–factor Xa activity of LMWH, although it may reduce bleeding. Dosing of 1 mg of protamine for every 100 antifactor Xa units of LMWH can normalize aPTT values, but antifactor Xa levels can only be reversed by 80%. Whereas the risk of HIT type 2 is lower in patients who receive LMWH, platelet counts should still be checked every 2 to 3 days from day 4 to day 14.
Fondaparinux
Fondaparinux is a synthetic heparin pentasaccharide that complexes with the antithrombin binding site for heparin to permit the selective inactivation of factor Xa but not thrombin. A major advantage of this medication is that there does not appear to be a risk of HIT type 2. It has comparable efficacy to both LMWH and UFH in nonpregnant patients. Although fondaparinux has been used in a small number of pregnant patients without complication, umbilical cord plasma concentrations at 10% of those of maternal plasma have been demonstrated, which suggests limited transplacental passage. Use of fondaparinux should be limited to women without other therapeutic alternatives, such as those with a history of HIT type 2 or heparin allergies and those who have failed other anticoagulants.
Coumarin
The coumarins are vitamin K antagonists that block the vitamin K–dependent, function-enhancing posttranslational modifications of prothrombin and factors VII, IX, and X as well as the anticlotting agents protein C and S. Coumadin (warfarin) has been shown to be effective for both the primary and secondary prevention of VTE, stroke, myocardial infarction, and systemic embolism due to artificial valves and atrial fibrillation. Warfarin is a pregnancy category X medication because fetal exposure can cause nasal and midface hypoplasia, microphthalmia, mental retardation, and other ocular, skeletal, and central nervous system malformations (see Chapter 8 ). Teratogenic potential is greatest between the 6th and 12th weeks of pregnancy, and the risk of warfarin also includes fetal hemorrhage. As a result, warfarin is rarely used during gestation, although the exceptional case of a patient requiring therapeutic anticoagulation who cannot be sufficiently treated with the heparins may require its use (e.g., certain patients with artificial heart valves). Warfarin is also favored in the postpartum period because it is safe to use during lactation.
Therapy is started at a dose of 5 to 10 mg for 2 days, with subsequent doses titrated to achieve an international normalized ratio (INR) of 2.0 to 3.0. Peak effects occur within 72 hours, and its half-life is from 36 to 42 hours. Warfarin metabolism is greatly affected by genetic polymorphisms, making the response to particular dosing regimens unpredictable. A pharmacogenetic approach to warfarin dosing includes testing for particular variations in vitamin K and warfarin metabolic enzymes and using this information to formulate a dosing regimen that expeditiously achieves a proper maintenance dose. A criticism of this approach is that pharmacogenetics only account for 30% to 50% of the variability in response because environment, diet, and comorbidities often have a significant influence as well. A nomogram for predicting warfarin dosing, based on patient-specific responses to warfarin doses as assessed by day 3 and day 5 INRs, has been proposed by Lazo-Langner and associates. Avoiding the need for costly genetic testing, this protocol accounts for individual variation of warfarin metabolism.
Because of protein C’s shorter half-life relative to the other vitamin K–dependent clotting factors, warfarin may initially create a prothrombotic state, particularly in pregnancy. Thus patients who are started on warfarin postpartum should be on therapeutic doses of UFH or LMWH for 5 days and until the INR is therapeutic for 48 hours. Coumadin effects can be reversed by vitamin K. The INR generally normalizes within 6 hours of a 5 mg dose of vitamin K and within 4 days with cessation of warfarin therapy. Fresh-frozen plasma can be used to achieve immediate reversal of effects. RCOG recommends avoiding warfarin until at least the fifth postpartum day and even longer in patients at increased risk for postpartum hemorrhage.
Treatment of Acute Deep Venous Thrombosis or Pulmonary Embolus
Women with new-onset VTE diagnosed during pregnancy should receive therapeutic anticoagulation, and this treatment should be continued for at least 20 weeks. If this period of time expires before the end of the postpartum period, prophylactic anticoagulation should be initiated in patients without highly thrombogenic thrombophilias and should be continued for 6 weeks to 6 months postpartum, depending on the severity of the thrombotic event and the underlying risk factors. During pregnancy, UFH and the LMWHs are the drugs of choice given their efficacy and safety profile. Therapeutic doses of the LMWH enoxaparin may start at 1 mg/kg subcutaneously twice daily . Dosing should be titrated to achieve anti–factor Xa levels of 0.6 to 1.0 U/mL when tested 4 hours after injection because of inconsistent efficacy of this weight-based regimen in pregnant patients. UFH for acute DVT or PE is initially given intravenously and is titrated to keep the aPTT at 1.5 to 2.5 times control (checked every 4 to 6 hours during the titration period), usually according to weight-based protocols ( Table 45-4 ). IV heparin should be continued for 5 to 10 days or until clinical improvement is noted. This regimen can be changed to subcutaneous injections every 8 to 12 hours to keep the aPTT at 1.5 to 2.0 times control when tested 6 hours after injection.
| Give a bolus of 80 U/kg, followed by a maintenance dosage of 18 U/kg/hr following aPTT values every 6 hr and with the following adjustments for aPTT values obtained: | |
| aPTT VALUE | ADJUSTMENT |
| <35 sec (<1.2× control value) | Repeat full bolus (80 U/kg) then increase infusion rate by 4 U/kg/hr |
| 35-45 sec (1.2-1.5× control) | Repeat half bolus (40 U/kg) then increase infusion rate by 2 U/kg/hr |
| 46-70 sec (1.6-2.3× control) | No change in infusion rate |
| 71-90 sec (2.4-3× control) | Decrease infusion rate by 2 U/kg/hr |
| >90 sec (>3× control) | Stop infusion for 1 hour then decrease to 3 U/kg/hr |
Prophylactic Anticoagulation Recommendations for Low-, Moderate-, and High-Risk Groups.
Risk stratification based on the likelihoods of recurrence, such as in Table 45-1 , is the essential foundation of the recommendations for antepartum and postpartum anticoagulation in patients without recent or active VTE. This issue has recently been addressed by ACOG and is summarized in Table 45-5 . The first principle is that the postpartum period represents a time of elevated risk, particularly in patients with risk factors. For this reason, anticoagulation recommendations in the postpartum period are typically a maintenance of antepartum recommendations or an increase. The second principle is that the various thrombophilias can be classified into high- and low-risk types, and anticoagulation recommendations for each type will be different depending on whether the patient has a personal history of VTE. Third, patients with recurrent VTE who have tested negative for the known inherited thrombophilias likely have an underlying pathology (e.g., an undiagnosed genetic malformation in a step in the coagulation cascade) and should be cared for cautiously.
