Elisabet E. Manasanch and Jay N. Lozier
Venous thromboembolism (VTE) is a major health problem in the United States with more than 900,000 estimated cases annually.1 The average yearly incidence is 117 cases per 100,000 population, with higher rates in women of childbearing age, males over 45 years old, and the elderly (where rates are up to five-fold higher). Pulmonary embolism (PE) occurs about 60% as often as deep venous thrombosis (DVT) and has a high mortality; the incidence may be higher since the diagnosis is often missed in hospitalized patients.2,3 Chronic VTE-induced pulmonary hypertension is a late complication of PE in 1% to 4% of cases after 4 years of follow-up, with all cases occurring before 2 years.4,5 Most cases of DVT occur in the lower extremities, but virtually any venous vascular bed can be involved. Upper extremity DVTs represent 1% to 5% of the total and are usually associated with long-term indwelling central venous access devices, thrombophilia, and/or cancer.6 When upper extremity DVTs occur, the potential for subsequent pulmonary embolization is estimated to be as high as 36% in some trials,7 though fatal pulmonary embolism is less common than for DVT of the lower extremities.
DEEP VENOUS THROMBOSIS AND PULMONARY EMBOLISM
DVT of the extremity is often the precursor of PE, though both may become symptomatic at the same time, and it is possible that PE may occur without symptomatic DVT, due to complete embolism of nascent thrombus before complete occlusion of the vein occurs. One autopsy series indicates that a substantial number of patients with PE have no pathologic evidence for DVT.8 The signs and symptoms of DVT and PE are listed in Table 22.1.7,9
The postthrombotic syndrome (PTS) is an important complication of DVT. It is caused by venous hypertension from outflow obstruction and valvular injury and varies from mild edema with little discomfort to incapacitating limb swelling with pain and ulceration. The severity of PTS can be assessed using the Villalta scale,10 which is based on the cumulative rates of signs and symptoms characteristic of the syndrome (Table 22.2).
The reported incidence of PTS after an acute episode of DVT has been reported to range from 23% to 60%11; severe, disabling PTS with skin breakdown and ulceration is seen much less commonly, perhaps in <10% of cases, while mild symptoms probably are experienced in the majority of DVT cases. Surprisingly, contralateral extremities may also develop postthrombotic manifestations without prior evidence of overt DVT11; perhaps occult obstruction of the inferior vena cava (IVC) may be to blame. Sized-to-fit graded compression stockings (typically 20–40 mm pressure) should be applied shortly after diagnosis of DVT to prevent acute dilatation that may result in permanent damage to the valves. If there is poor circulation in the leg due to complete obstruction of venous outflow, compression stockings should be used cautiously or withheld if the increased compression threatens to stop blood flow altogether. After definitive therapy (thrombolysis or mechanical clot extraction) and venous flow is restored to some extent, compression stockings may be reconsidered to mitigate acute symptoms and eventual PTS. A sized-to-fit compression stocking can reduce the rate of PTS by about 50%.12 Compression stockings do not substitute for adequate anticoagulation but are useful adjuncts to exercise-based rehabilitation programs. Stockings that are not fitted to provide 20 to 40 mm Hg compression do not confer equivalent benefits as seen from fitted stockings. A poorly fitted stocking can actually be detrimental, for instance, when the upper parts of the stocking roll down and form a loose “tourniquet” and thereby impair return blood flow from the leg. Unfortunately, compression stockings or sleeves for upper extremity DVTs have not proven to be as beneficial, perhaps due to the lesser hydrostatic pressures involved with the upper extremity.
Distal DVT is defined as thrombus found below the popliteal vein trifurcation, and occurs most commonly (71% of the time) in the peroneal vein.13 The risk for PE is almost negligible, whether treated with anticoagulation, or not; propagation to other distal calf veins does occur in about half of all patients, and propagation to more proximal deep veins is seen in ~5% of patients.13,14 Lysis of isolated calf DVT typically occurs within 3 months.13 The risk of anticoagulation under these circumstances (0%–6% bleeding risk) is approximately equal to the benefits realized (less propagation), so it remains controversial whether anticoagulation is necessary. A recent systematic review suggests that anticoagulation or serial imaging with noninvasive methods may be equally valid strategies.15 PTS occurs in ~5% of patients in the long term but is not characterized by severe changes such as skin ulceration.15
Table 22.2 The Villalta Scale for Evaluation of Postthrombotic Syndrome (PTS)
Symptoms
Pain
Cramps
Heaviness
Paresthesia
Pruritus
Clinical Signs
Edema
Venous dilatation/ectasia
Hyperpigmentation
Erythema
Lipodermatosclerosis (skin induration)
Pain during calf compression
Characteristics of PTS. Scoring is based on the cumulative rating of the signs and symptoms, with each rated a 0 (absent), 1 (mild), 2 (moderate), 3 (severe).Total score: 0–4, no PTS; 5–14, mild to moderate PTS; ≥15 or presence of a venous ulcer, severe PTS.
IMAGING OF DEEP VENOUS THROMBOSIS
Venography remains the standard for diagnostic imaging of DVT, but it is used less commonly than ultrasonography since it is an invasive procedure that uses radiocontrast dye and requires a skilled operator to perform the injection. Further, there is the need to bring a patient to a fluoroscopy suite, which may not be feasible in the acutely ill patient in an intensive care unit with other comorbidities. In contrast, ultrasonography is noninvasive, portable, and does not use contrast to which the patient may be allergic. Doppler ultrasound instruments are typically portable and can be brought to the bedside even in the most acutely ill patients. Venography retains an advantage for diagnosis of small distal DVTs that are not well imaged by ultrasound, as well as thrombosis of the vena cava or iliac veins of the pelvis that are not accessible to ultrasound examination because they are obscured by bowel gas. Venography may also be useful in instances where ultrasound is not feasible or an unequivocal diagnosis of DVT must be made. Ultrasonography is considerably more sensitive for the detection of proximal DVT than for distal DVT.
Sensitivity of compression ultrasound with venous imaging ranges from 89% to 96% when DVT is diagnosed by a combination of direct visualization of an occlusive thrombus and noncompressibility of a vein. The specificity of this finding for DVT ranges from 94% to 99%, but, unfortunately, the sensitivity may be substantially diminished (47%–62%) in patients with asymptomatic DVT.16 Serial ultrasound testing (which has little risk to the patient in contrast to venography) improves sensitivity, as a previously undiagnosed distal DVT may declare itself by proximal propagation. Furthermore, an ultrasound can accurately diagnose certain conditions that occasionally mimic DVT, such as Baker cyst. Impedance plethysmography may be useful in differentiating between a new or recurrent DVT, especially if the previous DVT has not resolved.
Magnetic resonance venography (MRV) and computed tomography venography (CTV) can diagnose DVT in a noninvasive manner. Prospective studies comparing CTV with venous ultrasound for diagnosis of DVT reported sensitivity rates of 100% and specificity of 96% to 100%.17,18 CTV can be easily combined with computed tomography angiography in patients suspected to have PE. However, it always requires the administration of IV contrast. One prospective blinded study reported the sensitivity to be >94% and the specificity to be >90% for the diagnosis of DVT using noncontrast MRV direct thrombus imaging.19 A major advantage for CTV and MRV is that deep abdominal, pelvic, and calf veins can be imaged. Furthermore, MRV can be successfully used avoiding contrast and its risks (such as gadolinium-associated systemic fibrosis in patients with chronic kidney disease20,21). Major disadvantages include cost, availability, expert reading, and possible need for IV contrast use when compared to ultrasound techniques.
PULMONARY EMBOLISM DIAGNOSIS: ECHOCARDIOGRAPHY, ELECTROCARDIOGRAPHY, AND X-RAY
PE is thought to be the consequence of clot breaking free from a lower extremity DVT. Upper extremity DVT embolization is less common, but the increasing use of central venous access devices for cancer chemotherapy or other long-term parenteral treatment may increase its frequency. Less commonly, PE may originate in the IVC (particularly in association with renal cell carcinoma) or the right ventricle of the heart from mural thrombus. There is intriguing evidence from the study of acute trauma patients8 that pulmonary “embolism” can be seen without concurrent DVT, raising the possibility that some pulmonary emboli may actually be in situ pulmonary thrombosis. Regardless of the mechanism, imaging approaches to diagnose pulmonary embolism include pulmonary angiography (the standard), CT angiography, ventilation/perfusion scanning, and more recently MRI. Echocardiographic confirmation of right ventricular hypertension may assist in the decision regarding thrombolysis with right heart failure. Both echocardiography and spiral CT have a low sensitivity for PE located in peripheral pulmonary vessels. In severe cases chest x-ray findings may include a Hampton hump (a wedge-shaped opacity with apex pointing to the hilum) or a focal paucity of blood vessel perfusion. Ancillary diagnostic chest radiography, ECG, and echocardiographic findings suggestive of PE are listed in Table 22.3.
LABORATORY DIAGNOSIS OF VENOUS THROMBOEMBOLISM
The D-dimer is a quantitative assay with good reproducibility that is often automated and can be used as an adjunct to imaging studies to exclude the presence of PE or DVT.22-24 When a thrombus is degraded by plasmin, D-dimers and other fibrin split products are formed from cross-linked chains of the fibrin clot—as such, they are markers for fibrin turnover. The use of D-dimer is preferable to the semi-quantitative fibrin split products that are measured by latex agglutination methods at varying dilutions of patient plasma. If a patient has a low pretest probability of VTE and the sensitive ELISA test for D-dimers is negative, as a practical matter VTE can be excluded. The low specificity for VTE requires that further diagnostic testing be conducted in the event of an elevated D-dimer to elucidate its etiology. Many other conditions can cause elevation of the D-dimer, including cancer, pregnancy, sepsis, sickle cell crisis, acute myocardial infarction, cardiopulmonary resuscitation, excessive bleeding, trauma, and recent surgery. Measurement of D-dimers may also guide the duration of anticoagulation for VTE, since the optimal course of therapy for such patients has not been established.25
The PROLONG Study indicated that patients with positive D-dimers 1 month after the completion of at least 3 months anticoagulation with vitamin K antagonists (VKAs) for idiopathic VTE had a significantly higher risk for recurrent VTE, which was mitigated by prolonged anticoagulation. Patients with abnormal D-dimers who did not resume anticoagulation experienced a 15% incidence of rethrombosis over the 18-month observation period compared to 2.9% if anticoagulation was restarted.26 The adjusted hazard comparing the rates of recurrence was 4.26 (95% CI 1.23–14.6, p = 0.02). A follow-up study, the PROLONG II, assessed the utility of repeated D-dimer testing in patients with a first unprovoked episode of VTE with normal D-dimer 1 month after stopping VKAs. D-dimer was tested at study initiation and every 2 months thereafter with a follow-up period of 13 months; 14% of patients with an initially negative D-dimer had a positive test at month 3 of evaluation. Furthermore, the D-dimer became abnormal at each subsequent time point in about 10–15% of patients up to 9 months of follow-up, at which time it decreased to 8–10%. An abnormal D-dimer at the first measurement or at day 30 usually remained abnormal over time in the majority of cases, and this pattern was associated with an increased risk of rethrombosis. The rates of recurrence for patients with an abnormal D-dimer at 3 months was 22.6% (95% CI 10%–41%) compared to 4.6% (95% CI 2%–9%) if D-dimer was normal (p= 0.003).27 Repeat D-dimer testing may help identify the subgroup of patients at lower risk for recurrence, in whom anticoagulation could be stopped. A multicenter prospective study (DULCIS, http://clinicaltrials.gov:NCT00954395) is currently underway to clarify this issue.
Table 22.3 Ancillary Findings Diagnostic for Pulmonary Embolism
Chest Radiography
Hampton hump
Focal paucity of blood vessel perfusion (Westermark sign)
Dilated pulmonary artery proximal to the thrombus
Atelectasis
Pleural effusion
Elevated diaphragm
Electrocardiography
New right bundle branch block
S1Q3T3 pattern (sign of acute cor pulmonale)
Supraventricular arrhythmias
Echocardiography
Right ventricular dilatation, often with myocardial hypokinesis
Pulmonary artery dilatation
Right ventricular mural thrombi
Tricuspid regurgitation
Loss of inspiratory collapse of the inferior vena cava
DEEP VENOUS THROMBOSIS IN SITES OTHER THAN THE DISTAL VEINS OF THE LOWER EXTREMITIES
Aside from typical presentations involving the lower extremities, DVT can occur in other sites such as veins of the upper extremities (particularly in conjunction with central venous catheters commonly used in cancer patients for chemotherapy), or in the veins of the chest or abdomen. Bilateral upper extremity DVTs are uncommon and should prompt a search for malignancy. The effects of DVT in these sites can be devastating, even in the absence of PE. Perioperative DVT of the splanchnic veins is common, occurring more frequently in laparoscopic procedures than with open ones. Unprovoked DVT of the splanchnic vessels should prompt a search for underlying abnormalities of hemostasis, such as deficiency of anticoagulant proteins, undiagnosed cancer, hematologic diseases such as myeloproliferative disorders (MPD), or paroxysmal nocturnal hemoglobinuria (PNH).
Portal vein thrombosis can commonly occur as a complication of surgical procedures, especially splenectomy, or during pregnancy, or with peritonitis. In cases that are not associated with a precipitating factor, antiphospholipid antibodies, deficiency of protein C, protein S, or (less commonly) anti-thrombin III may be found; the factor V Leiden and prothrombin 20210 gene polymorphisms may also be seen in patients at rates greater than in the general population (as is true in any group of patients with pathologic thrombosis). Clonal V617F point mutation in the Janus 2 kinase (JAK2V617F) tyrosine kinase gene occurs in a large proportion of MPD (particularly polycythemia rubra vera), and is found in 45% of BCS and 34% of portal vein thrombosis. Table 22.4 indicates the prevalence of various prothrombotic states in patients with portal or hepatic vein thrombosis.28-30
Treatment options for splanchnic vein thrombosis include anticoagulation, thrombolysis, and in the most extreme cases consideration should be given to orthotopic liver transplantation for those with liver failure. In addition, acute reduction of red blood cell mass and/or platelet count may be beneficial if the splanchnic vein thrombosis is due to polycythemia vera (PV) or essential thrombocythemia.
ACQUIRED THROMBOPHILIC STATES
Heparin-Induced Thrombocytopenia and Heparin-Induced Thrombocytopenia with Thrombosis
Heparin-induced thrombocytopenia (HIT) is a prothrombotic state caused by a drug reaction to unfractionated heparin (UFH) and less commonly to low molecular-weight heparin (LMWH), which can also exacerbate HIT due to cross reactivity with UFH. Even small exposures to UFH, including flushing of intravenous lines, may precipitate HIT and even thrombosis, in patients with antibodies. Antibodies of the IgG class are formed that produce a strong activation of platelets through their FcγIIa receptors. They recognize large multimolecular complexes of platelet factor 4 bound to heparin (PF4/H), although only about 10% of all anti-PF4/H antibodies have platelet-activating properties.31 This activation promotes thrombosis in vivo in both venous and arterial sites. HIT is a “clinical-pathologic syndrome,” requiring both a compatible clinical picture and positive laboratory test results. In general, platelet counts begin to decrease 5 to 9 days after the initiation of heparin. Thrombocytopenia and thrombosis may occur earlier in patients primed by heparin administration in the prior 100 days. Clinical scoring systems have been implemented to evaluate the pretest probability of HIT: 4Ts32 and most recently the HIT Expert Probability (HEP) score33 (Tables 22.5 and 22.6).
Low scores usually indicate a <2% probability of having a positive platelet activation assay. Two types of assays are used to establish diagnosis: enzyme immunoassays, such as an ELISA that detects PF4-heparin antibodies, and platelet activation/aggregation assays which detect spontaneous aggregation of platelets induced by the addition of heparin to the patient’s platelet rich plasma. Sensitivity of platelet aggregation assays can be increased when patient’s platelets are “loaded” with radioactive serotonin; serotonin release from platelets is detected as a marker of platelet activation in vitro, after the addition of heparin. The most sensitive test is the serotonin-loaded platelet aggregation assay. False positive results are common with the ELISA test, and its specificity varies depending on the pretest clinical probability. A negative result usually excludes HIT.
Two different classes of anticoagulants are used for the treatment of HIT in the presence or absence of thrombosis: direct thrombin inhibitors, lepirudin, bivalirudin, and argatroban, which are approved in the United States for the treatment of HIT and the factor Xa inhibitor fondaparinux, which is currently approved for DVT prophylaxis and treatment, but has growing evidence in literature to be an effective treatment for HIT.34 Fondaparinux binds anti-PF4 antibodies but does not activate platelets, unlike traditional LMWH preparations such as enoxaparin, dalteparin, and tinzaparin, which crossreact with the PF4-heparin antibodies and can activate platelets. Fondaparinux binds irreversibly to factor Xa and has a long active half-life (17 hours), which may prevent rebound hypercoagulability, but cannot be readily reversed if bleeding ensues. Dosing needs to be adjusted for impaired renal function. Lepirudin and argatroban require intravenous continuous infusions due to their short half-lives (<2 hours), and need PTT monitoring. Lupus anticoagulants (LAC) and/or high factor VIII levels may result in prolonged or shortened clotting times, respectively, thereby complicating monitoring. Algorithms exist for infusions of argatroban without monitoring when LAC interfere with monitoring.35 Levels of argatroban or other direct thrombin inhibitors can be monitored more directly by use of assays of the ecarin clotting time; ecarin is a snake venom that converts fibrinogen to fibrin but is not influenced by LAC or levels of factor VIII.36 Lepirudin is antigenic, and antibody formation results in excessive anticoagulation. Lepirudin is contraindicated in the presence of renal insufficiency. Bivalirudin (approved in HIT specifically for patients undergoing percutaneous coronary intervention) and argatroban are both hepatic-excreted, nonantigenic, and short-acting, and they should be used cautiously in patients with liver disease. Initiation of warfarin should be postponed at least until the platelet count is above 150,000/μL, since warfarin use during acute HIT is a major risk factor for venous limb gangrene, perhaps on the basis of rapid depletion of protein C, a vitamin-dependent anticoagulant protein with a short half-life (6–7 hours).
Nephrotic Syndrome
Patients with nephrotic syndrome are at higher risk for VTE, arterial thromboembolism (ATE), and renal vein thrombosis than is the general population [annual incidence, 9.85% (VTE) and 5.52% (ATE)].37Membranous glomerulonephritis confers the highest risk of VTE. Hypercoagulability appears due to alterations in plasma levels of proteins involved in coagulation and fibrinolysis. The anticoagulant proteins antithrombin III and protein S are lost in the urine, perhaps due to their relatively small size compared to procoagulant factors V, VIII, von Willebrand factor, and fibrinogen, which may be retained by the kidney.38 Further, fibrinogen, factor VIII, and von Willebrand factor are acute-phase reactants, and their levels may be increased by inflammation.39 The routine use of prophylactic anticoagulation for patients with nephrotic syndrome has not been established by randomized controlled trials, but some experts advocate this when additional risk factors such as membranous glomerulonephritis or the antiphospholipid syndrome are present. LMWH must be used with extreme caution in patients with renal insufficiency in addition to nephrotic syndrome.
Antiphospholipid Antibody Syndrome
Antiphospholipid antibodies (APA) may be associated with thrombosis on the basis of various mechanisms,40 summarized in Table 22.7. Anticardiolipin antibodies are APA associated with infections such as syphilis, but these are transient and not usually associated with thrombosis.
Tests for APA include specific assays to quantify levels directly by ELISA or indirectly through the detection of antiβ2-GPI antibodies. The APA that prolong phospholipid-dependent clotting assays (primarily the aPTT) are designated as LAC. Other, less commonly used assays that are affected by LAC include the dilute Russell viper venom time (DRVVT), the dilute prothrombin or tissue thromboplastin inhibition (TTI) test, and the kaolin clotting time (KCT). Confirmatory tests for LAC are performed by adding a source of phospholipids to a clotting reaction that is prolonged by the lupus anticoagulant to see if the abnormal clotting time is corrected. Phospholipids may include platelets (in the platelet neutralization procedure, or PNP), or more recently phospholipids with defined physical properties, such as hexagonal-phase lipids that are used in some commercial tests. The hematologist consultant should be aware of the specific screening and confirmatory tests that are used in the local laboratory when diagnosing APA syndromes or LAC. Also important is that LAC tests may be positive at the time of an acute thrombotic event but often are negative at subsequent time points only a few weeks later. In a study of 30 patients with lower extremity DVT that were treated with tPA, 19 of the 30 were initially positive for LAC at the time of presentation, but 11 of these were documented to be negative 6 months later (or sooner), three remained positive at later time points, and there were six lost to follow-up and one with an indeterminate value at later time points (unpublished observations on patients described in Refs. 41 and 42).
Table 22.7 Mechanisms of Thrombosis Associated with Antiphospholipid Antibodies (APA)
Binding of APA to β2-glycoprotein I (β2-GPI)/phospholipid complex exposed on the surface of injured or activated endothelial cells and monocytes
Overproduction of tissue factor by monocytes and endothelial cells
Activation of platelets that increases expression of glycoprotein IIb-IIIa and synthesis of thromboxane A2
Interaction with regulatory proteins such as annexin V, prothrombin, factor X, protein C, and plasmin
Two positive tests for APA (LAC, anticardiolipin antibody IgG or IgM, and anti-β2-glycoprotein I antibody IgG or IgM) obtained at least 12 weeks apart are required to fulfill the laboratory criteria for the APA syndrome. Clinical criteria for the diagnosis of the APA syndrome include detection of venous and/or arterial thrombotic events, autoimmune thrombocytopenic purpura, marantic endocarditis, multiple spontaneous abortions before the 10th week of gestation, or unexplained death of a morphologic normal fetus after the 10th week of gestation. Thrombosis can occur in virtually any vascular bed, and patients may present with a catastrophic syndrome with thrombosis in multiple vascular sites, including cerebrovascular accidents and DVT, with more than 50% mortality.
Treatment strategies should focus on modification or elimination of risk factors, such as smoking and oral estrogen contraceptives. For a thrombotic complication, systemic anticoagulation should be initiated. To prevent VTE recurrence, a randomized, double-blind, prospective study showed that dosing warfarin to an international normalized ratio (INR) of 2.0 to 3.0 was equally effective as higherintensity warfarin regimens to achieve an INR of 2.5 to 3.5.40 However, patients with recurrent VTE or arterial thrombosis may need more aggressive treatment with an INR goal 3.0 to 4.0 or combined antithrombotic therapy (warfarin plus low-dose aspirin). This is based on the observation that the rate of recurrent thrombosis was low with an INR >3.0, and is not endorsed by all experts.43
Hypercoagulability of Malignancy
VTE frequently complicates malignancy and results in significant morbidity and mortality. The estimated prevalence of VTE in patients with cancer is 10% to 15% and can be as high as 28% to 30% in pancreatic cancer or malignant gliomas.44 Certain sites have been associated with VTE associated with malignancy: intra-abdominal and bilateral lower extremity DVT (p< 0.05), which may justify cancer screening that otherwise might not be performed.45Malignancy promotes thromboses through a variety of mechanisms: release of tissue factor, activation of factor X by cancer procoagulants(s), endotheliumtumor cell interactions, and platelet activation. Hypercoagulability associated with malignancies is designated Trousseau syndrome and manifests as disseminated intravascular coagulation, nonbacterial thrombotic endocarditis, PE, DVT, and arterial thromboses. Occasionally, chemotherapy agents promote thrombosis, possibly through direct injury to the vascular endothelium. Of equal importance is the observation that adjuvant therapy with selective estrogen receptor modulators (SERMS), such as tamoxifen, or antiangiogenic medications, such as thalidomide and lenalidomide for the treatment of multiple myeloma, and bevacizumab for the treatment of breast, colon, brain, or lung malignancies, can all increase the potential thrombogenicity of the cancer type. Central venous indwelling catheters often complicate cancer care because of thrombus formation in the catheter itself and the vessel into which it has been inserted.46
Treatment of cancer-associated thrombosis is challenging. Warfarin, the current mainstay of long-term anticoagulation, may be difficult to manage due to concomitant medications and thrombocytopenia from chemotherapy or radiation; further, rethrombosis is common despite correct warfarin dosing. LMWH is the agent of choice for anticoagulation in cancer patients. Dalteparin is approved for the extended treatment of symptomatic VTE in patients with cancer based on the large clinical CLOT trial that showed a reduction in objectively confirmed, symptomatic DVT and/or PE during the 6-month study period when compared with oral anticoagulation (15.7% versus 8%).47Dalteparin appeared to reduce the incidence of rethrombosis most significantly in the first month of treatment compared to warfarin with a statistically significant (p = 0.03) improvement in overall survival (a secondary endpoint in the study) in patients with nonmetastatic disease. A similar trial randomized 138 patients to receive enoxaparin 1.5 mg/kg/day followed by warfarin for 3 months or enoxaparin 1.5 mg/kg/day for 3 months. There was a trend toward a decrease in VTE recurrence or major bleeding with enoxaparin versus warfarin (21.1% versus 10.5%; p = 0.09).48 LMWH may exert an antineoplastic effect through interference with tumor cell adhesion, invasion, metastasis formation, and angiogenesis, all of which are needed for tumor progression. However, it is unclear whether this is true for all tumor types and further studies are needed.49
Paroxysmal Nocturnal Hemoglobinuria
PNH causes intravascular hemolysis, bone marrow failure, and thrombotic events. Diagnosis can be made rapidly by detection of CD55- and CD59-deficient erythrocytes and neutrophils by flow cytometry. Patients with clones comprising greater than 50% of cells carry a high risk of thrombotic events (44% in 10 years), including unusual sites like the hepatic vain (Budd-Chiari syndrome) being a frequent manifestation.50 Thrombosis is the most common cause of death in PNH. The etiology of thrombosis may involve release of free hemoglobin (which activates the endothelium), complement-mediated damage of GPI-deficient erythrocytes, and/or deficiency of GPI-anchored fibrinolytic factors such as urokinase/plasminogen activator receptor.51 The humanized monoclonal antibody eculizumab targets the C5 terminal complement component and is FDA approved for treatment of PNH. Eculizumab significantly reduced the rate of VTE in eculizumab-treated patients in a non-randomized trial when compared with the same patients pretreatment [1.07/100 patient years, versus 7.37/100 patient-years (P< 0.001)].52
Surgery as a Risk for Acquired Thrombosis
Surgery is a major risk factor for thrombosis. Trauma to tissue results in endothelial injury, activation of the coagulation cascade (through the release of tissue factor), and platelet activation. The risk is modified by time of anesthesia, patient age, the presence of underlying heritable or acquired hypercoagulable states, and the nature of the surgical procedure. VTEs most frequently occur with hip or knee arthroplasty, hip fracture surgery, spinal cord injury, major trauma, and any surgery performed in the context of malignancy. Patients undergoing these procedures should receive thromboprophylaxis. Use of graduated pneumatic compression stockings plus LMWH, adjusted dose heparin, fondaparinux, and oral anticoagulation with warfarin to achieve an INR goal of 2 to 3 are all reasonable options. DVT prophylaxis should be individualized depending on bleeding risk, history of previous thrombosis, history of HIT, presence of renal insufficiency, and type of surgery; published guidelines can assist in management.53 Outpatient surgical procedures performed in patients younger than 40 years who can be made readily ambulatory do not require prophylactic anticoagulation. Prolonged prophylactic anticoagulation up to 30 days after surgery may be indicated for patients undergoing total hip replacement (at least until the patient is mobile) and for those in whom a malignancy persists.
Laparoscopic surgery has become increasingly popular as a substitute for conventional open surgical procedures. Although there is less tissue damage, shorter procedure times, and quicker recovery, patients may be subjected to induction of pneumoperitoneum and prolonged use of the reverse Trendelenburg position to visualize and manipulate internal organs that may result in venous stasis and increased risk of thrombosis in some patients.54 The American College of Chest Physicians’ (ACCP) Clinical Practice Guidelines (8th edition) recommend against routine prophylaxis for those undergoing laparoscopic surgery without additional thromboembolic risk factors, but recommends mechanical or pharmacologic prophylaxis in patients with any risk factors.53 The Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) guidelines for DVT prophylaxis during laparoscopic surgery stratifies inpatients into low, moderate, and high risk groups for thrombosis on the basis of a risk score imputed from the type of procedure and patient risk factors.55,56 Procedure-related risk factors include procedure lasting over 1 hour and pelvic procedures. Patient-related factors include age > 40 years, immobility, malignancy, thrombophilic states (protein C, protein S, or ATIII deficiency), obesity, peripartum state (or use of estrogens), heart failure, renal failure, varicose veins, inflammatory states, or infection. In the lowest risk group (procedure < 60 minutes in patients with no risk factors) elastic stockings and early ambulation suffice, and UFH or LMWH is optional. In the moderate-risk group (one patient risk factor in a procedure of less than 60 minutes, or any procedure >60 minutes with no patient risk factors) pneumatic compression devices or prophylactic heparin or LMWH are recommended. In the high-risk group (two or more risk factors in procedures >60 minutes) a combination of serial compression devices and prophylactic UFH or LMWH are recommended.55,56
Myeloproliferative Disorders
MPDs are paradoxically associated with increased risks for both hemorrhage and thrombosis, but thrombosis is the most common cause of death in MPDs. The thrombotic risk of PV is often exacerbated by the hyperviscosity produced by a markedly increased red cell mass. Treatments for PV include serial phlebotomy to decrease red cell volume/hyperviscosity and cytotoxic agents, such as hydroxyurea (HU), to reduce erythrocyte, leukocyte, and platelet production, with the ultimate goal to minimize the risk of thrombosis.
32P or alkylators such as chlorambucil or busulfan are rarely used, except in the elderly where long-term leukemia risk is not as concerning as for other patients. The goal of PV therapy by phlebotomy is to keep the hematocrit below 45% in males and below 42% in females, so as to prevent stroke, heart attack, or DVT, including the Budd-Chiari syndrome.57 PV-associated erythromelalgia is usually treated with low-dose aspirin 81 mg daily. Higher aspirin doses may be associated with thrombosis.
Therapy to prevent thrombosis is usually considered for patients with essential thrombocythemia who are older than 60 years of age or have a prior history of thrombosis. Guidance is derived from a randomized trial evaluating low-dose aspirin plus anagrelide versus HU. HU plus aspirin was associated with a lower risk of arterial thrombosis, serious hemorrhage, and transformation to myelofibrosis than anagrelide, a specific platelet-lowering noncytotoxic, and non-chemotherapy agent. VTE incidence was higher in the HU group; however, the totality of the data when analyzed for composite endpoints favored treatment with HU.58 Blood counts should be followed closely during therapy, and HU is contraindicated in pregnant women or for women who desire to become pregnant.
Inherited Hypercoagulable States
Most inherited hypercoagulable states associated with an increased risk of VTE are deficiencies of one or more anticoagulant proteins or defects in coagulation factors that increase their level of expression or make them no longer subject to inhibition or regulation by anticoagulant proteins. There may also be metabolic problems such as homocysteinemia where toxic levels of homocysteine may be toxic to the endothelium and promote thrombosis. Inherited hypercoagulable states are listed in Table 22.8.
Persistent elevations of coagulation factors VIII, IX, or XI have also been implicated as inherited hypercoagulable states.59-63 Extremely rare inherited hypercoagulable states include dysfibrinogenemias or other deficiencies of the fibrinolytic system (plasminogen deficiency). Screening for a thrombophilic hypercoagulable state should be considered for a young patient with a positive family history of thrombosis, unprovoked thrombosis, or recurrent thrombosis. The timing of testing is critical. Elevations of factor VIII or fibrinogen can occur at the time of an acute thrombotic event, perhaps from inflammation, and only a persistently increased level is likely to be the cause for thrombosis. Anticoagulant medications can interfere with measurements of proteins C and S (warfarin) or antithrombin (heparin), and testing should be done when the patient is not taking these medications to ensure an accurate assessment of levels. Pregnancy is a time when levels of factor VIII rise, and free protein S decreases, so the baseline assessment of hypercoagulability is best done when a woman is not pregnant.
Activated Protein C Resistance (Factor V Leiden)
A polymorphism in the gene for coagulation factor V (Arg506Gln; factor V Leiden) results in a factor V protein that is not inactivated by activated protein C. The factor V Leiden polymorphism is seen in ˜5% of the Caucasian population and is the most common known inheritable risk factor for DVT or PE.64 There are other rare factor V polymorphisms associated with resistance to activated protein C in Caucasians (factor V Cambridge, Arg306Thr; Ref. 65) and also in Chinese (factor V Hong Kong, Arg306Gly, Ref. 66) that will be missed by the DNA test for factor V Leiden. Most patients who carry the mutation do not develop thrombosis; however, additive risk factors, including estrogen use and the coexistence of the prothrombin mutation, greatly increase the risk of VTE. Laboratory testing includes direct detection of the characteristic mutation in the factor V gene, accomplished by the polymerase chain reaction (PCR) on peripheral blood leukocyte DNA. Indirect testing for inability of activated protein C to prolong the PTT (resistance to activated protein C) is done less commonly now, but may be useful if there is suspicion that a polymorphism other than the factor V Leiden may be present.
Table 22.8 Inherited Hypercoagulable States
Activated protein C resistance (factor V Leiden gene polymorphism, 506R→Q)
Prothrombin gene 20210 A→ T polymorphism
Protein C deficiency
Protein S deficiency
Antithrombin deficiency
Hyperhomocysteinemia
Prothrombin Mutation (G20210A)
The second most common gene mutation responsible for congenital hypercoagulability in Caucasians is the prothrombin G20210A polymorphism. As with factor V Leiden, VTE is more commonly seen than arterial thrombosis. This mutation is associated with ˜20% greater than normal prothrombin levels, seen almost exclusively in Caucasians, and was found to have a crude 2.8 odds ratio for the development of thrombosis.67 The polymorphism apparently leads to more stable factor II mRNA, and therefore greater prothrombin expression. G20210A can be detected directly by PCR analysis of the target site in the factor II gene; measurements of factor II are not particularly specific for the polymorphism, and should not be relied upon to make the diagnosis.
Deficiencies of Proteins S and C and Antithrombin (Antithrombin III)
Proteins C and S and antithrombin (previously known as antithrombin III) are synthesized in the liver and serve to modulate reactions of blood coagulation. Protein C is a serine protease, and when activated by thrombin cleavage has the ability to catalytically cleave and inactivate factors V and VIII, thereby shutting down further synthesis of thrombin. It is a vitamin K-dependent protein, and its activity level goes down in patients who are on warfarin. Protein S is a nonenzyme cofactor for activated protein C, and is also a vitamin K-dependent protein that is inactivated in patients on warfarin. Protein S binds to (complement) C4b-binding protein,68 which may mediate a decrease in protein S activity during acute inflammation. Antithrombin (previously known as antithrombin III) is a serine protease inhibitor that neutralizes thrombin, and activated forms of other serine proteases such as factors Xa, XIa, IXa, and XIIa.69Antithrombin’s activity as a serine protease inhibitor is accelerated by binding to heparin, a complex, sulfated polysaccharide that is a normal component of various tissues (lung, liver, intestine) and used in purified form as an anticoagulant medication. Deficiency of antithrombin is probably the most serious risk factor for thrombosis, especially VTE in association with surgery or invasive procedures.
Hyperhomocysteinemia
Homocysteine is an amino acid that does not appear in proteins, but is an intermediate in the metabolism of methionine, a sulfur-containing, essential amino acid. Reactions of the metabolic pathways in which homocysteine participates require folic acid, cyanocobalamine (vitamin B12), or pyridoxine (vitamin B6) as cofactors. Deficiencies of these vitamins may cause accumulation of homocysteine to high levels in the bloodstream. Extremely high levels of homocysteine are seen in homocystinuria, and in methylene tetrahydrofolate reductase gene defects. Homocysteine may be toxic to endothelial cells,70 and accelerated atherosclerosis can be a feature of the disease homocystinuria.70,71 Accordingly, many investigators have sought evidence for an association between elevated homocysteine and thrombosis in people who are not suffering from the extreme manifestations of homocystinuria. Evidence that elevated homocysteine levels (>95th percentile) is associated with VTE was found in a study comparing patients with first time DVT with matched controls, where the odds ratio for elevated homocysteine was 2.5.72 It is not clear if lowering homocysteine levels by vitamin therapy can alter the risk for VTE, however. Despite lowering homocysteine levels in the treatment group of the 5,222-patient HOPE-2 intervention trial, there was no difference in the incidence of VTE.73 A meta-analysis of 31 published studies indicated that presence of the homozygous TT form of the methylene tetrahydrofolate reductase (MTHFR) gene at amino acid 677 was only a very weak risk factor for thrombosis.74 Therefore, it is our practice not to routinely seek expensive molecular studies of polymorphisms in the MTHFR gene as risk factors for VTE, but instead to measure homocysteine levels in serum.
TREATMENT
Prophylaxis
The American College of Physicians (ACP) recently published guidelines for VTE prophylaxis in medical patients and patients with an acute stroke.75 Careful assessment of the risk for bleeding and thrombosis need to be performed prior to starting prophylaxis with heparin. The ACP guidelines recommend starting prophylaxis with heparin (either LMWH or UFH) in medical (including stroke) patients when the benefit outweighs the risk of bleeding. Pooled data from 18 trials showed that heparin prophylaxis was associated with a borderline statistically significant reduction in risk for mortality compared to no heparin prophylaxis (RR 0.93, CI 0.86–1.00), a statistically significant reduction in the risk of PE (RR 0.70, CI 0.56– 0.87), at the expense of an increase in bleeding events, which was also statistically significant (RR 1.28, CI 1.05–1.56). Interestingly, no improvements in clinical outcomes were seen in three studies of mechanical prophylaxis in patients with stroke, but more patients had lower extremity skin damage (RR 4.02, CI 2.34–6.91) an increase of 39 events per 1,000 patients treated. As such, the ACP recommends against the use of mechanical prophylaxis with graduated compression stockings for the prevention of VTE.75
The 9th edition guidelines of the American College of Chest Physicians address prophylactic treatment of VTE in nonorthopedic and orthopedic surgical patients.76 Recommendations for or against anticoagulation or mechanical prophylaxis in general surgery patients are based on the type of surgery and individual thrombotic and bleeding risk assessment. Major orthopedic surgery has an estimated risk of symptomatic VTE of 4.3% in 35 days in patients receiving no prophylaxis.76 Therefore, patients undergoing total hip arthroplasty, total knee arthroplasty, or hip fracture surgery who are not at increased risk of bleeding, should receive either thromboprophylaxis with LMWH (or alternative pharmacologic anticoagulation) or mechanical prophylaxis. Extended use of LMWH in the outpatient setting (up to 35 days post surgery) is recommended based on data from three systematic reviews including seven controlled trials showing a decrease in symptomatic VTE of 9 per 1,000 patients without an appreciable increase in major bleeding.76
Initial Treatment for Acute Venous Thromboembolism
Anticoagulation is the essential primary treatment and prophylaxis for VTE. Mechanical barrier devices (IVC filters) may be used in certain circumstances in lieu of anticoagulation, or as an adjunctive measure. Fibrinolytic therapy may also be beneficial in select patients as an adjunct to anticoagulation.
The 9th ACCP Guidelines for anticoagulation and treatment of VTE77 suggest that a DVT or PE should initially be treated with LMWH, fondaparinux, or rivaroxaban, that LMWH and fondaparinux are preferable to UFH, thrombolytic therapy may be advisable in the presence of hemodynamic compromise, and anticoagulation should be utilized for at least 3 months, in the setting of a provoked event, and longer for unprovoked events. The purpose of anticoagulation is to prevent additional clots from developing, to immediately stop further propagation of the existing clot, and to permit endogenous fibrinolysis to begin to dissolve the clot physiologically. If warfarin is used, it should be started after initiation of heparin, LMWH, or fondaparinux, typically at a 5-mg daily dose for most adults. A double-blind randomized study, however, showed that warfarin at 10 mg for 2 days followed by dose adjustment determined by a nomogram was also safe and effective.78 Parenteral anticoagulation is continued until there are successive INR values of between 2 and 3 after adjustment of the warfarin dose. In the absence of PE, or with PE that is not complicated by cardiovascular compromise, or other reasons to admit the patient, this can typically be accomplished as an outpatient. Subcutaneous (SQ) UFH versus LMWH was evaluated in the FIDO study, a randomized trial of 708 patients, showing that fixed dose SQ UFH is as effective and safe as LMWH in patients with acute VTE and is suitable for outpatient treatment.79 LMWH and fondaparinux are less likely than UFH to cause HIT, however. LMWH requires monitoring of anti-Xa levels in patients at high or low extremes of weight, or who have renal failure, or are pregnant. LMWH and fondaparinux are relatively contraindicated in renal failure and should be used with extreme caution (using dose reduction and monitoring), as the kidneys excrete LMWH. UFH may be preferred in a patient at risk for bleeding, due to the ability to neutralize with protamine and its shorter duration of effect. See Chapter 23 for discussion of new oral anticoagulant medications.
Extended Treatment to Prevent Recurrent Venous Thromboembolism
Warfarin has been the only oral anticoagulant available for long-term anticoagulation until recently. Despite the introduction of new oral anticoagulants such as rivaroxaban, dabigatran, and apixaban that have indications to prevent DVT or stroke in atrial fibrillation, it remains a useful drug due to its lower cost and established risk-benefit profile. Dalteparin sodium is FDA approved for the extended treatment and subsequent prevention of recurrent symptomatic VTE in patients with cancer. Unique risks of long-term LMWH use include osteopenia and HIT (the latter risk is uncommon in this setting).
The ideal intensity of warfarin anticoagulation for long-term (indefinite duration) therapy to prevent recurrent VTE has been addressed in two well-designed trials yielding conflicting results. The PREVENT trial concluded that low intensity anticoagulation with the target INR of 1.5–2.0 was successful in substantially reducing recurrent VTE risk.80 In a randomized, two-arm study of standard warfarin anticoagulation in Canada to achieve an INR between 2 and 3 versus a low-intensity INR arm with target INR between 1.5 and 2.0, the standard dosing regimen was over 60% more effective (p = 0.03) than low-intensity warfarin anticoagulation in reducing the cumulative probability of recurrent thromboembolism.81 There was no difference in bleeding complications between the two dosing intensities. The differences between the two studies may be related to trial design. For instance, the Canadian trial, in contrast to the PREVENT study, included cancer patients, who more likely would experience warfarin resistance.
At least 3 months of anticoagulation are recommended for idiopathic (unprovoked) DVT, and if there is no contraindication to anticoagulation, it should be continued indefinitely.77 Similarly, life-threatening PE requires indefinite anticoagulation, unless the risk of life-threatening bleeding on anticoagulation exceeds the risk of a fatal PE; in this rare circumstance an IVC filter may be utilized. The PREVENT trial corroborated that patients with idiopathic VTE have a high incidence of recurrent VTE and benefit from long-term anticoagulation.80 As previously discussed, the APA syndrome requires prolonged anticoagulation. Individuals with cancer remain hypercoagulable as long as the malignancy is present. Patients with transient risk factors (i.e., trauma) usually require anticoagulation for 3 to 6 months. Patients with increased D-dimers, elevated FVIII activities, or evidence of significant residual DVT at 1 month after discontinuing anticoagulation are at increased risk for rethrombosis and consideration should be given to resumption of long-term anticoagulation. The D-dimer can also be measured intermittently after discontinuation of therapy to assess for the risk of rethrombosis.27
Inferior Vena Cava Filters in the Treatment of Deep Venous Thrombosis or Pulmonary Embolus
IVC filters are placed to prevent large clots in the lower extremities from embolizing to the pulmonary circulation. Major reasons for the placement of an IVC filter include strong contraindications to the use of anticoagulants, intolerance to or noncompliance with anticoagulants, and recurrent PE despite adequate systemic anticoagulation. All of these factors are relative indications for use of an IVC filter, and the decision to place one is not to be taken lightly. A randomized study revealed that, in the short term, an IVC filter decreases the incidence of pulmonary embolus from 4.8% to 1.1%, but by 2 years the rate of recurrent DVT was 20.8% in the IVC filter group versus 11.6% in the nonfilter group, and overall mortality was no significantly different.82 Retrievable filters have the advantage that they may be removed after the risk of PE has passed, but they may be more prone to migration than permanent filters.
Fibrinolytic Therapy
Fibrinolytic therapy has typically been reserved for patients with massive PE associated with hemodynamic compromise.77 Patients with systolic BP < 90 mm Hg, or a BP drop > 40 mm Hg for > 15 minutes, not caused by cardiac arrhythmias, sepsis, or hypovolemia, may benefit from thrombolytic therapy with improved survival.83 Clinical studies do not show a survival advantage for thrombolytic agents in PE. Currently, tPA is the only fibrinolytic agent commonly available, and it has the advantage of a short half-life, and relative specificity for fibrin clot (as opposed to fibrinogen) when compared to urokinase or streptokinase. Fibrinolytic therapy is being investigated in the ongoing ATTRACT84 and recently reported CAVENT85 trials, in selected patients with massive iliac-femoral DVT. The CAVENT trial randomized 209 patients with iliofemoral DVT to anticoagulation with (101 patients) or without (108 patients) additional catheter-directed thrombolysis (CDT) with tPA. At 24 months, the patients given CDT had 41% incidence of PTS, versus 56% of the control group (p = 0.047). Patency of the iliofemoral system was seen in 66% of CDT patients versus 47% of the control group (p = 0.012). There was additional bleeding associated with CDT that included three major and five clinically significant bleeding episodes, consistent with typical estimates of bleeding risk from fibrinolytic therapy of ˜8%.86
At this time, catheter-directed fibrinolytic therapy with low doses of tPA may be considered for use in patients with DVT that is not responding to standard anticoagulation, and adjunctive measures such as stenting or balloon dilatation of venous segments with strictures may also be considered, but this is not yet standard practice.41,42,77
References