Abeloff's Clinical Oncology, 4th Edition

Part II – Problems Common to Cancer and its Therapy

Section B – Hematologic Problems

Chapter 46 – Diagnosis, Treatment, and Prevention of Cancer-Related Venous Thrombosis

Steven R. Deitcher




Venous thrombosis has an incidence of approximately 15% in cancer patients.



Venous thrombosis has been identified in up to 50% of cancer patients at autopsy.



Cancer patients with venous thrombosis at diagnosis have a worse prognosis.



Cancer patients are more likely to have large, proximal, persistent, and recurrent venous thromboses.



Venous thrombosis can complicate the management of any and all types of cancer.



Malignant tumors can cause venous stasis, vascular injury, procoagulant accumulation, natural anticoagulant deficiency, and impaired endogenous fibrinolysis.



Plasma markers of coagulation activation do not predict development of thrombosis but can assist in prognosis determination.



Cancer-associated venous thrombosis is likely underdiagnosed because of inadequate physician suspicion, underutilization of proper objective tests, and the asymptomatic nature of many thrombotic events.



Clinical diagnosis alone lacks sensitivity and specificity for acute venous thrombosis.



Diagnostic imaging studies such as duplex ultrasound and lung scintigraphy may provide more false-positive and false-negative results in cancer patients.



Cancer patients might not respond to or benefit from standard approaches to venous thrombosis management, as do patients without cancer.



Tendencies toward thrombocytopenia, osteopenia, malnutrition, brain metastasis, hepatic metastasis, and bleeding all complicate thrombosis care in cancer patients.



Cancer patients are more likely to manifest heparin resistance and warfarin failure and to have limited venous access to support therapeutic anticoagulant monitoring.



Low-molecular-weight heparins are safe and effective for the short- and long-term management of venous thrombosis in cancer patients.



Low-molecular-weight heparins are associated with lower recurrent thrombosis rates than is oral warfarin during up to 6 months of anticoagulation therapy.



Low-molecular-weight heparins may affect survival in subgroups of cancer patients with and without venous thrombosis.



Diagnosis should be sought by using duplex ultrasound with magnetic resonance venography, computed tomography, and venography reserved for confirmation.



D-dimer testing is likely less sensitive and less specific in cancer patients.



Low-molecular-weight heparins are likely to be the best drugs to minimize recurrence rates without significantly increasing bleeding complication rates.



It is prudent to continue anticoagulation for a minimum of 6 months and until all cytotoxic therapy has been completed and no evidence of active cancer is found.



Duplex ultrasound and venography are the mainstay of upper extremity thrombosis diagnosis.



Upper-extremity thrombosis should be treated like lower-extremity thrombosis, with the exception of catheter-related thrombosis, which may require catheter removal.



Helical computed tomography (CT) and lung scintigraphy are both means of objectively diagnosing pulmonary embolism. Helical CT is preferred when the chest radiograph is not normal.



Pulmonary embolism should be treated in a similar fashion and on the basis of the same principles as lower extremity deep venous thrombosis.



The risks of thrombolysis should be reserved for patients with hemodynamically unstable pulmonary embolism.



Incidental findings of intra-abdominal vein thrombosis on standard CT scans should be confirmed by a more appropriate diagnostic method.



Treatment should be the same as that for lower-extremity deep venous thrombosis.



Superficial venous thrombosis may herald the diagnosis of cancer and should be treated with local measures such as warm compresses and analgesics.



Patients with superficial thrombophlebitis may harbor asymptomatic deep venous thrombosis, so screening duplex ultrasound is recommended.



Inferior vena cava filters should be reserved for patients with acute lower extremity or pelvic thrombosis who have a contraindication to anticoagulation or documented anticoagulation failure.



Inferior vena cava filters may reduce the incidence of all pulmonary embolism in the short run after diagnosis of a deep vein thrombosis but may also increase the risk of recurrent deep vein thrombosis in the long run.



Greater attention to prevention would reduce our need to provide thrombosis treatment to this challenging population of at-risk patients.



Pharmacologic prophylaxis is safe and effective in a wide range of surgical and medical oncology patients and settings.


Venous thromboembolic disease (VTE), including deep venous thrombosis (DVT) and pulmonary embolism (PE), is a common but often underdiagnosed and underappreciated clinical problem in oncology that results in significant patient morbidity and mortality. Timely and accurate diagnosis of VTE is imperative because of the unacceptable outcomes associated with a misdiagnosis. VTE diagnosis based on clinical grounds alone is unreliable, so physicians should select an appropriate objective diagnostic test to confirm or refute their clinical suspicions. Compression duplex ultrasound remains the first-line imaging test for both suspected upper and lower extremity DVT. Magnetic resonance venography (MRV) is a valid alternative when ultrasound is inconclusive. Helical (spiral) computed tomography (CT) and ventilation-perfusion (V/Q) lung scintigraphy remain the first-line imaging modalities for suspected acute PE. Spiral CT is preferred in cases of obvious pulmonary or pleural-based disease. Indeterminate initial studies should prompt performance of additional tests, possibly including the “gold standard”: contrast venography and pulmonary angiography. Evidence to date suggests that D-dimer assays might be unreliable in excluding VTE in cancer patients.

Standard VTE treatment practices, including the use of intravenous unfractionated heparin (UFH) for initial anticoagulation, oral warfarin for chronic anticoagulation, and the prescription of only 3 to 6 months of total therapy, might not be optimal in the setting of active cancer and ongoing anticancer therapy. Challenges of VTE management in cancer patients include heparin resistance due to excess circulating acute-phase proteins, increased recurrence rates during and after standard-intensity warfarin therapy, limited venous access to support therapeutic monitoring, and anticoagulation intensity-independent increased bleeding rates during anticoagulation. Bleeding during anticoagulation is of particular concern in patients with disease- or chemotherapy-related thrombocytopenia, central nervous system (CNS) involvement with cancer, and recent invasive procedures. Low-molecular-weight heparins (LMWHs) have been shown to be at least as effective and safe for initial anticoagulation compared with UFH in patients with acute VTE and have gained popularity in the setting of VTE in cancer. The recently approved synthetic antifactor Xa pentasaccharide, fondaparinux, may also be an acceptable option in the cancer patient. LMWHs and fondaparinux have the advantage of less nonspecific protein binding, subcutaneous weight-based dosing without the need for monitoring in most cases, and less heparin-induced thrombocytopenia. Recent trials demonstrated efficacy superiority of select LMWHs in place of oral warfarin for long-term anticoagulation in the cancer patient.

The challenges of proper VTE diagnosis and treatment in the cancer patient are best averted by expanded use of VTE-prevention modalities. Trials have shown that medically ill patients have a high risk of VTE similar to that of high-risk surgical patients. LMWHs and fondaparinux have emerged as the premier pharmacologic agents for VTE prevention. Extended prophylaxis beyond acute-care hospitalization has a role in minimizing the total risk of VTE development. Newer oral agents may improve on the efficacy, safety, and convenience of LMWH-based prophylaxis.

This chapter addresses the diagnosis, treatment, and prevention of VTE specifically in the cancer patient population. [1] [2] [3] The potential for anticoagulant therapy to enhance cancer patient survival and the role of anticoagulants that are in development to affect cancer patient care also are addressed.


The incidence of clinically apparent (i.e., symptomatic) VTE in cancer patients has been reported to be approximately 15%, with reported incidence rates ranging from 3.8% to 30.7%. [4] [5] This is in comparison to an age-adjusted VTE incidence of 2.5% in the general population.[6] The wide range of reported incidence rates likely reflects differences in patient tumor histology, investigator level of VTE clinical suspicion, and VTE confirmation methods. Nonetheless, studies have uniformly demonstrated a higher incidence of VTE in patients with cancer of a specific organ compared with those with benign diseases of the same organ. In a Scandinavian retrospective study of more than 63,000 patients hospitalized for acute VTE, 18% had been diagnosed with cancer before the VTE.[7] A large Medicare registry study revealed that cancer patients were diagnosed with VTE at the time of initial hospitalization at a higher rate than were patients admitted for nonmalignant disorders.[8]

The risk of postoperative DVT in cancer patients after general surgery is as high as 36% and exceeds that of noncancer surgical patients by 1.5- to 3.6-fold.[9] VTE detection rates as high as 50% in autopsy series have been reported but might still not represent the true magnitude of VTE-related illness in cancer.[10] Limiting factors in the proper antemortem detection of VTE in cancer patients include an inadequate index of suspicion by many clinicians and idiosyncrasies of VTE diagnostic modalities, as will be described later.[2] The former is exacerbated by the fact that VTE is often asymptomatic, and even when symptoms are present, they are often nonspecific or mistakenly attributed to the underlying malignancy itself.


Thrombosis may be a presenting feature of occult malignancy, a life-threatening component of early or advanced cancer, a harbinger of cancer recurrence, and a complication of anticancer therapy itself. Patients with cancer diagnosed at the time of acute VTE detection have a greater likelihood of distant metastases at the time of diagnosis compared with individuals without concomitant VTE [44% versus 35.1%; prevalence ratio: 1.26; 95% confidence interval (CI): 1.13 to 1.40] and significantly lower 1-year survival rates (12% versus 36%; P < 0.001).[7] VTE presentations in the cancer patient include symptomatic DVT, PE, superficial thrombophlebitis, central venous access device–associated thrombosis, arterial thrombosis, and nonbacterial thrombotic endocarditis. Asymptomatic VTE, although often viewed as clinically insignificant, can evolve into symptomatic VTE, and both may be particularly deleterious in the cancer patient by promoting vascular endothelial growth factor expression.[11] VTE is likely to be a common proximate cause of death in patients with solid tumors.

The natural history of VTE in the cancer patient differs significantly from that in the noncancer patient. Cancer patients are more likely than noncancer patients to present with proximal DVT.[12] Cancer patients have been shown to present with a greater initial thrombus burden, to experience greater clinical deterioration despite anticoagulant therapy, and to have less venographic improvement in response to standard treatment when compared with noncancer patients.[12]

The perception of many physicians has been that cancer patients have a greater propensity toward recurrent VTE, both during and after completion of a course of antithrombotic therapy.[13] Recent studies have substantiated the high risk of recurrence. A population-based retrospective study of 404 individuals showed that people with cancer have a twofold to threefold increased risk of recurrent VTE.[14]Antineoplastic therapy further accentuated the risk of recurrence in the cancer patient population. The 5-year cumulative incidence of recurrent VTE was 21.5% in a prospective cohort study of 738 consecutive patients with a first or second DVT.[15] The relative risk of recurrence was 1.97 in patients with cancer. Another prospective study reported an overall VTE recurrence rate of 10.3% in 58 patients with cancer compared with 4.7% in 297 patients without cancer. [16] [17] The fact that cancer patients had a higher recurrence rate while reportedly “therapeutically” anticoagulated suggests that the usual target international normalized ratio (INR) range of 2.0 to 3.0 might not be therapeutic at all. Prandoni and colleagues[18] recently reported a 20.7% (95% CI: 15.6–25.8) 12-month incidence of recurrent VTE in cancer patients with VTE compared with a 6.8% (95% CI: 3.9–9.7) 12-month incidence of recurrence in noncancer patients with VTE. The greatest risk of recurrent VTE was observed in patients with genitourinary tract, gastrointestinal tract, and lung cancers and predominantly during the first month of anticoagulation.[18]

A recent analysis of malignancy status-specific and INR range-specific VTE recurrence rates revealed that patients with VTE and malignancy (n = 261) have an overall thromboembolism recurrence rate of 27.1 events per 100 patient years compared with 9 recurrent events per 100 patient years in individuals with VTE and no malignancy.[19] In both patient populations, the rate of VTE recurrence was greatest during periods when the INR was 2.0 or less (the lower boundary of the target INR range). Patients without cancer had 15.9 recurrent events per 100 patient years during such periods of suboptimal warfarin anticoagulation, whereas patients with underlying malignancy had a VTE recurrence rate of 54 events per 100 patient years during similar periods of inadequate anticoagulation. Thus, it can be surmised that cancer patients are exquisitely sensitive to periods during which the INR is less than a target level of 2.0 to 3.0.

Venous thrombosis has been traditionally associated with aerodigestive tract adenocarcinomas involving the pancreas, stomach, and lungs. A closer look at DVT and PE incidence rates for different tumor histologic types, though, reveals that ovarian carcinoma, primary brain tumors, and lymphomas are among the four tumor types with the highest VTE rates.[8] This is of particular interest, considering that hematologic malignancies such as lymphoma have traditionally been viewed as coagulation-inert histologic types and management of malignancy-associated VTE has traditionally been viewed as the sole domain of the solid tumor medical oncologist. Thrombosis can affect the clinical course of all histologic types, of all stages, of all grades, and during any and all treatments.


Different tumors of different extent, in different patients, with different comorbidities, and different inherited hypercoagulable states likely promote the development of VTE by different combinations of tumor-associated and non-tumor-associated procoagulant mechanisms ( Table 46-1 ). Solid tumor–mediated extrinsic vascular compression and invasion can obstruct venous return, resulting in blood flow stasis, endothelial cell injury, and coagulation activation.[1] Tumor cells can directly promote thrombin generation by producing tissue factor, expressing the coagulation factor X activator known as cancer procoagulant, and by displaying surface sialic acid residues that can support nonenzymatic factor X activation. Tumor cells also can indirectly promote thrombin generation by eliciting tissue factor expression by monocytes and endothelial cells. Selected tumors may mediate an accentuation of platelet activation and accumulation, whereas other tumor cells may express surface phospholipid species such as phosphatidyl serine, which can support prothrombin and factor X activation.

Table 46-1   -- Venous Thromboembolic Event Risk Factors in Cancer Patients


Extrinsic vascular compression and invasion

Tissue factor production

Cancer procoagulant production

Sialic acid residue support of nonenzymatic factor X activation

Promotion of tissue factor production by monocytes and endothelial cells

Accentuated platelet activation and accumulation

Expression of phosphatidyl serine, which supports prothrombinase and tenase activity

Inflammation-mediated increases in factor VIII, fibrinogen, and von Willebrand factor

Impaired endogenous fibrinolysis due to excess levels of plasminogen activator-inhibitor 1

Acquired deficiencies of natural anticoagulants


Central venous access devices

Antineoplastic agent–induced platelet activation and endothelial cell damage

Antiangiogenesis therapy

Anthracycline-induced congestive heart failure




Malignancy-associated inflammation can result in increased concentrations of acute-phase proteins such as factor VIII, fibrinogen, and von Willebrand factor. Whether in the setting of active malignancy or in otherwise normal patients, elevations of these acute-phase proteins are associated with an increased risk of thrombosis.[20] Tumor-associated increases in plasminogen activator inhibitor-1 can result in impaired endogenous fibrinolysis.[21] Malignancy-associated acquired deficiencies of natural anticoagulant proteins such as protein S have been described.[22] Baseline profiling of cancer patients with acute VTE found elevated levels of factor VIII, von Willebrand factor, plasminogen activator inhibitor-1, and the dilute Russell viper venom time (a specific lupus anticoagulant assay) in 17%, 71%, 34%, and 36% of patients, respectively. Deficiency of protein C, protein S, and antithrombin were found in 36%, 56%, and 6% of patients, respectively.[3] This study underscored the multifactorial pathogenesis of cancer-associated thrombosis. The variation between different cancer patients with regard to their hypercoagulable tendencies makes specific prognostic testing for cancer-associated thrombosis challenging and of limited clinical usefulness.

Although coagulation-marker testing may fall short of being able to predict which cancer patients are most apt to develop VTE, such testing might assist in prognosis determination. Beer and colleagues[23]performed a prospective evaluation of the predictive value of coagulation-activation markers for survival in cancer patients. They quantified thrombin-antithrombin complex, prothrombin fragment 1+2 (F1+2), D-dimer, fibrin monomer, and fibrinopeptide A. In general, patients with active malignancy, those with active adenocarcinoma, those with extensive disease, and those who died during follow-up (mean: 17 months) had significantly higher marker levels. A comparison of first and fourth quartiles in active cancer patients revealed significant odds ratios of death for fibrin monomer, thrombin-antithrombin complex, D-dimer, F1+2, and fibrinopeptide A of 4.1, 2.8, 2.7, 2.4, and 2.4, respectively.[23]

Non-tumor-derived VTE risk factors in the cancer patient include central venous catheters and antineoplastic agents themselves. Central venous catheters are the major risk factor for upper-extremity DVTs in the cancer patient population and can precipitate superior vena cava (SVC) thrombosis and SVC syndrome. [24] [25] These DVTs may result in PE, cause catheter dysfunction, and serve as a nidus for catheter-related infection.[26] Antineoplastic agents themselves, including cytotoxic chemotherapy, selective estrogen-receptor modulators, antiangiogenic agents, and especially combinations of these drugs, are associated with an increased risk of VTE. [1] [27] Thalidomide, an immunomodulating and antiangiogenic agent, has been linked to VTE in patients with multiple myeloma, especially when combined with anthracycline antineoplastic drugs. Theorized mechanisms of antineoplastic agent-induced hypercoagulability are varied and have not been completely elucidated. Medical complications of cancer and its therapy (including congestive heart failure, major infection, pathologic fractures, extended immobility, and pre-existent VTE) can exacerbate the tendency to venous thrombosis. Common inherited risk factors for VTE such as factor V Leiden and prothrombin G20210A have not been found to be more prevalent in cancer patients with thrombosis than in the general population.[3]


Selection of the most appropriate and effective treatment for a patient's cancer depends on a timely and accurate assessment of tumor histology, disease stage, and patient performance status. In similar fashion, prescription of the most appropriate and effective treatment for a cancer patient's disease- or treatment-associated VTE is dependent on a timely and accurate diagnosis. Failure to surpass a minimum threshold of anticoagulant intensity [activated partial thromboplastin time (aPTT) >1.5 times control for intravenous UFH] within 24 hours of acute DVT diagnosis is associated with a markedly increased risk of late thrombosis recurrence.[28] It thus makes sense that patients with acute DVT in whom a proper diagnosis is significantly delayed should have similar suboptimal outcomes. Classification of venous thromboses as superficial versus deep and distal versus proximal, differentiation between acute and remote thrombotic events, and distinction between a venous filling defect and extrinsic vessel compression are required to ensure that patients are appropriately treated.

It has long been recognized that the diagnosis of DVT and PE made on clinical grounds alone is notoriously unreliable. The severity of limb edema and pain is often unrelated to the location and extent of DVT, whereas the symptoms of PE vary depending on the degree and extent of vessel occlusion, as well as on a patient's cardiopulmonary reserve. The classically described Homans’ sign (calf discomfort triggered by passive dorsiflexion of the foot) has been found in only 8% to 60% of symptomatic patients with confirmed DVT and in up to 40% of symptomatic individuals without DVT.[2] Half of the patients with clinically suspected DVT do not have the diagnosis confirmed by objective testing.[2] Data from the Prospective Investigation of PE Diagnosis (PIOPED) Study revealed that dyspnea, pleuritic chest pain, cough, and lower-extremity edema, among other symptoms, were present in similar frequencies (30%–70%) among patients with or without angiographically confirmed PE. In cancer patients, the clinical diagnosis of DVT or PE is unlikely to be more accurate and may be even less accurate.

VTE is often asymptomatic or minimally symptomatic, and even when symptoms are present, they are nonspecific and can be easily attributable to the underlying malignancy. Surveys have shown that constitutional symptoms such as weakness and fatigue occur in 50% to 70% of patients in hospice or palliative care,[2] whereas dyspnea and cough occur in 25% to 50% of such patients and in more than 40% of patients with advanced cancer, respectively. The incidence of upper-extremity (UE) edema due to lymphedema in women after axillary lymph node dissection and/or radiotherapy for breast cancer has been reported to range from 6% to 30%.[2] In a series of more than 2000 palliative care patients with different types of pain syndromes, 22% and 11% had lower-extremity (LE) and UE pain, respectively.[2] None of these symptoms is unique to a particular disorder, and the etiology may be related to the underlying malignancy itself, a venous thrombotic complication, or both. Likewise, worsening dyspnea in a patient with primary or metastatic lung cancer; limb edema in a patient with bulky pelvic, axillary, or mediastinal tumor or adenopathy; and abdominal pain after colectomy or abdominal hysterectomy in patients with colon cancer or uterine cancer are a few examples in which VTE may mimic, be confused with, or be coexistent with the underlying disease. Therefore, awareness that VTE is a common complication of cancer and possession of a high index of suspicion are necessary to avoid missing a diagnosis of DVT or PE.

Compared with noncancer patients, those with cancer have a greater risk of bleeding while receiving oral anticoagulant therapy and a threefold to sixfold higher rate of recurrent VTE. [18] [19] Therefore, the diagnosis of DVT and PE should not be made on clinical grounds alone, and objective diagnostic confirmation is mandatory. Failure to make a timely diagnosis of VTE can result in significant morbidity and mortality because of recurrent VTE, whereas empirical anticoagulation therapy without a confirmed diagnosis may expose the cancer patient to unnecessary and potentially avoidable risk in the absence of any tangible benefit. With the exception of cases of superficial thrombophlebitis, any signs and symptoms that are suggestive of VTE should be used not as diagnostic endpoints, but simply as a compelling reason to pursue further testing.

Although the methods that are used to confirm the diagnosis of DVT or PE in noncancer and cancer patients are the same, particular features of the underlying malignant disease may, in some circumstances, reduce the accuracy of those diagnostic methods. The presence of direct tumor invasion of blood vessels or extrinsic venous compression by a bulky tumor or adenopathy, as well as rare primary vascular tumors, may all lead to false-positive diagnoses of DVT. Likewise, primary pulmonary artery tumors and compression of a pulmonary artery or vein by tumor or adenopathy may result in impaired regional pulmonary perfusion. This might be interpreted as “high probability” or “indeterminate” for PE on V/Q lung scanning (V/Q scan) or can lead to interpretative pitfalls on helical (spiral) CT. Prior radiation therapy to the chest wall also might lead to false-positive perfusion defects on V/Q scan, whereas fluctuating platelet counts might impair the ability of nuclear scintigraphy methods to detect LE-DVT. Moreover, increased baseline D-dimer levels in some cancer patients and impaired endogenous fibrinolysis in others might at least in part explain the limitation of D-dimer assays in excluding VTE in this population. Thus, an understanding of the limitations inherent in the various diagnostic modalities and of the circumstances in which these modalities may result in false-positive and false-negative diagnosis are of the utmost importance. Because the cancer population may be subject to a higher rate of nondiagnostic, false-positive, and false-negative noninvasive tests than are patients without cancer, it is possible that many cancer patients will require more than one diagnostic test to prove or rule out conclusively a diagnosis of DVT or PE.

Historically, the diagnostic approach to DVT and PE has shifted from purely clinical (insensitive and nonspecific) and angiography-based (invasive) to being dependent primarily on noninvasive or minimally invasive imaging techniques. These newer methods are less accurate to detect calf, pelvic, and intra-abdominal DVT, as well as PE in the subsegmental branches of the pulmonary artery. As a consequence, a number of clinical management models have been reported and validated. [2] [29] These models combine clinical assessment (pretest probability of DVT or PE) with noninvasive imaging tests and D-dimer assays, with the goal of reducing the need for repeated or invasive confirmatory tests but without compromising patient safety. The majority of patients who were included in these studies had a low-to-moderate pretest VTE risk, and the combination of a low pretest risk with a negative imaging test and a normal D-dimer has been shown to exclude DVT or PE safely. However, in these studies, patients with high pretest clinical probability of VTE and a negative or indeterminate initial test result almost invariably went on to undergo additional testing, including invasive studies.

Given the high incidence of VTE and the impact of such a diagnosis on cancer patients, it is our opinion that all patients with active cancer should be considered as having a “high pretest probability” or “high clinical suspicion” in models that rely on pretest clinical assessment. In addition, management models that include currently available D-dimer assays as part of the risk assessment should not be used to aid in the diagnosis of VTE in the cancer population for reasons that will be discussed later in this chapter.


DVT and PE warrant prompt institution of antithrombotic therapy to effectively prevent thrombus propagation, embolization, and recurrence; to ameliorate patient symptoms; and to allow thrombus organization, plasmin-mediated lysis, and restoration of venous patency.[30] Specific therapy and duration of therapy in the cancer patient depend on thrombus location (e.g., iliofemoral DVT versus calf DVT), thrombus extent (e.g., massive PE versus subsegmental PE), underlying thrombosis “trigger” (e.g., major abdominal surgery versus thalidomide-based therapy for multiple myeloma), and patient comorbidities (e.g., self-limited thrombocytopenia and hemorrhagic brain metastases). Despite the special attention afforded the cancer patient with regard to other disease-related complications, such as hypercalcemia, nausea, fatigue, and pain, the recommended and most commonly used treatment of VTE in patients with active cancer is not significantly different from the regimens that are prescribed to VTE patients without malignancy. Ideally, unique features of individual cancer patients, the natural history of VTE in cancer patients, and cancer patient response to various anticoagulant agents should be noted and should affect our approach to VTE treatment in this group.

Heparin Resistance

Cancer patients, like others with acute illness and inflammatory processes, have a propensity toward heparin resistance. “True” heparin resistance causes inadequate anticoagulant and antithrombotic responses from what would otherwise be perceived as an adequate dose of heparin. Some have deemed a requirement of more than 35,000 units of heparin per 24-hour period, regardless of patient weight, to reflect this form of heparin resistance.[31] With true heparin resistance, both a measurement of anticoagulant activity such as the aPTT and a measurement of antithrombotic activity such as the antifactor Xa activity assay demonstrate inadequate degrees of heparin activity. True heparin resistance most likely results from the nonspecific binding of heparin to mononuclear white cells, vascular endothelial cells, and acute-phase protein such as histidine-rich glycoprotein, vitronectin, and platelet factor-4, resulting in an inadequate quantity of free or antithrombin-bound heparin.[1] Another potential cause of heparin resistance in the cancer patient is disseminated intravascular coagulation–associated antithrombin deficiency.

Cancer patients can also manifest an “apparent” heparin resistance characterized by dissociation between the aPTT and heparin assays.[31] In these patients, the aPTT may be normal or near normal, while the antifactor Xa activity assay reveals a heparin activity level within the therapeutic range of 0.3 and 0.7I U/mL. Simply escalating the dose of heparin to achieve the desired aPTT without checking a heparin assay can result in a pronounced bleeding risk. Dissociation between the aPTT and heparin concentration likely reflects elevated levels of factor VIII that can shorten the in vitro aPTT without affecting the antithrombotic actions of the drug.

Warfarin Failure

Warfarin failure is the term that is often used to describe the development of an objectively documented recurrent VTE, despite an apparently stable INR between 2.0 and 3.0. Such an event suggests that this degree of anticoagulation was insufficient to neutralize the sum of hypercoagulable stimuli in a given individual. Warfarin failure must be distinguished from early thrombus extension during the initial period of acute parenteral anticoagulation. Underlying cancer, because of its potent prothrombotic nature, is often suspected in the setting of warfarin failure. Patients with VTE and known cancer are at an increased risk for recurrent thrombosis compared with noncancer patients. [12] [14] [15] [18] [19] [32] This may reflect cancer-associated hypercoagulability in excess of warfarin-induced anticoagulation or reflect less ability to keep cancer patients within the target INR range. Cancer patients have been shown to spend less time (43.3%) within the target INR of 2.0 to 3.0 than do control noncancer patients (56.9%) during standard warfarin anticoagulation.[1] It is likely that cancer patients spend approximately 30% of their time with an INR of 2.0 or less and 30% of their time with an INR of 3.0 or more. The remaining time is probably spent in transit between these out-of-target extremes. As Hutten and associates have shown,[19] cancer patients are at a particularly high risk for recurrent VTE when the INR is 2.0 or less. An increased frequency of therapeutic monitoring does not necessarily improve outcome in cancer patients with VTE. [1] [17]


Cancer patients can experience periods of excess catabolism, anorexia, corticosteroid-induced appetite stimulation, antimicrobial therapy, parenteral nutrition, and compromised hepatic function.[33] Each can affect either vitamin K supply or metabolism and thus vitamin K antagonist (i.e., warfarin) therapy. These often unpredictable aspects of the cancer patient may contribute to the greater degree of INR instability during oral warfarin therapy. An increasing number of cancer patients are relying on over-the-counter dietary supplements and herbal preparations to compensate for malnutrition, alleviate symptoms, and complement traditional anticancer therapy. Unfortunately, these preparations may contain vitamin K, vitamin K analogs, or compounds that are known to affect warfarin anticoagulation (e.g., Ginkgo biloba). Reliance on acetaminophen-containing narcotic and nonnarcotic analgesics also can affect the toxicity of oral warfarin.[34]


Thrombocytopenia in the cancer patient can develop for a multitude of underlying reasons. Some patients have decreased platelet production secondary to myelosuppressive chemotherapy or marrow infiltration by tumor. Others have peripheral consumption of platelets due to hypersplenism associated with Hodgkin's disease and lymphoproliferative disorders or congestive splenomegaly due to thrombosis and/or portal hypertension associated with extensive hepatic metastases. Other patients experience peripheral platelet destruction due to autoimmune clearance associated with low-grade non-Hodgkin's lymphomas and chronic lymphocytic leukemia, whereas others have consumption due to disseminated intravascular coagulation. Thrombocytopenia that develops during heparin administration for VTE prevention or treatment should be considered heparin-induced thrombocytopenia (HIT) until proven otherwise.[35] Detection of HIT in patients with pre-existent cancer-associated thrombocytopenia can be difficult. Because cancer patients may possess antiheparin:platelet factor-4 antibodies even in the absence of clinical HIT, testing should be reserved for the patient with new thrombocytopenia that develops during or shortly after a heparin exposure and without another possible explanation.[3]

The safety of administering systemic anticoagulation for any duration of time in patients with thrombocytopenia is likely dependent on the etiology and degree of thrombocytopenia, the presence or absence of concomitant platelet hypofunction, the location of primary and metastatic tumors, platelet transfusion responsiveness, and the anticipated duration of thrombocytopenia. No exact cutoff value has been established below which it is uniformly unsafe to administer heparin, LMWH, or oral anticoagulation. The fact that diverse cutoff values of 20,000/μL, 50,000/μL, and 100,000/μL are often used reflects a lack of physician consensus about the bleeding risk associated with thrombocytopenia and a wide range of comfort levels with anticoagulation in this setting. Recent cancer-associated VTE treatment trials comparing long-term dalteparin and enoxaparin with standard warfarin anticoagulation used minimum platelet counts of 75,000/μL and 50,000/μL, respectively, as study inclusion criteria. [3] [36] On the basis of thromboelastographic assessment of hematologic malignancy patients, some patients with chemotherapy-induced thrombocytopenia may even have evidence of overwhelming hypercoagulability.[37]

Bleeding Tendencies

Contemporary evidence on the risk of anticoagulation-related hemorrhage in cancer patients remains conflicting; primarily retrospective studies support a greater bleeding risk in cancer patients, and mainly prospective cohort studies suggest that the risk is no greater than that in noncancer patients. The true risk most likely depends on the temporal relationship between anticoagulation and major invasive and surgical procedures, concomitant thrombocytopenia or antiplatelet medication consumption, and the location and vascularity of cancerous lesions. Erosive, friable endobronchial, gastrointestinal, and genitourinary lesions are more likely to be prone to bleed during anticoagulation than at baseline. As in the noncancer population, older cancer patients and those with anemia, diabetes, recent myocardial infarction, renal insufficiency, history of stroke, and history of gastrointestinal bleeding are most likely at a greater risk for warfarin-associated bleeding. [38] [39]

A large, retrospective, population-based study from the Mayo Clinic demonstrated that malignancy was associated with major bleeding with a relative hazard ratio of 4.26 (95% CI: 1.61–11.33) in univariate analysis and 4.07 (95% CI: 1.53–10.87) in multivariate analysis.[40] A recent retrospective analysis of 1303 patients enrolled in two large randomized, prospective trials comparing intravenous UFH with subcutaneous LMWH for initial VTE management also demonstrated a greater likelihood of major bleeding in patients with VTE and cancer.[19] The incidence of bleeding was 4.2, 2.1, and 13.3 events per 100 patient years in all patients, noncancer patients (n = 1039), and cancer patients (n = 264), respectively. The bleeding rate correlated with anticoagulation intensity in the noncancer patients but was independent of INR level in the cancer patient group. The major limitation of this study was the small total number of major bleeding events (n = 12). One small, prospective study revealed an odds ratio of 2.4 for bleeding in cancer patients with VTE compared with noncancer patients with VTE.[41] A recently published study revealed that the 12-month cumulative incidence of major bleeding in cancer patients treated for symptomatic DVT was 12.4% (95% CI: 6.5–18.2) compared with 4.9% (95% CI: 2.5–4.1) in treated noncancer patients (P = 0.015).[18] Two other prospective studies have shown that the overall and major bleeding rates noted in DVT patients with and without cancer were not significantly different. Prandoni and coworkers [16] [17] reported major bleeding rates of 3.4% and 3.0%, respectively, in VTE patients with and without cancer during the first 3 months of oral anticoagulation. Overall bleeding rates in the same groups were 8.6% and 9.8%, respectively.

A recently published analysis, the largest study to date, suggests that the risk of major bleeding in cancer patients receiving oral warfarin after acute VTE is increased approximately sixfold and independent of the INR intensity.[19]

Osteopenia and Lytic Skeletal Lesions

Cancer patients, those affected by multiple myeloma and breast carcinoma in particular, are often plagued by osteopenia, lytic skeletal lesions, and pathologic fractures. Immobility as a result of chronic bone pain and fractures may promote DVT. Prolonged administration of UFH can result in osteopenia and osteoporosis in up to 30% of noncancer patients, is associated with a 1% to 2% incidence of vertebral fracture, and certainly has the potential to exacerbate cancer-associated bone loss.[42] Experimental models have demonstrated simultaneous increases in osteoclast activity and decreased osteoblast activity in response to UFH exposure. Less osteoclast activation is observed in response to LMWH exposure.[43]

Primary and Metastatic Central Nervous System Tumors

Patients with primary brain tumors and metastatic CNS lesions are at an increased risk for developing VTE and are viewed by many as having an absolute contraindication to systemic anticoagulation.[44]CNS hemorrhage is of particular concern in patients with highly vascular brain metastases, such as those from choriocarcinoma, melanoma, and renal cell carcinoma. VTE in these cancer patients may prompt the placement of an inferior vena cava (IVC) filter instead of systemic anticoagulation. Controlled CNS metastases (i.e., after radiation therapy) are often viewed as being less prone to bleed. Available limited data from several series suggest that the risk of spontaneous intracranial bleeding in patients with primary brain tumors (mainly glioblastoma multiforme) and CNS metastases might not be greater in those who receive anticoagulant therapy compared with those not receiving anticoagulant therapy.[45] A major limitation of these reports is that most studied patients received prophylactic, and not treatment, intensity of UFH or warfarin. Careful dosing and frequent therapeutic monitoring are imperative, especially in patients receiving concomitant phenytoin therapy. In the ONCENOX study, six patients with primary CNS cancer, CNS lymphoma, or CNS metastases were treated per protocol without developing minor or major bleeding or a recurrent VTE. [3] [46]

Limited or Compromised Venous Access to Support Therapeutic Monitoring

Initial management of acute VTE with continuous infusion UFH requires short-term but consistent intravenous access and frequent phlebotomy for therapeutic monitoring. With regard to oral warfarin, a narrow therapeutic index and wide intraindividual and interindividual variations in degree of anticoagulation achieved with a particular dose warrant frequent therapeutic monitoring and dose adjustment. Both UFH and warfarin therapeutic monitoring are associated with cost, patient inconvenience, and patient discomfort. Because of fluctuations in nutritional status and unpredictable fluctuations in the INR in cancer patients, frequent INR monitoring (once to twice weekly) might be required. The need for frequent phlebotomy and intravenous-access insertion can result in a cancer patient having limited to no usable peripheral veins. Venous sampling via central venous catheter for therapeutic monitoring may result in false prolongations of the aPTT and prothrombin time due to sample contamination.

Persistent Hypercoagulability

Active residual cancer of any extent and active anticancer therapy of any form represent persistent hypercoagulable states. The magnitude of the prothrombotic stimulus may actually intensify with time as disease burden and metastases mount. As in other persistent hypercoagulable states, including congenital deficiencies of natural anticoagulants and chronic antiphospholipid antibodies, active malignancy and ongoing therapy should prompt an extended duration of anticoagulant therapy.


Several prospective, randomized, controlled trials have demonstrated the efficacy and safety equivalency of intravenous, aPTT-adjusted, UFH and subcutaneous, weight-based LMWH for the treatment of acute LE-DVT. [47] [48] [49] [50] [51] The major advantage of subcutaneous LMWH is that it can be self-administered at home, without the need for therapeutic monitoring. This translates into a significant reduction in mean hospital length of stay compared with UFH initial therapy (1.1 versus 6.5 days).[47] Patients may be begun on LMWH in the hospital and then discharged in an accelerated fashion to continue the bridging to oral warfarin or may be treated exclusively in the outpatient setting. A reduction in length or avoidance of hospitalization may be of particular importance in immunocompromised cancer patients who are prone to nosocomial infections. LMWH treatment does require once- or twice-daily subcutaneous injection but does not require phlebotomy for therapeutic monitoring in the majority of patients. Cancer patients might be even more adept and accepting of subcutaneous injections than are noncancer patients because of experience with self-administered hematopoietic growth factor therapy.

LMWHs are associated with less HIT than is UFH, especially in heparin-naive patients.[52] This statistic makes LMWHs particularly attractive for VTE prevention and treatment in cancer patients with disease- and chemotherapy-related thrombocytopenia in whom HIT detection may be hindered by pre-existent low platelet counts. LMWHs are associated with less osteopenia in animal models and could offer a theoretical treatment advantage in cancer patients with osteolytic lesions.[43] Osteoporosis, though, has been reported with long-term LMWH use, and an increase in spontaneous fracture has been described during long-term LMWH exposure in pregnant women.[42]

LMWHs display less nonspecific binding to acute-phase plasma proteins, platelets, mononuclear leukocytes, and endothelial cells.[51] Active cancer patients treated with LMWHs are thus theoretically less likely to experience true heparin resistance. LMWHs also have been shown to promote a small but significantly greater degree of thrombus regression and restoration of venous patency than does UFH.[12]Disadvantages of LMWHs include the inability to be completely reversed by protamine sulfate in the event of bleeding or unanticipated surgery. LMWH accumulation in patients with severe (creatinine clearance: <30 mL/min) renal insufficiency precludes its predictable use in cancer patients with renal failure.

Two early meta-analyses of randomized controlled clinical trials comparing LMWHs with UFH in patients with acute DVT demonstrated a reduction in short-term mortality in cancer patients treated with LMWH (relative risks: 0.44 and 0.33, respectively). [53] [54] More recent meta-analyses demonstrated an overall survival advantage in VTE patients treated with LMWHs compared with UFH. [55] [56] Much of this observed advantage was derived from a single trial comparing tinzaparin with UFH. [49] [57] A total of 97 cancer patients were included in the analysis, 47 of whom received tinzaparin. Death rates at 3 months were 10.6% in those randomized to tinzaparin and 28% in those who received UFH (P = 0.041).[49] Proposed mechanisms for the survival advantage include tumor growth retardation, metastasis prevention, tumor neovascularization inhibition, and fatal VTE prevention.[58] If real, the survival advantage associated with LMWH is likely due to a combination of effects. Properly powered prospective randomized trials are needed to confirm the meta-analysis and preclinical experimental observations.

Kakkar and associates[59] reported on a randomized, placebo-controlled trial of dalteparin, 5000 anti-Xa units daily, in patients with advanced solid tumor malignancy without evidence of underlying thrombosis, with the primary objective of determining effect on survival at 1 year. The Kaplan-Meier survival estimates at 1, 2, and 3 years after randomization were 42%, 19%, and 13%, respectively, for placebo and 45%, 27%, and 21%, respectively, for the dalteparin group (P = 0.29). Although no significant early impact on survival was seen, a post hoc analysis of those surviving more than 17 months demonstrated survival estimates at 2 and 3 years after randomization of 56% and 37%, respectively, for placebo versus 77% and 59%, respectively, for dalteparin (P = 0.04).[59] A recent analysis of the impact of dalteparin on survival in patients with thromboembolism reported no difference in mortality at 12 months between those treated for 6 months with dalteparin and those treated with oral warfarin (56% versus 58%).[60] A post hoc subgroup analysis suggested that long-term treatment-intensity dalteparin might reduce mortality in cancer patients with nonmetastatic disease and acute VTE compared with oral anticoagulation with warfarin (20% versus 35%). In a double-blind study in 302 patients with metastatic or locally advanced solid tumors, 6 weeks of subcutaneous nadroparin resulted in a median survival of 8 months in comparison to 6.6 months in those treated with placebo.[61] The major benefit from the nadroparin was observed in the prespecified subgroup of patients with a life expectancy of greater than or equal to 6 months at enrollment.[61] In a study of 84 patients with limited or extensive stage small cell lung cancer randomized to chemotherapy alone versus chemotherapy plus once-daily dalteparin 5000 U for 18 weeks, both progression-free survival (10 months versus 6 months) and overall survival (13 months versus 8 months) were longer in the chemotherapy plus LMWH arm (P = 0.01 for both).[62] Larger-scale trials are needed to confirm these findings.

LMWH preparations differ in manufacturing methods, mean molecular weight, molecular weight distribution, effect on tissue factor pathway inhibitor expression, and possibly clinical effect. In part, for these reasons, LMWHs should not be viewed as interchangeable, and favorable data on one preparation do not necessarily apply to any other.


Diagnosis of Lower-Extremity Deep Venous Thrombosis in Cancer Patients

The LE deep venous segments that can be affected by thrombosis include, in ascending order from the ankle, the paired calf veins (posterior tibial, anterior tibial, peroneal, gastrocnemius, and soleal), the popliteal, superficial femoral, deep femoral (profunda femoral), common femoral, and external and common iliac veins. The superficial and deep femoral veins converge to form the common femoral vein in the proximal thigh. Despite its terminology, the superficial femoral vein is actually a deep vein, not a superficial vein; isolated superficial femoral vein thrombosis should therefore be treated as a DVT and not viewed as a superficial thrombophlebitis. Currently available diagnostic methods for the objective diagnosis and exclusion of lower-extremity DVT include several imaging techniques and biochemical assays ( Table 46-2 ). Although contrast venography remains the gold standard method for lower-extremity DVT diagnosis, duplex ultrasound is the most appropriate initial diagnostic test owing to the combination of accuracy, noninvasiveness, short examination time, portability of newer equipment, and lower cost.

Table 46-2   -- Tests for the Diagnosis or Exclusion of Upper and Lower Extremity Deep Venous Thrombosis

Contrast venography

Duplex ultrasound

Impedance plethysmography

Venous-phase helical computed tomography

Magnetic resonance venography

[125I]-fibrinogen scan

[99mTechnetium]-apcitide scintigraphy

D-dimer semiquantitative and quantitative assays



Contrast Venography

Contrast venography has the ability to outline the deep venous system after the injection of radiopaque contrast medium into a dorsal foot vein. A central deep venous access (e.g., popliteal or common femoral vein) is usually necessary for adequate opacification of the pelvic (iliac) veins.[2] Available contrast agents include ionic and nonionic iodinated aromatic-acid salts of varying osmolality and CO2. CO2 may provide an uneven intravascular distribution but is not nephrotoxic and is quite inexpensive to use. The presence of a constant intraluminal filling defect in at least two distinct projections is themost reliable diagnostic criterion for acute DVT.[63] Nonfilling of one or more venous segments proximal to the site of injection, abrupt termination of the column of contrast at a constant site, and the presence of flow in collateral veins are indirect signs that can be caused by artifacts. Artifacts can result from improper contrast administration and flow artifacts, particularly at the common femoral vein level, when nonopacified blood from the deep femoral vein mixes with opacified blood from the superficial femoral vein.[2]

The greatest limitations of venography are primarily related to its invasiveness and need for intravenous injection of contrast medium. The procedure also requires meticulous technique as well as experienced radiologists to avoid misinterpretation of inadequate studies. Unsuccessful venograms have been reported in 2% to 20% of examinations, because of either inadequate technique with failure to outline a venous segment properly or inability to perform the test,[2] and interobserver disagreement ranges from 4% to 21%.[63] Although the deep femoral (profunda) vein is visualized only 50% of the time, isolated deep femoral vein DVT appears to be very rare.[2] Indeterminate findings also may result from the lack of visualization of the calf veins, which may occur because of nonfilling of the calf veins either because of improper test technique or occlusive DVT.[2] Other limitations include the fact that examinations cannot be performed at the bedside and are restricted to the limb where venous access has been gained.

Complications of contrast venography include postprocedure DVT in approximately 2% to 10% of examinations and superficial thrombophlebitis at the site of contrast injection.[2] Rare complications include tissue necrosis due to extravasation of contrast during injection and hypersensitivity reactions.

Nonfilling of the iliac veins during contrast venography in cancer patients with known or suspected pelvic masses neither rules in nor rules out DVT. Such finding may be due to DVT, extrinsic venous compression by tumor, or both. In one study, 51% of all high-grade non-Hodgkin's lymphoma patients who were diagnosed with VTE also had concomitant pelvic venous compression by bulky lymphadenopathy.[2] In this situation of nonvisualization of the pelvic veins, an alternative imaging method, such as CT or magnetic resonance imaging (MRI), should be considered to assess for the presence of extrinsic vascular compression.[2]

Duplex Ultrasound

A venous duplex ultrasound study combines real-time B-mode ultrasound with pulsed-color Doppler flow imaging. The former provides direct visualization of the vessels and surrounding tissues, whereas the latter detects blood flow when the emitted ultrasound energy is reflected by red blood cells and sensed at a different frequency (Doppler shift).[2] The lower-extremity deep venous segments that can routinely be examined by duplex ultrasound include the very distal external iliac, common femoral, superficial femoral, popliteal, and calf veins.

The diagnosis of DVT with duplex ultrasound relies on a combination of the following findings: vein noncompressibility, vein dilatation, visualization of echogenic intraluminal material, and lack of spontaneous and augmented blood flow. [2] [64] The most widely used and reliable, validated criterion is lack of vein wall compressibility on B-mode ultrasound. The other criteria are relatively inaccurate when applied individually. Most important, however, the use of these criteria in combination does appear to increase the accuracy of the examination.[65]

Although the sensitivity and specificity of compression duplex ultrasound for acute femoropopliteal DVT have been reported to be quite high in symptomatic patients (92%–100% and 94%–100%, respectively), this method lacks adequate sensitivity when used for screening of asymptomatic individuals.[2] In addition, duplex ultrasound has been reported to lack adequate sensitivity (36%–95%) for the detection of calf DVT, despite comparable specificity (89%–100%). [2] [65] A number of conditions can impair the ability of duplex ultrasound to detect calf vein thrombi, such as edema, large calf size, and the presence of open wounds, surgical bandages, or even large collateral veins, all of which lead to a high rate (10% to 40%) of inadequate examinations that impair interpretation.[2] This technical inadequacy is the likely explanation for the lower sensitivity of duplex ultrasound to detect calf DVT. Two studies comparing the sensitivities of all (adequate plus inadequate) calf examinations versus adequate-only examinations have shown sensitivities that increased from 73% and 85% to 95% and 99%, respectively.[2] In addition, the use of color Doppler has been shown to facilitate the localization and visualization of the calf veins, yielding fewer indeterminate examinations than venography in patients with isolated calf DVT.[2] Lack of color Doppler examination might explain the lower accuracy in early studies. Visualization of the calf veins by duplex ultrasound also has been enhanced by the use of an intravenous contrast agent, which in a small series led to reduction in the number of inadequate calf vein examinations from 55% to 20%.[66] Contrast ultrasound has the potential to improve the accuracy of this diagnostic modality even further.

The advantages of duplex ultrasound include the facts that it is noninvasive and that bilateral examinations can be performed in a timely fashion. Unlike venography, duplex ultrasound might also increase the overall diagnostic yield of calf examinations because of its ability to detect extravascular pathology, such as hematomas and Baker's cysts.[2]

Pitfalls of duplex ultrasound include the risk of false-positive incompressibility of the distal superficial femoral vein at the level of Hunter's (femoral) canal; the risk of setting the color gain inappropriately high, leading to “color-blossoming” that might obscure nonocclusive DVTs; and the potential to miss a DVT diagnosis in patients with duplicate venous systems or with the rare cases of isolated deep femoral DVTs.[2]

An important limitation of duplex ultrasound is the inability to perform compression maneuvers adequately in the veins above the inguinal ligament (common iliac and proximal external iliac veins). The lowest accuracy of duplex ultrasound was reported by a study that attempted to visualize and interpret iliac vein compression maneuvers routinely.[65] Although these venous segments may be visualized and interrogated for the presence of spontaneous blood flow and normal respiratory phasicity, these criteria are insensitive because normal flow may be present in cases of nonocclusive DVT. Only the distal 3 cm of the external iliac veins could be adequately visualized by one study, partial visualization of the external and common iliac veins being accomplished in 79% and 47% of the examinations, respectively.[2] Therefore, a negative LE duplex ultrasound does not rule out iliac vein thrombosis, whereas attempts to diagnose iliac vein DVT by compression ultrasound alone may lead to a number of false-positive DVT diagnoses. Conversely, the use of Doppler flow imaging may assist in the diagnosis of pelvic vein DVT or extrinsic compression: In a study of 37 cancer patients with leg edema and negative compression ultrasound, a 100% correlation was found between a monophasic waveform in the common femoral vein by spectral Doppler and the presence of either more proximal, not directly visualized DVT or extrinsic pelvic venous compression by a mass.[67] However, this finding does not differentiate iliac DVT from extrinsic venous compression by tumor.

A controversial issue is the need for bilateral LE duplex ultrasound in patients with unilateral LE symptoms. Six studies have shown rates of isolated DVT in the asymptomatic, contralateral leg ranging from 0% to 5%, with another 2% to 22% of patients having bilateral DVT despite unilateral symptoms. [2] [68] [69] One study that was performed exclusively in cancer patients found that 1% of patients had DVT in the asymptomatic contralateral leg, and an additional 7% had bilateral DVT.[68] A further study that used bilateral duplex ultrasound examinations in patients with unilateral symptoms found that eight (53%) patients who had DVT in the asymptomatic limb had active cancer, and in seven of those eight patients, symptoms developed in the previously asymptomatic limb during the first month of anticoagulation therapy.[69] Of these seven, four were diagnosed with a new DVT in a previously unaffected venous segment, but three were found not to have a new DVT, illustrating the relevance of detecting symptomless DVT in the cancer population.[69] Lack of documentation of acute DVT in an asymptomatic limb could negatively affect patient management by leading to a false diagnosis of recurrent VTE and “warfarin failure” in future duplex ultrasound examinations. Thus a cancer patient should always undergo bilateral duplex ultrasound examinations for suspected LE-DVT, even if the symptoms are unilateral.[2]

Another limitation of duplex ultrasound is that the accuracy of compression maneuvers to detect recurrent DVT is uncertain. The ability to distinguish acute from remote (chronic) DVT is hampered by the fact that 50% of patients with prior LE-DVT will have some degree of residual vein obstruction.[2] Old, organized thrombi may appear hyperechoic and heterogeneous sonographically, but this might not help to differentiate acute from remote DVT.[64] Unless the new DVT is found in a previously normal venous segment, compression ultrasound could be unreliable. Although serial duplex ultrasound evaluation of vein diameters and comparison with prior ultrasound measurements has been proposed as an alternative means of diagnosing recurrent DVT (with an incremental increase in vein diameter being attributed to recent DVT),[70] this method has not been validated by a large-scale prospective comparison with contrast venography. It also is unclear whether the same protocol would be as useful in cancer as in noncancer patients.

Contrast-Enhanced Computed Tomography

CT has not been validated as a method for DVT diagnosis (i.e., no accuracy studies have been performed comparing it with the gold standard contrast venography, no interobserver variability has been formally assessed, and no validation (outcome) studies have been performed). [2] [71] Therefore, it should not be routinely used for this purpose. However, because cancer patients frequently undergo body CT as a component of initial tumor staging, assessment of cancer recurrence and surveillance, or a workup for persistent fever, physicians are frequently faced with the dilemma of an incidental finding of filling defects involving the pelvic veins or common femoral veins. In this situation, it is imperative that the diagnosis be confirmed by a validated method, such as duplex ultrasound or contrast venography.[2]

The use of combined helical CT pulmonary angiography with CT venography (venous phase CT) of the pelvis and proximal legs has been studied as a means to detect DVT in patients with suspected PE.[2]The examination starts 2 to 4 minutes after contrast is injected for helical CT of the chest, and images are obtained either at 4- to 5-cm intervals or contiguously from the diaphragm to the ankles. Criteria for acute DVT include visualization of a filling defect in an opacified vein, a nonopacified segment between normally opacified proximal and distal segments, venous dilatation (when compared with the contralateral side), and venous wall ring enhancement. The sensitivity and specificity range from 71% to 100% and 87% to 100%, respectively, when compared with duplex ultrasound.

Advantages of helical CT venography include its ability to visualize the IVC, portal, ovarian, and renal veins, as well as the soft tissues, to detect the presence or absence of intra-abdominal DVT and concomitant extrinsic venous compression.[72] No formal studies have been performed to assess the usefulness of this modality in differentiating acute from chronic DVT, although possible signs of chronicity include venous wall calcification and “shrunken” vessels. A disadvantage of helical CT venography is that it requires the use of iodinated contrast and if contiguous image acquisition is used the effective radiation dose is exponentially increased. This possibility has led some to state that the use of this technique should be limited in individuals younger than 30 years.[2]

False-positive DVT diagnoses have occurred because of flow artifacts, particularly at the level of the calf and pelvic veins and also because of muscle hematomas and abscesses.[72] False-negative findings have occurred in cases of extensive bilateral DVTs, in which no normal vein enhancement in either limb is seen, so no contralateral normal vein is present for comparison, and a case of a thrombosed left-sided IVC that was misinterpreted as necrotic lymph nodes.

Although observational and accuracy studies of helical CT venography do exist, no comparisons with venography and no clinical outcome studies have been performed. The appropriateness of this technique in cancer clinical practice is unclear at this time.

Magnetic Resonance Venography

Magnetic resonance angiography (MRA) to evaluate the venous system (MRV) can be performed with time-of-flight and phase-contrast techniques. The most commonly used is an axial, two-dimensional time-of-flight technique based on standard two-dimensional gradient-echo imaging, with or without gadolinium enhancement.[2] The observed signs of acute DVT as seen by MRV include total obstruction of the vein and venous dilatation, as well as the presence of a rim of increased signal intensity surrounding the thrombus. Although one study suggested that the pattern of rim enhancement was useful in differentiating acute from chronic DVT, no correlation with contrast venography was performed.[73]

MRV has been shown to be as accurate as duplex ultrasound for diagnosing LE-DVT in three prospective studies that compared the two methods with contrast venography (100% sensitivity, 95%–100% specificity). False-positive diagnoses occurred in patients with extrinsic compression of the iliac veins. The method is less accurate when evaluating for calf DVT.

Advantages of MRV include the lack of need for intravenous iodinated contrast and the ability to assess for the presence of extrinsic venous compression when combined with soft tissue-weighted MRI and the fact that, unlike duplex ultrasound, MRV can routinely visualize the pelvic veins.[74] Disadvantages include cost and lack of wide availability.

Nuclear Scintigraphy

Activated platelets expressing glycoprotein IIb/IIIa (GP IIb/IIIa) that become incorporated into acute evolving venous thromboses serve as the physiologic target for [99mtechnetium]-apcitide (AcuTect) in the diagnosis of LE-DVT. A phase III prospective clinical trial in 280 patients who underwent both [99mTc]-apcitide scintigraphy and contrast venography demonstrated an overall sensitivity and specificity of 76% and 73%, respectively, for imaging acute LE-DVT.[75] Higher sensitivity and specificity (91% and 84%, respectively) were seen in patients with their first DVT event and signs and symptoms of less than 3 days’ duration, and false-negative diagnoses occurred more frequently in patients with calf DVT.[75] Because [99mTc]-apcitide is a functional rather than anatomic diagnostic imaging method, it has the theoretical potential to discern acute from chronic (without activated platelets) thrombus. In cancer patients, thrombocytopenia of any etiology and tumor-associated platelet activation likely limit the usefulness of this technique for the diagnosis of acute LE-DVT.

[125I]-Fibrinogen leg scanning was used mostly in the 1970s and early 1980s for the diagnosis of acute LE-DVT. It was highly sensitive (90%) for calf DVT but insensitive (60%–80%) for DVT involving the proximal veins because the isotope is a relatively low gamma emitter and because the bladder frequently contained radioactive urine. The use of this method has largely been abandoned because of its attendant risk of transmissible viral diseases and the wide availability of duplex ultrasound. In addition, [125I]-fibrinogen leg scanning was never used alone for the diagnosis of acute DVT because it can take up to 72 hours to become positive. Other radiolabeled peptides under investigation for DVT diagnosis include [99mTc]-DMP 444 and [99mTc]-FDB (fibrin domain of fibronectin). A radiolabeled antibody targeted against the DD domain of fibrin also is under development.


D-dimer is a cross-linked degradation product resulting from the plasmin-mediated lysis of cross-linked fibrin. Three readily available techniques are used to assay D-dimer: enzyme-linked immunosorbent assays (ELISA), plasma-based latex agglutination, and whole-blood hemagglutination (WBA) assays. Qualitative latex agglutination assays are relatively inexpensive to perform, widely available, and rapidly performed but are not sufficiently sensitive to exclude VTE. Both ELISA and WBA assays have been prospectively studied in clinical management trials and found to have high negative predictive values in outpatients with suspected VTE.[2] These studies suggest that lack of D-dimer elevation combined with a negative noninvasive imaging test reliably excludes LE-DVT, but D-dimer elevation alone neither rules in nor rules out thrombosis. In general, these D-dimer assays are excellent screening tests for VTE, but the negative predictive value of the WBA D-dimer assay has been shown to be lower in cancer patients (78.9%) than in those without cancer (96.5%).[76] Thus, it is not as useful for excluding VTE in cancer patients. Conversely, false-negative D-dimer results have been described in cancer patients with PE and baseline impaired endogenous fibrinolysis due to excessive levels of plasma plasminogen activator inhibitor-1.[21] More recently, however, both a subgroup analysis of a retrospective cohort study and a prospective cohort study showed that the negative predictive value of a new latex agglutination assay remained high and reliably excluded DVT in cancer patients. [77] [78]However, confidence intervals in the latter study were somewhat wide; therefore, additional studies are needed to determine whether the D-dimer assay is truly reliable in excluding DVT in the cancer population.[78] Until conclusive evidence from larger studies is available, D-dimer assays should not be used to exclude VTE in cancer patients.

Lower-Extremity Deep Venous Thrombosis

Patients with clinically suspected LE-DVT or an incidental finding of DVT on CT should be initially evaluated with duplex ultrasound. A nondiagnostic ultrasound examination should prompt further evaluation with contrast venography. If the patient is known to have a pelvic mass and has either no opacification of the pelvic veins by venography or a monophasic Doppler signal during sonographic interrogation of the common femoral vein, MRV or CT should also be considered to rule out concomitant extrinsic venous compression. Given the high rate of VTE recurrence and “warfarin failure” in cancer patients, it is advisable to pursue contrast venography in cases of suspected DVT recurrence in which the duplex ultrasound does not detect a new DVT in a previously normal segment. Unfortunately, however, venography also has limitations in diagnosing recurrent DVT involving previously involved venous segments, and neither MRV nor CT has been validated or conclusively shown to differentiate acute from chronic DVT reliably.

Treatment of Lower-Extremity Deep Venous Thrombosis in Cancer Patients

General guidelines for VTE management have been published elsewhere. [30] [79] Specific, evidence-based guidelines for VTE management in the cancer population are lacking. Analysis of recently completed and ongoing clinical trials could provide the basis for future detailed guidelines. In the meantime, cancer patients with VTE should be treated in a manner that takes the previously described unique features, subgroup analyses of prospective, randomized VTE treatment trials, and recently completed LMWH trials into account ( Box 46-1 ).

Box 46-1 


Initial Phase Anticoagulant



Acute treatment with a parenteral agent (unfractionated heparin or low-molecular-weight heparin) should be initiated in the inpatient or outpatient setting.



Intravenous unfractionated heparin should initially be dosed on the basis of patient weight and adjusted to achieve an activated partial thromboplastin time that corresponds to an antifactor Xa activity level of 0.3 to 0.7 U/mL.



Subcutaneous low-molecular-weight heparins should be dosed on the basis of patient weight, avoided in individuals with calculated or actual creatinine clearance rates less than 30 mL/min, and not monitored in the majority of cases.



When using enoxaparin sodium for initial phase anticoagulation, I prefer to use 1 mg/kg twice daily instead of 1.5 mg/kg once daily.



I prefer using a low-molecular-weight heparin rather than unfractionated heparin to facilitate outpatient management and because of evidence suggesting a possible survival advantage.

Subacute Phase Anticoagulant: Up to 6 Months



Subacute management can consist of oral warfarin with a target international normalized ratio (INR) between 2.0 and 3.0 or once daily subcutaneous low-molecular-weight heparin without therapeutic monitoring.



Warfarin therapy requires a minimum of 4 days overlap with the parenteral agent used for initial phase anticoagulant and can be difficult to maintain between an INR of 2.0 and 3.0 because of variable nutrition, concomitant medications, and hepatic dysfunction.



I favor low-molecular-weight heparin because of excellent patient compliance, reduced on-treatment thrombosis recurrence rates, and a possible survival advantage.

Chronic Phase Anticoagulant: Beyond 6 Months



Patients with persistent hypercoagulability from their malignancy, anticancer therapy, antiphospholipid antibody, or underlying inherited prothrombotic state such as deficiency of a natural anticoagulant have their subacute phase therapy extended until the persistent hypercoagulable state has resolved or completed.



Attenuated intensity therapy with warfarin (INR 1.5 to 2.0) and low-molecular-weight heparin (primary prophylaxis intensity) are not recommended for patients with active malignancy and those receiving anticancer therapy.

Initial Anticoagulation Therapy

The mainstay of pharmacologic therapy for VTE in all patients remains anticoagulation. Initial (acute-phase) therapy typically consists of parenteral UFH or LMWH for a minimum of 4 days and until a stable, target intensity of warfarin treatment has been achieved.[79] Treatment with UFH or LMWH should be begun as soon as possible after VTE diagnosis unless an absolute contraindication exists. Whenever possible, therapy should actually begin as soon as VTE is suspected and even before diagnostic tests are obtained. A delay in achieving a therapeutic intensity of initial parenteral therapy may negatively affect a patient's long-term VTE recurrence rate. [28] [80] Weight-based initial dosing of UFH (80-U/kg bolus followed by 18 U/kg/hr) with subsequent dose adjustments based on a standardized nomogram achieves a therapeutic aPTT within 24 hours of treatment commencement.[81] Because of problems with heparin resistance, greater than usual doses of UFH may be required in the cancer patient.

LMWH may be preferred for both initial inpatient and outpatient treatment of acute VTE in the stable cancer patient for the safety, efficacy, and survival reasons already discussed. The optimal LMWH preparation, dose, and dosing frequency remain to be determined. Acute DVT treatment safety and efficacy data exist for enoxaparin, 1.5 mg/kg once daily, and enoxaparin, 1.0 mg/kg twice daily, in cancer patients.[82] No statistically significant difference in VTE recurrence rate was observed between patients randomized to initial treatment with intravenous UFH, once-daily enoxaparin, and twice-daily enoxaparin. Of the 47 cancer patients randomized to the twice-daily dosing (total daily dose of 2.0 mg/kg), 3 (6.4%) developed recurrent VTE compared with 6 (12.2%) of the 49 cancer patients allocated to once-daily dosing (total daily dose of 1.5 mg/kg). This trend has led many physicians to advocate twice-daily dosing of enoxaparin for initial VTE treatment in cancer patients. The observed difference probably reflects the difference in total daily dose rather than an inherent inadequacy of once-daily dosing. This fact is important to note in considering the use of once-daily LMWHs such as tinzaparin and dalteparin. Tinzaparin, 175 IU/kg once daily, has an excellent safety and efficacy track record in cancer patients.[49] Dalteparin is given at a dose of 200 IU/kg up to a maximal dose of 18,000 IU once daily. Whether these or higher total daily doses of these two agents given as divided twice-daily injections in cancer patients would be superior to the standard dosing is not known. Concerns about underdosing cancer patients weighing more than 90 kg with standard-dose dalteparin might be justified.

Despite the excellent safety profiles of LMWHs in cancer and noncancer patients alike, initial inpatient UFH treatment may be preferred in patients who are at very high risk for bleeding and in those who are likely to require urgent invasive procedures. Such patients include those with recent surgery, gastrointestinal lesions, any past gastrointestinal or neuraxial bleeding, significant anemia, and marked thrombocytopenia. At present, cancer patients with severe renal dysfunction and most weighing more than 120 kg should be treated with adjusted-dose, monitored UFH. On the basis of a recent pharmacokinetic analysis in obese patients up to 165 kg, tinzaparin appears to be able to be dosed on the basis of actual weight without dose adjustment.[83] Patients with nonhemorrhagic CNS primary tumors or metastatic CNS lesions and VTE should be considered for anticoagulation, preferably begun in the hospital. The randomized study of enoxaparin sodium alone versus initial enoxaparin sodium followed by warfarin for a 180-day period as secondary prevention of venous thromboembolic events in patients with active malignancy (ONCENOX) trial enrolled six patients with confirmed CNS malignancy, all of whom were treated exclusively in the outpatient setting and in none of whom did recurrent VTE or bleeding complications develop.[3] Patients with hemorrhagic CNS lesions are probably best treated with IVC filter placement.

Subcutaneous fondaparinux, 7.5 mg once daily, has been shown to be as safe and effective as subcutaneous enoxaparin for the initial treatment of patients with acute DVT and adjusted-dose intravenous UFH for the initial treatment of patients with acute PE. [84] [85] Specific data in patients with active cancer and VTE have not been published.

Chronic Anticoagulation Therapy

Chronic-phase anticoagulation for VTE has traditionally consisted of oral warfarin dosed to achieve an INR between 2.0 and 3.0. Cancer patients with lupus anticoagulants and baseline elevated prothrombin times may require alternative warfarin therapeutic monitoring in place of the INR.[86] Chromagenic factor X activity assays or assessment of individual vitamin K-dependent factor activity levels on dilute plasma samples are acceptable alternatives. Warfarin therapy can be started as soon as a therapeutic-intensity aPTT has been achieved with UFH or an initial weight-based dose of LMWH has been given. Bolus dosing of warfarin does not help to achieve a stable, target INR faster and might actually delay achievement of a stable INR and prolong hospitalization.[87] Initial dosing with 2.5 to 7.5 mg per day (based on patient weight and nutritional status) seems prudent. Frequent (weekly) INR monitoring might not actually facilitate a more stable INR in cancer patients but still seems prudent. Warfarin therapy alone is contraindicated in the setting of acute thrombosis because of the inherent delay in achieving therapeutic anticoagulation and the theoretical transient exacerbation of hypercoagulability caused by a rapid reduction in protein C functional activity.[87] This warfarin-induced paradoxical hypercoagulability may contribute to warfarin-induced limb gangrene in patients with HIT and to warfarin-induced skin necrosis and may be particularly troublesome in patients with hypercoagulability of malignancy.

Warfarin Failure

Patients in whom objectively confirmed recurrent VTE develops during periods of subtarget INR (≤2.0) should be restarted on treatment-intensity UFH, LMWH, or fondaparinux until a stable INR between 2.0 and 3.0 is achieved. Patients in whom objectively confirmed recurrent VTE develops despite an INR between 2.0 and 3.0 (warfarin failure) can either be treated with UFH or LMWH until a higher intensity of oral anticoagulation (INR: 3.0–4.0) is attained or be switched to primary long-term therapy with LMWH. [32] [88] LMWH therapy is gaining popularity in patients with warfarin failure because of the challenges of warfarin therapy regulation at any target intensity, and data from early randomized trials suggest efficacy and safety comparability to warfarin.[1] The optimal long-term anticoagulation dose of any LMWH is not known and may be less than the dose used during initial VTE treatment. The ONCENOX trial evaluated the feasibility, safety, and efficacy of long-term enoxaparin at 1.5 mg/kg once daily and 1.0 mg/kg once daily based on this supposition.[3] In the event of warfarin failure followed by LMWH or fondaparinux failure, adjusted-dose, subcutaneous direct-thrombin inhibitor therapy with lepirudin should be considered. [89] [90]

Long-Term Anticoagulation Therapy with Low-Molecular-Weight Heparin

Each of the three commercially available LMWHs in the United States has been recently studied as a substitute for oral warfarin in the management of cancer patients with acute VTE ( Table 46-3 ). [3] [36] [91] [92] In the randomized trial of long-term dalteparin LMWH versus oral anticoagulant therapy in cancer patients with VTE (CLOT), 8.0% of LMWH-treated patients experienced recurrent VTE during 6 months of treatment compared with 15.8% of those treated with warfarin (target INR: 2.5).[36] In the ONCENOX trial, in 3.3% of the patients treated for 180 days with one of the two once-daily doses of enoxaparin, recurrent VTE developed, compared with 6.7% of those treated with warfarin (target INR: 2.0–3.0).[3] In the comparison of LMWH and warfarin for the secondary prevention of VTE in patients with cancer (CANTHANOX) trial, in 3.0% of the patients treated for 3 months with enoxaparin, 1.5 mg/kg once daily, recurrent VTE developed, compared with 4.2% of those treated with oral warfarin.[92] In the cancer subgroup of the randomized trial evaluating long-term LMWH therapy for 3 months versus intravenous heparin followed by warfarin sodium (LITE), in 5.9% of those treated with tinzaparin, recurrent VTE developed, compared with 10.5% of those treated with heparin followed by warfarin.[91] Major bleeding rates were not statistically different between the treatment groups in these clinical trials. These studies support the use of once-daily subcutaneous LMWH in place of oral warfarin in cancer patients with acute VTE to minimize recurrent VTE rates.

Table 46-3   -- Long-Term Low-Molecular-Weight Heparin Therapy in Cancer-Associated Venous Thromboembolic Events






CLOT (dalteparin)[36]

27/336 (8.0)

53/336 (15.8)

LITE (tinzaparin)[91]

6/101 (5.9)

11/105 (10.5)

CANTHANOX (enoxaparin)[92]

2/67 (3.0)

3/71 (4.2)

ONCENOX (enoxaparin)[3]

2/61 (3.3)

2/30 (6.7)


37/565 (6.5)

69/542 (12.7)



Duration of Anticoagulation

With regard to the optimal duration of anticoagulation in patients with VTE in the setting of cancer, it seems prudent to treat for a minimum of 6 months and at least until all cancer therapy has been completed and the patient has been deemed to have no residual malignancy. Residual malignancy constitutes a persistent hypercoagulable state, which typically warrants long-term anticoagulant treatment. Warfarin therapy itself has been shown to improve survival in patients with extensive-stage small cell lung carcinoma and a longer duration of oral anticoagulation (6 months versus 6 weeks) after acute VTE has been shown to reduce the risk of developing genitourinary tract cancers. [93] [94]

The recently published study of long-term, low-intensity warfarin therapy for the prevention of recurrent VTE (PREVENT) demonstrated a 64% reduction in recurrent VTE without increased risk of major bleeding compared with placebo when warfarin with a target INR of 1.5 to 2.0 is prescribed to patients after completion of a standard course of anticoagulation for idiopathic VTE.[95] This study specifically excluded patients with known active malignancy and therefore should not be applied to this population.


Diagnosis of Upper-Extremity Deep Venous Thrombosis in Cancer Patients

The UE deep venous segments that can be affected by thrombosis include, in ascending order from the elbow, the brachial, axillary, subclavian, and brachiocephalic veins, as well as the SVC. Similar to LE-DVT, the prevalence of confirmed DVT is less than 50% among symptomatic patients who are suspected of having UE-DVT.[25] Therefore diagnostic confirmation by an imaging method is mandatory. In general, the same modalities that are used for the diagnostic approach of LE-DVT apply for patients with suspected UE-DVT. However, no formal studies of helical CT venography, nuclear scintigraphy, or impedance plethysmography have been performed to diagnose DVT in the upper extremities, and as was previously discussed, D-dimer assays should not be considered reliable for the exclusion of DVT in a patient with cancer.

Contrast Venography

A number of anatomic variants and the converging nature of the central venous anatomy in the UE can lead to flow turbulence artifacts that can complicate the interpretation of UE venography.[2] In addition, because the technique is usually performed by contrast injection in an antecubital vein, it does not visualize the internal jugular veins, which can be easily visualized by duplex ultrasound. Similar to its applicability in the lower extremities, the single most reliable criterion for acute DVT is the presence of a constant, intraluminal filling defect in at least two different projection views. The complications related to the use of contrast medium are infrequent and similar to those described for LE-DVT. However, no data are available on the incidence of DVT after venography in the UEs.

Duplex Ultrasound

The internal jugular, subclavian, axillary, and brachial veins can be routinely visualized in the neck and arms. However, the medial two thirds of the subclavian veins are not easily compressible because of their anatomic location behind the clavicles.[2] In addition, it is not technically feasible to perform compression maneuvers at the level of the brachiocephalic veins and the SVC; in fact, the right brachiocephalic vein and the SVC are usually not visualized by ultrasound examination. Even if these segments of the subclavian veins and left brachiocephalic veins are visualized, duplex ultrasound has not been validated for the detection or exclusion of DVT in these locations because the diagnosis cannot be established by the compression method. Rather, DVT in those locations is suggested by indirect Doppler-flow criteria, such as lack of diameter change with inspiration and incomplete color filling of the lumen, and by the presence of echogenic material on B-mode ultrasound imaging.[2]

The sensitivity and specificity of duplex ultrasound for the diagnosis of acute symptomatic axillosubclavian DVT range from 96% to 100% and from 94% to 100%, respectively. [2] [25] [96] [97] Studies of the accuracy associated with other sonographic criteria than vein compressibility have been found to possess a lower sensitivity of 50% to 73%.[96] The presence of an indwelling catheter in an UE vein has been shown not to alter significantly the Doppler-flow dynamics compared with the contralateral vein that does not have a catheter in place. Nevertheless, duplex ultrasound appears to be unreliable as a screening method to diagnose asymptomatic, catheter-related UE-DVT.

Other Imaging Modalities

Neither contrast-enhanced CT nor MRV has been validated in evaluating the UEs, including the SVC. CT is not an ideal method because the convergence of the central veins in the upper chest is frequently associated with flow artifacts. Thus, the diagnosis of any suspected SVC or innominate DVT found by CT should ideally be confirmed by contrast venography. No studies have reported on the use of helical CT venography for central chest veins.

Two early studies (28 and 25 arms examined, respectively) correlating MRV with contrast venography for the diagnosis of acute axillary-subclavian DVT found a sensitivity and specificity of 80% and 100%, respectively. [98] [99] However, in one study, the sensitivity for nonocclusive DVT was quite low (20%) in comparison with the one for occlusive DVT (80%).[98] Nevertheless, MRV appears to be very accurate in detecting conditions that may be associated with UE or central chest vein extrinsic compression or stenosis. Two studies in selected patients with UE central venous abnormalities showed 100% correlation between 3D, gadolinium-enhanced MRV, and contrast venography for detecting SVC and brachiocephalic vein stenosis or compression.

Treatment of Upper-Extremity Deep Venous Thrombosis in Cancer Patients

Central venous catheter–associated DVT has been described in up to 56% of patients with indwelling catheters.[100] These “UE” DVTs may result in SVC thrombosis, SVC syndrome, and PE. The true risk of PE from central venous catheter–associated thromboses is often debated. Prandoni and coworkers and Bernardi[25] reported a 36% rate of PE complicating UE DVT; Monreal and colleagues[101]reported a 16% rate of PE in this setting. Ault and Artal,[102] in contrast, suggested that PE is uncommon in patients with central venous catheters based on a lack of any PE in 33 of their patients with DVT due to peripherally placed central catheters. In patients with symptomatic central venous catheter–related DVT and a functional catheter, line removal is often unnecessary. Anticoagulation management following the same guidelines as for LE-DVT in the cancer patient is recommended. Continuation of anticoagulation for the life of the catheter seems reasonable. This management approach allows uninterrupted cancer treatment and prevents the need for additional vascular-access surgery. Central venous catheter–related DVT in conjunction with a dysfunctional catheter usually warrants anticoagulation and line removal. Thrombolytic therapy followed by anticoagulation may alleviate thrombosis-related symptoms more quickly and completely than anticoagulation alone. Randomized trials comparing thrombolysis with anticoagulation for central venous catheter–associated DVTs are lacking.


Diagnosis of Pulmonary Embolism in Cancer Patients

Currently available tests for the diagnostic confirmation or exclusion of PE include various imaging methods and blood-based biochemical assays ( Table 46-4 ). Although many believe that pulmonary angiography is no more accurate than helical CT, pulmonary angiography remains the gold standard diagnostic test for PE. However, helical CT and V/Q lung scanning are the most appropriate initial tests because they are noninvasive and widely available.

Table 46-4   -- Tests for the Objective Diagnosis or Exclusion of Pulmonary Embolism



Pulmonary angiography



Ventilation-perfusion lung scintigraphy



Helical (spiral) computed tomography



Magnetic resonance angiography



D-dimer semiquantitative and quantitative assays



Pulmonary Angiography

Pulmonary angiography is typically performed via common femoral vein access (or internal jugular or brachial vein) and can include selective catheterization of either or both the right and left main pulmonary arteries. The pulmonary vascular tree is then visualized after an injection of iodinated contrast medium.

On the basis of clinical-outcome studies, the sensitivity and specificity of pulmonary angiography have been estimated to be 98% and 97%, respectively. [2] [30] [103] The specificity of pulmonary angiography approaches 100% when a filling defect or abrupt “cutoff” of a pulmonary artery branch is present.[103] Ancillary findings that may be present but are not specific for PE include abnormal distribution of flow to the different lobes, delayed venous return, and partially opacified, tortuous vessels.[103] The accuracy of the test is not influenced by the presence of chronic obstructive pulmonary disease.[103] In the PIOPED study, interobserver disagreement occurred more often for the exclusion of PE (17%) than for PE confirmation (8%) by pulmonary angiography.[104] Experts agreed 98%, 90%, and 66% of the time on the presence of lobar, segmental, and subsegmental PE, respectively.[105]

Complications reported with pulmonary angiography include minor and major complications as well as death. These three endpoints were observed in 5%, 1%, and 0.5%, respectively, of 1111 patients who underwent pulmonary angiography in the PIOPED study.[105] A review of more than 7000 patients undergoing pulmonary angiography revealed a 0.1% death rate with the procedure.[106]

Ventilation and Perfusion Lung Scintigraphy

V/Q lung scanning combines ventilation and perfusion nuclear medicine imaging techniques. The ventilation (V) study involves the inhalation of a radioactive gas (e.g., xenon), which provides an image of all ventilated portions of the lung. The perfusion (Q) study consists of an intravenous injection of [99mTc]-labeled macroaggregated human serum albumin particles with the patient in supine position, and the particles become trapped in approximately 0.1% of the pulmonary capillary bed.[103] Any obstruction to arterial flow is viewed as an area of hypoperfusion and is called a “perfusion defect” on gamma-camera images. The presence of multiple segmental perfusion defects increases the test specificity for PE.[103] On the basis of the presence and extent of matched (absence of both perfusion and ventilation) and unmatched (absence of perfusion but preserved ventilation) defects, the V/Q scan can be interpreted by using PIOPED published criteria as either normal or low probability, intermediate probability, or high probability for PE. [104] [107]

Because they provide indirect evidence of PE, V/Q scans are most clinically useful when considered in combination with an assessment of pretest clinical suspicion.[104] On the basis of PIOPED data, a normal V/Q scan essentially rules out clinically significant PE, and a high-probability V/Q scan alone has a high positive predictive value (88%) for PE, with 96% of patients with high pretest clinical suspicion and a high-probability V/Q scan having PE documented by pulmonary angiography.[104] Only 4% of patients with both low-probability V/Q scan and low clinical suspicion had PE on angiography.[104] Low- and intermediate-probability scans are now considered together as being “indeterminate” scans.

A plain chest radiograph is necessary before interpretation of a V/Q scan. Pleural effusions, bullous disease, pulmonary infiltrates or masses, and atelectasis have been associated with a higher frequency of indeterminate-probability scans and with a lower positive predictive value of high-probability findings.[103]

V/Q scan has the advantage of not using iodinated contrast. Its greatest limitation as a diagnostic tool for PE is that it provides a definitive result in a minority of patients. In the PIOPED series, only 13% had a normal study, and 14% had a high-probability scan.[104] Therefore, it can be expected that as many as 73% of all patients undergoing a V/Q scan will have a nondiagnostic, indeterminate scan that warrants further testing such as pulmonary angiography to confirm or exclude PE. In addition, it has been demonstrated that the rates of nondiagnostic V/Q scan findings increase significantly, from 21% to 63% in patients without chronic obstructive pulmonary disease to 46% to 91% in those with chronic obstructive pulmonary disease and from 9% in patients with normal chest radiographs to 48% in patients with abnormal chest radiographs. [108] [109]

Pitfalls in the interpretation of V/Q scintigraphy may represent a significant problem in the cancer population, particularly when primary or secondary (metastatic) pulmonary involvement is present. Patients with a history of PE may have a high-probability V/Q scan that does not reflect a new, acute event but rather the remnants of old PE. [103] [104] False high-probability scans may occur as a consequence of an abnormal perfusion scan due to pulmonary artery invasion or compression, regional hypoventilation secondary to bronchial compromise, pulmonary vein obstruction by hilar masses or adenopathy, and pulmonary leukostasis, which have been described in lymphoma, osteosarcoma, neuroblastoma, lymphangitic carcinomatosis, lung carcinoma, carcinoid, left atrial leiomyosarcoma, metastatic renal cell carcinoma, pulmonary artery sarcomas, and acute myelogenous leukemia.[2] Areas of V/Q mismatch also have been described as a result of prior radiation therapy to the chest in patients with breast and lung carcinoma. [2] [110]

Helical (Spiral) Computed Tomography Angiography

Contrast-enhanced spiral CT is performed by scanning a distance of 10 to 12 cm from the aortic arch to 2 cm below the inferior pulmonary veins during a single 30-second breath hold while the pulmonary vasculature is opacified by the automated injection of iodinated contrast medium. Technical parameters such as collimation, rate and timing of contrast administration, and scanning delay, as well as breathing, motion artifacts, and even central venous catheters, can influence the timing and quality of opacification of the pulmonary arteries, thus having the potential to compromise the quality of the study. [2] [103] The rate of studies that are inadequate for interpretation ranges from 2% to 13%. [2] [108]

Spiral CT has an overall sensitivity of 53% to 100% and specificity of 78% to 100% for the diagnosis of PE. [2] [30] These rates approach 95% to 100% for the central and segmental pulmonary arterial branches, but are lower (sensitivity: 53% to 63%) for PE involving the subsegmental branches. The accuracy of helical CT is highly influenced by, and dependent on, the equipment being used. Earlier CT scanners included 5-mm collimation that resulted in an effective section thickness of 6.57 mm and imaging reconstruction at 3-mm intervals. [2] [111] Many currently available CT scanners include 2- to 3-mm collimation, resulting in effective section thickness of 2 to 4 mm and improved visualization of the subsegmental arterial bed. The newest-generation multislice CT scanners allow subsecond scanning with 1.25-mm collimation, 1.25-mm section thickness, and image reconstruction at 0.6-mm intervals. Although these modern scanners also have been shown to improve visualization of subsegmental arteries significantly, two recent studies disagreed on whether this improved visualization led to higher detection rates of subsegmental PE. [112] [113] In addition, the use of workstations for image viewing and primary interpretation has been shown to increase the detection of PE by 25% in comparison to hard-copy image viewing.

Studies with the highest frequency of isolated subsegmental PE, and hence the lowest spiral CT sensitivity, have been reported in patients who also had nondiagnostic V/Q scans, and studies have shown that the prevalence of subsegmental PE ranges from 6% to 36%. [2] [114] Whether subsegmental PE is clinically relevant and should definitively be detected remains a controversial matter. A total of six prospective management studies have been performed with spiral CT in patients with indeterminate V/Q scan findings. [115] [116] [117] [118] [119] [120] In five of these studies, patients with a negative helical CT also underwent LE duplex ultrasound, anticoagulation being held after a negative result. Follow-up periods ranged from 3 to 6 months, and rates of recurrent VTE ranged from 2% to 4%. The only study that withheld anticoagulation on the basis of a negative helical CT alone reported a 1% rate of VTE recurrence in patients with low pretest probability for PE. Because these studies did not rely on a negative CT alone before withholding anticoagulation and because the sensitivity of helical CT is low in patients with nondiagnostic V/Q scans, the current evidence is insufficient to rely on a negative helical CT alone to justify withholding anticoagulation and even to support the hypothesis that subsegmental PE is not clinically significant.

The major advantages of spiral CT include the rapid nature of data acquisition and the lower percentage of nondiagnostic studies due to its ability to evaluate vascular as well as nonvascular intrathoracic structures simultaneously and provide an alternative diagnosis that either suggests or supports the final clinical diagnosis in 26% to 67% of examinations.[2] Although some studies have suggested that the presence of such alternative diagnoses adequately excludes PE, other studies have shown that the frequency of alternative diagnoses was the same in patients with and without confirmed PE.[121] Therefore, even the presence of an alternative explanation of the patient's symptoms does not necessarily imply that it is safe to withhold anticoagulation, and pursuit of pulmonary angiography might still be appropriate in such patients.

Magnetic Resonance Angiography

MRA is a promising technique in patients with suspected PE, particularly with the use of 3D, gadolinium-enhanced imaging. [2] [103] [122] The sensitivity and specificity of MRA for the detection of PE are 50% to 100% and 95%, respectively. [122] [123] The sensitivity is less than 50% for subsegmental PE, and overall accuracy is highly dependent on the technique and experience of the interpreting physicians.[123] In cases of suspected pulmonary artery sarcoma, MRI appears to be more useful than CT because the presence of gadolinium enhancement suggests tumor instead of thrombus.


The sensitivity and specificity of the transesophageal echocardiogram for the detection of central pulmonary artery PE range from 76% to 97% and from 77% to 100%, respectively, when compared with helical CT.[124] However, the sensitivity for peripheral PE is lower.[124] Echocardiography currently appears to be most useful in assessing patients who are hemodynamically unstable, are unable to undergo helical CT or V/Q scan, and need to have a workup initiated at the bedside.

Pulmonary Embolism Diagnosis in Cancer Patients

When a cancer patient is suspected of having PE, a chest radiograph should be performed to exclude other conditions that might require immediate intervention, such as a central line–related tension pneumothorax, pulmonary hemorrhage, and malignant pleural effusion. However, the chest radiograph should not be used as a means of supporting or refuting the need for specific diagnostic testing for PE. Helical CT is probably best indicated when the chest radiograph is abnormal and in individuals with known pulmonary disease. Alternatively, if the chest radiograph is normal, a V/Q-scan approach also is appropriate, but it is important that physicians clearly establish and document their pretest clinical suspicion of PE. Otherwise, PIOPED probability criteria may not be applicable. Regardless of the initial diagnostic approach, an indeterminate or negative initial test result should be followed by a duplex ultrasound of the legs and possibly of the arms as well, particularly in patients with symptoms of pain and swelling or who have an indwelling catheter. A positive duplex ultrasound does not confirm PE, but detection of acute DVT will prompt and justify systemic anticoagulation. In the cancer patient who is suspected of having PE, the combination of a nondiagnostic helical CT or V/Q scan with the absence of acute DVT by duplex ultrasound warrants the performance of pulmonary angiography. In this clinical setting, the risks of pulmonary angiography are sufficiently low and are outweighed by the benefits. Pulmonary angiography also is indicated in the setting of a normal spiral CT alone, particularly with a high pretest suspicion.

Treatment of Pulmonary Embolism in Cancer Patients

Most cancer patients with symptomatic PE should be treated with anticoagulation by following the same guidelines for treatment that are applied to DVT. [1] [51] [85] [125] Initial inpatient intravenous UFH is recommended in symptomatic patients with extensive PE. Some patients, including cancer patients, with PE may derive benefit from thrombolytic therapy to degrade actively the thrombus obstructing the pulmonary vasculature. Clear indications for PE thrombolysis are debated.[126] Thrombolysis has been demonstrated to improve survival in patients with massive PE plus shock and is probably indicated in these patients regardless of cancer status. When compared with anticoagulation alone, thrombolytic therapy results in more rapid thrombus lysis, an early improvement in pulmonary blood flow, and improvement of right ventricular function.[127] However, these improvements in cardiopulmonary function alone have not resulted in decreased mortality in stable patients without significant hemodynamic compromise. [128] [129] Thrombolysis is contraindicated in any cancer patient with significant transfusion-refractory thrombocytopenia, active bleeding, and CNS lesions.


Diagnosis of Inferior Vena Cava and Intra-abdominal Deep Vein Thrombosis in Cancer Patients

Many cancer patients undergo serial imaging with CT as a means of assessing cancer-therapy efficacy, disease stage or progression, and nonspecific abdominal symptoms. Incidental findings of what appears to be a DVT should not affect patient treatment and prompt the placement of an IVC filter. Prompt and proper diagnostic imaging is especially needed in such a situation.

Contrast Venography

Popliteal or common femoral vein access is the most appropriate approach for performance of contrast cavography. Access through an internal jugular vein might be necessary in some circumstances. Contrast venography is the reference standard to show the presence or absence of IVC or renal vein thrombosis conclusively, and the same criteria that are used to diagnose LE- and UE-DVT by venography apply to the IVC. However, in cancer patients not all intraluminal filling defects represent thrombus, and the distinction between intravascular tumor and thrombus might require further investigation with CT or MRI and, in some selected cases, with transvenous catheter-guided biopsy. Likewise, if the column of contrast does not opacify the IVC, CT or MRI is indicated to assess for the presence of extrinsic compression or invasion by tumor. The diagnosis of portal, mesenteric, or ovarian vein thromboses by venography is more challenging because of limited ability to perform selective contrast injections.

Duplex Ultrasound

No studies have validated the use of duplex ultrasound for the detection of IVC thrombosis (i.e., duplex ultrasound has never been compared with the gold standard contrast venography, nor has it been subject to accuracy or management studies in this setting). Although visualization of the IVC and interrogation of its lumen for Doppler-flow measurements have been described in studies that imaged the IVC before IVC filter placement, [2] [130] compression maneuvers are not technically feasible in the abdomen, and the indirect signs of impaired flow and loss of flow phasicity are not specific for IVC thrombosis. Moreover, duplex ultrasound cannot differentiate IVC thrombus from tumor, except in cases of suspected portal vein invasion by tumor. In this setting, color Doppler ultrasound can be a reliable diagnostic tool.[131]

Contrast-Enhanced Computed Tomography

CT remains a nonvalidated method to assess for the presence of thrombus in the IVC. Indirect signs that have been described in cases of IVC thrombosis include IVC enlargement, reduced IVC lumen density compared with that of the aorta, and rim enhancement. However, these signs may occur as a result of contrast flow phenomena mimicking an intraluminal filling defect and also have been described in patients with renal cell or adrenal cortical carcinoma extending into the renal veins and IVC, producing a “tumor thrombus.” [132] [133] [134] [135] [136] Although spiral CT venography has been shown to visualize the intra-abdominal veins accurately, no formal studies have been published. Any incidental finding of an IVC or renal vein “filling defect” by CT should ideally be confirmed by venography and not prompt the initial placement of an IVC filter.

The previously mentioned features also are used in the diagnosis of portal and mesenteric vein thrombosis by CT.[137] These signs are of little value in a patient with a history of portal vein thrombosis who is suspected to have a recurrent event. Although the true sensitivity of contrast-enhanced CT in diagnosing portal vein thrombosis is unknown, its specificity has been suggested to be quite high. The presence of cavernous transformation—a “masslike” network of collateral veins—is suggestive of remote portal vein thrombosis.[137]

Only one retrospective case series pertaining to the diagnostic imaging of ovarian vein thrombosis in cancer patients has been published.[138] In this small study, none of the six patients with ovarian DVT had the related CT findings of uterine enlargement and other pelvic masses that are typically described in larger series of patients with puerperal ovarian DVT. Similar to portal DVT, these indirect CT signs are not useful in a patient with a history of ovarian vein thrombosis, because at least half of the patients will not have normalization of the original CT findings after 3 months to 2 years after the index event.[139]

Magnetic Resonance Venography

MRV is currently considered the diagnostic method of choice for diagnosing IVC, renal, and portal vein thrombosis, particularly when spin-echo and cine MRI techniques are used in combination. In a small series of 26 patients with puerperal ovarian DVT, MRV had 100% sensitivity and specificity.[139] Spin-echo MRI may help to differentiate acute from chronic thrombus on the basis of differences in patterns of signal intensity.[140]

Artifacts created by flow phenomena may cause signal voids at the junction of the renal vein and the IVC because this is an area of slow and convergent blood flow. This can lead to a false-positive diagnosis of IVC thrombosis.[141] MRI also is limited in distinguishing true portal DVT from tumor invasion of the portal vein unless an adjacent mass is seen.

The rare IVC leiomyosarcoma appears to be equally demonstrated by CT or MRI, although MRI is superior because it seems to be capable of differentiating tumor (homogeneous, intermediate signal intensity on T1-weighted images, and high signal intensity on T2-weighted images) from thrombus (hyperintense on T1- and T2-weighted images). [142] [143] [144]

Treatment of Inferior Vena Cava and Intra-abdominal Deep Venous Thrombosis in Cancer Patients

No specific guidelines exist for the treatment of IVC and other intra-abdominal DVT in the cancer patient or noncancer patient. Whether to use standard anticoagulant therapy as is used for proximal LE-DVT or catheter-directed thrombolytic therapy depends on the extent of thrombosis, patient symptoms, and patient bleeding risk. Patients with acute, complete IVC occlusion may develop significant bilateral LE swelling and pain and are at risk for phlegmasia cerulean dolens (venous limb gangrene). Such patients could benefit most from pharmacologic or mechanical thrombolysis. Long-term anticoagulation is likely warranted.


Diagnosis of Superficial Thrombophlebitis in Cancer Patients

Superficial venous thrombophlebitis (SVT) is the only manifestation of venous thromboembolic disease that does not require objective diagnostic imaging. The diagnosis can be made clinically by the detection of a palpable tender cord in the course of a superficial vein; the induration of the vein is usually associated with erythema of the overlying skin. The clinical differential diagnosis includes sarcoidal granulomas, Kaposi's sarcoma, and lymphangitis.

Duplex ultrasound should be considered to rule out concomitant DVT, particularly when the greater or lesser saphenous veins are involved. The most common location for progression from SVT to DVT to occur is at the junction between the greater saphenous vein (a superficial leg vein) and the common femoral vein (a proximal deep leg vein). Proximity of an SVT to the junction between the involved superficial vein and its connection to the deep venous system does not seem to affect the likelihood of PE. The rates of DVT in patients with clinical signs and symptoms of isolated SVT have been reported to range from 6% to 57%, and the rates of symptomatic PE have been reported to range from 4% to 10%. [145] [146] [147] [148] [149] [150]

Trousseau's syndrome or migratory thrombophlebitis was initially described in patients with mucin-secreting carcinoma of the gastrointestinal tract. It is unclear whether the “thrombophlebitis” in many of the reported cases was manifested as SVT or DVT. The SVT related to Trousseau's syndrome typically was first seen as multiple tender nodules, which progressed to form palpable cords, usually involving the bilateral LEs and, on occasion, the UE veins and the abdominal wall veins as well.

Treatment of Superficial Thrombophlebitis in Cancer Patients

Typically, nonsteroidal anti-inflammatory drugs and warm compresses are adequate treatment for SVT symptom control. When the deep system is involved or symptomatic pulmonary embolism is diagnosed, standard anticoagulant therapy is indicated. Because of the reported high rate of progression from SVT to DVT, many physicians treat SVT with anticoagulants for a variable time. Anticoagulants such as UFH and LMWH may help to relieve symptoms related to vessel inflammation but are probably best reserved for individuals with recurrent SVT or documented DVT. Serial ultrasound to detect meaningful SVT progression seems prudent in selected cases such as those with SVT already at the saphenofemoral junction.


IVC filters are often placed in cancer patients with acute VTE, especially in the settings of thrombocytopenia, active bleeding, and CNS malignancy. [151] [152] [153] [154] Limitations to IVC filters include technical difficulties during insertion, insertion-site hemorrhage or thrombosis, caval thrombosis, and obstruction below the filter, filter change of position (migration or tilting), caval erosion and perforation, and filter failure.[154] IVC filters obviously play no role in the management of UE-DVT. SVC filter placement has been shown to be technically feasible, but outcome data are lacking.[155] A recent study published by Decousus and associates[155] addressed the impact of IVC filter placement on PE prevention and DVT recurrence rate. All patients had proximal DVT, and all received anticoagulation. Placement of an IVC filter conferred a significant benefit in preventing PE within the first 12 days after DVT, with PE developing in 4.8% of patients without filters compared with 1.1% of patients with filters (P = 0.03). At 2 years’ follow-up, however, the benefit was no longer statistically significant with regard to symptomatic PE prevention. Moreover, at 2 years, those patients with prophylactic filter placement had a higher risk of DVT recurrence than did those who did not have a filter placed (20.8% versus 11.6%; P = 0.02).[155] These findings underscore the need for caution in placing filters, especially when patients are receiving cancer treatment with a curative intent, and long-term survival is contemplated. When a filter is placed because of a transient contraindication to anticoagulation, appropriate pharmacologic therapy should be commenced once the contraindication has passed.


The relation between cancer and clinical thrombosis has been recognized for more than 150 years. The concept of and an appreciation of thromboprophylaxis have been widely accepted, primarily in the surgical setting, for more than 25 years. However, only in the last decade has the widespread application of thromboprophylaxis to patients with cancer begun to receive significant attention from the medical community. One of the observations driving this level of interest is the expanding body of data suggesting that the connection between thrombosis and cancer may well be bidirectional.

In the not-too-distant past, the prevailing opinion was that cancer caused thrombosis and that, although it was a regrettable complication, thrombosis did not significantly affect the overall clinical course of the patient with cancer. The interactions between cancer and thrombosis might not be as simple as that. Take the time-honored observation that in an inordinate number of patients with idiopathic VTE, cancer develops in the subsequent several months[156] and extend the follow-up for several years, and one finds that the rate of malignancy continues to increase inordinately in this population for at least 6 years, with no evidence of a plateau developing in the curve.[157] Although it is conceivable that some aspect of an occult malignancy (even one that does not become clinically apparent for 6 years) promotes thrombosis, it is equally conceivable that some pathophysiologic element related to the thrombosis itself promotes the development or progression of the cancer. If this were true, then it would logically follow that alterations in the biology of thrombosis might alter the likelihood of subsequent malignancy.

Data from randomized clinical trials support this hypothesis. Patients who received oral warfarin secondary prophylaxis for 6 months after an idiopathic VTE had a lower rate of subsequent malignancy over the ensuing 6 years than did patients who were randomized to receive only 6 weeks of anticoagulation.[157] Not only does it appear that thrombosis might, through mechanisms unknown, predispose to the development of cancer, but the presence of thrombosis[158] or even evidence of activation of coagulation without overt thrombosis[23] also is associated with more aggressive behavior of the associated malignancies. This body of data, taken to its logical conclusion, allows the development of a hypothesis that thromboprophylaxis could be used for both primary prevention of and treatment of some forms of cancer. Add to this the facts that standard anticoagulation treatment of acute VTE is less efficacious and more toxic in cancer patients than in noncancer patients (see the previous sections) and that the risk of death from PE might be higher in cancer patients than in noncancer patients,[1] and it becomes clear that prevention of thrombosis is clinically important in patients with cancer.

To date, the available data on thromboprophylaxis specifically in cancer patients is derived mainly from studies whose major endpoints included the incidence of any (primarily asymptomatic) DVT rather than more meaningful endpoints such as disease progression, fatal PE, and survival. However, with meta-analysis, heparin thromboprophylaxis that was successful in preventing asymptomatic DVT in patients without cancer has been shown to prevent both symptomatic DVT and fatal PE.[159] Therefore, it is reasonable to infer that any intervention that prevents asymptomatic DVT also prevents clinically significant thromboembolic disease in patients with cancer. Data discussed earlier also raise the possibility that thromboprophylaxis in the cancer patient could modify the course of the malignancy in a favorable direction. Consequently, comments in the remainder of this section are based on the assumption that thromboprophylaxis (operationally defined as any intervention that prevents asymptomatic DVT) is clinically useful and should be used in cancer patients in several common, clinically defined situations.

Thromboprophylaxis in Surgical Oncology

Most studies of surgical thromboprophylaxis, such as those focusing on elective joint replacement, have not differentiated between patients with and without underlying malignancy. Consequently, the quantitative risk added by the presence of cancer to each and every type of surgical procedure is unknown. Despite this, the American College of Chest Physicians Consensus Conference on Antithrombotic Therapy categorizes any patient older than 40 years undergoing any major surgery in the setting of active or prior cancer as being in the “highest risk” group of patients, with an estimated risk of proximal DVT of 10% to 20% and fatal PE of 0.2% to 5.0% without thromboprophylaxis.[160]

Two basic methods of thromboprophylaxis have been widely studied in surgical patients: mechanical and pharmacologic. For practical purposes in surgical oncology, mechanical forms of prophylaxis are limited to external pneumatic compression (EPC) boots, whereas pharmacologic methods are limited to heparin-related compounds.

EPC has been found to provide a modest degree of efficacy in surgical oncology, ultrasound-defined failure rates varying from 1% in gynecologic oncology patients[161] to 14% in cancer patients undergoing orthopedic procedures.[162] Assuming that they are applied correctly 100% of the time, EPC boots are a cost-effective form of prophylaxis during and after high-risk gynecologic oncology procedures.[163] Unfortunately, the assumption of proper application of EPC devices 100% of the time is generally incorrect in practice, rates of 33% being more common.[164] In evaluation of the efficacy of EPC in gynecologic surgery, the presence of cancer has been found to be an independent risk factor for failure, with a relative risk of DVT of 4.9 compared with noncancer patients.[165] EPC as the sole method of prophylaxis is not recommended for the “highest risk” patients[160] and should not be relied on for most surgical procedures performed on cancer patients.

Heparin and heparin-related compounds form the mainstay of pharmacologic methods of thromboprophylaxis, although other medications, such as oral direct thrombin inhibitors and factor Xa inhibitors, could begin to play a major role in the future. The efficacy of UFH and its derivatives has been studied primarily in noncancer patients undergoing surgery. The first major investigation of the efficacy of heparin in curative cancer surgery was a prospective, randomized, double-blind study comparing subcutaneous UFH, 5000 units three times daily, with subcutaneous enoxaparin, 40 mg once daily, begun 2 hours before surgery.[166] Contrast venography within 24 hours of the last drug injection (10 ± 2 days) was scheduled to be performed in all patients. Of 1116 randomized patients, 319 UFH-treated patients and 312 enoxaparin-treated patients were evaluable. Total VTE, symptomatic DVT, and any DVT were detected in 18.2%, 1.9%, and 17.6%, respectively, of the UFH-treated patients and in 14.7%, 1.3%, and 14.4%, respectively, of the enoxaparin-treated group. These rates of thrombosis are more than twice those generally seen in noncancer patients undergoing general surgery receiving the same medications and evaluated with the same endpoints.[160] As the data demonstrate, most detected thromboses were asymptomatic (any DVT minus symptomatic DVT). To deal with the high failure rate of pharmacologic thromboprophylaxis in this group of patients, two different, and not mutually exclusive, approaches could be taken: (1) find a way to improve in-hospital efficacy and/or (2) find a way to prevent these asymptomatic thrombi from propagating and posing a risk of death after hospital discharge. The first approach might be accomplished by combining mechanical and pharmacologic methods of prophylaxis. This approach has been shown to be valid in noncancer patients by using EPC and graduated-compression stockings combined with UFH or LMWH [167] [168] and has shown promise in patients undergoing craniotomy for brain tumors.[169] The second approach might be accomplished by extending the duration of thromboprophylaxis sufficiently that the asymptomatic thromboses fail to propagate, fail to embolize, and successfully undergo spontaneous thrombolysis. This approach has been recently found to be valid as well.

A similar group of patients undergoing planned curative open surgery for abdominal or pelvic cancer were all given enoxaparin, 40 mg once daily for 6 to 10 days, and subsequently randomized to 21 additional days of enoxaparin at the same dose or 21 days of placebo.[170] Bilateral venography was performed between days 25 and 31 after surgery, and patients were clinically followed up for a total of 3 months after surgery. A total of 501 patients were randomized with 332 patients included in the efficacy analysis. The group that was randomized to enoxaparin had a 4.8% rate of any DVT at 4 weeks compared with 12.0% in the placebo group (P = 0.02). The majority of detected DVTs were asymptomatic and involved calf veins. Approximately 1.5% of patients in both groups had clinically apparent VTE in the 2 months of follow-up with no prophylaxis (an annualized rate of 9%, roughly 90 times that of the normal population).[170] A recent clinical outcome-based study of venous thrombosis following cancer surgery reported that 40% of confirmed thrombotic events occurred more than 21 days after the operation.[171] In patients undergoing high-risk abdominal surgery, including cancer surgery, postoperative subcutaneous fondaparinux was at least as effective as perioperative subcutaneous dalteparin in preventing venous thromboembolism.[171] In the 1408 evaluated subjects in the PEGASUS study who underwent cancer surgery, 4.7% of those who were randomized to fondaparinux were diagnosed with postoperative venous thrombosis compared to 7.7% of those who were randomized to dalteparin.[172] Taken in total, these data suggest that the following scheme should be strongly considered for all cancer patients undergoing major surgery:



All should be given combined EPC and prophylactic-intensity LMWH or fondaparinux during hospitalization (unless contraindicated by renal insufficiency or a history of heparin-induced thrombocytopenia)



All should be considered for prophylactic dose anticoagulation for 3 weeks after hospitalization (with the same caveats)



Extended thromboprophylaxis with prophylactic anticoagulation should be considered for those with continuing risk factors for VTE, such as chemotherapy, infection, paralysis, and use of central venous catheters

Prevention of Central Venous Access Device–Associated Deep Venous Thrombosis

The quantitative scope of the problem of central venous catheter–related UE, internal jugular, and thoracic vein thrombosis (a clinical entity distinct from catheter dysfunction due to a fibrin sheath or small, nonocclusive tip thrombosis) depends on the method of detection. The incidence ranges from 2.4% to 35% if only symptomatic events are considered and from 36% to 66% if surveillance venography is used to detect all thromboses. [160] [173] [174] [175] [176] Because of the reported incidences of catheter-associated thrombosis, cancer patients with central venous catheters should always be considered as being at high risk for DVT.[160] These thrombi are not benign, having at least a 25% incidence of asymptomatic PE.[177] Given that PE is a significant cause of death among cancer patients, central venous catheter–related DVT must be considered a potentially lethal problem.

In addition to their shared relation with PE, central venous catheter–related DVTs also pose clinical problems that are different from those seen with LE-DVT. As with leg vein DVT, not all central venous catheter–related thromboses undergo complete physiologic thrombolysis. An unknown fraction undergo organization and remain as a permanent obstruction to the involved veins. These organized thrombi prevent insertion of subsequent catheters at that vascular site, such as at the time of tumor relapse, from 14% to 30% of the time.[175] Central venous catheter–related thrombi also can be a nidus for infection. The risk of sepsis in patients with central venous catheter–related DVT is 2.62 times that of patients with catheters but without thrombosis.[178]

By consensus, the gold standard for determining the efficacy of any method of thromboprophylaxis is the randomized clinical trial using contrast venography-documented thrombosis as an endpoint.[160]Studies using purely clinical endpoints are generally not used in this context. Relatively small studies that examined fixed, low-dose warfarin and fixed-dose subcutaneous LMWH have demonstrated successful catheter-related DVT prevention. [173] [177] Reported rates of thrombosis in the active therapy groups were in the 6% to 10% range. Neither study showed toxicity with these methods of prophylaxis, although while receiving antibiotic therapy, patients taking fixed, low-dose warfarin often had sufficiently long prothrombin times that a bleeding risk was probable.[174] During times of acute illness, especially when antibiotics are used, prothrombin time monitoring is recommended in patients receiving low-dose warfarin for thromboprophylaxis. Despite the results of small, provocative studies, larger-scale and more rigorously executed studies of low-dose oral warfarin and LMWHs have failed to demonstrate greater thromboprophylactic activity than placebo. [179] [180] [181] In fact, these recently reported studies demonstrated much lower thrombosis rates than were previously reported. This might reflect improvements in catheter material and placement. At this time, routine prescription of thromboprophylaxis in cancer patients with indwelling central venous catheters is not recommended.[160]

Prevention of Central Venous Access Device–Associated Thrombotic Occlusion

Very little work has been directed at finding methods of prevention of central venous catheter thrombotic occlusion. Thrombotic occlusion can manifest as the inability to infuse, inability to withdraw, or a combination of the two, known as total occlusion. The methods that are used to prevent central venous catheter–associated DVT have not been evaluated for efficacy in prevention of catheter obstruction. Although it is widely held that heparin flushing and meticulous catheter care prevent this problem, no experimental data support this contention. Heparin flushing of catheters does, however, increase the risk of HIT in this already hypercoagulable population. Fortunately, a readily available and highly effective therapy is available for patients who are experiencing thrombotic catheter obstruction. Thrombolytic therapy with low-dose recombinant tissue–type plasminogen activator instilled into the obstructed catheter has roughly a 90% likelihood of restoring catheter function at 4 hours. [182] [183]Consequently, in the face of limited data on efficacy, infrequent but potentially severe toxicity, and highly effective salvage therapy, routine flushing of catheters with UFH to prevent occlusion cannot be recommended. Flushing of catheters with normal saline after each use is suggested as a preferable alternative to heparin flushing.

The only parameter that has been found to affect rates of thrombotic catheter occlusion is the anatomic location of the catheter tip. Catheters whose tips have been placed in the innominate vein have an inordinate incidence of thrombosis, especially compared with those whose tips are in the lower portion of the SVC or right atrium. Use of intraoperative guide-wire measurement or alternative techniques to ensure that the catheter tip is placed in the lowest third of the SVC is recommended.

Thromboprophylaxis during Chemotherapy

Chemotherapy is an independent risk factor for both VTE [1] [184] [185] and death within 1 week of VTE,[159] above and beyond that conferred by the presence of cancer alone. Unfortunately, the absolute risk of VTE that is conferred has been studied in only a relative few of the many chemotherapy regimens used to treat the wide spectrum of known malignant diseases. The absolute risk of VTE varies from 1.3% per year with tamoxifen only to 7.9% per year with tamoxifen combined with CMF (cyclophosphamide, methotrexate, and 5-fluorouracil) as adjuvant therapy for breast cancer [1] [27] to as much as 43% during shorter-duration therapy of a variety of malignancies with thalidomide-containing regimens.[186] The 0.1% per year risk in the general population pales in comparison to these rates of VTE.

Despite the growing recognition of chemotherapy as a risk for VTE, little has been done to evaluate the role of primary thromboprophylaxis. The only trial to evaluate prophylaxis formally was done in the setting of systemic therapy for randomized breast cancer patients on active therapy with either warfarin, 1 mg daily for the first 6 months, followed by adjusted-dose warfarin to achieve an INR between 1.5 and 1.9 or placebo. The study compared the rates of clinically detected VTE over a mean of 6 months.[187] Active therapy reduced the incidence of VTE from 4.4% to 0.7%, with no differences in the rates of major bleeding complications. This intervention did not increase the costs of medical care for these patients.[188] Secondary prophylaxis (prevention of recurrence after the initial episode of VTE) by using warfarin in therapeutic doses has been effective in a limited number of patients taking thalidomide, suggesting that standard methods of thromboprophylaxis may be effective in a spectrum of chemotherapy regimens. In short, this phenomenon of chemotherapy-related VTE has not been given the attention it deserves, given the potential scope of the problem. Despite this, because of the low toxicities of current thromboprophylactic measures, it seems reasonable to consider either warfarin in doses to prolong the INR to 1.5 to 1.9 or prophylactic doses of LMWH for patients receiving chemotherapy regimens associated with a meaningful increased risk of VTE.

Thromboprophylaxis for the Hospitalized Cancer Patient

Being sufficiently ill to require hospitalization for any reason is associated with a 100-fold increase in the risk of VTE[189] and an 18-fold increase in the risk of death within 1 week of VTE[159] compared with community residents. Patients who have been hospitalized for medical illness account for almost 23% of all cases of VTE.[185] Hospitalization for medical illness complicating cancer further increases the risk of VTE. [185] [190] Patients with cancer who have been hospitalized for nonsurgical illness account for almost 30% of all cases of VTE,[184] making this population a prime target for the use of thromboprophylaxis.

Thromboprophylaxis has not been studied specifically in medically ill cancer patients. Inferences must be made from studies of a wide spectrum of patients with medical illnesses, some of whom have cancer. In this population, meta-analysis has shown that thromboprophylaxis with UFH or LMWH is associated with a more than 50% reduction in symptomatic DVT and PE without an increase in bleeding complications.[191] The first prospective, randomized, controlled trial of thromboprophylaxis in this group of patients using the gold standard of contrast venography endpoint randomized 866 patients (14% of whom had cancer) to receive either placebo or one of two doses of enoxaparin (20 mg daily or 40 mg daily) during their hospitalization.[192] At the time of discharge, DVT had developed in approximately 15% in the placebo group and in the group randomized to receive 20 mg of enoxaparin daily but only in 5.5% in the group randomized to 40 mg of enoxaparin daily, a risk reduction of more than 60% that was highly statistically significant. This population of medically ill patients was at high risk of bleeding complications. As a group, 1.3% experienced major hemorrhage during their hospitalization (a median of 7 days). The use of enoxaparin was not associated with an increase in this basal rate of major hemorrhage. This therapy was associated with a small increase in the cost of hospitalization but could actually be cost effective if one considers the savings from not having to treat as many new VTEs.[193] These data strongly suggest that all hospitalized cancer patients receive thromboprophylaxis with an LMWH unless contraindicated by either severe renal insufficiency or a history of HIT.

Chemotherapy-related thrombocytopenia, intracranial malignancy, and gastrointestinal lesions have been perceived by some physicians as contraindications to the use of prophylactic doses of LMWH. Whereas no data formally evaluate the safety of prophylactic LMWH in these high-risk situations, inferences can be made from several sources. Patients undergoing elective neurosurgery, including surgery for intracranial malignancy, have an incidence of bleeding similar to that of hospitalized medical patients (approximately 2%–3%). [169] [194] This rate of bleeding was not increased with the use of 40 mg of enoxaparin daily.[194] Patients undergoing curative abdominal and pelvic surgery for cancer do not have an increased rate of bleeding complications with the use of this dose of enoxaparin.[166] From these data, it appears that the use of prophylactic doses of LMWH is unlikely to pose an additive risk of bleeding over that inherent with the underlying disease. Consequently, unless the extent of the intracranial or gastrointestinal disease or the degree and duration of thrombocytopenia are great, these comorbidities pose only a relative, and probably small, contraindication to the use of pharmacologic thromboprophylaxis. In situations in which this type of prophylaxis is perceived to be contraindicated, use of EPC and/or periodic ultrasound surveillance[195] is strongly recommended.

In the previously mentioned thromboprophylaxis trial,[192] all patients were monitored for the development of symptomatic VTE for 3 months after discharge. In approximately 1% of the patients, symptomatic DVT or PE developed during this time, which is roughly 40 times the incidence of VTE in the general population. This incidence is similar to that seen in the 3-month follow-up of patients undergoing curative surgery for cancer,[170] in which prolonged use of enoxaparin after discharge has been shown to be safe and effective in prevention of DVT. These data suggest that the risk of VTE does not end as the patient passes out through the doors of the hospital but continues for a protracted period. Extended thromboprophylaxis, as used in cancer surgery, should be strongly considered, especially in patients whose in-hospital cancer therapy has not been completely curative.


As more and more oncologists become aware of the importance and challenges of clinical thrombosis in patients with malignancy, we will likely see more attention paid to earlier venous thrombosis diagnosis, optimization of acute and chronic venous thrombosis management, and greater compliance with thromboprophylaxis recommendations. We hope that venous thrombosis will be viewed less as simply a nuisance during the care of cancer patients and more as a major source of patient morbidity, treatment delay, and mortality.

Continued improvements in duplex ultrasound, CT, and MRI are expected. The ability to scan the pulmonary vasculature, abdominal vasculature, pelvic veins, and lower extremities with one contrast injection and one imaging session will likely be perfected. This will provide a more comprehensive approach to thrombosis confirmation and will limit intravenous contrast exposure. Nuclear medicine scans that are capable of whole-body thrombus imaging are on the horizon and might even assist with the detection of occult malignancy in patients with an initial idiopathic VTE. Integration of diagnostic algorithms into one's oncology practice could assist the busy clinician with decision making in the case of suspected DVT and PE.

Continued interest in developing improved strategies to treat VTE specifically in the cancer patient is anticipated. Further reduction in both thrombosis recurrence rates and bleeding rates is needed. Prospective clinical trials are needed to support the long-term anticoagulation of the cancer patient with thrombosis. In particular, optimal management of women with thrombosis and the need for years of adjuvant hormonal therapy for breast cancer must be clarified. Despite excellent compliance with daily subcutaneous LMWH for up to 6 months of VTE therapy, newer oral anticoagulants could make long-term anticoagulation more palatable to all patient populations. Newer oral compounds might prove to be effective for primary and secondary venous thromboprophylaxis specifically in the cancer patient. Careful assessment for hepatic toxicity will be required before use in cancer patients who are prone to liver metastases and those receiving hepatically metabolized chemotherapeutic agents. Further research into the anticancer and survival prolongation properties of selected anticoagulants might identify a particular agent or class of agents as being ideal in cancer patients in general or in those with specific responsive tumor histologies.

Prevention will continue to be of paramount importance. It is always easier and less potentially toxic to prevent thromboses with low doses of anticoagulation than to treat life-threatening thromboses with longer courses of higher doses of the same drugs. Ongoing research attempting to link thrombosis with accelerated tumor neovascularization, growth, and metastases may highlight the importance and value of aggressive primary thrombosis prevention in all cancer patients. Time and significant research efforts will tell.


  1. Deitcher SR: Cancer-related deep venous thrombosis: clinical importance, treatment challenges, and management strategies.  Semin Thromb Haemost2003; 29:247-258.
  2. Gomes MPV, Deitcher SR: Diagnosis of venous thromboembolic disease in cancer patients.  Oncology (Huntingt)2003; 17:126-135.139
  3. Deitcher SR, Kessler CM, Merli G, et al: Secondary prevention of venous thromboembolic events in patients with active cancer: enoxaparin alone versus initial enoxaparin followed by warfarin for a 180-day period.  Clin Appl Thromb Hemost2006; 12:389-396.
  4. Prandoni P, Piccioli A, Girolami A: Cancer and venous thromboembolism: an overview.  Haematologica1999; 84:437-445.
  5. Brill-Edwards P, Ginsberg JS, Johnston M, Hirsh J: Establishing a therapeutic range for heparin therapy.  Ann Intern Med1993; 119:104-109.
  6. Hansson PO, Welin L, Tibblin G, et al: Deep vein thrombosis and pulmonary embolism in the general population: “The Study of Men Born in 1913.”.  Arch Intern Med1997; 157:1665-1670.
  7. Sorensen HT, Mellemkjaer L, Steffensen FH, et al: The risk of a diagnosis of cancer after primary deep venous thrombosis or pulmonary embolism.  N Engl J Med1998; 338:1169-1173.
  8. Levitan N, Dowlati A, Remick SC, et al: Rates of initial and recurrent thromboembolic disease among patients with malignancy versus those without malignancy: risk analysis using Medicare claims data.  Medicine1999; 78:285-291.
  9. Piccioli A, Prandoni P, Ewenstein BM, Goldhaber SZ: Cancer and venous thromboembolism.  Am Heart J1996; 132:850-855.
  10. Luzzatto G, Schafer AI: The prothrombotic state in cancer.  Semin Oncol1990; 17:147-159.
  11. Deitcher SR, Goldman CK, Ruiter K: Vascular endothelial growth factor (VEGF) levels in patients with idiopathic acute proximal deep venous thrombosis. Thromb Haemost 86(suppl):1517.
  12. Breddin HK, Hach-Wunderle V, Nakov R, Kakkar VV: Effects of a low-molecular-weight heparin on thrombus regression and recurrent thromboembolism in patients with deep-vein thrombosis.  N Engl J Med2001; 344:626-631.
  13. Krauth D, Holden A, Knapic N, et al: Safety and efficacy of long-term oral anticoagulation in cancer patients.  Cancer1987; 59:983-985.
  14. Heit JA, Mohr DN, Silverstein MD, et al: Predictors of recurrence after deep vein thrombosis and pulmonary embolism: a population-based cohort study.  Arch Intern Med2000; 160:761-768.
  15. Hansson PO, Sorbo J, Eriksson H: Recurrent venous thromboembolism after deep vein thrombosis incidence and risk factors.  Arch Intern Med2000; 160:769-774.
  16. Prandoni P, Lensing AW, Cogo A, et al: The long term clinical course of acute deep venous thrombosis.  Ann Intern Med1996; 125:1-7.
  17. Prandoni P: Antithrombotic strategies in patients with cancer.  Thromb Haemost1997; 78:141-144.
  18. Prandoni P, Lensing AW, Piccioli A, et al: Recurrent venous thromboembolism and bleeding complications during anticoagulant treatment in patients with cancer and venous thrombosis.  Blood2002; 100:3484-3488.
  19. Hutten BA, Prins MH, Gent M, et al: Incidence of recurrent thromboembolic and bleeding complications among patients with venous thromboembolism in relation to both malignancy and achieve international normalized ratio: a retrospective analysis.  J Clin Oncol2000; 8:3078-3083.
  20. Deitcher SR, Carman TL, Sheikj M, Gomes M: Hypercoagulable syndromes: evaluation and management strategies in acute limb ischemia.  Semin Vasc Surg2001; 14:74-85.
  21. Deitcher SR, Lucore C, Eisenberg PR: Impaired resolution of a massive pulmonary embolus associated with an impaired fibrinolytic response.  Am J Med1994; 96:483-484.
  22. Deitcher SR, Erban JK, Limentani SA: Acquired free protein S deficiency asociated with multiple myeloma.  Am J Hematol1996; 51:319-323.
  23. Beer JH, Haeberli A, Vogt A, et al: Coagulation markers predict survival in cancer patients.  Thromb Haemost2002; 88:745-749.
  24. Bona RD: Thrombotic complication of central catheters in cancer patients.  Semin Thromb Hemost1999; 25:147-157.
  25. Prandoni P, Polistena P, Bernardi E, et al: Upper extremity deep vein thrombosis: risk factors, diagnosis, and complications.  Arch Intern Med1997; 157:57-62.
  26. Raad II, Luna M, Khalil SA, et al: The relationship between the thrombotic and infectious complications of central venous catheters.  JAMA1994; 271:1014-1016.
  27. Deitcher SR, Gomes MP: The risk of venous thromboembolic disease associated with adjuvant hormone therapy for breast carcinoma: a systematic review.  Cancer2004; 101:439-449.
  28. Hull RD, Raskob GE, Brant RF, et al: The importance of initial heparin treatment on long-term clinical outcomes of antithrombotic therapy: the emerging theme of delayed recurrence.  Arch Intern Med1997; 157:2317-2321.
  29. Wells PS, Anderson DR, Ginsberg J: Assessment of deep vein thrombosis or pulmonary embolism by the combined use of clinical model and noninvasive diagnostic tests.  Semin Thromb Hemost2000; 26:643-656.
  30. Deitcher SR, Carman TL: Deep venous thrombosis and pulmonary embolism.  Curr Treat Options Cardiovasc Med2002; 4:223-238.
  31. Levine MN, Hirsh J, Gent M, et al: A randomized trial comparing activated thromboplastin time with heparin assay in patients with acute venous thromboembolism requiring large daily doses of heparin.  Arch Intern Med1994; 154:49-56.
  32. Chan A, Woodruff RK: Complications and failure of anticoagulation therapy in treatment of venous thromboembolism in patients with disseminated malignancy.  Aust N Z J Med1992; 22:119-122.
  33. Deitcher SR: Interpretation of the international normalized ratio in patients with liver disease.  Lancet2002; 359:47-48.
  34. Wells PS, Holbrook AM, Crowther NR, et al: The interaction of warfarin with drugs and food: a critical review of the literature.  Ann Intern Med1994; 121:676-683.
  35. Deitcher SR, Carman TL: Heparin induced thrombocytopenia: natural history, diagnosis, and management.  Vasc Med2001; 6:113-119.
  36. Lee AY, Levine MN, Baker RI, et al: Randomized comparison of low-molecular-weight heparin versus oral anticoagulant therapy for the prevention of recurrent venous thromboembolism in patients with cancer.  N Engl J Med2003; 349:146-153.
  37. Shah JJ, Gomes MPV, Deitcher SR: Thromboelastographic (TEG) characterization of patients with hematologic malignancy and chemotherapy induced thrombocytopenia (CIT).  Blood2004; 104:2729.
  38. Beyth RJ, Quinn LM, Landefeld S: Prospective evaluation of an index for predicting the risk of major bleeding risk outpatients treated with warfarin.  Am J Med1998; 105:91-99.
  39. Levine MN, Raskob G, Beyth RJ, et al: Hemorrhagic complications of anticoagulant treatment.  Chest2004; 126(suppl):287S-310S.
  40. Gitter MJ, Jaeger TM, Petterson TM, et al: Bleeding and thromboembolism during anticoagulant therapy: a population-based study in Rochester, Minnesota.  Mayo Clin Proc1995; 70:725-733.
  41. Wester JPJ, de Valk HW, Nieuwenhuis HK, et al: Risk factors for bleeding during treatment of acute venous thromboembolism.  Thromb Haemost1996; 76:682-688.
  42. Sanson B-J, Lensing AW, Prins MH, et al: Safety of low-molecular-weight heparin in pregnancy: a systematic review.  Thromb Haemost1999; 81:668-672.
  43. Bhandari M, Hirsh J, Weitz J, et al: The effects of standard and low molecular weight heparin on bone nodule formation in vitro.  Thromb Haemost1998; 80:413-417.
  44. Carman TL, Kanner AA, Barnett GH, Deitcher SR: Neurosurgery thromboprophylaxis following tumor surgery: a survey.  South Med J2003; 96:17-22.
  45. Schiff D, DeAngelis LM: Therapy of venous thromboembolism in patients with brain metastases.  Cancer1994; 73:493-498.
  46. Kessler CM, Deitcher SR, Merli G, et al: Low molecular weight heparin (enoxaparin) provides safe and effective antithrombotic therapy in active primary brain malignancies.  Proc Am Soc Clin Oncol2003; 22:741.(abstract 2980)
  47. Levine M, Gent M, Hirsch J, et al: A comparison of low-molecular-weight heparin administered primarily at home with unfractionated heparin administered in the hospital for proximal deep-vein thrombosis.  N Engl J Med1996; 334:677-681.
  48. Koopman MM, Prandoni P, Piovella F, et al: Treatment of venous thrombosis with intravenous unfractionated heparin administered in the hospital as compared with subcutaneous low-molecular-weight heparin administered at home.  N Engl J Med1996; 334:682-687.
  49. Hull RD, Raskob GL, Pineo GF, et al: Subcutaneous low-molecular-weight heparin compared with continuous intravenous heparin in the treatment of proximal vein thrombosis.  N Engl J Med1992; 326:975-982.
  50. Lindmarker P, Holmstrom M, Granqvist S, et al: Comparison of once-daily subcutaneous Fragmin with continuous intravenous unfractionated heparin in the treatment of deep vein thrombosis.  Thromb Haemost1994; 72:186-190.
  51. The Columbus Investigators : Low-molecular-weight heparin in the treatment of patients with venous thromboembolism.  N Engl J Med1997; 337:657-662.
  52. Warkentin TE, Levine MN, Hirsh J, et al: Heparin-induced thrombocytopenia in patients treated with low-molecular-weight heparin or unfractionated heparin.  N Engl J Med1995; 332:1330-1335.
  53. Lensing AW, Prins MH, Davidson BL, Hirsh J: Treatment of deep venous thrombosis with low molecular weight heparins: a meta-analysis.  Arch Intern Med1995; 155:601-607.
  54. Siragusa S, Cosmi B, Piovella , et al: Low molecular weight heparins and unfractionated heparin in the treatment of patient with acute venous thromboembolism: results of a meta-analysis.  Am J Med1996; 100:269-277.
  55. Gould MK, Dembitzer AD, Doyle RL, et al: Low-molecular-weight heparins compared with unfractionated heparin for treatment of acute deep venous thrombosis: a meta-analysis of randomized, controlled trails.  Ann Intern Med1999; 130:800-809.
  56. Dolovich LR, Ginsberg JS, Douketis JD, et al: A meta-analysis comparing low-molecular-weight heparins with unfractionated heparin in the treatment of venous thromboembolism: examining some unanswered questions regarding location of treatment, product type, and dosing frequency.  Arch Intern Med2000; 160:181-188.
  57. Green D, Hull RD, Brant R, Pineo GF: Lower mortality in cancer patients treated with low-molecular-weight versus standard heparin.  Lancet1992; 339:1476.
  58. Niers TM, Klerk CP, Dinisio M, et al: Mechanisms of heparin induced anticancer activity in experimental cancer models.  Crit Rev Oncol Hematol2007; 61:195-207.
  59. Kakkar AK, Levine MN, Kadziola Z, et al: Low molecular weight heparin, therapy with dalteparin, and survival in advanced cancer: the fragmin advanced malignancy outcome study (FAMOUS).  J Clin Oncol2004; 22:1944-1948.
  60. Lee AY, Rickles FR, Julian JA, et al: Randomized comparison of low molecular weight heparin and coumarin derivatives on the survival of patients with cancer and venous thromboembolism.  J Clin Oncol2005; 23:2123-2129.
  61. Klerk CP, Smorenburg SM, Otten HM, et al: The effect of low molecular weight heparin on survival in patients with advanced malignancy.  J Clin Oncol2005; 23:2119-2120.
  62. Altinbas M, Coskun HS, Er O, et al: A randomized clinical trial of combination chemotherapy with and without low-molecular-weight heparin in small cell lung cancer.  J Thromb Haemost2004; 2:1266-1271.
  63. Lensing AWA, Buller HR, Prandoni P, et al: Contrast venography, the gold standard for the diagnosis of deep vein thrombosis: improvement in observer agreement.  Thromb Haemost1992; 67:8-12.
  64. Dauzat M, Laroche JP, Deklunder G, et al: Diagnosis of acute lower limb deep venous thrombosis with ultrasound: trends and controversies.  J Clin Ultrasound1997; 25:343-358.
  65. Killewich LA, Bedford GR, Beach KW, Strandness Jr DE: Diagnosis of deep venous thrombosis: a prospective study comparing duplex scanning to contrast venography.  Circulation1989; 79:810-814.
  66. Bucek RA, Kos T, Schober E, et al: Ultrasound with Levovist in the diagnosis of suspected calf vein thrombosis.  Ultrasound Med Biol2001; 27:455-460.
  67. Bach AM, Hann LE: When the common femoral vein is revealed as flattened on spectral Doppler sonography: is it a reliable sign for diagnosis of proximal venous obstruction?.  AJR Am J Roentgenol1997; 168:733-736.
  68. Giess CS, Bach AM, Hann LE: Lower extremity venous sonography in the high-risk cancer population: one leg or two?.  AJR Am J Roentgenol2001; 176:1049-1052.
  69. Prandoni P, Lensing AWA, Piccioli A, et al: Ultrasonography of contralateral vein in patients with unilateral deep-vein thrombosis.  Lancet1998; 352:786.
  70. Prandoni P, Lensing AWA, Berbardi E, et al: The diagnostic value of compression ultrasonography in patients with suspected recurrent deep vein thrombosis.  Thromb Haemost2002; 88:402-406.
  71. Mayo JR, Ketai LH: Combined CT venography and pulmonary angiography.  Radiographics2002; 22:S20-S22.
  72. Ghaye B, Szapiro D, Willems V, Dondelinger RF: Pitfalls in CT venography of lower limbs and abdominal veins.  AJR Am J Roentgenol2002; 178:1465-1471.
  73. Froehlich JB, Prince MR, Greenfield LJ, et al: “Bull's-eye” sign on gadolinium-enhanced magnetic resonance venography determines thrombus presence and age: a preliminary study.  J Vasc Surg1997; 26:809-816.
  74. Spritzer CE, Norconk Jr JJ, Sostman HD, Coleman RE: Detection of deep venous thrombosis by magnetic resonance imaging.  Chest1993; 104:54-60.
  75. Taillefer R, Edell S, Innes G, et al: Acute thromboscintigraphy with 99mTc-apcitide: results of the phase 3 multicenter clinical trial comparing 99mTc-apcitide scintigraphy with contrast venography for imaging acute DVT.  J Nucl Med2000; 41:1214-1223.
  76. Lee AYY, Julian JA, Levine MN, et al: Clinical utility of a rapid whole-blood D-dimer assay in patients with cancer who present with suspected acute deep venous thrombosis.  Ann Intern Med1999; 131:417-423.
  77. Bates SM, Grand'Maison A, Johnston M, et al: A latex D-dimer reliably excludes venous thromboembolism.  Arch Intern Med2001; 161:447-453.
  78. Bates SM, Kearon C, Crowther M, et al: A diagnostic strategy involving a quantitative latex D-dimer assay reliably excludes deep venous thrombosis.  Ann Intern Med2003; 138:787-794.
  79. Büller HR, Agnelli A, Hull RD, et al: Antithrombotic therapy for venous thromboembolic disease.  Chest2004; 126(suppl):401S-428S.
  80. Anand SS, Bates S, Ginsberg JS, et al: Recurrent venous thrombosis and heparin therapy: an evaluation of the importance of early activated partial thromboplastin times.  Arch Intern Med1999; 159:2029-2032.
  81. Raschke RA, Reilly BM, Guidry JR, et al: The weight-based heparin dosing nomogram compared with a standard care nomogram: a randomized, controlled trial.  Ann Intern Med1993; 119:874-881.
  82. Merli G, Spiro TE, Olsson C-G, et al: Subcutaneous enoxaparin once or twice daily compared with intravenous unfractionated heparin for treatment of venous thromboembolic disease.  Ann Intern Med2001; 134:191-202.
  83. Hainer JW, Barrett JS, Assaid CA, et al: Dosing in heavy-weight/obese patients with the LMWH, tinzaparin: a pharmacodynamic study.  Thromb Haemost2002; 87:817-823.
  84. Büller HR, Davidson BL, Decousus H, et al: Fondaparinux or enoxaparin for the initial treatment of symptomatic deep venous thrombosis: a randomized trial.  Ann Intern Med2004; 140:867-873.
  85. Büller HR, Davidson BL, Decousus H, et al: Subcutaneous fondaparinux versus intravenous unfractionated heparin in the initial treatment of pulmonary embolism.  N Engl J Med2003; 349:1695-1702.
  86. Moll S, Ortel TL: Monitoring warfarin therapy in patients with lupus anticoagulants.  Ann Intern Med1997; 127:177-185.
  87. Choueiri T, Deitcher SR: Why shouldn't we solely use warfarin to treat acute venous thrombosis?.  Cleve Clin J Med2002; 69:546-548.
  88. Monreal M, Lafoz E, Olive A, et al: Comparison of subcutaneous unfractionated heparin with a low molecular weight heparin (Fragmin) in patients with venous thromboembolism and contraindications to coumarin.  Thromb Haemost1994; 71:7-11.
  89. Deitcher SR: Clinical utility of subcutaneous hirudins.  Am J Health Syst Pharm2003; 60(suppl):S27-S31.
  90. Deitcher SR, Ngengwe R, Kaplan R, et al: Subcutaneous lepirudin for heparin induced thrombocytopenia and when other anticoagulants fail: illustrative cases.  Clin Adv Hematol Oncol2004; 2:382-384.
  91. Hull RD, Pineo GF, Mah AF, Brant RF: A randomized trial evaluating long-term low-molecular-weight heparin therapy for three months versus intravenous heparin followed by warfarin sodium.  Blood2002; 100:148a.
  92. Gruel Y, Meyer G, Marjanovic Z, et al: Canthanox, a randomized controlled study comparing low molecular weight heparin and warfarin for the prevention of recurrent venous thromboembolism in patients with cancer.  Blood2001; 98:707a.
  93. Zacharski LR, Henderson WG, Rickles FR, et al: Effect of warfarin on survival in small cell carcinoma of the lung: Veterans Administration Study No. 75.  JAMA1981; 245:831-835.
  94. Schulman S, Lindmarker P: Incidence of cancer after prophylaxis with warfarin against recurrent venous thromboembolism: duration of anticoagulation trial.  N Engl J Med2000; 342:1953-1958.
  95. Ridker PM, Goldhaber SZ, Danielson E, et al: Long-term, low-intensity warfarin therapy for the prevention of recurrent venous thromboembolism.  N Engl J Med2003; 348:1425-1434.
  96. Baarslag H-J, van Beek EJR, Koopman MMW, Reekers JA: Prospective study of color duplex ultrasound compared with contrast venography in patients suspected of having deep venous thrombosis of the upper extremities.  Ann Intern Med2002; 136:865-872.
  97. Mustafa BO, Rathbun SW, Whitsett TL, Raskob GE: Sensitivity and specificity of ultrasonography in the diagnosis of upper extremity deep vein thrombosis: a systematic review.  Arch Intern Med2002; 162:401-404.
  98. Haire WD, Lynch TG, Lieberman RP, et al: Utility of duplex ultrasound in the diagnosis of symptomatic catheter-induced subclavian venous thrombosis.  J Ultrasound Med1991; 10:493-496.
  99. Hansen ME, Spritzer CE, Sostman HD: Assessing the patency of mediastinal and thoracic inlet veins: value of MR imaging.  Am J Radiol1990; 155:1177-1182.
  100. Balestreri L, De Cicco M, Matovic M, et al: Central venous catheter-related thrombosis in clinically asymptomatic oncology patients: a phlebographic study.  Eur J Radiol1995; 20:108-111.
  101. Monreal M, Raventos A, Lerma R, et al: Pulmonary embolism in patients with upper extremity DVT associated to venous central lines: a prospective study.  Thromb Haemost1994; 72:548-550.
  102. Ault M, Artal R: Upper extremity DVT: what is the risk?.  Arch Intern Med1998; 158:1950-1952.
  103. Gotway MB, Edinburgh KJ, Feldstein VA, et al: Imaging evaluation of suspected pulmonary embolism.  Curr Probl Diag Radiol1999; 28:132-184.
  104. The Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) Investigators : Value of the ventilation/perfusion scan in acute pulmonary embolism.  JAMA1990; 263:2753-2759.
  105. Stein PD, Athanasoulis C, Alavi A, et al: Complications and validity of pulmonary angiography in acute pulmonary embolism.  Circulation1992; 85:462-468.
  106. Smith TP: Pulmonary embolism: what's wrong with this diagnosis?.  AJR Am J Roentgenol2000; 174:1489-1497.
  107. Gottschalk A, Sostman HD, Coleman RE, et al: Ventilation-perfusion scintigraphy in the PIOPED study, Part II: evaluation of the scintigraphic criteria and interpretations.  J Nucl Med1993; 34:1119-1124.
  108. Hartmann IJC, Hagen PJ, Melissant CF, et al: Diagnosing acute pulmonary embolism: effect of chronic obstructive pulmonary disease on the performance of D-dimer testing, ventilation/perfusion scintigraphy, spiral computed tomographic angiography, and conventional angiography.  Am J Respir Crit Care Med2000; 162:2232-2237.
  109. Forbes KPN, Reid JH, Murchison JT: Do preliminary chest X-ray findings define the optimum role of pulmonary scintigraphy in suspected pulmonary embolism?.  Clin Radiol2001; 56:397-400.
  110. Chin BB, Welsh JS, Kleinberg L, et al: Nonsegmental ventilation-perfusion scintigraphy mismatch after radiation therapy.  Clin Nucl Med1999; 24:54-56.
  111. Garg K: CT of pulmonary thromboembolic disease.  Radiol Clin North Am2002; 40:111-122.
  112. Remy-Jardin M, Tillie-Leblond I, Szapiro D, et al: CT angiography of pulmonary embolism in patients with underlying respiratory disease: impact of multislice CT on image quality and negative predictive value.  Eur Radiol2002; 12:1971-1978.
  113. Schoepf UJ, Holzknecht N, Helmberger TK, et al: Subsegmental pulmonary emboli: improved detection with thin-collimation multi-detector row spiral CT.  Radiology2002; 222:483-490.
  114. Stein PD, Henry JW: Prevalence of acute pulmonary embolism in central and subsegmental pulmonary arteries and relation to probability interpretation of ventilation/perfusion lung scans.  Chest1997; 111:1246-1248.
  115. Ferretti GR, Bosson J-L, Buffaz P-D, et al: Acute pulmonary embolism: role of helical CT in 164 patients with intermediate probability at ventilation-perfusion scintigraphy and normal results at duplex US of the legs.  Radiology1997; 205:453-458.
  116. Blachere H, Latrabe V, Montaudon M, et al: Pulmonary embolism revealed on helical CT angiography: comparison with ventilation-perfusion radionuclide lung scanning.  AJR Am J Roentgenol2000; 174:1041-1047.
  117. Ost D, Rozenshtein A, Saffran L, Snider A: The negative predictive value of spiral computed tomography for the diagnosis of pulmonary embolism in patients with nondiagnostic ventilation-perfusion scans.  Am J Med2001; 110:16-21.
  118. Tillie-Leblond I, Mastora I, Radenne F, et al: Risk of pulmonary embolism after a negative spiral CT angiogram in patients with pulmonary disease: 1-year clinical follow-up study.  Radiology2002; 223:461-467.
  119. Nilsson T, Olausson A, Johnsson H, et al: Negative spiral CT in acute pulmonary embolism.  Acta Radiol2002; 43:486-491.
  120. van Strijen MJL, de Monyé W, Schiereck J, et al: Single-detector helical computed tomography as the primary diagnostic test in suspected pulmonary embolism: a multicenter clinical management study of 510 patients.  Ann Intern Med2003; 138:307-314.
  121. Shah AA, Davis SD, Gamsu G, Intriere L: Parenchymal and pleural findings in patients with and patients without acute pulmonary embolism detected at spiral CT.  Radiology1999; 211:147-153.
  122. Meaney JFM, Weg JG, Chenevert TL, et al: Diagnosis of pulmonary embolism with magnetic resonance angiography.  N Engl J Med1997; 336:1422-1427.
  123. Oudkerk M, van Beek EJR, van Ooijen PMA, et al: Comparison of contrast-enhanced magnetic resonance angiography and conventional pulmonary angiography for the diagnosis of pulmonary embolism: a prospective study.  Lancet2002; 359:1643-1647.
  124. Pruszczyk P, Torbicki A, Kuch-Wocial A, et al: Diagnostic value of transesophageal echocardiography in suspected haemodynamically significant pulmonary embolism.  Heart2001; 85:628-634.
  125. Simonneau G, Sors H, Charbonnier B, et al: A comparison of low-molecular-weight heparin with unfractionated heparin for acute pulmonary embolism.  N Engl J Med1997; 337:663-669.
  126. Goldhaber SZ: Thrombolysis in pulmonary embolism: a debatable indication.  Thromb Haemost2001; 86:444-451.
  127. Konstantinides S, Tiede N, Geibel A, et al: Comparison of alteplase versus heparin for resolution of major pulmonary embolism.  Am J Cardiol1998; 82:966-970.
  128. Konstantinides S, Geibel A, Heusel G, et al: Heparin and alteplase compared with heparin alone in patients with submassive pulmonary embolism.  N Engl J Med2002; 347:1143-1150.
  129. Nambi V, Gomes MP, Deitcher SR: Goal-oriented thrombolytic therapy for venous thromboembolic disease.  Clin Adv Hematol Oncol2004; 2:169-172.
  130. Conners MS, Becker S, Guzman RJ, et al: Duplex scan-directed placement of inferior vena cava filters: a five-year institutional experience.  J Vasc Surg2002; 35:286-291.
  131. Ricci P, Cantisani V, Biancari F, et al: Contrast-enhanced color Doppler US in malignant portal vein thrombosis.  Acta Radiol2000; 41:470-473.
  132. van Breda A, Rubin BE, Druy EM: Detection of inferior vena cava abnormalities by computed tomography.  J Comput Assist Tomogr1979; 3:164-169.
  133. Stanley JH, Sanchez F, Vujic I, Schabel SI: Failure to visualize the abdominal portion of the inferior vena cava with computed tomography: a clue to underlying abdominal pathology.  Br J Radiol1986; 59:1163-1166.
  134. Marks WM, Korobkin M, Callen PW, Kaiser JA: CT diagnosis of tumor thrombosis of the renal vein and inferior vena cava.  Am J Roentgenol1978; 131:843-846.
  135. Glazer GM, Callen PW, Parker JJ: CT diagnosis of tumor thrombus in the inferior vena cava: avoiding the false-positive diagnosis.  Am J Roentgenol1981; 137:1265-1267.
  136. Didier D, Racle A, Etievent JP, Weill F: Tumor thrombus of the inferior vena cava secondary to malignant abdominal neoplasms: US and CT evaluation.  Radiology1987; 162:83-89.
  137. Valla D-C, Condat B: Portal vein thrombosis in adults: pathophysiology, pathogenesis and management.  J Hepatol2000; 32:865-871.
  138. Jacoby WT, Cohan RH, Baker ME, et al: Ovarian vein thrombosis in oncology patients: CT detection and clinical significance.  Am J Roentgenol1990; 155:291-294.
  139. Yassa NA, Ryst E: Ovarian vein thrombosis: a common incidental finding in patients who have undergone total abdominal hysterectomy and bilateral salpingo-oophorectomy with retroperitoneal lymph node dissection.  Am J Roentgenol1999; 172:45-47.
  140. Soler R, Rodriguez E, Lopez MF, et al: MR imaging in inferior vena cava thrombosis!.  Eur J Radiol1995; l19:101-107.
  141. Cheng HC, Chu WC, Chai JW: Convergent flow phenomenon mimics the appearance of venous thrombosis in gradient-echo images with or without the presence of a contrast agent.  Magn Reson Imaging1997; 15:863-867.
  142. Blum U, Wildanger G, Windfuhr M, et al: Preoperative CT and MR imaging of inferior vena cava leiomyosarcoma.  Eur J Radiol1995; 20:23-27.
  143. Roy C, Beaujeux R, Mutter D: Leimyosarcoma of the femoral vein: imaging findings.  AJR Am J Roentgenol1993; 160:1125-1126.
  144. van Zanten TEG, Golding RP: CT and MR demonstration of leiomyosarcoma of inferior vena cava.  J Comput Assist Tomogr1987; 11:670-674.
  145. Kalodiki E, Nicolaides AN: Superficial thrombophlebitis and low-molecular-weight heparins.  Angiology2002; 53:659-663.
  146. Lutter KS, Kerr TM, Roedersheimer LR, et al: Superficial thrombophlebitis diagnosed by duplex scanning.  Surgery1991; 110:42-46.
  147. Jorgensen JO, Hanel KC, Morgan AM, Hunt JM: The incidence of deep venous thrombosis in patients with superficial thrombophlebitis of the lower limbs.  J Vasc Surg1993; 18:70-73.
  148. Chengelis DL, Bendick PJ, Glover JL, et al: Progression of superficial venous thrombosis to deep vein thrombosis.  J Vasc Surg1996; 24:745-749.
  149. Bouanameaux H, Reber-Wasem M-A: Superficial thrombophlebitis and deep vein thrombosis: a controversial association.  Arch Intern Med1997; 157:1822-1824.
  150. Blumemberg RM, Barton E, Gelfand ML, et al: Occult deep venous thrombosis complicating superficial thrombophlebitis.  J Vasc Surg1998; 27:338-343.
  151. Schwarz RE, Marrero AM, Conlon KC, Burt M: Inferior vena cava filters in cancer patients: indications and outcomes.  J Clin Oncol1996; 14:652-657.
  152. Ihnat DM, Mills JL, Hughes JD, et al: Treatment of patients with venous thromboembolism and malignant disease: should vena cava filter placement be routine?.  J Vasc Surg1998; 28:800-807.
  153. Streift MB: Vena cava filters: a comprehensive review.  Blood2000; 95:3669-3677.
  154. Gomes MPV, Kaplan KL, Deitcher SR: Patients with inferior vena caval filters should receive chronic thromboprophylaxis.  Med Clin N Am2003; 87:1189-1203.
  155. Decousus H, Leizorovicz A, Parent F, et al: A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis.  N Engl J Med1998; 338:409-415.
  156. Prandoni P, Lensing AW, Buller HR, et al: Deep-vein thrombosis and the incidence of subsequent symptomatic cancer.  N Engl J Med1992; 327:1128-1133.
  157. Schulman S, Lindmarker P: Incidence of cancer after prophylaxis with warfarin against recurrent venous thromboembolism.  N Engl J Med2000; 342:1953-1958.
  158. Sørensen HT, Mellemkjær L, Olsen JH, et al: Prognosis of cancers associated with venous thromboembolism.  N Engl J Med2000; 343:1846-1850.
  159. Collins R, Scrimgeour A, Yusuf S, et al: Reduction in fatal pulmonary embolism and venous thrombosis by perioperative administration of subcutaneous heparin: overview of results of randomized trials in general, orthopedic, and urologic surgery.  N Engl J Med1988; 318:1162-1173.
  160. Geerts WH, Pineo GF, Heit JA, et al: Prevention of venous thromboembolism.  Chest2004; 126:338S-400S.
  161. Maxwell GL, Synan I, Dodge R, et al: Pneumatic compression vs low molecular weight heparin in gynecologic oncology surgery: a randomized trial.  Obstet Gynecol2001; 98:989-995.
  162. Lin PP, Graham D, Hann LE, et al: Deep venous thrombosis after orthopedic surgery in adult cancer patients.  J Surg Oncol1998; 68:41-47.
  163. Maxwell GL, Myers ER, Clarke-Pearson DL: Cost-effectiveness of deep venous thrombosis prophylaxis in gynecologic oncology surgery.  Obstet Gynecol2000; 95:206-214.
  164. Comerota AJ, Katz ML, White JV: Why does prophylaxis with external pneumatic compression for deep vein thrombosis fail?.  Am J Surg1992; 164:265-268.
  165. Clarke-Pearson DL, Dodge RK, Synan I, et al: Venous thromboembolism prophylaxis: patients at high risk to fail intermittent pneumatic compression.  Obstet Gynecol2003; 101:157-163.
  166. ENOXACAN Study Group : Efficacy and safety of enoxaparin vs unfractionated heparin for prevention of deep vein thrombosis in elective cancer surgery: a double-blind randomized multicentre trial with venographic assessment.  Br J Surg1997; 84:1099-1103.
  167. Ramos R, Salem BI, De Pawlikowski MP, et al: The efficacy of pneumatic compression stockings in the prevention of pulmonary embolism after cardiac surgery.  Chest1996; 109:82-85.
  168. Agnelli G, Piovella F, Buoncristiani P, et al: Enoxaparin plus compression stockings compared with compression stockings alone in the prevention of venous thromboembolism after elective neurosurgery.  N Engl J Med1998; 339:80-85.
  169. Goldhaber SZ, Dunn K, Gerhard-Herman M, et al: Low rate of venous thromboembolism after craniotomy for brain tumor using multimodality prophylaxis.  Chest2002; 122:1933-1937.
  170. Bergqvist D, Agnelli G, Cohen AT, et al: Duration of prophylaxis against venous thromboembolism with enoxaparin after surgery for cancer.  N Engl J Med2002; 346:975-980.
  171. Agnelli G, Bolis G, Capussotti L, et al: A clinical outcome-based prospective study on venous thromboembolism after cancer surgery: the @RISTOS project.  Ann Surg2006; 243:89-95.
  172. Agnelli G, Bergqvist D, Cohen AT, Gallus AS, Gent M: Randomized clinical trial of postoperative fondaparinux versus perioperative dalteparin for prevention of venous thromboembolism in high-risk abdominal surgery.  Br J Surg2005; 92:1212-1220.
  173. Bern MM, Lokich JJ, Wallach SR, et al: Very low doses of warfarin can prevent thrombosis in central venous catheters.  Ann Intern Med1990; 112:423-428.
  174. De Cicco M, Matovic M, Balestreri L, et al: Central venous thrombosis: an early and frequent complication in cancer patients bearing long-term Silastic catheter. A prospective study.  Thromb Res1997; 86:101-113.
  175. Horne MK, May DJ, Alexander R, et al: Venographic surveillance of tunneled venous access devices in adult oncology patients.  Ann Surg Oncol1995; 2:174-178.
  176. Monreal M, Alastrue A, Rull M, et al: Upper extremity deep venous thrombosis in cancer patients with venous access devices: prophylaxis with a low molecular weight heparin (Fragmin).  Thromb Haemost1996; 75:251-253.
  177. Monreal M, Lafoz E, Ruiz J, et al: Upper-extremity deep venous thrombosis and pulmonary embolism: a prospective study.  Chest1991; 99:280-283.
  178. Timsit JF, Farkas JC, Boyer JM, et al: Central vein catheter-related thrombosis in intensive care patients: incidence, risk factors and relationship with catheter-related sepsis.  Chest1998; 114:207-213.
  179. Couban S, Goodyear M, Burnell M, et al: Randomized placebo-controlled study of low-dose warfarin for the prevention of central venous catheter-associated thromboses in patients with cancer.  J Clin Oncol2005; 23:4063-4069.
  180. Verso M, Agnelli G, Bertoglio S, et al: Enoxaparin for the prevention of venous thromboembolism associated with central venous catheters: a double-blind, placebo-controlled, randomized study in cancer patients.  J Clin Oncol2005; 23:4057-4062.
  181. Karthaus M, Kretzschmar A, Kroning H, et al: Dalteparin for prevention of catheter-related complications in cancer patients with central venous catheters: final results of a double-blind, placebo-controlled phase III trial.  Ann Oncol2006; 17:289-296.
  182. Deitcher SR, Fesen MR, Kiproff PM, et al: Safety and efficacy of alteplase for restoring function in occluded central venous catheters: results of the Cardiovascular Thrombolytic to Open Occluded Lines trial.  J Clin Oncol2002; 20:317-324.
  183. Semba CP, Deitcher SR, Li X, et al: Treatment of occluded central venous catheters with alteplase: results in 1,064 patients.  J Vasc Interv Radiol2002; 13:1199-1205.
  184. Heit JA, Silverstein MD, Mohr DN, et al: Risk factors for deep vein thrombosis and pulmonary embolism.  Arch Intern Med2000; 160:809-815.
  185. Heit JA, O'Fallon WM, Petterson TM, et al: Relative impact of risk factors for deep vein thrombosis and pulmonary embolism.  Arch Intern Med2002; 162:1245-1248.
  186. Desai AA, Vogelzang NJ, Rini BI, et al: A high rate of venous thromboembolism in a multi-institutional phase II trial of weekly intravenous gemcitabine with continuous infusion fluorouracil and daily thalidomide in patients with metastatic renal cell carcinoma.  Cancer2002; 95:1629-1636.
  187. Levine M, Hirsh J, Gent M, et al: Double-blind randomised trial of very-low-dose warfarin for prevention of thromboembolism in stage IV breast cancer.  Lancet1994; 343:886-889.
  188. Rajan R, Gafni A, Levine M, et al: Very low-dose warfarin prophylaxis to prevent thromboembolism in women with metastatic breast cancer receiving chemotherapy in economic evaluation.  J Clin Oncol1995; 13:42-46.
  189. Heit JA, Melton LJ, Lohse CM, et al: Incidence of venous thromboembolism in hospitalized patients vs community residents.  Mayo Clin Proc2001; 76:1102-1110.
  190. Samama MM, for the Sirius Study Group : An epidemiologic study of risk factors for deep vein thrombosis in medical outpatients.  Arch Intern Med2000; 160:3415-3420.
  191. Mismetti P, Laporte-Simitsidis S, Tardy B, et al: Prevention of venous thromboembolism in internal medicine with unfractionated or low-molecular-weight heparins: a meta-analysis of randomised clinical trials.  Thromb Haemost2000; 83:14-19.
  192. Samama MM, Cohen AT, Darmon J-Y, et al: A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients.  N Engl J Med1999; 341:793-800.
  193. de Lissovoy G, Subedi P: Economic evaluation of enoxaparin as prophylaxis against venous thromboembolism in seriously ill medical patients: a US perspective.  Am J Manag Care2002; 8:1082-1088.
  194. Agnelli G, Piovella F, Buoncristiani P, et al: Enoxaparin plus compression stockings compared with compression stockings alone in the prevention of venous thromboembolism after elective neurosurgery.  N Engl J Med1998; 339:80-85.
  195. Estrada CA, McElligott J, Dolezal JM, et al: Asymptomatic patients at high risk for deep venous thrombosis who receive inadequate prophylaxis should be screened.  South Med J1999; 92:1145-1150.