Cancer is a well-established risk factor for VTE. Thrombosis is a major cause of morbidity in patients with neoplastic disease, and is reported to be clinically evident in 11–15% of patients being treated for malignancy. Risk for thrombosis increases with disease progression. One study of hospice patients with cancer revealed a 50% prevalence of thrombosis, often associated with poor mobility and low albumin level (4). Thrombosis has been observed in up to 50% of patients with cancer at autopsy (2, 3). The most thrombogenic tumors include ovarian, brain, pancreatic, gastric, and colorectal neoplasms. Breast carcinoma has a relatively low risk, but the risk rises with certain hormonal manipulations (5).

Mechanisms of cancer-induced activation of the coagulation mechanism have been attributed to aberrant expression of tissue factor by tumors or reactive endothelium, tumor-derived procoagulant factors that can activate factor X on malignant cells, dysfunctional prothrombotic hematopoietic clones, hyperviscosity, and inflammatory mechanisms (6, 7). Proinflammatory cytokines such as tumor necrosis factor-α and interleukin-1β induce expression of tissue factor on endothelial cells (8). These mechanisms may explain why patients with cancer are at increased risk for development of disseminated intravascular coagulation (DIC), which sometimes presents as localized thrombosis. Finally, a host of chemotherapeutic drugs, hormonal manipulations, and stasis resulting from vascular compression add to thrombotic risk (5).

Many other advanced disease states are complicated by thrombosis. The incidence of stroke and venous thrombosis is as high as 4% in severe heart failure (9). Hepatic dysfunction increases risk of thrombosis in part through decreased hepatic clearance of activated coagulation factors. Inflammatory reactions and acute phase responses are common to many disease states; the actions of inflammatory cytokines and elevated levels of factor VIII, fibrinogen (10, 11), and platelets all tip the hemostatic balance toward thrombosis. Additional risk factors include the presence of central venous catheters (6, 12) and venous stasis associated with immobilization due to pain and debilitation.

Although most thrombotic events that complicate the care of debilitated patients are venous, arterial thrombosis is also a potential problem. Arterial events are generally attributed to atherosclerotic disease, with formation of platelet-fibrin thrombi. Hypotension may worsen the progression of vascular ischemia in the face of preexisting arterial disease. Polycythemia vera and essential thrombocythemia are the predisposing factors in arterial and venous thrombosis. Finally, embolic venous thrombi may cross into the arterial circulation
through cardiac shunts and present as “paradoxical” arterial emboli in patients with patent foramen ovale.

The presenting signs and symptoms of VTE are often nonspecific, and the problem may be particularly pronounced in the palliative care setting. Alternative causes of extremity swelling include nonthrombotic vascular obstruction, heart failure, renal insufficiency, hypoalbuminemia, lymphatic obstruction, neurologic factors, and hypothyroidism. Similarly, the sensation of breathlessness may stem from anxiety, cardiac failure, tumor invasion, infection, and obstructive pulmonary disease. Conversely, edema in unusual sites may indicate venous thrombosis in palliative care patients. Upper extremity edema may be due to axillary or mediastinal metastasis, catheter-related thrombosis, or venous thrombosis. Hepatic vein thrombosis may present as worsening hepatic failure or sudden onset of ascites (2). Detection of VTE disease is important because it can be successfully treated. Treatment not only reduces the risk of fatal pulmonary embolism (PE) but it can also reduce leg pain, immobility, and symptoms of breathlessness (13). Bilateral asymmetric leg edema was the most common presenting finding in one study of hospice patients with advanced cancer who later developed VTE (13). Investigation of symptoms consistent with thrombosis is strongly recommended in patients in whom antithrombotic therapy may be considered.

Noninvasive studies are the diagnostic tools of choice, because contrast venography and conventional pulmonary angiography (the reference standards) are inconvenient, costly, and associated with substantial morbidity. Quantitation of fibrin D-dimer (the plasmin-derived degradation product of cross-linked fibrin clot) may be insufficient for exclusion of VTE in patients with cancer (14). Although one study observed a satisfactory negative predictive value in patients with cancer (15), another study reported that the negative predictive value of the SimpliRED bedside D-dimer test was only 79% in patients with cancer versus 97% in a more general population of ambulatory patients (16, 17).

Clinical approaches that include noninvasive imaging studies are recommended for the evaluation of suspected VTE (18, 19). Compression ultrasonography may be the best study for the diagnosis of proximal deep vein thrombosis (DVT) in the terminal patient. It is simple, highly accurate, and fast when done by experienced personnel. The sensitivity for proximal leg DVT is reported at over 97%, with specificity reported at 92–100%. Compression ultrasonography is significantly less useful in the evaluation of thrombosis below the knee (2, 19).

Spiral computed tomography (CT) (computed tomographic pulmonary angiography) is rapidly replacing lung scintigraphy (also known as ventilation/perfusion scan) as the diagnostic procedure of choice for noninvasive evaluation of patients with suspected PE. Although, a “normal” scintigraphy report effectively excludes the diagnosis of PE and a “high-probability” report is sufficient for diagnosing a patient with a moderate to high clinical suspicion of PE, the technique is nondiagnostic in approximately two thirds of patients (20). Pulmonary disease and tumor in the lung degrade the diagnostic yield of lung scintigraphy studies.

Spiral CT scan has a high sensitivity for detection of PE in central pulmonary vessels (sensitivity and positive predictive value approach 95%), and advances using multidetector CT scan have improved visualization of subsegmental pulmonary vessels (21). Recent evidence supports the safety of withholding anticoagulation therapy based on a negative spiral CT scan study (22). CT scan is also useful for uncovering alternative sources of pulmonary symptoms in patients with advanced disease, because the images provide details of lung parenchyma, mediastinum, and pleura. Concerns about CT scan include the lingering questions regarding sensitivity for detection of embolism in subsegmental pulmonary arteries, the occasional misinterpretation of studies, the requirement for intravenous injection of a significant iodine-contrast dye load, and the high cost of the study (23).

Utilization of these diagnostic techniques in the palliative care setting has not been extensively evaluated. A survey of palliative care physicians in the United Kingdom revealed that only 60–80% of responding physicians would use tests to confirm a clinically suspected VTE (24). One palliative care group’s protocol is to establish the degree of clinical suspicion, and if high, to obtain leg ultrasonography. When PE is suspected and the leg ultrasound is inconclusive, spiral CT scan is obtained. Pulmonary scintigraphy is reserved for patients in whom dye load is contraindicated (25).

The American College of Chest Physicians periodically updates its recommendations for the management of VTE in the nonpalliative care setting (26). The goals of treatment of VTE are to prevent death from progressive PE and to minimize the postphlebitic symptoms of pain, swelling, and dyspnea. Thrombolytic therapy is usually considered overly aggressive for palliative care. Anticoagulation is the mainstay of therapy and is instituted immediately to inhibit new clot formation while intrinsic fibrinolytic mechanisms reopen obstructed blood vessels. Anticoagulation is then continued on a long-term basis to prevent recurrent thrombosis. In a general population of patients, the duration of anticoagulation therapy is stratified according to the patient’s risk for recurrence (26).

The main complication of anticoagulation therapy is hemorrhage, and assessment of the risks of hemorrhage should be undertaken before instituting anticoagulation (Table 31.1). Absolute contraindications include significant active bleeding or severe bleeding tendency. Relative contraindications include recent bleeding, recent surgery, moderate to severe bleeding tendency, thrombocytopenia, active peptic ulcer disease, uncontrolled hypertension, and severe renal or liver disease. Central nervous system hemorrhage is a particular concern in patients with metastatic cancer in the brain, especially from melanoma, choriocarcinoma, or renal cell carcinoma. However, several authors advocate the safety of anticoagulation in the setting of nonhemorrhagic metastatic disease to the central nervous system when close control of anticoagulation is maintained (27, 28). When hemorrhagic risk contraindicates anticoagulation therapy or anticoagulation has been proved to be insufficient to prevent thrombotic progression, inferior vena cava (IVC) filter devices can be inserted to preserve lung function and prevent death due to acute PE.

Heparin drugs have been the mainstay of initial anticoagulation therapy owing to their immediate onset of action (29). Their anticoagulant effect is achieved by promoting the inhibitory activity of antithrombin. Heparin drugs have been subdivided into unfractionated heparin (UFH), low-molecular-weight heparins (LMWH) derived by depolymerization of UFH, and synthetic pentasaccharides (fondaparinux).

LMWH preparations offer several important pharmacologic advantages over UFH (29) and have become the primary medication used for initial management of VTE disease. Many large clinical trials have demonstrated that LMWH is equivalent in safety and efficacy to UFH in the management of acute
VTE and that it is safe for use in outpatient community-based care (30). A recent meta-analysis of randomized trials revealed that LWMH is more effective and safer then UFH, but whether LMWH is superior in the subset of patients with cancer has not been formally evaluated (31). Similar to UFH, LMWH is a parenteral medication, but depolymerization results in a longer half-life, greater bioavailability, and more predictable pharmacodynamics. For the average patient with VTE, these pharmacologic advantages translate into weight-adjusted dosing once or twice daily without a requirement for laboratory monitoring. These advantages also render LMWH an appropriate agent for outpatient use. Other potential advantages include reduced risks for development of osteoporosis (32) and heparin-induced thrombocytopenia (HIT) (33). The main disadvantages of LMWH are increased cost and a more prolonged anticoagulant effect that is less reversible with protamine sulfate. Among patients with cancer, one study found a trend toward increased thrombus recurrence with once-daily dosing of enoxaparin compared to twice-daily dosing (34); however, the potential benefit of twice-daily dosing must be balanced against inconvenience and cost when considering the care of patients in a palliative care setting. Multiple LMWH preparations are available and the dosing schedules for each were largely empirically determined. The dosing schedule should appropriately match the LMWH preparation being used. Because LMWH is cleared by the kidney, in patients with renal insufficiency (creatinine greater than 2 mg per dL or estimated creatinine clearance of under 30 mL per minute) use of UFH or dose modification of LMWH with monitoring of levels is advisable. Although target levels of LWMH have not been established through therapeutic trials, a target peak level of 0.6–1.0 antifactor Xa units per mL measured 3–4 hours after subcutaneous administration of LMWH has been recommended (26).

UFH may still be used for the initial management of VTE in some patients. Its main advantages are low cost, short half-life, and reversibility by administration of protamine sulfate. The main disadvantages of UFH are its wide dose–response variability, narrow therapeutic window, and the need for parenteral administration. Other complications include the rare but serious immunologic condition HIT (35) and the risk of osteoporosis with very long term heparin therapy (32). UFH is usually given by continuous infusion (36), but subcutaneous administration every 12 hours has also been used. The therapeutic dose is determined empirically. Algorithms for prescriptive dose adjustment are based on frequent monitoring of anticoagulant effect (29). The sensitivity of the activated partial thromboplastin time (aPTT) to heparin effect varies widely between laboratories; it is advisable to consult the local laboratory to learn the recommended therapeutic range. The aPTT is a less useful yardstick of heparin effect in patients who have “heparin resistance,” which is defined as requirement of over 40,000 units of UFH per day to achieve a therapeutic range aPTT. Either direct heparin assessment by “antifactor Xa” assays (therapeutic range 0.35–0.70 U per mL) or use of LMWH is recommended in this setting.

The first synthetic pentasaccaride anticoagulant, fondaparinux, was recently approved for initial treatment of VTE, with randomized trials demonstrating noninferiority to LMWH and UFH (37, 38). Like LMWH, fondaparinux requires subcutaneous administration and is cleared by the kidney. Potential advantages include a longer half-life (17–21 hours) allowing once-daily administration, and emerging experience suggests a potential use in patients with a history of HIT.

After achieving initial anticoagulation, long-term anticoagulation (secondary prophylaxis) is undertaken to prevent recurrence of VTE. Warfarin sodium derivatives (e.g., Coumadin) are frequently chosen for this phase of care (26), but LMWH should be considered in some settings (see subsequent text). Coumadin and other oral vitamin K antagonists inhibit hepatic synthesis of multiple coagulation factors. The onset of oral anticoagulant effect is delayed until previously synthesized coagulation factors are cleared. Therefore, “loading” doses do not overcome the long half-life of circulating clotting factors. Because of this delay, heparin drugs are usually used concurrently with initial oral anticoagulation to provide protection before the onset of oral anticoagulation effect. Current recommendations suggest that heparin drugs be maintained for at least 5 days and continued for 2 days after laboratory studies confirm that adequate oral anticoagulation has been established.

Management of oral anticoagulation is complex owing to the narrow therapeutic window, multiple drug interactions, interindividual differences in hepatic metabolism, and the shifting intensity of anticoagulation due to changes in diet (39). The therapeutic intensity of oral anticoagulation requires laboratory monitoring. A prothrombin time (PT)-based international normalized ratio (INR) target of 2.5 is suggested in most settings, but a higher target of 3.0 is suggested for patients with many types of mechanical heart valves (40). The typical initial dose of warfarin sodium is 5 mg per day in acute care patients, but initial doses may need to be lower in chronically ill patients, patients with poor nutrition, patients on medications that are known to increase oral anticoagulant effect, and in the elderly. Owing to individual variation in anticoagulant effect, ongoing monitoring of individual patient response and dose adjustment are required. Initially, INR monitoring and dose adjustment are performed daily, with monitoring intervals lengthened as the dose requirement is empirically established. Weekly evaluation may be prudent for at least the first 6–12 weeks of therapy, the time when the highest rate of hemorrhage occurs (41). In general patient populations, the risk of bleeding with INRs in the therapeutic range is between 2 and 3%, but patients with cancer are at increased risk for bleeding complications (2). Adverse events may be avoided through more frequent monitoring (28, 42). Outcome data are scant in the palliative care literature. In this setting, oral
anticoagulation may be more problematic owing to changes in diet, gastrointestinal (GI) or hepatic disturbances, changes in medications, or the need to discontinue anticoagulation to accomplish other invasive interventions without untoward risk of bleeding. One small hospice audit revealed a high incidence of oral anticoagulation–related hemorrhagic events, and found that external bleeding was quite distressing to the patients and their caregivers (1, 43). Tight INR control was somewhat helpful, but required frequent INR monitoring (averaging once every 2.4 days), adding a considerable burden to dying patients.