Thromboembolic Complications of Malignancy: Part 1
Thromboembolic Complications of Malignancy: Part 1
Thromboembolism affects many patients with solid tumors and clonal hematologic malignancies. Pathogenetic mechanisms include inflammatory- and tissue factor-mediated coagulation, natural anticoagulant deficiencies, fibrinolytic alterations, hyperviscosity, and activation of platelets, endothelial cells, and leukocytes. High rates of venous thromboembolism (VTE) occur with advanced pancreatic, breast, ovarian, germ cell, lung, prostate, and central nervous system cancers. Hodgkin disease, non-Hodgkin's lymphoma, myeloma, paroxysmal nocturnal hemoglobinuria, and certain leukemias also predispose to venous thromboembolism. Arterial and venous events occur with polycythemia vera and essential thrombocythemia. Central venous catheters and prothrombotic antitumor regimens augment the risk in some patients. Part 1 of this two-part article addresses pathophysiology, clinical presentations, and risk of malignancy-associated thrombosis. Part 2, which will appear in next month's issue, covers prophylaxis and treatment of these thromboembolic complications.
Over half of all adult cases of venous thromboembolism (VTE), including deep venous thrombosis and pulmonary embolism, occur in the setting of acquired, situational, hereditary, pharmacologic, and/or iatrogenic prothrombotic risk factors.[1,2] Recent or active cancer is one of the most powerful independent and additive acquired hypercoagulable conditions, approximating or exceeding the attributable risks associated with advanced age, acute infectious illness, or a prior history of VTE. The annual incidence of VTE among all cancer patients has been estimated to be 0.5%, with 10- to 20- fold higher rates among those with advanced ovarian, breast, pancreatic, and lung cancer, brain tumors, myeloproliferative disorders (ie, polycythemia vera and essential thrombocythemia), myeloma, and those treated with thrombogenic antitumor regimens.[3-5] Thromboembolic complications of malignancy significantly affect quality of life and may profoundly compromise overall management and outcome. Compared to VTE in individuals without cancer, malignancy-associated deep vein thrombosis and pulmonary embolism are associated with longer hospitalization times, a twofold higher 28-day mortality rate, a threefold higher rate of readmission and/or death at 6 months, and two- to fourfold greater risks of subsequent bleeding (on warfarin anticoagulation) and/or recurrent thrombosis (with or without continued anticoagulation).[5-7] Moreover, concurrent cancer increases the likelihood of perioperative venous and arterial thrombotic complications by two- to threefold. The 1-year survival rate of patients with cancer-associated VTE is onethird lower than the survival rate of cancer patients who remain thrombosis- free (12% vs 36%), reflecting, in part, the more advanced stage of disease among thrombotic patients. These thromboembolic complications translate into higher health-care expenditures[ 7] and spotlight the need for more selective, aggressive, and cost-effective interventions in this high-risk population. Part 1 of this two-part review summarizes pathophysiology, clinical presentation, and risk of malignancyassociated thrombosis. In part 2, which will appear in next month's issue, we will explore the management of thromboembolic complications in cancer patients, including the specifics of prophylaxis and treatment, cost considerations, and future perspectives. Pathophysiology of Malignancy- Associated Thrombosis Venous vs Arterial Vascular Events
The prothrombotic mechanisms that affect the vascular systems in patients with malignancies mirror those that are important in individuals without cancer. In general, venous thrombosis results from abnormalities that affect blood flow, the integrity of the venous endothelium and/or the hemostatic balance of activated procoagulants, natural anticoagulants, fibrinolytic mediators and, in some cases, platelets. Most acquired and congenital prothrombotic conditions alter those hemostatic mechanisms and, thereby, predominantly lead to deep venous thrombosis and pulmonary embolism.[1,2,9] Pharmacologic agents that inhibit one or more steps in the coagulation cascade, including unfractionated heparin, low-molecular-weight heparin, warfarin, or more specific inhibitors of factor Xa (eg, fondaparinux [Arixtra]) or thrombin (eg, argatroban, lepirudin [Refludan], bivalirudin [Angiomax]) are required to prevent and treat VTE. Given the minor role of platelets in this process, drugs that reversibly or irreversibly inhibit platelet function (eg, aspirin, clopidogrel [Plavix], ticlopidine, dipyridamole) are not effective for VTE prophylaxis or therapy. Arterial vascular events generally occur as a result of underlying atherosclerotic disease that ultimately leads to acute platelet-fibrin thrombotic occlusion and/or downstream emboli. In some cases, arterial emboli originate in the heart (eg, from a left atrial thrombus or valvular vegetations) or, paradoxically, from VTEs that traverse an intracardiac shunt into the arterial circulation. Antiplatelet agents decrease the chance of primary and secondary atherothrombotic complications and they may partially reduce the risk of cardioembolic events. Acute arterial thrombosis is treated with systemic anticoagulants and, in some cases, thrombolytic agents (eg, recombinant tissue-type plasminogen activator [tPA]). Malignancy-associated thromboembolic complications usually occur in the venous system; however, some conditions predispose to arterial thromboemboli. Patients with polycythemia vera, essential thrombocythemia, and disorders complicated by disseminated intravascular coagulation (DIC), heparin-induced thrombocytopenia with thrombosis, or nonbacterial thrombotic endocarditis (also known as "marantic" endocarditis), are at increased risk of arterial events.[3,4,10] In addition, a recent retrospective cohort study observed an overall 1.5% incidence of arterial thromboemboli among 66,106 patients hospitalized for malignancy and neutropenia. Interestingly, onehalf of those arterial events involved patients with leukemia and lymphoma, and the annual frequency more than doubled from 1995 to 2002. Prothrombotic Mechanisms and Altered Laboratory Values
The prothrombotic mediators and mechanisms implicated in malignancy- associated hypercoagulability are summarized in Table 1. The presence and severity of these abnormalities relate, in part, to the underlying histology and stage of disease, associated comorbid conditions, and antitumor treatments. Importantly, some of these mechanisms may also participate in tumor progression by facilitating microinvasion and metastasis. In addition, neoangiogenesis may be stimulated through tissue factor-induced upregulation of vascular endothelial growth factor (VEGF). The hypothetical link between malignancy-associated hypercoagulability and tumor biology has been supported by clinical data showing a decreased incidence of subsequent primary cancer diagnosis among patients with primary VTE who received longer-duration warfarin therapy and improved survival among a subset of cancer patients who received low-molecular-weight heparin for VTE prophylaxis. Alterations in routine hemostatic laboratory markers may reflect direct effects of cancer-associated prothrombotic factors, activation of downstream mediators, and/or the products of the malignant cells themselves (Table 1).[12,15,16] Among the most common examples of causal association are acute promyelocytic leukemia and advanced gastrointestinal (GI) ad- enocarcinomas, both of which are associated with high rates of thrombosis and laboratory markers of DIC (ie, elevated fibrin degradation products and/ or D-dimer, increased prothrombin time, increased activated partial thromboplastin time, increased thrombin time, decreased fibrinogen, and/or decreased platelet count). By comparison, patients with polycythemia vera and essential thrombocythemia are at increased risk of venous and arterial thrombosis, and endogenous platelet aggregation defects are frequently demonstrable by in vitro assays. The presence or absence of qualitative platelet defects, however, does not correlate with the thrombotic risk observed in patients with polycythemia vera or essential thrombocythemia. In most patients with malignancies, routine hemostatic assays are normal or only nonspecifically altered, and these findings do not predict clinical complications. More specialized or research-based assays for cellular-derived and plasma coagulation factors and hemostatic activity detect abnormalities in subsets of cancer patients (Table 1).[15-17] Increased levels of tissue factor (the major activator of factor VII), platelet factor 4 (a marker of platelet activation), procoagulants and/or markers of thrombin/fibrin generation (ie, prothrombin fragment 1 + 2 [F1+2], thrombin-antithrombin complex, fibrinopeptide A, and tPA), have been observed in patients with adenocarcinomas (particularly of the GI or genitourinary [GU] tract) and advanced solid tumors, and after acute thrombotic complications or recurrent thrombosis on anticoagulant therapy.[ 17] Additional abnormalities include antiphospholipid antibodies, primary fibrinolysis, and decreased activities of natural anticoagulants, including protein S, antithrombin III (ATIII), protein C, and/or the activated protein C (APC) complex.[16,17] The question of whether specific laboratory alterations might identify high-risk individuals who would benefit from prophylactic anticoagulation has been addressed in various studies. One prospective myeloma treatment trial observed an increased risk of VTE among patients with acquired APC resistance. Other recent studies have assessed coagulation markers in cancer patients at the time of VTE and compared those with markers in cancer patients without VTE. These trials have found correlations between acute VTE and acquired APC resistance or levels of thrombin-antithrombin III complex (TAT), F1+2, tPA, protein C activity, and/or von Willebrand factor antigen. Recurrent VTE has been associated with Ddimer and TAT levels. Before these hemostatic markers can be used to guide thromboprophylaxis in the clinic, confirmatory evidence is needed from well-designed prospective trials among patients with similar disease types and treatment courses. Clinical Presentations of Malignancy-Associated Thrombosis Occult Malignancy in Patients With VTE
Thrombosis may be the first clinical sign of an underlying malignancy. Roughly 7% to 10% of adults with idiopathic VTE (ie, not predisposed by an identifiable preexisting risk factor), will be diagnosed with cancer at the time of presentation or within the next 6 to 24 months.[19,20] This is especially true for older individuals. Because the prevalence of cancer is low in younger adults, the relative risk of an associated malignancy is actually highest among those under age 60. The malignancies most likely to be discovered include non-Hodgkin's lymphoma (NHL) and carcinomas of the pancreas, ovary, liver, brain, GI tract, lung, breast, and GU system. Approximately 10% to 25% of patients with unrecognized polycythemia vera or essential thrombocythemia will be diagnosed at the time of an arterial or venous thromboembolic event and/or will have a history of previous thrombosis. Similarly, up to 20% of patients with occult paroxysmal nocturnal hemoglobinuria (PNH), an acquired hematopoietic clonal disorder, will present with venous or, less commonly, arterial thrombosis. Of clinical impor- tance, the thrombotic complications of PNH, polycythemia vera, or essential thrombocythemia may involve unusual anatomic sites, such as the hepatic veins (Budd-Chiari syndrome), mesenteric, portal, and splenic veins, and cerebral sinuses. Retrospective studies and small prospective trials have suggested that extensive screening for occult malignancy in adults with VTE is not beneficial or cost-effective.[19,20] Either routine testing was felt to be sufficient to uncover cancer-related abnormalities in most patients and/or the prognosis would not have been affected by earlier detection. However, a more recent randomized screening trial and a prospective cohort study found that roughly half of preexisting cancers in patients with VTE are not detected by routine screening (ie, a complete physical examination, routine blood counts, blood chemistries, urine testing, and chest x-ray), but can be found with more extensive studies. Prostate-specific antigen, carcinoembryonic antigen, CA-125, alpha-fetoprotein, abdominopelvic imaging (by ultrasound and/or computed tomography), or additional imaging studies and endoscopy (as indicated) identified the occult cancer in 50% to 90% of cases. In addition, many of the malignancies found with extensive screening at the time of VTE were at earlier stages than the tumors diagnosed 8 to 11 months later among patients who had undergone initial routine testing, suggesting that some patients might benefit from extensive screening. No studies have yet determined whether extensive screening and earlier detection of occult malignancies in patients with VTE affect overall prognosis and survival. The one recent randomized clinical trial designed to address these questions could not complete accrual. Thus, until evidence- based guidance is available, the clinician must be cognizant of important clinical and laboratory indicators to prompt additional evaluation (Table 2).[21-23] Based on recent studies, imaging of the abdomen, pelvis, and chest would be an efficient next step in the search for an occult solid tumor. Thrombosis in Patients With Known Malignancies
The risks of VTE among patients with known malignancies have been estimated by retrospective analyses of treatment cohorts, patient registries, and Medicare claims databases.[5,24,25] The highest incidence rates, ranging from 76 to 120 events per 10,000 patient admissions, occur among patients with ovarian, brain, pancreatic, gastric, renal, and colorectal cancers (Figure 1). Increased incidence is also seen with lymphoma, leukemia, myeloma, liver, lung, prostate, gynecologic, and breast cancers. Some retrospective studies, but not others, have identified advanced tumor stage (compared to limited stage) and/or recent chemotherapy (compared to no therapy) as strong independent risk factors. From the clinician's perspective, hospitalizations for deep venous thrombosis or pulmonary embolism most frequently involve malignancies that are more prevalent in the community (such as lung, colon, prostate, and breast cancers, leukemia, and lymphoma) than malignancies that are most highly thrombogenic (Figure 1). Prospective treatment trials have provided additional important data regarding the incidence rates of thrombosis among patients with certain solid tumors. The 5-year rate of VTE in women with stage I/II breast cancer on no adjuvant treatment is roughly 0.2%, but is fourfold or 20-fold higher for women on tamoxifen or on chemotherapy plus tamoxifen, respectively. In addition, roughly 4% to 18% of women with advanced-stage breast cancer on chemotherapy suffer a thromboembolic event (reviewed in ). Similarly, VTE affects 11% of women during treatment for ovarian cancer,[ 26] 8% of men on chemotherapy for germ cell cancer (especially those with liver metastasis and receiving high-dose corticosteroids), up to 26% of patients with malignant glioma (especially those with lowerextremity paresis, reviewed in ), 4% to 7% with lung cancer (especially adenocarcinoma and metastatic disease),[ 29,30] 2% to 25% with prostate cancer (particularly among those receiving estrogenic agents), and 15% to 28% with pancreatic cancer (especially those with metastatic disease). Among the hematologic malignancies, roughly 6% to 13% of patients with Hodgkin's disease or NHL develop VTE.[33,34] Many of these events are related to tumor-associated vessel compression, catheter-associated thrombosis, more advanced disease stage, and/or earlier course of treatment. Notably, deep venous thrombosis and/or pulmonary embolism develop in 60% of patients with central nervous system (CNS) lymphoma. VTE occurs in 4.8% of patients following stem cell transplantation for hematologic malignancies, including 0.7% who develop arterial events. Predisposing risks include indwelling central catheters, line infection, sepsis, and pulmonary disease. Recent treatment trials for myeloma have demonstrated VTE in 10% to 15% of patients, with up to 35% incidence rates among some cohorts on thalidomide (Thalomid)- containing regimens. Polycythemia vera and essential thrombocythemia are associated with a 5% to 10% annual risk of thromboembolic events.[4,38] The majority of those events affect the arterial system (ie, stroke and myocardial infarction) and occur most commonly in older individuals (ie, age over 65 years) and/or among those with a prior thrombotic history. Thrombosis may also occur in patients with acute leukemias and concurrent disseminated intravascular coagulation, especially among those with acute promyelocytic leukemia. Roughly 3% to 38% of patients with acute lymphoblastic leukemia on asparaginase (Elspar)- and/or prednisone- containing regimens suffer deep venous thrombosis, CNS venous or arterial events, and catheter-related thrombosis.[39,40] PNH is associated with a 10-year risk of VTE ranging from 4% to 44%, with the highest risk among nonaplastic patients who circulate a high proportion of clonal PNH granulocytes.[21,41]