|Home | About | Journals | Submit | Contact Us | Français|
Changes in the hemostatic system and chronic hemostatic activation are frequently observed in patients with cancer, even in the absence of venous thromboembolism (VTE). VTE is a leading cause of death among patients with cancer and contributes to long-term mortality in patients with early as well as advanced-stage cancer. Mounting evidence suggests that components of the clotting cascade and associated vascular factors play an integral part in tumor progression, invasion, angiogenesis, and metastasis formation. Furthermore, there are intriguing in vitro and animal findings that anticoagulants, in particular the low molecular weight heparins (LMWHs), exert an antineoplastic effect through multiple mechanisms, including interference with tumor cell adhesion, invasion, metastasis formation, angiogenesis, and the immune system. Several relatively small randomized controlled clinical trials of anticoagulation as cancer therapy in patients without a VTE diagnosis have been completed. These comprise studies with LMWH, unfractionated heparin, and vitamin K antagonists, with overall encouraging but nonconclusive results and some limitations. Meta-analyses performed for the American Society of Clinical Oncology VTE Guidelines Committee and the Cochrane Collaboration suggest overall favorable effects of anticoagulation on survival of patients with cancer, mainly with LMWH. However, definitive clinical trials have been elusive and questions remain regarding the importance of tumor type and stage on treatment efficacy, the impact of fatal thromboembolic events, optimal anticoagulation therapy, and safety with differing chemotherapy regimens. Although the LMWHs and related agents hold promise for improving outcomes in patients with cancer, additional studies of their efficacy and safety in this setting are needed.
Patients with cancer have a substantially increased risk of venous thromboembolism (VTE) compared with patients without cancer.1–8 Changes in the hemostatic system and evidence of chronic hemostatic activation, including disseminated intravascular coagulation, are frequently observed in patients with cancer, even in the absence of VTE.9–14 There are many links between the hemostatic system and tumor cells, including fibrin or fibrinogen surrounding the tumor tissue bed.15–18 It has been postulated that components of the clotting cascade and associated vascular factors play an integral part in tumor progression, invasion, angiogenesis, and metastasis formation.19–23 Furthermore, intriguing in vitro and animal findings suggest that anticoagulants, in particular the low molecular weight heparins (LMWH), exert an antineoplastic effect, most likely by interfering with metastasis formation.24–42
VTE is associated with significant morbidity, including hospitalization, reduced pulmonary function, and post-thrombotic syndrome. Compared with patients without cancer, patients with cancer experience increased VTE recurrences and bleeding complications on anticoagulation therapy.7,43 In addition, a venous thrombotic event may impact the chemotherapy schedule and potential future therapeutic choices. The most concerning sequela remains the increased risk of mortality with VTE. Several reports have shown that VTE is a leading cause of death among patients with cancer.44,45 Sorensen et al46 were among the first to define the worse prognosis and reduced survival of patients with cancer with VTE compared with control patients matched by age, sex, cancer type, and year of diagnosis. Chew et al47 evaluated retrospectively the impact of VTE among patients in the California Cancer Registry and found that after adjusting for stage, VTE patients with cancer continue to experience increased mortality. In a subsequent analysis restricted to patients with breast cancer, increased mortality again was observed even after adjusting for comorbid conditions. This appeared most pronounced in early-stage patients and the first few months after VTE diagnosis.48 Based on a prospectively accrued, and closely followed cancer cohort, including more than 4,000 patients with solid tumor and lymphoma starting a new chemotherapy regimen at 115 oncology practices throughout the United States, Kuderer et al49 determined that VTE is an independent risk factor for mortality during the initial months of chemotherapy in patients with cancer of all stages. This remained the case after adjusting for all major risk factors including age, sex, ethnicity, cancer type and stage, chemotherapy type and relative dose intensity, performance status, body mass index, comorbid conditions, and major laboratory abnormalities.49 Additional research is needed to clarify if excess death in VTE patients with cancer is mainly due to more aggressive disease versus due to the occurrence of fatal thromboembolic events.
Evidence of hemostatic activation is a common finding in most patients with cancer.10,12,50,51 Patients with cancer exhibit numerous patient-, disease-, and treatment-related predisposing factors for venous thrombosis. There are many ways that the tumor may contribute to a hypercoagulable state, including acute phase reactions, hemodynamic changes, tissue necrosis, and aberrant protein metabolism.51,52 However, probably the most important factors contributing to the general prothrombotic state in patients with cancer derive from the tumor cells themselves. Their specific prothrombotic properties and their ability to further induce a hypercoagulable environment have been reviewed previously, in particular by Falanga et al51,52 (Fig 1), and will be addressed in further detail elsewhere in this special issue.
Briefly, tissue factor (TF), cancer procoagulant (CP), and to a lesser extent tumor mucins are the best described tumor procoagulants. TF is a transmembrane protein that forms a macromolecular complex with factor VII to activate both factor IX and X.53 CP is mainly found in malignant tissue and is a 68kDa cysteine protease that activates factor X directly, independent of factor VII.54 Both TF and CP are expressed in large numbers in human and animal tumors including leukemic cells, particularly the promyelocytic leukemic subtype.54–62
Most components of the fibrinolytic system are found to be expressed in tumor cells. These include tissue plasminogen activator (t-PA), urokinase plasminogen activator (u-PA), as well as plasminogen activator inhibitor 1 (PAI-1) and 2 (PAI-2).63–65 In addition to their role in hemostasis, fibrinolytic factors participate in the process of tumor invasion necessary for metastasis formation and may contribute to the increased risk of bleeding.55,64,65 Tumor cells also release cytokines, such as tumor necrosis factor (TNF), interleukin-1(IL-1), and vascular endothelial growth factor (VEGF), which in turn promote prothrombotic endothelial changes and angiogenesis.18,66–69 IL-1β and TNF-α together with bacterial lipopolysaccharides (endotoxins) increase the endothelial expression of TF and PAI-1, while downregulating thrombomodulin (TM), with resulting decrease in protein C activation, one of the principal anticoagulant defense systems.18,66,67,69–73 Cytokine release and endotoxins can also rapidly increase t-PA, which tends to be followed by an even larger rise in PAI-1, resulting in an overall procoagulant state.68 At the same time, endothelial activation by IL-1β and TNF-α leads to increases in expression of endothelial adhesion molecules. These enable tumor-endothelial cell binding and likely facilitate tumor cell extravasation and invasion.74–77 In addition, tumor cells have the ability to interact either directly or through soluble mediators with other blood cells, especially monocytes, macrophages, and platelets, which mainly promote thrombosis through the clotting cascade or platelet activation.
Patients presenting with idiopathic venous thrombosis are at increased risk of developing cancer in subsequent years.78–81 Thrombosis may be an early manifestation of an occult tumor. However, there is mounting evidence that the hemostatic system itself contributes to tumor cell survival and cancer progression. Several mechanisms have been implicated, in which various factors of the hemostatic system may aid in cancer cell survival, proliferation, invasion, metastasis formation, and tumor blood vessel formation (Fig 2).21,82,83
Fibrin deposits surrounding tumors and tumor cells may not only protect against immunologic attack,15–17 but also form a necessary matrix or support stroma for tumor tissue,21,22,84,85 and stimulate angiogenesis.86 Next to its dual role in fibrin formation and platelet activation, thrombin appears to contribute in concert with TF to tumor cell invasion and metastasis formation.87,88 Furthermore, thrombin is believed to act as an autocrine growth and proinflammatory stimulant through cleavage and activation of protease activated receptors (PARs), and can induce the expression of c-myc proto-oncogene and TF.21,89–96
In its role as primary initiator of hemostasis, TF results in thrombin and fibrin matrix formation. Furthermore, TF contributes to angiogenesis, tumor cell migration, and tumor progression by various pathways that include clotting-dependent and clotting-independent mechanisms and can result in direct or indirect induction of growth hormones, for example, vascular endothelial growth factor (VEGF).83,97,98 TF and VEGF closely colocalize and correlate with microvessel density, disease progression, and more advanced disease in breast and other cancers.97,99–102 The TF-VIIa complex directly induces tumor production of VEGF, other growth factors, and cytokines through signaling, mainly via PARs as well.103–106 TF expression and activity is an important determinant of tumor metastatic potential and TF inhibition decreases metastasis development in animal models.99,107–110 The TF-VIIa complex inhibits apoptosis and prolongs tumor cell survival through thrombin-independent cellular pathways, including signaling via p44/42 MAPK and Jak/STAT pathways.111,112
Platelets may contribute to tumor cell survival in the bloodstream and overall tumor growth. Tumor cells entering blood vessels quickly recruit platelets to form tumor-platelet aggregates and later microthrombi that also include fibrin and leukocytes.113,114 These tumor microthrombi help cancer cells survive the hostile environment in blood vessels and facilitate lodging at distant sites.113,114 Platelets also contribute to angiogenesis and potentially to tumor growth by secreting a variety of growth and stimulatory factors, including VEGF, platelet-derived growth factor, transforming growth factor, thrombin, and fibrinogen.113–115
P- and L-selectins play a critical role in cancer cell interaction with endothelial cells, platelets, and leukocytes and are important for adhesion and metastasis formation.30,35,82,116–118 Neutrophils and monocytes that are commonly found around tumor blood vessels contribute to metastasis formation by facilitating tumor cell extravasation.119,120 More importantly, knockout mice for P- or L-selectin demonstrate a marked attenuation of metastasis formation, and double-deficient mice have a near abrogation of metastases.30,82,116–118
As early as the 1930s, heparins were reported to interfere with several vital steps of tumor progression and in particular metastasis formation in laboratory animals, including tumor growth, tumor cell motility, migration, invasion, metastases development, and cancer survival.26,121–127 Numerous studies confirmed that heparins prolong survival of laboratory animals after tumor cell inoculation.26–34,36 This antineoplastic effect of heparin likely involves multiple mechanisms, as has been investigated by numerous preclinical studies.
In addition to inhibiting or decreasing fibrin and thrombin formation and their secondary stimulatory effects,21,84,128 heparins have significant antiangiogenic properties.129–136 In vitro studies have shown that LMWHs in contrast to heparin can inhibit the activity of VEGF and basic fibroblast growth factor (bFGF), resulting in reduced endothelial cell growth.132,133,136 This phenomenon appears to be dependent on the specific molecular weight of heparins and was not observed with the pentasaccharide fondaparinux.132,133,136 Heparins also interfere with angiogenesis indirectly by reducing TF expression.137
In most preclinical studies, the antiproliferative effects of LMWH and unfractionated heparin (UFH) did not directly affect primary tumor growth,31,33,34 but rather interfered with metastasis formation.26,28–31,33–36 Heparins' main antiproliferative effect is on noncancer cells, such as endothelial cells, epithelial cells, fibroblasts, and vascular smooth muscles.133,138–140 Heparins also have been implicated in inducing differentiation and apoptosis in in vitro studies.32,37,141
Tumor cell migration and vascular adhesion is an integral part of metastasis development. Heparins intervene in several steps of the metastatic process. For any cellular movement through tissue and endothelial cell layers, the extracellular matrix needs to be modulated and degraded. The cancer cell achieves this with the secretion of heparinases and other proteolytic enzymes. Heparins reduce tumor invasiveness by inhibiting heparinases and other extracellular matrix components.36,142,143 The binding and inhibition of P- and L-selectins by heparin is a key mechanism by which it interferes with crucial cancer cell interactions with endothelium, platelets, and leukocytes, vital for metastasis development as previously discussed.30,35,82,116–118,144
Natural killer (NK) cells are key players in the destruction of circulating tumor cells before lodging at a distant site. The NK cell tumoricidal activity appears to be enhanced by both LMWH and UFH through increased activity of TNF and interferon in mice experiments.145 As mentioned previously, heparin also blocks crucial selectin binding for leukocyte-mediated tumor cell extravasation.82,117
There is less evidence to suggest that vitamin K antagonists exert an antineoplastic effect. Nevertheless, some evidence indicates that warfarin may reduce metastasis formation in part by activating macrophages.146 Warfarin may also have an inhibitory effect on carcinoma-induced and bFGF-associated angiogenesis.100,147
While heparins have been implicated in potentially enhancing the therapeutic effects of surgery and chemotherapy,38,148–153 this has never been conclusively confirmed. Evidence in support of the hypothesis that anticoagulation may improve cancer survival comes otherwise from two major types of clinical studies. The first category was of patients treated for established VTE with different anticoagulants to determine which agent was most effective in preventing recurrent thrombosis. In these trials, survival was either not assessed at all or evaluated as a secondary outcome. In the second category of studies, the primary goal was to study the impact of anticoagulants on cancer survival in patients without a diagnosis of VTE. The results of both types of studies support the hypothesis that the administration of anticoagulants may improve survival independent of thrombosis.
Clinical trials of anticoagulation in patients with documented VTE also included patients with cancer. Subgroup analyses of patients with cancer in these trials have suggested that initial LMWH can improve survival compared with UFH. Prandoni et al154 compared initial treatment of patients with proximal deep vein thrombosis (DVT) with either LMWH or UFH. During the 6 months of follow-up, 44% of the patients with cancer in the UFH group died compared with 7% of patients on LMWH (P = .02). Meta-analyses of this and several subsequent randomized trials of symptomatic VTE demonstrated improved survival in patients with cancer on LMWH compared with UFH.155–158 Hettiarachchi et al157 compared the mortality rates of LMWH versus UFH treatment for DVT in patients with cancer across nine randomized controlled trials in general medical patients with cancer representing 18% of the trial population. Among patients with cancer receiving LMWH, fewer deaths were reported in the first 3 months than in those receiving UFH (odds ratio, 0.61; 95% CI, 0.40 to 0.93). However, it is surprising that such a difference would occur with only brief initial exposure differences to either LMWH or UFH, which were both subsequently followed by the same vitamin K antagonist. Given the post hoc analyses of cancer subgroups in these trials, their results remain hypothesis generating only.
The CLOT study, a large randomized controlled trial of VTE treatment in patients with cancer, was conducted comparing LMWH (dalteparin) for 6 months to dalteparin for a brief initial period followed by a vitamin K antagonist.159 Whereas the primary outcome of this trial was recurrent thrombosis, survival was assessed as a secondary outcome and not found to be improved in the dalteparin arm compared with controls. A post hoc subgroup analysis, focusing only on the patients with nonmetastatic disease, reported the 12-month all-cause mortality at 20% in the dalteparin group compared with 35% in vitamin K antagonist treated patients (P = .04).160 However, in patients with metastatic cancer, no survival difference was observed between dalteparin (72%) and oral anticoagulant (69%) study arms (P = .46).
Several randomized controlled trials have been undertaken to directly study the impact of anticoagulant therapy (LMWH, UFH, and warfarin) on overall survival in patients with cancer without VTE.
In an early anticoagulation survival study, Zacharski et al161 randomly assigned patients with lung, colon, head and neck, and prostate cancer to standard treatment with or without the addition of warfarin for an average of 26 weeks. Significant improvements in time to tumor progression and in overall survival were reported among the 50 patients with small-cell lung cancer, while no difference in survival was observed in other tumor types. In another study, 328 patients with small-cell lung cancer receiving chemotherapy were randomly assigned to concurrent warfarin or not.162 A significant improvement in objective tumor response was observed with a trend toward improved overall survival. A subsequent study of warfarin in 347 patients with limited-stage small-cell lung cancer demonstrated no significant improvement in response rate, disease-free, or overall survival.163 A recent randomized controlled trial of LMWH in 84 patients with small-cell lung cancer demonstrated median progression-free survival of 10 months with LMWH compared with 6 months in chemotherapy control patients (P = .01) with improvements observed in patients with both limited and extensive disease.164 In a multicenter trial of 277 patients with small-cell lung cancer randomly assigned to chemotherapy alone or with UFH for 5 weeks,165 patients receiving UFH experienced improved response rates (P = .04), and median survival (P = .01) with greatest benefit in limited-stage disease. Whereas most studies of anticoagulants as cancer treatment have demonstrated improved survival in small-cell lung cancer, such studies have been limited by small sample size, heterogeneous cancer patient populations potentially leading to an imbalance of important prognostic factors between randomization arms, outdated chemotherapy, and limited data on thromboembolic and bleeding complications.
The impact of LMWH on survival in other tumor types has also been studied in randomized trials. In a study of 385 patients with advanced malignancy receiving standard treatment with or without dalteparin or placebo for 1 year, Kakkar et al166 observed no significant differences in survival up to 3 years. However, in a post hoc landmark analysis of 102 patients still alive at 17 months, significant improvement in survival was found for patients receiving dalteparin. Although the improved outcome in patients with less advanced cancer is consistent with studies discussed above, these results must be interpreted with caution. In another LMWH trial, 302 patients with locally advanced or metastatic solid tumors were randomly assigned to nadroparin or placebo for 6 weeks.167 With a mean follow-up of 1 year, a significant improvement in overall survival was observed (relative risk, 0.75; 95% CI, 0.59 to 0.96; P = .02) among patients receiving nadroparin. Sideras et al168 could not confirm these findings in a trial with 144 advanced solid tumor patients randomly assigned to standard treatment with or without dalteparin, but results were limited by small sample size.
Before 2004, there were several meta-analyses suggesting improved survival in patients with cancer on anticoagulation therapy in the absence of venous thrombosis.169–172 However, results of these studies were confounded by including nonrandomized studies or favorable post hoc subgroup analyses from trials in general medical patients. They also preceded the publication of several more recent trials,164,167,168,173–175 and reporting on bleeding events was limited.
As part of the activities of the American Society of Clinical Oncology Venous Thromboembolism Guidelines Panel, Kuderer et al176 performed a meta-analysis and systematic review of all RCTs of the efficacy and safety of anticoagulant trials (LMWH, UFH, vitamin K antagonists) in the treatment of patients with cancer without venous thrombosis. Across the 11 eligible trials, anticoagulation significantly decreased overall 1-year mortality with a relative risk of 0.905 (95% CI, 0.85 to 0.97; P = .003; Fig 3). For LMWH, the relative risk for mortality was 0.88 (95% CI, 0.79 to 0.98; P = .015), compared with a nonsignificant effect of warfarin, resulting in an absolute risk reduction in mortality of 8% for LMWH. Major bleeding episodes occurred less frequently in patients who received LMWH compared with those who received warfarin (Fig 4), with an absolute risk increase of 1% in bleeding with LMWH therapy compared with 11.5% with warfarin. Although this difference in major bleeding between LMWH and warfarin could be in part a result of differences in study design and trial population, these results are consistent with previous findings.159,177–181 Overall, fatal bleeding was a rare event. Most trials did not report on thrombosis outcomes.176 These findings were confirmed in subsequent reviews, including the Cochrane Collaboration with their separate reports on heparins and vitamin K antagonists.182–184
Mounting evidence suggests that components of the clotting cascade play an integral part in tumor progression, invasion, angiogenesis, and metastasis formation. Furthermore, there are intriguing experimental findings suggesting that especially LMWHs exert an antineoplastic effect through multiple mechanisms. When combined in meta-analyses, available randomized controlled trials of anticoagulation as cancer therapy in patients without VTE suggested favorable effects of LMWH on mortality of patients with cancer. However, definitive large clinical trials remain elusive. Only one relatively small trial with unfractionated heparin has been reported, not allowing any firm conclusions. Numerous additional questions remain regarding the importance of tumor type and stage on treatment efficacy, the impact of prevention of fatal venous and arterial thromboembolic events, optimal anticoagulation therapy, bleeding complications with differing chemotherapy regimens, and potential special patient monitoring needs.
To better elucidate the clinical benefits of anticoagulation therapy in patients with cancer, ongoing and future trials should focus on homogenous patient populations with the same tumor type, extent of disease, and treatment regimen, modeled after other well-designed cancer trials. In addition, these trials should report on standard survival and treatment response outcomes as well as clinically relevant thromboembolic events and bleeding complications. Despite some remaining hurdles, LMWH and related agents hold promise for improving cancer outcomes.
We thank Nancy Thomasson for her invaluable administrative assistance.
Supported by Grants No. NIH 5T32 CA009307-30 from the National Cancer Institute (N.M.K.), 1R01HL095109-01 from the National Heart, Lung and Blood Institute (C.W.F.), U01 HL072289 and U54 HL77878; and U01 DD000014 from the Centers for Disease Control and Prevention (T.L.O.).
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Charles W. Francis, Eisai (C), Boehringer-Ingelheim (C) Stock Ownership: None Honoraria: Charles W. Francis, Eisai Research Funding: Thomas L. Ortel, Eisai; Charles W. Francis, Takeda Expert Testimony: None Other Remuneration: None
Conception and design: Nicole M. Kuderer, Thomas L. Ortel, Charles W. Francis
Administrative support: Thomas L. Ortel, Charles W. Francis
Provision of study materials or patients: Nicole M. Kuderer
Collection and assembly of data: Nicole M. Kuderer, Charles W. Francis
Data analysis and interpretation: Nicole M. Kuderer, Thomas L. Ortel, Charles W. Francis
Manuscript writing: Nicole M. Kuderer, Thomas L. Ortel, Charles W. Francis
Final approval of manuscript: Nicole M. Kuderer, Thomas L. Ortel, Charles W. Francis