Coagulation and cancer: biological and clinical aspects


Anna Falanga, Division of Immunohematology and Transfusion Medicine and “Hemostasis and Thrombosis” Center, Ospedali Riuniti, Largo Barozzi 1, 24128 Bergamo, Italy.
Tel.: +39 035 266540; fax: +39 035 266659.


Summary.  Malignancy affects the hemostatic system and the hemostatic system affects malignancy. In cancer patients there are a number of coagulation abnormalities which provide the background for an increased tendency of these patients to both thrombosis and hemorrhage. The causes of this coagulation impairment rely on general risk factors which are common to other categories of patients, and other factors which are specific to cancer, such as tumor type and disease stage. In addition, data from basic research indicate that the hemostatic components and the cancer biology are interconnected in multiple ways. Notably, while cancer cells are able to activate the coagulation system, the hemostatic factors play a role in tumor progression. This opens the way to the development of bifunctional therapeutic approaches that are both capable of attacking the malignant process and resolving the coagulation impairment. On the other hand, the management of thrombosis and hemorrhages in cancer patients can be different. To approach these problems, some guidelines have been released by prominent international scientific societies. Also actively investigated is the issue of identifying new biomarkers to classify the subjects at a higher risk, thus improving the prevention of thrombohemorrhagic events in these patients. Finally, novel prophylactic and therapeutic approaches are currently under development. This review provides an overview of the hemostatic complications in cancer, together with new insights into the interaction between hemostasis and cancer biology. We also review the assessment of the risk of thrombohemorrhagic events in cancer patients, and the prophylaxis and treatment of such manifestations.


Malignancy is characterized by a derangement of the hemostatic system which predisposes cancer patients to both thrombosis and hemorrhage. Although traditionally described more in patients with solid tumors, thrombotic events are now also recognized as important complications in hematological malignancies, with a thrombotic rate similar to that observed in solid tumors at a high risk of thrombosis. Hemorrhages and uncompensated disseminated intravascular coagulation (DIC) further complicate the spectrum of hemostatic complications in malignancy and can be fatal, as observed in acute leukemia [1].

The close relationship between cancer and thrombosis has been known since 1865, when Armand Trousseau first described a clinical association between thrombosis and a yet undiagnosed cancer. Cancer favors the activation of blood coagulation with the appearance of a hypercoagulable state or chronic DIC in these patients. Abnormalities in one or more coagulation tests are common in cancer patients, even without overt thrombotic and/or hemorrhagic manifestations. The results of laboratory tests demonstrate that a process of fibrin formation and fibrinolysis parallels the development of malignancy, increasingly in those with metastases [2]. Particularly, subtle hemostatic alterations are detected, such as high levels of plasma by-products of clotting reactions (i.e. prothrombin fragment 1+2 [F1+2], fibrinopeptide A [FPA], thrombin–antithrombin complex [TAT] and D-dimer), or an acquired protein C resistance, as well as high levels of circulating microparticles (MP) shed by tumor cells and platelets [3]. On the other hand, hemostatic proteins and reactions interdigitate the process of tumor growth and dissemination [4].

The pathogenesis of blood coagulation activation in cancer is complex and multifactorial. However, a unique feature in malignancy is the role played by the expression of tumor cell-associated clot promoting properties. These properties lead to the activation of the clotting cascade, with the generation of thrombin and fibrin, and the stimulation of platelets, leukocytes and endothelial cells which expose their cellular procoagulant features. Several of these mechanisms can contribute to tumor development and progression [4], particularly MP-enriched prothrombotic and proangiogenic factors are new important players in supporting tumor growth [5]. In the last decade, our knowledge grew after the discovery of a complex scenario in which oncogenic events drive the procoagulant conversion of tumor cells [6].

Although cancer patients can experience both venous and arterial thrombosis, the thrombotic venous occlusions have been studied more extensively. The risk of developing venous thromboembolism (VTE) in this population is increased up to seven-fold compared with the general population [7]. This rate has grown in recent years as a consequence of improved oncology outcomes, the use of more thrombogenic therapy regimens and an aging population. Finally, thrombosis can be the first sign of a malignant disease, preceding the clinical detection of cancer by months or even years [8]. Many factors can increase the thrombotic risk in cancer patients, including general risk factors, such as immobility, old age and surgery, and risk factors specific of cancer, such as the cancer type, advanced disease stage and anti-cancer therapies. To help clinicians in the prevention and management of thrombotic and hemorrhagic events in cancer patients, a number of guidelines have been released from international scientific societies [9,10].

In this review, we provide an overview of the hemostatic complications in cancer and new insights into the interactions between the hemostatic factors and cancer biology. Finally, the assessment of the patient risk factors for thrombohemorrhagic manifestations, and the prevention and treatment of such manifestations will also be summarized.

The spectrum of hemostatic complications in cancer

Although Trousseau’s original report was about venous thrombosis in patients with gastric cancer, now the term ‘Trousseau’s syndrome’ is used to indicate any type of thromboembolic manifestation occurring in cancer. These include arterial and venous thrombosis, non-bacterial thrombotic endocarditis (NBTE), thrombotic microangiopathy (TMA) and veno-occlusive disease (VOD).

In the venous bed, deep venous thrombosis (DVT) of the lower limbs is the most common manifestation, followed by upper-limb DVT, pulmonary embolism (PE), cerebral sinus thrombosis and migratory superficial thrombophlebitis. Large retrospective and prospective population studies show an overall VTE incidence ranging from 0.6% up to 7.8% [11]. This wide range is because of the presence of many and different factors that contribute to the global VTE risk, the most important being cancer type [11]. An unusual high prevalence of splanchnic (i.e. Budd–Chiari syndrome and portal vein thrombosis) and cerebral vein thrombosis is reported among patients with myeloproliferative neoplasms (MPN), and these events are often the presenting feature of the disease, before diagnosis [12].

Data regarding arterial thromboembolic events (ATE) are more limited [13,14]. In cancer patients, a variety of arterial thrombotic syndromes have been reported and the sites most commonly involved are the peripheral blood circulation of the upper and lower extremities and cerebral vessels [15]. The mesenteric vessels, kidney and liver represent other unusual sites of ATE. Nevertheless, ATE incidence in cancer may be estimated around 2–5%, accounting for 10–30% of total thrombotic complications. In a recent retrospective analysis in ambulatory cancer patients receiving chemotherapy, the incidence of symptomatic ATE was 0.27% [14].

In patients with MPN, major thromboses occur in 10–40% of patients, of these up to 70% are as a result of ATE events, including ischemic stroke, acute myocardial infarction and peripheral arterial occlusion [16]. In addition to ATE and VTE, MPN patients may present with perturbances of the microcirculation, leading to erythromelalgia, TIA, visual or hearing transitory defects, recurrent headache and peripheral paresthesia [16]. NBTE is particularly common in MPN but can also be observed in solid tumors. NBTE, detected in 0.9–1.3% of patients dying of cancer, is the cardiac manifestation of systemic hemostatic activation resulting in the formation of platelet and fibrin vegetations on cardiac valves [13]. These vegetations can cause ATE after dislodgement, leading to strokes, splenic infarctions and acute limb ischemia.

Other thrombotic manifestations observed in cancer are TMA and VOD. TMA represents a heterogeneous group of diseases characterized by microangiopathic hemolytic anemia, peripheral thrombocytopenia and organ failure of variable severity. TMA in cancer is a rare but severe complication with a short-term life-threatening prognosis, and can manifest as thrombotic thrombocytopenic purpura (TTP) and hemolytic uremic syndrome (HUS). TMA has been described in association with the use of specific chemotherapeutic agents, particularly mitomycin, gemcitabine and, recently, also with some targeted cancer agents, for example immunotoxins, monoclonal antibodies and tyrosine kinase inhibitors [17]. Cancer-associated TMA displays typical features at presentation, which should alert clinicians of the possibility of an underlying malignancy in a patient with a newly diagnosed TMA. [18].

VOD is a serious liver disease characterized by obstruction of small intrahepatic central venules by microthrombi and fibrin deposition [19], and is observed in approximately 50–60% of allogeneic hematopoietic stem cell transplanted patients. In its severe form, whose incidence varies from 0% to 77% of all VOD cases, this complication is associated with a mortality rate close to 85%, as a consequence of multi-organ failure [20]. Risk factors for VOD include hepatic damage, high-dose chemotherapy drugs, abdominal irradiation, female gender and donor-recipient HLA disparity. The deoxyribonucleic acid derivative defibrotide has proven successful for the prevention and treatment of VOD: the underlying mechanisms of action include a protective role of this drug on the microvascular endothelium, as suggested by in vitro studies [21].

Thrombosis can also represent the earliest clinical manifestation of an occult cancer [22]. Indeed, patients with idiopathic VTE show a four- to seven-fold increased risk of being diagnosed with cancer in the first year after thrombosis, as compared with patients with VTE secondary to known causes, and the risk of cancer is even higher in patients with recurrent thromboembolism and in those with bilateral VTE. Whether extensive screening for occult malignancy in patients with idiopathic VTE is recommendable or not is still being debated [23]. Indeed, only two previous studies have compared limited vs. extensive screening, with no conclusive results [24,25].

Finally, on the other side of the spectrum are the hemorrhagic manifestations. Indeed, in cancer patients, abnormal bleeding represents an important cause of mortality, observed in about 10% of patients with solid tumors and in a higher proportion in patients with hematologic malignancies [26]. Bleeding manifestations include melena, hematuria, hematemesis, hematochezia, hemoptysis, epistaxis, vaginal bleeding or ulcerated skin lesions. Ecchymoses, petechiae or bruising are also described. Hemorrhage may occur as an acute catastrophic event, episodic major bleeds, or an ongoing low-degree emission. Potential causes include thrombocytopenia, decreased synthesis of coagulation factors owing to liver dysfunction or vitamin K deficiency, oral anticoagulation, pre-existing mild coagulation factor deficiencies, congenital von Willebrand disease (VWD), vessel wall erosion, DIC and, rarely, acquired inhibitors against blood clotting factors. In MPN, high platelet counts (i.e. above 1000 × 109 per L) are associated with acquired VWD [27], this results from the adsorption of von Willebrand factor (vWF) multimers onto platelets. A similar mechanism is observed in lymphoproliferative diseases, where vWF multimers bind onto malignant cells [27].

DIC is a syndrome characterized by systemic intravascular activation of coagulation, leading to widespread occurrence of (micro)vascular thrombosis contributing to organ failure. In addition, laboratory assessments show activation of fibrinolysis and non-specific proteolysis systems. The ongoing consumption of clotting factors and platelets, together with hyperfibrinolysis and other proteolytic effects result in bleeding from various sites. Important in this context is the measurement of D-dimer, the lysis product of cross-linked fibrin, which shows hyperfibrinolysis in response to clotting activation and fibrin formation. Severe DIC is particularly associated with acute leukemias, causing severe hemorrhages. Intracerebral and pulmonary hemorrhages are relatively common life-threatening complications in acute promyelocytic leukemia (APL), and are the most frequent cause of early death during induction therapy, although can also occur before APL diagnosis and the start of therapy. APL patients with concomitant risk factors have a higher risk of developing a fatal hemorrhage [10]. In contrast to leukemias, scarce data exist on the impact of DIC in patients with solid tumors. One study that evaluated DIC occurrence in 1117 patients with solid tumors reported a 7% incidence [28]; however, this single study is not sufficient to draw any definitive conclusion.

Finally, acquired hemophilia is a rare but potentially fatal bleeding complication caused by the development of autoantibodies directed against plasma coagulation factors, most frequently FVIII (i.e. acquired hemophilia A). As acquired hemophilia can result in significant morbidity and mortality, the differential diagnosis when evaluating the cancer patient with unexplained bleeding should always be considered [26].

Pathogenesis of coagulation alterations

The pathogenesis of the coagulation system imbalance in cancer is complex and involves multiple factors, both clinical and biological.

Clinical factors

Clinical factors can be grouped in three main categories (Table 1): (i) patient-related characteristics, (ii) cancer-related features and (iii) anti-cancer therapies. The first category comprises many risk factors that are not exclusive of but frequent amongst cancer patients, and include advanced age, prolonged immobility, a prior history of thrombosis, high leukocyte and platelet counts, obesity, immobility, and also comorbid conditions such as heart disease, acute infection and respiratory disease [29–32].

Table 1. Thrombotic risk factors in cancer patients
Patient relatedCancer relatedTreatment related
  1. CVC, central venous catheters.

Older age
Bed rest
Previous thrombosis
Prothrombotic mutations
High leukocyte and platelet counts
Site of cancer: brain, pancreas, kidney, stomach, lung, bladder, gynecologic, hematologic malignancies
Stage of cancer: advanced stage and initial period after diagnosis
Chemo- and hormonal therapy
Anti-angiogenic therapy
Erythropoiesis stimulating agents
Blood transfusions

Considering the second category, large epidemiological studies have recognized malignant brain tumors, hematological malignancies and adenocarcinoma of the pancreas, stomach, ovary, uterus, lung and kidney as having the highest VTE risk [33]. Among hematological malignancies, multiple myeloma, non-Hodgkin’s lymphoma and Hodgkin’s disease showed the highest VTE incidence [34]. Moreover, advanced, metastatic cancer has been shown to be associated with an increased risk of VTE compared with localized tumors [33]. Last, active anticancer treatments, including chemotherapy, hormonal therapy, antiangiogenic agents, combination regimens and surgery have a pro-thrombotic effect [35]. The direct injury of endothelial cells by chemotherapeutic agents, or by tumor-derived products, leading to a loss of antithrombotic properties is thought to play a role in the increased VTE risk.

An important finding is also the elevation in the expression of procoagulant tissue factor (TF) and/or phosphatydilserine (PS) exposure and the release of MPs after treatment caused by different chemotherapeutic agents [36]. The appearance of several novel anti-cancer agents carrying a thrombogenic effect brings this issue to the forefront of cancer medicine.

Biological mechanisms

Beyond clinical factors, biological pathways probably play an important role in the pathogenesis of hemostatic alterations in cancer. Cancer cells can activate the hemostatic system through the expression of procoagulant proteins, exposure of procoagulant lipids, release of inflammatory cytokines and MPs, and adhesion to host vascular cells [37] (Fig. 1). The most characterized tumor procoagulant is TF.

Figure 1.

 Tumor-hemostatic system interactions. Tumor cells activate the hemostatic system in multiple ways. Tumor cells may release procoagulant tissue factor, cancer procoagulant and microparticles (MP) that can directly activate the coagulation cascade. Tumor cells may also activate the host’s hemostatic cells (endothelial cells and platelets), by either release of soluble factors or by direct adhesive contact, thus further enhancing clotting activation.

Constitutively expressed on the malignant cell surface, TF can lead to the formation of both localized, as well as systemic procoagulant states. TF activity on tumor cells can be potentiated by the expression of anionic phospholipids (i.e. PS) on the outer leaflet of the cell membrane [38,39] and the secretion of heparanase. The main function of heparanase is to degrade heparan sulfates of the extracellular matrix, thereby promoting tumor invasion and metastasis. However, heparanase can also interact with the TF pathway inhibitor (TFPI) on the cell surface, leading to dissociation of TFPI from the cell membrane of endothelial and tumor cells which results in an increased cell surface TF activity [40].

Another tumor procoagulant is cancer procoagulant (CP) that, unlike TF, directly activates FX independently of FVII. CP has been detected in various tumor cells, and in amnion-chorion tissues but not in normally differentiated cells. In patients with APL, CP, expressed by bone marrow blast cells at the onset of disease, disappears when remission is reached [37]. CP has also been studied also in patients with different types of solid tumors [41,42].

TF can be actively released by tumor cells in a cell membrane-associated form represented by TF-bearing MPs [43]. MPs are plasma membrane vesicles of 0.1–1 μm in diameter produced by active vesiculation of virtually all type of cells [5]. PS expressed on MP surfaces provides a suitable anionic phospholipid surface for the assembly of tenase and prothrombinase complexes, thereby promoting the coagulation cascade. This capacity can be further enhanced by the concomitant expression of TF (Fig. 2). Low plasma MP levels are present in healthy subjects, the majority being of a platelet origin (> 80%), but in pathological conditions an overall increment in MP occurs and a significant amount of MP of other vascular sources, including tumor cells, can be detected. Elevated levels of MPs (bearing or not TF) have been described in patients with both solid [5] and hematological malignancies [44–46]. MPs of a platelet origin were found higher in stage IV vs. I and II/III gastric cancer, showing the highest diagnostic accuracy for metastasis prediction [47]. Finally, the pathogenetic role of MPs in cancer-associated thrombosis has been demonstrated by the development of a DIC-like syndrome in mice after intravenous injection of highly TF-positive MPs of a tumor origin [5].

Figure 2.

 Microparticle (MP) production and activities in cancer. Tumor cells actively release MP but also promote MP formation by platelets. Tissue factor (TF) and phosphatidylserine (PS) expression on the surfaces of both platelet- and tumor-derived MP are involved in blood clotting activation and thrombus formation. On the other hand, the elevated content of proangiogenic factors in platelet-derived MP (VEGF, vascular endothelial growth factor, FGF, fibroblast growth factor, PDGF, platelet-derived growth factor), render these elements also important mediators of the neangiogenesis process. Finally, intracellular transfer of MP may occur between cancer cells, leading to a horizontal propagation of oncogenes and amplification of their angiogenic phenotype.

Tumor cells are also capable of interacting with the host fibrinolytic system, owing to the expression of plasminogen activators (uPA and t-PA), their inhibitors (PAI-1 and PAI-2) and receptors such as uPAR [37], and of annexin II, a co-receptor for plasminogen and tissue plasminogen activator (tPA). In APL, the increased annexin II expression has been linked to an excessive activation of fibrinolysis [48]. Likely, depending on which side, pro- or anti-fibrinolytic, prevails, the clinical manifestations of fibrinolysis system alterations may be quite different, from bleeding symptoms as observed in leukemia, to VTE evidenced in solid tumors.

Furthermore, tumor cells release a variety of soluble pro-inflammatory (i.e. tumor necrosis factor-alpha [TNF-α] and interleukin-1beta [IL-1β]) and proangiogenic (i.e. vascular endothelial growth factor [VEGF] and basic fibroblast growth factor [bFGF]) factors [37], which can stimulate the prothrombotic features of vascular cells. In endothelial cells, these molecules induce TF expression, stimulate PAI-1 production, downregulate thrombomodulin and promote cell-adhesion molecules upregulation [37]. The capacity of tumor cells to adhere to vascular endothelium favors localized blood clotting activation and thrombus formation.

In the last decade, molecular studies of experimental models of human cancer have demonstrated that oncogene and repressor gene-mediated neoplastic transformation activate clotting as an integral feature of neoplastic transformation [49]. Genome-wide expression profiling of mouse hepatocytes expressing the MET oncogene have demonstrated a significant upregulation of PAI-1 and cyclooxygenase-2 (COX-2). The inactivation of the tumor suppressor gene PTEN was linked to TF gene induction in human astrocytoma cells. In non-small cell lung cancer specimens, TF mRNA expression was found to be increased in samples presenting either tumor protein p53 (TP53) and PTEN mutations; the concomitant presence of mutations in three tumor suppressor genes further dramatically increased TF expression [50]. Finally, in a colorectal cancer model, activation of K-ras and inactivation of p53 resulted in increased tumor cell TF expression and enhanced release of TF-bearing MPs [49].

Role of hemostatic proteins in tumor progression

Coagulation activation and tumor progression are closely linked. Tumor growth and aggressiveness largely rely on cancer cell capacity to promote neoangiogenesis and metastasis [37]. In these processes, a contribution of different components of the hemostatic system, including thrombin, TF and FVIIa, FXa, fibrinogen and vascular cells, has been clearly documented by both in vitro and in vivo tumor models. Mechanisms described include both clotting-dependent activities of these molecules, culminating in platelet activation and fibrin deposition, as well as clotting-independent properties.

Clotting-dependent mechanisms

Fibrin deposited within the tumor vasculature facilitates angiogenesis by providing an excellent scaffold for new vessel formation. Furthermore, this fibrin matrix binds and sequesters a number of growth factors (bFGF, VEGF and insulin-like growth factor-1) and protects them from proteolytic degradation [37]. Localized fibrin deposition favors the metastatic process by stabilizing tumor cell adhesion or tumor cell-associated emboli to the endothelium.

However, the fibrin matrix present in tumors is in a dynamic state, undergoing simultaneous deposition and dissolution by the fibrinolytic system to be replaced by mature connective tissue stroma. This highlights a crucial role of fibrinolysis in tumor biology and the relevance of fibrinolytic protein expression by neoplastic cells in promoting cell mobility and motility. Although fibrinogen deficiency does not impair growth and angiogenesis of the primary tumor in animal models, it significantly reduces lung metastases, probably owing to diminished adhesion and stability of metastatic cells [51].

Independently from thrombin, platelets can be directly activated by tumor cells through the release of pro-aggregating substances or adhesion mechanisms. For example, the binding of podoplanin, expressed on tumor cells’ surfaces, to CLEC-2, expressed by platelets, induces platelet activation and aggregation [52]. The formation of tumor cell-platelet thrombi may support metastasis formation by preventing interactions between tumor and innate immune cells [53].

Clotting-independent mechanisms

Thrombin and TF participate to tumor progression through clotting-independent mechanisms, as a result of interactions with specific receptors belonging to the family of protease-activated receptors (PAR) expressed by platelets, tumor cells, endothelial cells, vascular smooth muscle cells and macrophages [37]. Thrombin cleaved-PAR-1 stimulates growth factors, chemokines, and extracellular proteins release that promote tumor cell proliferation and migration [54]. In endothelial cells, thrombin upregulates many angiogenesis-related genes, including VEGF, VEGFR, bFGF and metalloproteinase (MMP)-2. Both thrombin and FXa can induce the angiogenesis-promoting gene Cyr61 and connective tissue growth factor (CTGF) expression [55]. Finally, thrombin-activated platelets become highly proangiogenic by releasing proangiogenic factors from their granule contents, including VEGF [56] and platelet-derived growth factor (PDGF) [57]. The pro-metastatic activity of thrombin has been demonstrated by experimental models, in which lung metastasis was increased by pre-treatment of tumor cells with thrombin [37].

Binding of FVII to TF generates downstream signaling cascades that promote increased endothelial cell adhesion and migration. Furthermore, TF in complex with FVIIa and FXa activates one or more PARs, including endothelial PAR-2 to support angiogenesis in vivo [58]. Recent studies have focused on a possible role of the alternatively spliced TF (asTF), a soluble isoform of circulating TF, in cancer progression [59]. The suggestive hypothesis coming from research is that both asTF and flTF contribute to tumor progression, but through different ways: the first acting directly on the proliferative tumor side, and the second on the angiogenic side with the contribution of FVIIa and PAR-2.

Platelet MPs display high proangiogenic activity owing to their unique content of angiogenesis-stimulating agents derived from platelet alpha-granules, together with the expression of adhesion molecules, and the capacity to induce pro-angiogenic factor release by tumor cells [5]. TF expression by MPs might represent another important mechanism involving MPs in tumor progression [5]. Moreover, intracellular transfer of MPs may occur between cancer cells, leading to a horizontal propagation of oncogenes and their associated transforming phenotype [60] (Fig. 2).

Finally, leucocytes have been involved in tumor growth and metastasis [37]. Neutrophils, activated in the tumor microenviroment by locally generated inflammatory mediators (i.e. TNF-α), can adhere to tumor cells and facilitate their migration across the endothelial barrier. By releasing metalloproteinases, they can hydrolyze endothelial cell matrix components and favor cancer cell motility [61].

Anticoagulation and tumor progression

The above evidence taken together suggests that reducing coagulation activation in cancer patients may not only prevent hemostatic complications but also improve cancer survival [62]. Indeed, clinical trials with anticoagulant drugs were started in the early 1980s to test the effects of anticoagulation on cancer survival. A recent meta-analysis [63] overall suggested a favorable impact of anticoagulation on the survival of cancer patients, particularly with low-molecular-weight heparins (LMWH). In recent years, prospective randomized controlled trials (RCTs) have been designed to address as a primary endpoint the survival of cancer patients receiving LMWH, but the results available so far have not provided clear conclusions yet [64,65]. Ongoing studies on VTE prevention in cancer patients incorporate as a secondary end-point the impact of LMWH on cancer survival [66–68]. These results will be available in the next few years.

Several previous in vitro and in vivo studies have demonstrated that the potential beneficial effects of heparins on survival may rely on some direct anticancer properties. These include: inhibition of tumor-induced angiogenesis, tumor cell heparanase, cell surface selectin-mediated tumor cell metastasis and blood coagulation activation [62].

The impact of natural anticoagulant proteins, such as activated protein C and thrombomodulin, on tumor biology is under active investigation as well [69,70].

Prevention and treatment of thrombosis and bleeding


With regard to arterial thrombosis, cancer patients share the same general risk factors with non-cancer patients (i.e. arterial hypertension, diabetes, dislipidemia and obesity); additional risk factors are co-morbidities including pulmonary disease, renal disease, infection, blood transfusion and chemotherapy. Older age (> 60 years) and previous thrombosis are well-established cardiovascular risk factors for thrombosis in MPN which identify the so-called high-risk patients. There is great deal of attention in moving beyond these recognized risk factors, particularly in the young or asymptomatic low- and intermediate-risk individuals at risk of thrombosis. Recently, the impact of new risk factors, such as leukocytosis and JAK2V617F mutational status and/or mutational burden, is under active investigation [71]. Today, no guidelines or recommendations are available for ATE prophylaxis in the cancer setting. An exception is represented by MPN, in which aspirin, in association or not with hydroxyurea, anagrelide and/or phlebotomy, has proven useful to significantly reduce the risk of cardiovascular events [72].

More certainty is available for VTE primary prophylaxis. Thromboprophylaxis with either unfractionated heparin or LMWH has been shown to be safe and effective in high-risk settings such as hospitalization for medical illness and the postsurgical period. Current European and American guidelines found a good level of consensus on VTE prevention with LMWH in the surgical oncology setting [73–77], particularly in cancer patients undergoing ‘high-risk’ major abdominal or pelvic surgery [78,79]. However, the duration of post-surgical prophylaxis is still being debated.

The international consensus is lower in the medical setting, in which two high-risk situations can be identified: the first involves patients hospitalized or bedridden for prolonged periods of time, and the second relates to ambulatory patients receiving chemotherapy or radiation. Cancer patients hospitalized for an acute medical illness are at risk for thrombosis. They should be considered for thromboprophylaxis with prophylactic doses of LMWH or fondaparinux, but the duration of thromboprophylaxis is yet undefined. Moreover, to date, no evidences exist that the highest risk patients should receive adjusted drug doses for prophylaxis.

Most VTE occur in the outpatients setting; however, the guideline panels agree on not recommending routine thromboprophylaxis in ambulatory cancer patients. Nevertheless, in recent years, RCTs have been conducted to evaluate the impact of thromboprophylaxis in ambulatory patients with solid tumors receiving systemic chemotherapy [80–82]. Overall, these studies suggest that outpatient thromboprophylaxis is feasible, safe and effective. Notably, the VTE low rates observed in these studies suggest the importance of patient selection and disagree on a wide application of prophylaxis. In line with this, the most recent guidelines of the National Comprehensive Cancer Network emphasize the need for VTE risk assessment in ambulatory cancer patients and the need of RCTs for patients with a favorable risk–benefit ratio [83]. Other areas of uncertainty include prophylaxis in patients who receive chemo-, radio- or hormone therapy and have a history of VTE, patients with cerebral cancer, patients undergoing surgery other than abdominal or pelvic procedures, or patients undergoing laparoscopy procedures lasting more than 30 min [79].

Unlike solid tumor experience, no ad hoc studies or guidelines are available to help clinicians with best practices for prophylaxis of VTE in hematologic malignancies [84]. In patients with acute leukemia and lymphoma, some information on thromboprophylaxis comes from two studies [85,86], while more data are available in multiple myeloma [87]. Given the higher VTE risk during treatment of multiple myeloma patients with thalidomide or lenalidomide in combination with dexamethasone or multi-agent chemotherapy, thromboprophylaxis is recommended. However, which is the best prophylactic scheme in this setting is an issue of debate. The results of prospective randomized trials of GIMEMA studies, comparing the efficacy of LMWH, warfarin (fixed low-dose or full dose) and aspirin for prophylactic anticoagulation, showed only a trend for a more effective thromboprophylaxis with LMWH [88,89]. The guidelines of the American Society of Clinical Oncology (ASCO), the European Society of Medical Oncology (ESMO) and the Italian Society for Haemostasis and Thrombosis (SISET) recommend prophylaxis with LMWH or adjusted-dose warfarin (International Normalized Ratio 2-3) [79,90,91]. The International Myeloma Working Group proposes different thromboprophylactic strategies based on stratification of the patient’s risk of VTE [92].

Current research is focusing on the design of VTE risk models to identify high-risk patients who might benefit from primary thromboprophylaxis. The score-model proposed by Khorana et al. is based on five simple predictive clinical and laboratory parameters in cancer patients: cancer site, platelet count, hemoglobin level or the use of erythropoiesis-stimulating agents, leukocyte count and body mass index. This model has been shown to predict the short-term risk of symptomatic VTE accurately in patients undergoing chemotherapy [93]. Recently, the inclusion of P-selectin and D-Dimer quantification to this model improved the risk stratification strategy [94]. Emerging biomarkers are represented by plasma TF and MPs [95].

Treatment of thrombosis

The treatment of ATE in cancer patients relies on anti-platelet and anticoagulant/fibrinolytic agents according to the same protocols recommended for secondary prophylaxis for stroke and myocardial infarction in the non-cancer population. However, specific protocols have been developed for acute VTE treatment, which has replaced the traditional regimens based on initial therapy with UFH, LMWH or fondaparinux followed by long-term therapy with a vitamin K antagonist (VKA). Data from various RCTs, comparing LMWH with VKA in the long-term VTE therapy in cancer [96–99], show the superiority of LMWH monotherapy, which is now endorsed by international guidelines [90,100]. For patients who develop a recurrence while on LMWH, dose escalation of LMWH is often effective, whereas for patients who develop a recurrence on VKA therapy, the recommended practice is to switch to LMWH. Raising the intensity of VKA therapy is not recommended because of the potential for increasing bleeding [101,102]. There are limited data on the use, safety and long-term outcome of vena cava filters. Today, the use of these devices can be considered as an alternative to prevent PE (as filters are not effective in reducing DVT risk) only in patients who have a contraindication to anticoagulation [103].

Novel oral anticoagulants (i.e. dabigatran, rivaroxaban and apixaban) may change the therapeutic scenario in patients with cancer. These agents, which achieve rapid inhibition of activated factor X or thrombin, may offer an easier solution than LMWH but studies focusing on treatment of cancer-associated thrombosis with these agents are lacking. To date, some of these agents have shown comparable efficacy and safety compared with traditional anticoagulants in RCTs that included primarily patients without cancer [104,105]. Given the higher risk of recurrent thrombosis and bleeding in cancer patients, further research is needed to understand the antithrombotic impact of these new agents in this setting [106].

Prophylaxis and treatment of bleeding

On the bleeding side, the most important issue is represented by the prophylaxis and treatment of the fatal hemorrhagic syndrome in APL. However, the management of this syndrome is particularly difficult. Recent recommendations indicate that three simultaneous actions must be immediately undertaken when a diagnosis of APL is suspected: the start of ATRA therapy, the administration of supportive care with transfusions of plasma and platelets, and the confirmation of molecular diagnosis [107,108]. According to current recommendations, a prophylactic platelet transfusion is an essential part of supportive care, with the aim to maintain a platelet count above 20 × 109 per L in non-bleeding patients and above 50 × 109 per L in those with active bleeding [71,84]. The role of heparin in the treatment of the coagulopathy is undefined. No systematic studies have been reported on the use of LMWH or any of the newer anticoagulants (i.e. FXa and IIa inhibitors, hirudin, fondaparinux) to treat the thrombohemorrhagic syndrome of APL. Other types of therapeutic regimens, including antifibrinolytic agents or protease inhibitors (i.e. aprotinin), have been suggested, but no data from RCTs are available. Interestingly, the occurrence of thromboembolic events was reported during antifibrinolytic agent administration in combination with ATRA. A lack of efficacy of tranexamic acid on the hemorrhage-associated mortality in APL was shown in the large PETHEMA trial [109].


Cancer disease is associated with a high risk of both thrombotic and hemorrhagic complications. Preventing these complications is clinically relevant because they significantly contribute to the morbidity and mortality of these patients. The efforts of the research have significantly elucidated the clinical and the biological mechanisms underlying the hemostatic derangement caused by cancer. Interestingly, direct oncogene activation and/or tumor suppressor gene impairment, may result in the stimulation of blood clotting and/or suppression of fibrinolysis, which can produce thrombosis and/or DIC in some in vivo models.

Some of the factors involved in the multiple interconnections existing between the hemostatic system and cancer biology may prove useful for establishing risk assessment models, not only for thrombosis but also for cancer prognosis. The hope is that our accrued knowledge in the field will help us to establish more accurate and appropriate interventions for the management of hemostatic complications. Meanwhile, we can rely on the guidelines issued by the relevant international scientific societies.


The authors wish to thank Associazione Italiana per la Ricerca sul Cancro (A.I.R.C., grants IG10558 and ‘5 per mille’ 12237) and Regione Lombardia (grant ‘Piano Regionale Sangue 2010–2012’) for their support.

Disclosure of Conflict of Interest

The authors state that they have no conflicts of interest.