Breast Cancer as an Acquired Thrombophilic State

Pathophysiology of cancer-related thrombosis

In a normal coagulation-fibrinolysis system, there is a natural balance between activation and inhibition of procoagulants and anticoagulants. Cancer cells can alter this balance by mechanisms such as the production of procoagulant and fibrinolytic agents, the release of proinflammatory and proangiogenetic cytokines, and direct interaction with the vessels and blood cells through adhesion molecules [9].

The most studied cancer-related procoagulants are tissue factor (TF) and cancer procoagulant (CP). TF is a transmembrane glycoprotein and the primer of physiological coagulation cascade. It is produced by monocytes and vascular endothelium and; is active both in the cell surface and as soluble plasma component. Elevated levels of TF are detected in tumor cells through the activation of oncogenes (e.g., k-ras, MET) and/or the failure of tumor suppressor genes (e.g., p53); they may also be detected in endothelial cells surrounding the tumor, due to secretion of interleukin 1b (IL-1b) and tumor necrosis factor alpha (TNFα) by tumor cells. Genetic alterations related to neoplastic transformation, such as TF, cyclooxygenase 2 (COX-2) and plasminogen activator inhibitor-1 (PAI-1) [10], have been shown to be involved in regulating genes that control hemostasis.

CP, a cysteine protease, is a direct activator of factor X and is encountered only in malignant tissues. CP, with the presence of factor V, can enhance up to three times the production of thrombin. Elevated levels of CP have been reported in patients with acute promyelocytic leukemia, melanoma, breast, lung, and kidney cancer; however, there is in vitro evidence that it is implicated in the metastatic process [11].

Cancer cells can express fibrinolytic proteins, namely tissue plasminogen activator (tPA), urokinase plasminogen activator (uPA), uPA receptor (uPAR) and inhibitors of plasminogen activators-1,-2 (PAI-1, PAI-2). Recent data suggest that these fibrinolytic factors are involved in the development and spread of tumor [12]. Proinflammatory cytokines, such as TNFα and IL-1b, secreted by cancer cells, induce TF expression from monocytes and endothelial cells, reduce the expression of thrombomodulin, a key factor in the protein C pathway [9], suppress the fibrinolytic activity of endothelium and increase the endothelial production of e-selectin and von Willebrand factor (vWF) [11].

The production of vascular endothelial growth factor (VEGF) by cancer cells plays an important role in angiogenesis. VEGF has chemotactic properties for macrophages and induces the expression of TF from monocytes and endothelium. In turn, the expression of TF enhances the expression of VEGF from cancer cells. This correlation between VEGF and TF explains the close relationship between thrombosis, inflammation, growth and metastasis of tumor [13].

Cancer cells have on their surface adhesion molecules that allow their direct interaction with host cells, such as platelets, endothelial cells and leukocytes. Malignant cells that bind to the vascular wall likely play a role in the local induction of coagulation and clot formation, while released cytokines promote leukocyte and platelet adhesion. Fibrin and activated platelets overlap cancer cells and protect them from natural killer cells (NK cells), enhancing their survival [9].

Microparticles are small membrane sections, 0.1 to 1.0 µm in diameter, released from various cells (platelets, leukocytes, endothelial cells, tumor cells) under conditions of apoptosis and cellular activation. They also have hemostatic action due to the expression of TF, phosphatidyloserin, p-selectin and its ligand (PSGL-1), in addition; to the binding capacities of factors V, VIII, IX, X and prothrombinase complex on them. Elevated platelet microparticles have been described in gastrointestinal and breast cancer. They are involved in the induction of angiogenesis, tumor metastasis, and thrombosis through their phospholipids and TF, while a cancer cell itself may also have particles with a procoagulant phenotype [14, 15].

Surgery-related thrombosis in breast cancer

The impact of VTE on women undergoing surgery for breast cancer has not yet been determined. In a study by Andtbacka et al. [16], overall incidence of VTE was 0.16%, per surgery, within 60 days after a breast operation. This incidence was lower than the figures published by European Organization for Research and Treatment of Cancer (EORTC), where the incidence, including cases of superficial thrombophlebitis, reached 0.8% within 6 weeks after surgery. Without those cases the incidence was 0.3% [17].

Clahsen et al. [17] reported an increased incidence of VTE in women undergoing mastectomy and postmenopausal women (2.3% and 2%, respectively). Other studies found an increasing trend of VTE incidents in surgical procedures of longer duration and complexity, but the increase was not always statistically significant. In most studies, the incidence of VTE after breast surgery increased in advanced stages of disease [16, 17].

The damage of vascular endothelium is a mechanism that triggers the development of prothrombotic phenotype, and vascular lesion may be caused by the mechanical local destruction of the vascular wall by the tumor itself, or during surgical interventions, especially in combination with lymph node dissection [18].

Cytotoxic therapy-related thrombosis in breast cancer

In many breast cancer patients, treated with various combined chemotherapy regimens, based on cyclophosphamide or methotrexate or 5-flourouracil, incidence of VTE during treatment was 5% to 7%, and there was no significant difference between regimens [18]. In contrast, during the 2,413-month treatment-free follow-up, no incident of VTE was reported. There is no correlation between treatment-associated VTE and expression of hormonal receptors, patient's age, number of affected lymph nodes, VTE recurrence, and previous history of VTE. The incidence of VTE increased to 18% in women taking cyclophosphamide, methotrexate, 5-flourouracil, vincristine, prednisolone (CMFVP) regimen for stage IV breast cancer [19]. In a randomized study, patients with early stage breast cancer, who received chemotherapy during the perioperative period, had higher risk of VTE within 6 weeks after surgery, compared with patients who did not receive chemotherapy (2.1% vs. 0.6%) [17].

Wall et al. [20] recorded a 1.3% incidence of arterial thrombosis, either cerebral or peripheral, in a series of 1,014 women with stage II and III disease, during chemotherapy. A retrospective study of 2,352 patients showed a significant increase in arterial thrombosis in premenopausal women treated with combined chemotherapy, whereas risk was significantly limited in postmenopausal women treated with either tamoxifen alone or in combination with chemotherapy [21].

The underlying pathophysiology mechanism of thrombotic complications in women with breast cancer, receiving chemotherapy, has not been fully elucidated. The destruction of cells by chemotherapy and radiotherapy causes the release of procagulant agents and direct damage to vascular endothelium. In breast cancer patients receiving the regimen cyclophosphamide, methotrexate, 5-fluorouracil (CMF), a significant reduction of protein C, protein S, fibrinogen and factor VII has been observed [22]. The levels of protein C returned to normal after completion of chemotherapy. Other chemotherapy agents can induce the direct expression of TF in macrophages and monocytes, or cause disruption to the vitamin K cycle, reduce protein synthesis in the liver due to inhibition of DNA and RNA in cells and provoke DIC [18]. Angiogenic factors, such as thalidomide, lenalidomide and bevacizumab, have also been associated with an increased risk of VTE, especially in combination with other chemotherapy drugs [9].

However, recent in vitro studies report that the antineoplastic drug paclitaxel inhibits platelet aggregation induced by collagen, has antiplatelet activity through inhibition of thromboxane A2 synthase, and reduces the expression of TF in breast cancer cell lines [23, 24].

Hormone therapy-related thrombosis in breast cancer

Tamoxifen, either alone or in combination with chemotherapy, is associated with increased risk of thrombosis. In a randomized study of 1,312 patients [25], those receiving adjuvant treatment with tamoxifen have a 2.5-fold higher risk of thrombosis. In a study of Eastern Cooperative Oncology Group (EGOG), the premenopausal women receiving chemotherapy and tamoxifen had significantly more thrombotic complications than those receiving only chemotherapy (2.8% vs. 0.8%, p=0.03). Similarly, postmenopausal women undergoing chemotherapy with tamoxifen had more episodes of VTE compared with those receiving tamoxifen alone (8% vs. 0.4%, p=0.0001). The incidence of arterial thrombosis was also significantly higher in women undergoing chemotherapy with tamoxifen, compared with those receiving chemotherapy alone (1.6% vs. 0%, p=0.004) [21].

There are a number of studies regarding the prothrombotic effects of tamoxifen. Some studies have found a mild reduction in levels of antithrombin (AT) and protein C [18]. Furthermore, some studies have shown that the risk of thrombosis during treatment with tamoxifen is higher in women with FV Leiden mutation and resistance to activated protein C (APCR) [26]. In a small number of patients with breast cancer, there was a decrease of tissue factor pathway inhibitor (TFPI), an increase of APCR, platelet (CD41), and leukocyte (CD45) microparticles during treatment with tamoxifen [27].

Medroxyprogesterone is used in women with metastatic breast cancer, and several studies have demonstrated the prothrombotic action of the agent. Medroxyprogesterone induces the activation of coagulation, as indicated by measurements of fibrin and thrombin products. The levels of AT, D-dimers, thrombin-antithrombin (TAT) complexes, factor II, factor IX, and plasminogen remained elevated throughout treatment with medroxyprogesterone [18].

Aromatase inhibitors inhibit the synthesis of estrogens from androgens. The risk of VTE during monotherapy with anastrazol is significantly lower, compared with tamoxifen monotherapy (2.1% vs. 3.5%, p=0.0006). The risk of thrombosis significantly increases in a case of combination of two factors, compared with tamoxifen monotherapy [28].

Venous catheter-related thrombosis in breast cancer

Several types of central venous catheters (CVC) and venous implants are used in cancer patients, facilitating the administration of drugs and their mobility. The incidence of catheter-related thrombotic events varies greatly among both symptomatic (0.3-28.3%) and asymptomatic (27-65%) patients. Recent studies have reported a reduction of catheter-related VTE, which can be attributed to improved care methods of central venous lines [29].

Catheters can cause endothelial damage and might impede blood flow, while the fibrin sheath generates in about 87% of catheters, as shown on venography within 24 hours of catheter placing. The sheath is then colonized by cocci, but this is not always indicative of subsequent thrombosis [30]. The incidence of intralumen thrombosis ranges between 13.2% and 93% with a frequency of 0.81 per 1,000 catheter-day [31].

In a study of 300 breast cancer patients [32] with CVC, the incidence of catheter-related thrombotic complications was estimated at 8.3%. All cases were episodes of DVT, although about 6% of cancer patients with arm DVT develop symptomatic PE; asymptomatic PE can be diagnosed by scintigraphy in 11% of cases. Patients with the FV Leiden mutation had a six-fold greater risk of DVT. Furthermore, the study showed significantly higher incidence of DVT in the cohort with advanced metastatic disease, a finding which not confirmed in other studies. Catheter-related thrombosis is a serious complication causing morbidity and delaying anticancer therapy. In a study by Mandalà et al. [32], the median number of chemotherapy cycles in patients experiencing thrombotic complication was significantly lower than in those without DVT (p<0.001).

Thromboprophylaxis in breast cancer

There are two main forms of perioperative thromboprophylaxis: 1) mechanical, which involves calf stimulation, intermittent pneumatic compression devices, graduated static compression stockings, and venous foot pump devices; and 2) pharmacologic, including unfractionated heparin (UFH), low molecular weight heparin (LMWH) and oral anticoagulants. The administration of thromboprophylaxis in breast cancer increases the risk of postoperative hematoma and bleeding. In particular, 11% to 18% of women receiving LMWH, while only 7% of patients under mechanical thromboprophylaxis, present with hemorrhagic complications [16].

Given the relative low incidence of VTE in breast cancer, most studies suggest a mechanical perioperative prophylaxis. In addition to early mobilization, the use of tight socks should not be omitted, even in low-risk surgery; it should be continued throughout hospital stay. It is recommended that the discontinuation of tamoxifen either before any surgery, or in immobilized patients until full mobilization [18].

According to a randomized controlled trial, thromboprophylaxis with low dose warfarin (international normalized ratio [INR] range, 1.3-1.9) was effective and safe in reducing the incidence of thrombosis in women, receiving chemotherapy, with stage IV metastatic breast cancer. However, the need for laboratory monitoring of patients receiving warfarin has not allowed its use on a large scale [3]. Systemic thromboprophylaxis is not recommended for all patients receiving cytotoxic or hormonal therapy. Generally, thromboprophylaxis is recommended 1) during chemotherapy after mastectomy in postmenopausal women with increased BMI (>25 kg/m²) and preexisting coagulation disorders, 2) in the presence of long-term CVC, 3) in congenital thrombophilia (e.g., FVLeiden, AT deficiency), 4) in positive VTE history, 5) in a combination of above states.

The anticoagulant modality includes LMWH or low-dose warfarin, which is begun 3 to 4 days before chemotherapy and is continued throughout cytotoxic therapy [18]. There is relatively little data about the incidence of thromboembolism in patients undergoing radiotherapy. The radiation can lead to endothelium damage, which in turn may activate the platelets and the clotting mechanism. Thus, in high risk patients, the use of thromboprophylaxis should be discussed.

In cases where the use of tamoxifen is contraindicated due to its thrombogenic propetries, administration of third generation aromatase inhibitor, such as anastrozole, should be considered; the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial resulted in a significant reduction of venous thrombosis [28]. Currently, however, all pharmaceutical agents used in endocrine therapy or chemoprevention of breast cancer, including aromatase inhibitors, correlate with an increased risk of thromboembolic events.

The use of thromboprophylaxis in patients with CVC has not been established and is not applicable in practice. Thrombus formation can not be avoided despite regular care of the catheter with heparin. Fibrinolytic agents, such as streptokinase and urokinase, can be used to dissolve the occlusion of catheters [29].

Treatment of acute cancer-related thromboembolism

The standard treatment regimen for thrombotic events consists of an initial phase with UFH or LMWH, followed by a long-term phase with vitamin K antagonists (VKA). This two-phase treatment approach provides an immediate anti-coagulant effect while waiting for VKA action. Multiple randomized trials have confirmed that LMWHs are at least as efficacious as UFH in reducing recurrent thrombosis, bleeding complications and 3-month mortality; they also appeared to have an advantage against coumarin anticoagulants in cancer patients [37].

In a cancer patient population, warfarin is associated with increased risk of bleeding and recurrent VTE. Moreover, several other problems, such as thrombocytopenia, medication, invasive procedures, infections, malnutrition, and hepatic dysfunction, may influence the anticoagulant therapy and lead to closer coagulation monitoring; often, it leads to treatment discontinuation. In the CLOT trial [38], 602 cancer patients with documented VTE were randomly assigned to receive either deltaparin or deltaparin followed by VKA, for 6 months. Recurrent VTE occurred in 9% of patients in the deltaparin group, compared to 17% in the second group.

These data have established LMWH, for at least 3 to 6 months, as the standard care for cancer-related thrombosis, as recommended by the American Society of Clinical Oncology (ASCO), National Comprehensive Cancer Network (NCCN), European Society of Medical Oncology (ESMO), and other guidelines [39]. The optimal duration of VTE treatment in breast cancer patients has not been accurately determined. Generally, heparin is administered in cases of active malignant disease and throughout cancer treatment [16].

In symptomatic patients with acute CVC-related thrombosis, such as limb pain and swelling, it is often necessary to administer long-term anticoagulation therapy. Catheter removal is not necessary at the initial phase and is reserved for cases of catheter malfunction, extension of thrombosis, and persistent pain or inflammation [18].

Future perspectives in anticoagulation therapy

Heparins and VKAs have been the mainstay of anticoagulant therapy for over 50 years. There are data for UFH, LMWH, fondaparinux and warfarin for primary prophylaxis, but contemporary studies focus on LMWH in a population of cancer patients. A new generation of anticoagulants are available or under development, and these offer a favorable balance of anticoagulation to hemorrhagic effects. One phase II study [40] evaluated apixaban in outpatients receiving chemotherapy, and a new ultra LMWH (AVE5026) [41] showed a favorable safety-to-efficacy ratio, being suitable for cancer patients at increased risk of both bleeding and VTE. Dabigatran and rivaroxaban are orally administered drugs effective in reducing VTE incidence in orthopedic surgery [42].

The novel drugs are of significant interest in cancer patients as they are orally administered and do not require systemic laboratory monitoring. However, cancer patients with thrombotic complications are a special population, and they substantially differ from non-cancer patients with thrombosis. Thus, larger randomized trials are needed in order to prove the efficacy and safety of new anticoagulation agents in the oncology setting.