Venous Thromboembolism (VTE) is a potentially devastating and yet frequently under-recognised complication of active cancer and cancer therapy. The risk of VTE is thought to be at least 5-fold higher among patients receiving cancer therapy than that which is observed in the general population and it is estimated that up to 20% of all patients with cancer will experience some form of venous thrombosis during the course of their disease. 1-3 Acute VTE among patients with cancer is associated with a particularly high burden of morbidity and mortality risk. Cancer patients who develop VTE have a risk of recurrent thrombosis (despite optimal anticoagulant therapy) which may be as high as 20% at 12 months. 4, 5 Patients with cancer also experience more anticoagulant-related bleeding events than that observed in the general population. 5 The risk of death following VTE recurrence is also substantially higher among patients with cancer, estimated to be in the region of 15% in some subgroups. 6 In fact it is widely recognised that VTE represents the second-leading cause of death among patients with cancer, second only to the progression of cancer itself. 7 While significant advances have been made in recent years, several important clinical questions have yet to be addressed. The development of strategies for individualised prediction of VTE risk has proven to be challenging. Consequently, there is a degree of uncertainty among clinicians regarding the optimal approach to the use of primary pharmacological thromboprophylaxis for potentially ‘high-risk’ patients. Similarly, it remains to be determined if novel drug targets may represent a means of addressing the high rates of VTE recurrence and bleeding which continue to arise in this patient population despite current best therapy.
Mechanisms of hypercoagulability in cancer
The precise molecular mechanisms underlying the increased thrombosis risk among patients with cancer remain to be fully elucidated. However, a number of investigators have demonstrated clear evidence of a complex relationship between tumour cells and the haemostatic system. 8, 9 In particular, the aberrant expression of tissue Factor (TF, a procoagulant glycoprotein) on the surface of tumour cells and on circulating tumour-derived microvesicles has been implicated in the pathogenesis of hypercoagulability in this population. Tumour cells have also been shown to directly activate coagulation through a number of other distinct mechanisms including through the expression/release of various other pro-coagulant substances which directly activate elements of the haemostatic system including coagulation factors and platelets. 8 Tumour sub-type may also influence thrombosis risk due to the varying propensity of different tumours to activate coagulation, with particularly high-rates of thrombosis observed with tumours such as pancreatic carcinoma and multiple myeloma. The importance of this direct tumour effect in driving hypercoagulability is reflected in the observation that the risk of thrombosis appears to be greatest during periods of increased tumour activity, with the incidence of cancer-associated thrombosis being greatest in the first 6-months following diagnosis (as well as also being increased during periods of relapsing disease and at end-of-life).
Patient-specific and treatment specific factors also appear to significantly modulate thrombosis risk in the cancer population. Personal risk factors including prior VTE and elevated body mass index have been reported to be associated with an increased risk of thrombosis following a cancer diagnosis. 3 Numerous specific cancer treatments including traditional chemotherapies such as certain platinum-based agents (e.g. cisplatin, carboplatin) as well as some more novel treatments, including novel immunotherapies (such as the immune checkpoint inhibitors) have also been reported to increase risk. 10 Cancer surgery, radiation therapy and the use of central venous catheters have also all been shown to contribute to the VTE risk profile of patients with cancer requiring therapy. 3
The interaction of these tumour specific, patient-specific and treatment-specific factors largely determine the risk of thrombosis in an individual patient. However, using this information to accurately risk-stratify an individual or to identify a patient who might derive benefit from primary thromboprophylaxis is challenging and remains an area of significant controversy and debate.
Prevention of cancer-associated thrombosis
The provision of education relating to thrombosis risk should represent an integral component of the supportive care programme provided to patients who are about to commence cancer therapy. In recent years we have seen numerous reports of patients’ experiences with cancer associated thrombosis where the diagnosis and time to treatment initiation had been delayed as a consequence of patients being unaware of the symptoms of VTE (and the need to seek urgent medical attention). 11 This lack of awareness may lead to adverse clinical outcomes as well as substantial psychological distress. 11 The HSE have recently launched the VTE Alert Card initiative which provides patients with written information regarding VTE risk and the signs of symptoms which should prompt them to seek medical advice (Figure 1 on page 43). 12 These cards have been issued to all hospitals in Ireland and current HSE recommendations state that all patients being discharged from hospital should be provided with this written information. Incorporating the VTE alert card initiative into current patient education programs relating to cancer therapy may assist clinicians in ensuring that at-risk patients are sufficiently educated with regard to these risks.
Given the magnitude of VTE risk associated with cancer and cancer therapy, many experts believe that primary thromboprophylaxis should be offered to selected patients with cancer during periods of cancer treatment or in other high-risk scenarios. Current national and international clinical practice guidelines recommend that all patients with cancer who are admitted to hospital should be risk assessed with regards to thrombosis and bleeding risk and that pharmacological thromboprophylaxis with low molecular weight heparin (LMWH) be considered for the period of hospitalisation unless contraindicated due to a competing risk of bleeding. 13, 14 Similarly, the use of pharmacological thromboprophylaxis with LMWH, aspirin or a low-dose direct oral anticoagulant (DOAC) has become standard of care for out-patients receiving therapy for myeloma (a haematological cancer associated with a high risk of thrombosis). 13,14 The role of the pharmacological thromboprophylaxis for other cancer out-patients is an area of significant debate. A key obstacle to determining whether prophylaxis is warranted is the difficulty in accurately identifying the patients who are most at risk of VTE (and for whom the potential risks of pharmacological thromboprophylaxis could be justified). Several VTE risk predication models have been proposed for use in the cancer population, with the Khorana score being the most widely validated (Table 1). This algorithm stratifies cancer patients who are about to commence cancer therapy based on specific clinical and laboratory parameters as being at low (score 0), intermediate (score 1-2) or high risk (score 3+) for VTE. In validation studies, these thresholds have corresponded to predicted 6-month cumulative VTE risks in the region of 5%, 6.6% and 11% respectively – although absolute risk appears to vary considerably across tumour types and a key limitation of the Khorana score is that it seems to perform less well in some populations (e.g. patients with lung cancer). Consequently the reported risk estimates in different validation studies have been variable. In addition, up to 50% of patients who eventually experience VTE would not have been identified as being at intermediate (or higher) risk, which suggests that efforts to enhance the sensitivity of the Khorana score may also be warranted. 15,16
Most recently, the Khorana score has been utilised in large randomised trials which have evaluated the safety and efficacy of low dose DOACs for patients with cancer commencing chemotherapy. The AVERT and CASSINI randomised trials (which sought to investigate the safety and efficacy of apixaban 2.5mg BD and rivaroxaban 10mg OD respectively in patients with Khorana scores of 2 or more) demonstrated that prophylaxis with these reduced the risk of thrombosis in patients commencing cancer therapy (although this did not achieve statistical significance in the CASSINI trial). 17, 18 A systematic review and meta analysis incorporating data from both these DOAC trials as well as several earlier LMWH trials (using a post-hoc application of the Khorana Score), suggested that the use of pharmacological thromboprophylaxis in patients with a Khorana score of 2 or more significantly reduces the risk of thrombosis in that population (relative risk 0.51; 95% CI 0.34-67) with a number needed to treat (NNT) of 25; without significantly increasing the risk of bleeding (but also without any significant difference in all-cause mortality). 19 These and other data have subsequently shaped international clinical practice guidelines with recent guidance from bodies such as the American Society of Hematology, American Society of Clinical Oncology and the National Comprehensive Cancer Network suggesting that primary prophylaxis be considered for cancer outpatients who are about to commence chemotherapy and who are deemed to be at intermediate to high risk of thrombosis based on Khorana score. 13, 14, 20
Notwithstanding these recent recommendations, the widespread use of primary thromboprophylaxis in patients with cancer has yet to be adopted into routine clinical practice. It is clear that concerns remain among clinicians regarding the potential risks and challenges associated with implementing this approach in a ‘real-world setting’ as opposed to in the highly selected clinical trial setting where certain patient groups, such as the elderly, are generally underrepresented. Moreover, some experts point to the fact that while primary thromboprophylaxis does appear to reduce thrombosis risk, this has not yet been shown to deliver a survival benefit, at least in the studies conducted to date.
While the debate regarding the optimal use of primary thromboprophylaxis in the cancer population is likely to continue over the coming years, all stakeholders appear to be united in agreement that improving awareness of thrombosis risk among patients and healthcare providers should be prioritised. Taking simple steps, such as providing patients with written information relating to their VTE risk and the associated symptoms may reduce the risk of delaying treatment of a potentially life-threatening event.
Diagnosis of cancer-associated VTE
When patients with cancer present with suspected VTE it is crucial that diagnostic testing and initiation of appropriate anticoagulant therapy is not delayed. Clinical decisions rules for exclusion of acute VTE, such as the Wells score for pulmonary embolism or deep vein thrombosis, have been widely validated in the general population and lead to reductions in unnecessary use of diagnostic imaging tests. The value of these pre-test probability predictive algorithms in the cancer population is less clear due to the higher prevalence of VTE and due to concerns regarding the predictive value of the d-dimer in this group. 21 Consequently, clinicians must rely heavily on their clinical judgement when assessing cancer patients and will frequently need to resort to invasive diagnostic imaging tests. Efforts to determine the efficacy of pre-test probability algorithms in the cancer population are ongoing and most notably the HYDRA Study, a randomised trial comparing the use of the YEARS algorithm (which comprises of specific clinical signs in combination with a two-level d-dimer cut-off positivity threshold) versus diagnostic imaging with computed tomography for rapid exclusion of PE is currently recruiting worldwide. 22 The results of this trial are eagerly awaited.
Treatment of cancer-associated thrombosis
Therapeutic-intensity anticoagulation with LMWH was established as the standard of care for cancer-associated thrombosis following the publication of the CLOT randomised controlled trial in 2003 which confirmed the superiority of this agent over warfarin therapy in the cancer population. 23 More recently the publication of the HOKUSAI cancer VTE trial, the SELECT-D trial and the CARAVAGGIO trial have confirmed the non-inferiority of the DOACs edoxaban, rivaroxaban and apixaban to LMWH for the treatment of cancer-associated thrombosis. 24-26 All major international clinical practice guidelines now endorse the use of these agents as first-line alternatives for the management of cancer-associated thrombosis. 13, DOACs do appear to confer a higher risk of bleeding in some sub-groups of patients with cancer, specifically patients with luminal gastrointestinal tumours or genito-urinary tumours. Most clinicians continue to favour LMWH in these patient subgroups. LMWH may also be favoured for patients when there are potentially significant drug drug interactions present between DOACs and cancer therapies, in the setting of treatment or disease-related thrombocytopenia and where there are concerns regarding the potential for impaired gastrointestinal DOAC absorption.
Despite the advances in thrombosis therapy which have been achieved in recent years, outcomes for patients with cancer remain inferior to those which are observed in the general population. Even with the use of appropriate anticoagulant therapy, the risk of recurrent thrombosis remains unacceptably high in the cancer patient population with 6-month recurrence rates in the region of 6-10% being reported in the recent large DOAC clinical trials. 24-26 Similarly, the risk of bleeding complications on anticoagulant therapy is higher in this population, with rates of major bleeding of approximately 4-6% at 6-months also reported. Given the clear danger associated with both recurrent thrombosis and major bleeding (which may have case-fatality rates in the region of 10-15% in the cancer population), more therapeutic options are urgently required to mitigate these competing risks. The contact pathway of coagulation activation (comprising of the clotting factors pre-kallikrein, high-molecular weight kininogen, factor XII and factor XI) has emerged as an attractive novel drug target for the prevention and treatment of thromboembolism. 27 In preclinical models, the inhibition of the contact pathway does not appear to impair normal physiological haemostasis but does appear to inhibit pathological coagulation activation. Therefore, a drug which inhibits the contact pathway might be predicted to reduce the risk of thrombosis without significantly increasing the risk of bleeding. 27 Coagulation factor XI appears to be the most promising drug-target for this novel class of anticoagulant and clinical trials evaluating factor XI inhibitors for the treatment of cancer-associated thrombosis are ongoing, including the ASTER & MAGNOLIA randomised trials which are comparing the FXI inhibitor abelacimab to apixaban or dalteparin in the management of cancer-associated thrombosis (ClinicalTrials.gov Identifier NCT05171049).
VTE remains a common complication of cancer therapy despite the advances in preventative and therapeutic strategies which have been made in recent decades. As our population ages and as more people with cancer survive longer with their disease as a result of advances in cancer therapy, it is likely that the number of patients at risk of cancer-associated thrombosis will continue to rise. Given the burden of morbidity associated with VTE as well as the potential mortality risk, strategies for improved prevention, diagnosis and treatment of cancer-associated thrombosis are clearly needed. The level of awareness of VTE risk among patients with cancer appears to be low. Patient organisations such as Thrombosis Ireland have recognised this deficit and have championed efforts to prioritise patient education in our hospitals. This can only succeed with the support of healthcare providers, healthcare administrators and all other stakeholders involved in cancer care in this country. As we look to the future it is clear that a number of other key issues need to be addressed. Firstly, while current evidence appears to support primary pharmacological thromboprophylaxis for selected ambulatory cancer patients, we have yet to determine the optimal means of identifying the most ‘at-risk’ patients who would likely derive benefit from this strategy. Moreover, the absolute balance of benefit versus risk for this intervention in the ‘real-world’ setting is uncertain. Secondly, while the recent clinical trials of DOAC therapy for cancer associated thrombosis have increased the number of potential therapeutic options available to clinicians for managing VTE, rates of VTE recurrence and bleeding remain high. It remains to be seen if new strategies aimed at novel coagulation targets may help to bridge this gap. Addressing these important clinical questions will represent a major challenge for clinicians and researcher alike but will be crucial if we are to continue to improve outcomes for patients, their families and care-givers.
References available on request
Read our Latest News