Benefit risk profile of non-vitamin K antagonist oral anticoagulants in the management of venous thromboembolism

Transcription

1 Review Article 231 Benefitrisk profile of non-vitamin K antagonist oral anticoagulants in the management of venous thromboembolism Jan Beyer-Westendorf 1 ; Walter Ageno 2 1 Center for Vascular Diseases and Medical Clinic III, Dresden University Hospital Carl Gustav Carus, Dresden, Germany; 2 Department of Clinical and Experimental Medicine, University of Insubria, Varese, Italy Summary The prevention and treatment of venous thromboembolism (VTE) remains a clinical challenge, primarily owing to drawbacks associated with the use of heparins and vitamin K antagonists (VKAs). These and other factors, including a growing elderly population, mean that VTE presents a continuing burden to patients and physicians. Anticoag - ulant therapy is a fundamental approach for VTE management. Non- VKA oral anticoagulants, including the factor Xa inhibitors apixaban, edoxaban and rivaroxaban, and the thrombin inhibitor dabigatran, have been studied in phase III trials across a spectrum of thromboembolic disorders. These agents offer simplified care, with similar or improved efficacy and safety outcomes compared with heparins and vitamin K antagonists. There are several factors a physician must consider when prescribing an anticoagulant. An important consideration with all anticoagulant use is bleeding risk, especially in high-risk groups such as the elderly or those with renal impairment or cancer. In orthopaedic patients, other risks include a need for surgical revision or blood transfusion, or wound complications. Therefore, the clinical benefits of an anticoagulant should ideally be balanced with any risks associated with the therapy. Quantitative benefitrisk assessments are lacking, and owing to differences in trial design the non-vka oral anticoagulants cannot be compared directly. Based on trial and reallife data, this review will summarise the clinical data for the non-vka oral anticoagulants in the prevention and treatment of VTE, focusing on the balance between the benefits and risks of anticoagulation with these drugs, and their potential impact on VTE management. Keywords Benefitrisk assessment, non-vka oral anticoagulants, real-life, thromboprophylaxis, venous thromboembolism Correspondence to: Jan Beyer-Westendorf Center for Vascular Diseases and Medical Clinic III Dresden University Hospital "Carl Gustav Carus" Fetscherstraße Dresden, Germany Tel.: , Fax: jan.beyer@uniklinikum-dresden.de Financial support: Editorial support was funded by Bayer HealthCare Pharmaceuticals and Janssen Scientific Affairs, LLC. Received: June 2, 2014 Accepted after major revision: September 3, 2014 Epub ahead of print: October 16, Thromb Haemost 2015; 113: Introduction Deep-vein thrombosis (DVT) and pulmonary embolism (PE), which collectively comprise venous thromboembolism (VTE), are associated with considerable morbidity and mortality and remain a challenging healthcare burden (1, 2) because of the risks of recurrent VTE or chronic sequelae such as post-thrombotic syndrome (3) or chronic thromboembolic pulmonary hypertension (1, 4), which also affect patients quality of life. Anticoagulant therapy is a fundamental approach for the prevention and treatment of VTE. Low-molecular-weight heparin (LMWH), unfractionated heparin (UFH) and fondaparinux are typically recommended for VTE prevention (5), whereas a dualdrug approach of LMWH overlapping with and followed by a vitamin K antagonist (VKA), such as warfarin, is a common approach for VTE treatment (6, 7). Non-VKA oral anticoagulants (NOACs) include the factor Xa inhibitors apixaban, edoxaban and rivaroxaban, and the thrombin inhibitor dabigatran. Classically, NOACs have been known as new or novel oral anticoagulants; however, because dabigatran and rivaroxaban, for instance, gained approval for anticoagulant-related indications as early as 2008, a new and consistent terminology is demanded and non-vka is the recommended terminology of today. NOACs have been investigated across a spectrum of thromboembolic disorders and individually offer a simplified approach for patient care compared with traditional drugs. All of the NOACs are associated with a rapid onset of action and predictable pharmacokinetics and pharmacodynamics, obviating the need for routine coagulation monitoring (8). Drugdrug and drugfood interactions are also minimal with the NOACs compared with LMWH/VKAs, but caution should be exercised, or co-administration avoided, with inhibitors and inducers of P- glycoprotein or cytochrome P450 3A4 owing to a propensity for changing plasma concentrations (911). Each agent undergoes renal excretion to some extent; dabigatran has substantially higher renal clearance than apixaban, edoxaban and rivaroxaban (~85 % Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

2 232 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE vs ~2535 %) (912); this increased renal clearance is an important clinical differentiator between dabigatran and the three factor Xa inhibitors. There are several factors a physician must consider when prescribing an anticoagulant, and any potential risks must be weighed against the clinical benefits, particularly in certain patient subgroups, such as the elderly, those with renal impairment or cancer, or those who are on an antiplatelet regimen. Based on trial and real-life data, this review will summarise the clinical data for the NOACs in the prevention and treatment of VTE, with a focus on the benefits versus the risks of anticoagulation with these agents and their potential impact on VTE management. NOACs in the prevention and treatment of VTE VTE prevention in orthopaedic surgery Orthopaedic surgery is a common risk factor for VTE: without effective prophylaxis, ~50 % of patients who undergo total hip or knee replacement (THR or TKR, respectively) surgery develop VTE, but symptoms may only manifest in ~5 % of cases (13). All four NOACs showed acceptable benefitrisk profiles compared with standard therapy in phase III studies (1426), and the outcomes of these trials are summarised in Table 1. Differences between phase III studies of the NOACs, including different treatment durations, dosing regimens and lack of standardised bleeding definitions, make head-to-head comparisons difficult. Network meta-analyses and indirect comparisons of NOACs have been undertaken and generally show similar or improved efficacy outcomes with rivaroxaban, apixaban and dabigatran when compared with enoxaparin, but outcomes vary regarding bleeding risk; results have also been mixed when the agents are compared with each other ( Figure 1) (2729). Because the defined endpoints differ markedly between studies, bias may be introduced and, therefore, indirect comparisons are not suitable in this setting. To the best of our knowledge, there are no comparative data with edoxaban available in the literature. Valid benefitrisk evaluations for the group of NOACs are generally lacking. However, calculations of the relative risk based on symptomatic VTE, major bleeding and death in phase II and III studies of VTE prevention do not show any significant differences between the group of NOACs and enoxaparin. More extensive benefitrisk assessments have been conducted for dabigatran and rivaroxaban (30, 31). Dabigatran has been found to have a favourable benefitrisk profile over LMWH owing to the simple route of administration and a low risk of bleeding and other adverse events (31), and a quantitative benefitrisk assessment of RECORD14 pooled data has shown that the benefits (reductions in VTErelated events) of rivaroxaban exceeded the risks (increased bleeding events) (30). One of the main differences between the phase III studies of VTE prevention is the variation in bleeding definitions. The definition for major bleeding used in the RECORD programme did not include bleeding leading to treatment cessation or surgical-site bleeding events unless they were fatal or required re-operation (32), which was thought to allow better assessment of clinically important bleeding events. This RECORD definition was used in all of the rivaroxaban phase III orthopaedic studies, and a similar definition was used in studies involving edoxaban (32, 33); the results pointed towards similar or higher rates of bleeding among the NOACs compared with standard of care (SOC). In comparison, the International Society on Thrombosis and Haemostasis (ISTH) definition for major bleeding, including surgical-site bleeding, was used in the apixaban studies, for which no increase in the incidence of the principal safety outcome was found (21), and the European Medicines Agency (EMA) definition, including bleeding leading to treatment cessation and surgical-site bleeding, was used in the dabigatran studies, for which no increase in the rate of bleeding events was noted (24). There was a consistent trend for an increase in major bleeding events for rivaroxaban for each of the three definitions considered; however, none of the trends reached significance (32) (RECORD: hazard ratio [HR]=1.84 [95 % confidence interval (CI) ]; ISTH: HR=1.31 [95 % CI ]; EMA: HR=1.24 [95 % CI ] for the total treatment duration pool). Of note, bleeding leading to treatment cessation is no longer considered major by the EMA, owing to its subjectivity (34). When the North American regimen of enoxaparin (30 mg twice daily [bid]) was used as a comparator, both dabigatran and apixaban failed to reach non-inferiority (18, 21). Although edoxaban showed superiority to enoxaparin (20 mg bid) in Japanese and Taiwanese patients, this dose is not commonly used outside of Japan; therefore, such findings may not apply to a wider patient population (35). Rivaroxaban was the only agent to show superiority to both North American and EU enoxaparin regimens. Evidence from real-life studies with NOACs A prospective, international, observational study has evaluated the safety and efficacy of dabigatran (220 mg od) in prespecified subgroups (5,292 patients) with different body mass index (BMI), renal function and age undergoing THR/TKR surgery ( Table 1) (36). The results of the study suggested consistent efficacy and safety outcomes with dabigatran in patients at high risk of VTE in real life, and supported outcomes from phase III studies. In patients with moderate renal impairment (e. g. creatinine clearance [CrCl] 3049 ml/min) who received dabigatran, there was a threefold increase in symptomatic VTE and all-cause mortality compared with those with normal renal function. A twofold increase in the incidence of major bleeding was noted in severely obese patients (BMI > 35 kg/m 2 ) compared with patients with BMI < 25 kg/m 2, but this did not reach significance. Multicentre experience of thromboprophylaxis with dabigatran has shown that it presents high levels of satisfaction in patients and staff, is easy to use, and is associated with a good safety profile (37). However, in the same study, several disadvantages associated with dabigatran in an orthopaedic setting were identified, including time-consuming implementation into clinical practice, inaccuracy of laboratory monitoring parameters and a lack of long- Thrombosis and Haemostasis 113.2/2015 Schattauer 2015

3 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE 233 term experience (37). Of note, these disadvantages would not be related to dabigatran specifically, but to the group of NOACs in general. Preliminary data evaluating post-procedural dalteparin administration followed by outpatient use of dabigatran showed that this could be an effective and well-tolerated approach in patients undergoing THR/TKR surgery (38). There are currently no phase IV studies collecting real-life experience with apixaban or edoxaban in the setting of VTE prevention after orthopaedic surgery. XAMOS, a phase IV, non-interventional study in 17,701 patients across 37 countries, investigated the effectiveness and safety of rivaroxaban compared with SOC for the prevention of VTE after major orthopaedic surgery in routine care (39). The incidence of symptomatic thromboembolic events was significantly lower in patients who received rivaroxaban compared with SOC (0.9 % vs 1.4 %; relative risk reduction [RRR]=36 %; for reasons of comparability, all RRR values have been calculated using the same method, e. g =0.5; 0.5/1.4=36 %). Treatment-emergent Table 1: Overview of phase III, phase IV and registry data for the NOACs in the prevention of VTE after major orthopaedic surgery. Drug Phase III studies Dabigatran Apixaban Edoxaban Rivaroxaban Real-world studies Dabigatran Rivaroxaban Rivaroxaban Study name or programme RE-MODEL, RE- MOBILIZE and RE-NOVATE I/II (18, 2224) ADVANCE-13 (1921) STARS E-3 and J-5 (25, 26) RECORD14 (1417) Real-world study (36) XAMOS (39) ORTHO-TEP (40, 41) Phase and design Phase III RCT vs. enoxaparin Phase III RCT vs. enoxaparin Phase III RCT vs. enoxaparin Phase III RCT vs. enoxaparin Prospective, observational single-arm study Phase IV multinational registry vs. SOC Single-centre, retrospective registry vs. SOC (LMWH/ fondaparinux) Number of patients randomised 10,265 11,659 1,326 12,729 5,292 17,701 5,061 Regimen Dabigatran: 220 mg od or 150 mg a,b Enoxaparin: 40 mg od or 30 mg c Apixaban: 2.5 mg bid Enoxaparin: 40 mg od or 30 mg c Edoxaban: 30 mg od Enoxaparin: 20 mg bid Rivaroxaban: 10 mg od Enoxaparin: 40 mg od or 30 mg bid c Dabigatran: 220 mg od Rivaroxaban: 10 mg od SOC: LMWH in 82 % of cases Rivaroxaban: 10 mg od LMWH: 3,000 to 6,000 anti-factor Xa units od Fondaparinux: 2.5 mg od Results for primary efficacy outcome with study drug RE-MODEL, RE-NOVATE I and II: non-inferior RE-MOBILIZE: did not reach non-inferiority ADVANCE-1: did not reach non-inferiority ADVANCE-2 and -3: superior Superior in both studies Superior in all studies Consistent symptomatic VTE and all-cause mortality irrespective of BMI, renal function or age Significantly fewer thromboembolic events Significantly fewer inhospital symptomatic VTE events Results for primary safety outcome with study drug Similar incidence of major bleeding d Similar incidence of major bleeding e Similar incidence of major bleeding Similar incidence of major bleeding f Consistent major bleeding outcomes irrespective of BMI, renal function or age Similar incidence of treatment-emergent major bleeding f Significant reduction in severe g (vs fondaparinux) or major h bleeding (vs LMWH), but potential for bias a Only dabigatran 220 mg od was investigated in RE-NOVATE II; b dabigatran 75 mg or 110 mg was administered on the day of surgery in all four studies; c enoxaparin 30 mg bid was used in RE-MOBILIZE, ADVANCE-1 and RECORD4; d clinically overt bleeding associated with 20 g/l fall in haemoglobin; leading to transfusion of 2 units packed cells or whole blood; fatal, retroperitoneal, intracranial, intraocular or intraspinal bleeding; bleeding warranting treatment cessation or leading to reoperation; e acute, clinically overt bleeding accompanied by one or more of the following events: fatal bleeding; bleeding that occurred at a critical site (e. g. retroperitoneal, intracranial, intraocular, pericardial or intraspinal); bleeding into the operated joint that required additional operation or intervention; intramuscular bleeding with compartment syndrome; decrease in haemoglobin level of 20 g/l within a 24 hours period; a transfusion of 2 units of packed red cells; f clinically overt bleeding that was fatal; bleeding that occurred in a critical organ (e. g. retroperitoneal, intracranial, intraocular or intraspinal); bleeding that required reoperation; clinically overt extrasurgical-site bleeding associated with a fall in haemo globin level of 20 g/l or that required transfusion of 2 units of whole blood or packed cells; g bleeding leading to surgical revision or death, occurring in a critical site, or transfusion of 2 units of packed red blood cells; h overt bleeding with either a documented transfusion of 2 units of red blood cells, a surgical revision due to bleeding or a bleeding into critical site (intracranial, intraocular, intra-articular, retroperitoneal or overt gastrointestinal bleeding) or fatal bleeding. bid, twice daily; BMI, body mass index; LMWH, low-molecularweight heparin; NOAC, non-vitamin K antagonist oral anticoagulant; od, once daily; RCT, randomised controlled trial; SOC, standard of care; VTE, venous thromboembolism. Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

4 234 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE Figure 1A: Meta-analysis of NOAC studies in orthopaedic VTE prevention trials. Cl, confidence interval; NOAC, non-vka oral anticoagulant; VTE, venous thromboembolism. Adapted with permission from BMJ Publishing Group Limited (29). major bleeding events as defined in the RECORD programme were similar between the rivaroxaban and SOC groups at 0.4 % and 0.3 %, respectively. When using the EMA definition, treatment-emergent major bleeding events were generally higher, yet similar for rivaroxaban and SOC (1.7 % and 1.4 %, respectively) (39). Non-major bleeding as defined by RECORD, EMA nonmajor bleeding, and any bleeding were significantly higher in the rivaroxaban group compared with the SOC group (4.2 % vs 2.8 %, 2.9 % vs 1.7 %, and 4.7 % vs 3.2 %, respectively). Overall data from XAMOS confirmed the favourable benefitrisk profile of rivaroxaban when used in routine clinical care. The single-centre, retrospective ORTHO-TEP registry (~5,000 patients) evaluated the efficacy and safety of rivaroxaban for the prevention of VTE vs fondaparinux or LMWH (40, 41), and demonstrated a significant reduction in symptomatic VTE with rivaroxaban compared with both LMWH (2.1 % vs 4.2 %; RRR=50 %; p=0.005) (41) and fondaparinux (2.1 % vs 5.6 %; RRR=63 %; p< 0.001) (40). The incidence of PE was numerically lower with rivaroxaban compared with LMWH (0.2 % vs 0.5 %; p=0.213) (41). Rivaroxaban was associated with a significant reduction in the incidence of major bleeding compared with LMWH (2.9 % vs 7.0 %; RRR=59 %; p< 0.001) and fondaparinux (2.9 % vs 4.9 %; Thrombosis and Haemostasis 113.2/2015 Schattauer 2015

5 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE 235 Figure 1B: Meta-analysis of NOAC studies in orthopaedic VTE prevention trials. Cl, confidence interval; NOAC, non-vka oral anticoagulant; VTE, venous thromboembolism. Adapted with permission from BMJ Publishing Group Limited (29). RRR=41 %; p=0.010). The number of surgical revisions due to bleeding complications was also significantly lower with rivaroxaban in both comparisons (LMWH: 0.4 % vs 1.3 %; p=0.02; fondaparinux: 0.4 % vs 1.1 %; p=0.041), and the requirement for blood transfusions was reduced with rivaroxaban compared with LMWH (28.4 % vs 40.9 %; p< 0.001) (41). The length of hospital stay was significantly shorter in patients who received rivaroxaban compared with LMWH (8.3 days vs 11.1 days; p< 0.001) and fondaparinux (8.3 days vs 9.3 days; p< 0.001), which was driven mainly by the reduction in post-surgical VTE and bleeding complications (41). Of note, as with every retrospective cohort study, these findings have to be interpreted with caution. The tendency to have limited information on the incidence of bleeding events and their intensity is a potential source for bias. However, data from ORTHO-TEP confirmed that the efficacy of rivaroxaban was at least not offset by any increase in surgical complications (40, 41). VTE prevention in patients with acute medical illness Hospitalised patients with acute medical illness are at increased risk of VTE (42), and include a heterogeneous population such as Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

6 236 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE Figure 1C: Meta-analysis of NOAC studies in orthopaedic VTE prevention trials. Cl, confidence interval; NOAC, non-vka oral anticoagulant; VTE, venous thromboembolism. Adapted with permission from BMJ Publishing Group Limited (29). those with active cancer, heart failure or acute infectious or rheumatic diseases. Guideline recommendations for thromboprophylaxis in such patients include low-dose UFH, LMWH or fondaparinux (43). Although anticoagulants are effective in medically ill patients (44, 45), they remain underused (46, 47). Phase III studies have been undertaken with apixaban (ADOPT) and rivaroxaban (MAGELLAN) in acutely ill medical patients (48, 49). Rivaroxaban was effective in reducing the composite of VTE or VTErelated death (non-inferior at Day 10; superior at Day 35) compared with enoxaparin/placebo, while superiority was not reached for apixaban. In both studies, the rates of major bleeding were significantly higher in patients treated with NOACs compared with SOC ( Table 2). It is, therefore, more challenging to achieve an acceptable balance between benefit and risk in these patients and, currently, none of the NOACs are approved for this indication. Treatment of acute VTE A traditional dual-drug approach with LMWH overlapping with and followed by a VKA is effective for VTE treatment (6, 7). Owing to the slow onset of action of VKAs, LMWHs are administered initially because they exert a more rapid and predictable Thrombosis and Haemostasis 113.2/2015 Schattauer 2015

7 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE 237 Table 2: Overview of phase III data for the NOACs in the prevention of VTE in medically ill patients. Drug Apixaban Rivaroxaban Study name ADOPT (49) MAGELLAN (48) Number of patients randomised 6,528 8,101 Regimen Apixaban: 2.5 mg bid for 30 days Enoxaparin: 40 mg od for 614 days Rivaroxaban: 10 mg od for 35 days (subcutaneous placebo for 10 days) Enoxaparin: 40 mg od for 10 days (oral placebo for 35 days) bid, twice daily; NOAC, non-vitamin K antagonist oral anticoagulant; od, once daily. Results for primary efficacy outcome with study drug Did not reach superiority 10 days: non-inferior 35 days: superior Results for primary safety outcome with study drug Significant increase in major bleeding Significant increase in clinically relevant bleeding anticoagulant effect; however, they require parenteral administration (50), which can be burdensome. VKAs are associated with multiple food and drug interactions, and require coagulation monitoring and dose adjustment (50), which may contribute to low treatment adherence. Furthermore, major bleeding complications and poor patient outcomes are not uncommon with VKAs (51, 52). The NOACs offer simplified and effective alternatives to SOC in the treatment of patients with VTE, as shown in phase III studies (summarised in Table 3 and Table 4). Dabigatran Dabigatran showed non-inferiority in the reduction of recurrent VTE compared with warfarin in the RE-COVER and RE-COVER II studies (pooled analysis 2.3 % vs 2.2 %; p< 0.001), and was associated with similar risk of major bleeding events (1.2 % vs 1.7 %) (53, 54). In RE-COVER, more adverse events leading to discontinuation occurred in patients who received dabigatran and this number showed a trend towards significance (53). However, this was not seen in a pooled analysis of RE-COVER and RE-COVER II (54). When investigating dabigatran in various patient subgroups, including those with renal impairment and those 75 years of age, efficacy was consistent (safety outcomes in subgroups were not reported) (53). Apixaban Apixaban was non-inferior to enoxaparin/warfarin for the reduction of recurrent symptomatic VTE in the AMPLIFY study (2.3 % vs 2.7 %; p< 0.001), and was associated with a significant reduction in both major bleeding (0.6 % vs 1.8 %; RRR=67 %; p< 0.001) and the composite of major and non-major clinically relevant bleeding (4.3 % vs 9.7 %; RRR=56 %; p< 0.001) compared with enoxaparin/ warfarin. The incidence of non-fatal major bleeding was numerically lower in those patients treated with apixaban; this included critical-site bleeding (such as intracranial and retroperitoneal bleeding) and gastrointestinal bleeding. The composite of VTE, VTE-related death or major bleeding was significantly lower with apixaban compared with enoxaparin/warfarin (2.8 % vs 4.5 %; HR=0.62; 95 % CI ; p=0.001). Efficacy with apixaban was consistent across subgroups, whereas safety was significantly improved in some subgroups, including those 75 years of age, patients with extensive PE or those weighing > 100 kg (55). Edoxaban Data from Hokusai-VTE showed that edoxaban was non-inferior to warfarin in the reduction of recurrent VTE or VTE-related death (3.2 % vs 3.5 %; p< 0.001) and superior to warfarin in the reduction of the composite of major or non-major clinically relevant bleeding (8.5 % vs 10.3 %; RRR=17 %; p=0.004) (56). The incidence of major bleeding was similar between treatment groups, whereas the incidences of fatal and critical-site bleeding events were lower in patients who received edoxaban compared with those who received warfarin. A significant reduction in recurrent VTE was observed in patients with right ventricular dysfunction (defined as N-terminal pro-brain natriuretic peptide levels 500 pg/ml). Approximately 18 % of patients received a reduced dose of edoxaban; outcomes were similar or improved in this group compared with warfarin. Edoxaban also showed a significant reduction in recurrent VTE in fragile patients, defined as age 75 years, body weight 50 kg or CrCl 3050 ml/min. The composite of major and non-major clinically relevant bleeding events was significantly lower with edoxaban in patients weighing 60 kg compared with warfarin. In other subgroups, e. g. patients 75 years old or with CrCl 3050 ml/min, there was a numerical reduction in bleeding with edoxaban that did not reach significance (56). Rivaroxaban In EINSTEIN DVT, rivaroxaban was shown to be non-inferior to enoxaparin/vka for the reduction of recurrent VTE (2.1 % vs 3.0 %; p< 0.001), with a similar incidence of major and non-major clinically relevant bleeding (8.1 % vs 8.1 %) (57). Rates of major bleeding, including critical-site bleeding, were similar between the rivaroxaban and enoxaparin/vka groups (57). A net clinical benefit was shown for rivaroxaban compared with enoxaparin/ VKA for the composite of symptomatic recurrent VTE plus major bleeding (2.9 % vs 4.2 %; HR=0.67; 95 % CI ; p=0.03) (57). Adverse events and other safety outcomes were similar between treatment groups (57). In EINSTEIN PE, rivaroxaban was also non-inferior to enoxaparin/vka for the reduction of recurrent VTE (2.1 % vs 1.8 %; p=0.003), showing a similar incidence of major and non-major clinically relevant bleeding (10.3 % vs 11.4 %; p=0.23) and a significant reduction in major bleeding with rivaroxaban (RRR=50 %; p=0.003) (58). The incidence of critical-site bleeding was substan- Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

8 238 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE Table 3: Overview of the phase III studies of the NOACs in the treatment of acute VTE. Study name Study drug Comparator Study design Number of patients randomised Pre-randomisation heparin Heparin lead-in Study drug treatment regimen Treatment duration (months) Primary efficacy outcome Primary safety outcome b Type of index event (%) DVT only PE only PE with DVT PE with or without DVT Moderate or severe renal impairment (n) c Unprovoked VTE (%) Active cancer (%) Anatomical extent of PE (%) Limited Intermediate Extensive Results for primary efficacy outcome with study drug Results for primary safety outcome with study drug RE-COVER, RE-COVER II (53, 54) Heparin/dabigatran Heparin/warfarin Double-blind 5,132 At least 5 days 150 mg bid 6 Recurrent symptomatic VTE and related deaths Not defined; safety endpoints included bleeding events, acute coronary syndrome and other adverse events d 4.8 d Non-inferior Similar incidence of major bleeding AMPLIFY (55) Apixaban Enoxaparin/warfarin Double-blind 5, hours for 77 % None 10 mg bid for 7 days followed by 5 mg bid 6 Recurrent symptomatic VTE or VTE-related death Major bleeding alone Non-inferior Significant reduction in major bleeding Hokusai-VTE (56) Heparin/edoxaban Heparin/warfarin Double-blind a Edoxaban 30 mg od was used in patients perceived to be at higher risk for bleeding by predefined criteria (those with CrCl 3050 ml/min, weighing 60 kg or those receiving concomitant treatment with potent P-glycoprotein inhibitors); b all bleeding definitions adapted from ISTH criteria. Bleeding was defined as major if it was clinically overt and associated with a fall in the haemoglobin level of 20 g/l, or if it led to transfusion of 2 units of red cells, or if it was retroperitoneal, intracranial, occurred in a critical site or contributed to death. Clinically relevant non-major bleeding was defined as overt bleeding not meeting the criteria for major bleeding but associated with medical intervention, unscheduled contact with a physician, interruption or discontinuation of study treatment, or associated with any other discomfort such as pain or impairment of activities of daily life. Parameters for non-major clinically relevant bleeding for apixaban, edoxaban and rivaroxaban included macroscopic gastrointestinal haemorrhage, haemoptysis, intramuscular haematoma and the incidence of subcutaneous haematoma > 100 cm 2, if provoked, or because of a traumatic cause (in addition to subcutaneous haematoma > 25 cm 2 ). For apixaban and edoxaban, nonmajor clinically relevant bleeding also included any bleeding compromising haemodynamics; c moderate renal impairment defined as CrCl 3049 ml/min; severe renal impairment defined as CrCl < 30 ml/min; equivalent CrCl categories for AMPLIFY were: 30 ml/min; > ml/min, and for Hokusai-VTE ml/ min (no lower categories described); d only reported for RE-COVER; e Hokusai-VTE included patients with history of cancer; f EINSTEIN PE only. bid, twice daily; CrCl, creatinine clearance; DVT, deep-vein thrombosis; ISTH, International Society on Thrombosis and Haemostasis; NOAC, non-vitamin K antagonist oral anticoagulant; od, once daily; PE, pulmonary embolism; VKA, vitamin K antagonist; VTE, venous thromboembolism. 8,292 At least 5 days 60 mg od or 30 mg od a 312 Recurrent symptomatic VTE: composite of DVT or non-fatal or fatal PE Clinically relevant bleeding: composite of major and non-major clinically relevant bleeding e Non-inferior Significant reduction in major or clinically relevant non-major bleeding EINSTEIN DVT, EINSTEIN PE (57, 58) Rivaroxaban Enoxaparin/VKA Open-label, assessor-blind 8, hours for 83 % None 15 mg bid for 21 days followed by 20 mg od 3, 6 or 12 Recurrent symptomatic VTE: composite of fatal or non-fatal PE or DVT Clinically relevant bleeding: composite of major and nonmajor clinically relevant bleeding f 58.3 f 24.3 f Non-inferior Similar incidence of major or clinically relevant non-major bleeding Thrombosis and Haemostasis 113.2/2015 Schattauer 2015

9 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE 239 tially lower with rivaroxaban compared with enoxaparin/vka, and included fewer intracranial and retroperitoneal bleeding events (58). Unlike EINSTEIN DVT, there was no statistical difference in the net clinical benefit between patients receiving rivaroxaban or enoxaparin/vka (3.4 % vs 4.0 %; HR=0.85; 95 % CI ; p=0.28). Adverse events and other safety outcomes were again similar between patients receiving rivaroxaban or enoxaparin/vka (58). Rivaroxaban showed marginally better efficacy in EINSTEIN DVT, whereas bleeding outcomes were better in EIN- STEIN PE; subtle differences between DVT and PE patients may influence treatment effects. Rivaroxaban had an improved benefitrisk profile in both patient groups. The EINSTEIN DVT and EINSTEIN PE pooled analysis was pre-specified before the individual studies began (59), and the results confirmed the non-inferiority of rivaroxaban to enoxa - parin/vka (2.1 % vs 2.3 %; p< 0.001) and showed a 41 % RRR in major bleeding (1.0 % vs 1.7 %; p=0.002). This included fewer cases of critical-site bleeding and fatal bleeding with rivaroxaban. Clinical outcomes with rivaroxaban were consistent across key patient subgroups, including fragile patients (patients > 75 years old or CrCl < 50 ml/min or body weight 50 kg) or those with cancer or extensive VTE. A 71 % RRR in major bleeding was observed with rivaroxaban in fragile patients (59). Several meta-analyses have examined in more detail the topic of acute VTE treatment and NOACs. In 2012, Fox et al. conducted an adjusted indirect comparison between rivaroxaban, dabigatran, ximelagatran and apixaban and VKAs for this indication, assessing each NOAC with VKAs separately and additionally in a stratified analysis by pharmacological class, namely direct thrombin inhibitors or factor Xa inhibitors. The NOACs showed a similar risk of acute VTE recurrence and all-cause mortality (60), and rivaroxaban was associated with a reduced risk of bleeding. The factor Xa inhibitor-stratified analysis showed the same reduction, which is likely to be driven by a predominance of rivaroxaban data (60). A metaanalysis performed in 2013 also compared the efficacy and safety of NOACs with VKA in acute VTE patients and concluded, similarly to the 2012 study, that the efficacy of the NOACs was similar to that of the VKAs. Furthermore, NOACs were associated with a significantly lower risk of bleeding-related complications (61). Although the NOACs have very similar pharmacological characteristics, slight differences in efficacy and safety cannot be ruled out; however, large, randomised clinical trials based on tens of thousands of patients would be required to assess this. In the absence of such trials, pooling data from NOACs and comparing them with VKA regimens may enable a solid assessment of whether NOACs are preferable for the treatment of acute VTE. A meta-analysis conducted by Kang and Sobieraj that indirectly compared NOAC treatment with VKAs for acute VTE found that apixaban may have a better safety profile than dabigatran, rivaroxaban and edoxaban (62). On the other hand, the apixaban study included a different spectrum of VTE patients and excluded patients with provoked VTE in the absence of a persistent risk factor for recurrence. As a consequence, indirect comparisons between the NOACs should be avoided. Another important aspect that requires consideration is that VTE patients with impaired renal function have a higher risk of bleeding complications during anticoagulant therapy, but these patients also have a higher risk of recurrent VTE compared with patients with normal renal function. This has been demonstrated in the EINSTEIN DVT, EINSTEIN PE and AMPLIFY studies. A recent meta-analysis pooled the data from patients with renal impairment in studies of NOACs (including rivaroxaban, edoxaban, dabigatran and ximelagatran) and compared their efficacy and safety with VKA treatment in patients stratified according to their level of renal impairment (63). The results demonstrated that both the efficacy and safety profiles of the NOACs (as a class) are preserved in patients with impaired renal function. In addition, the proportion of renal excretion of the NOACs seemed to modify the safety profile but not the efficacy. Although indirect comparison using pooled populations may strengthen the statistical power and offer new conclusions, the authors emphasised the hypothesisgenerating aspect of their study. Extended anticoagulation to prevent recurrent VTE Some patients are at significant risk of VTE recurrence, and require prolonged anticoagulation for an indefinite period (64). In other cases, the risk of bleeding may outweigh the benefits of anticoagulation, and anticoagulation should ideally cease after 36 months (6, 65, 66). However, in most VTE patients in daily care there is clinical equipoise for stopping (e. g. bleeding risk or complete clot resolution) or continuing anticoagulation (e. g. persistent weak VTE risk factors, male gender, incomplete clot resolution) (6). The primary challenge that physicians face in this situation is maintaining an acceptable benefitrisk balance, in terms of preventing recurrent venous thromboembolic events without increasing the risk of bleeding. Most long-term phase III studies of NOACs (summarised in Table 4) have used placebo as a comparator, because most patients in daily care would have discontinued anticoagulation at this stage. Dabigatran RE-SONATE and RE-MEDY were two double-blind, randomised trials comparing dabigatran (150 mg bid) with placebo or warfarin, respectively. The treatment duration was up to six months in the placebo-controlled study and up to 36 months in the activecontrol trial. Outcomes from RE-SONATE showed a 92 % RRR in recurrent VTE for dabigatran compared with placebo (0.4 % vs 5.6 %; p< 0.001), and substantial reductions in symptomatic DVT and non-fatal PE (67). The incidence of major bleeding was similar between groups, whereas the composite of major or clinically relevant bleeding was significantly higher with dabigatran compared with placebo (5.3 % vs 1.8 %; p=0.001) (67). Other safety outcomes were similar between treatment groups, and liver function remained normal (67). In RE-MEDY, dabigatran was non-inferior to warfarin in the reduction of recurrent VTE, and the number of major bleeding events was numerically lower in dabigatran patients (0.9 % vs 1.8 %; RRR=50 %; p=0.06) (67). The composite of major or clinically relevant bleeding was significantly lower in patients who received dabigatran compared with warfarin (5.6 % vs 10.2 %; Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

10 240 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE Table 4: Overview of phase III data for the NOACs in the extended treatment of VTE. Drug Study name Number of patients randomised Regimen Treatment duration Results for primary efficacy outcome with study drug Results for primary safety outcome with study drug Dabigatran RE-MEDY (67) 2,866 Dabigatran: 150 mg bid Warfarin 636 months Non-inferior No significant increase in major bleeding Dabigatran RE-SONATE (67) 1,353 Dabigatran: 150 mg bid Placebo 12 months Superior No significant increase in major bleeding Apixaban AMPLIFY-EXT (68) 2,486 Apixaban: 2.5 mg bid or 5 mg bid Placebo 12 months Superior No significant increase in major bleeding Rivaroxaban EINSTEIN EXT (57) 1,197 Rivaroxaban: 20 mg od Placebo 6 or 12 months Superior No significant increase in major bleeding bid, twice daily; NOAC, non-vitamin K antagonist oral anticoagulant; od, once daily. RRR=45 %; p< 0.001). The overall number of adverse events did not differ significantly between treatment groups, yet the number of acute coronary syndrome events was significantly higher with dabigatran compared with warfarin (67). Apixaban In AMPLIFY-EXT, both doses of apixaban (2.5 mg and 5 mg bid) reduced the risk of recurrent VTE compared with placebo (2.5 mg bid: 3.8 % vs 11.6 %; RRR=64 % and 5 mg bid: 4.2 % vs 11.6 %; RRR=67 %), without increasing the incidence of major bleeding (68). A net clinical benefit (the composite of a reduction in symptomatic recurrent VTE, death related to VTE, myocardial infarction, stroke, death related to cardiovascular disease, or major bleeding) was observed with both doses of apixaban compared with placebo (2.5 mg vs placebo: 2.4 % vs 10.4 %; HR=0.23 [95 % CI ]; 5 mg vs placebo: 2.5 % vs 10.4 %; HR=0.24 [95 % CI ]). Other safety outcomes were similar between patients who received apixaban and placebo. Subgroup analyses showed that both doses of apixaban were effective and well-tolerated, regardless of age, sex, body weight, renal impairment, or whether the index event was DVT or PE (68). The effectiveness of dose reduction seen with apixaban in AMPLIFY-EXT has indicated the need to evaluate the efficacy and safety of prolonged VTE treatment in further dose-reduction studies, such as EINSTEIN CHOICE for rivaroxaban. Rivaroxaban The EINSTEIN EXT study showed an 82 % RRR in recurrent VTE with rivaroxaban (20 mg od) compared with placebo (1.3 % vs 7.1 %; p< 0.001), with no significant increase in major bleeding and a significant improvement in net clinical benefit (2.0 % vs 7.1 %; HR=0.28 [95 % CI ]; p< 0.001) (57). Most cases of nonmajor clinically relevant bleeding with rivaroxaban were nasal or rectal in origin, or involved blood in the urine; only one case was gastrointestinal. There were substantially fewer non-fatal PE and recurrent DVT events with rivaroxaban, and mortality and cardiovascular outcomes were similar between patients who received rivaroxaban or placebo, with no changes to liver function in either treatment group (57). Evidence from real-life studies A number of studies of NOACs for the treatment of VTE in routine clinical settings are ongoing. The Dresden NOAC registry is evaluating the effects of NOAC-based therapy in daily care, and will investigate the acute and long-term efficacy and safety of rivaroxaban, apixaban and dabigatran in ~2,500 patients with AF or acute or chronic VTE (clinicaltrials.gov identifier: NCT ). Preliminary data from ~400 VTE patients have already confirmed the effectiveness of rivaroxaban for acute and extended VTE treatment, with low rates of recurrent VTE and acceptably low rates of major bleeding and cardiovascular events (69). An assessment of overall bleeding complications showed that only a small proportion of reported bleeding events seen with NOACs were major (70). Gastrointestinal bleeding was found to be the most common site of major bleeding, whereas skin and mucosal bleeding events were the most common manifestations of non-major bleeding. Surgical treatment was required only in a small number of cases, suggesting that current practice, i. e. endoscopy, local compression, tamponade or transfusion, is sufficient to manage bleeding events with these agents in routine care (70). XALIA is a non-interventional cohort field study that will investigate the long-term safety of rivaroxaban for the treatment of patients with acute DVT in clinical practice, compared with standard therapy (initial LMWH, UFH or fondaparinux, overlapping with and followed by a VKA) for a period of 3 months (clinicaltrials.gov identifier: NCT ). Enrolment for XALIA is complete and the follow-up ongoing. PREFER in VTE is a multicentre, prospective observational disease registry planning to enrol more than 4,000 patients across Europe. It is the first patient registry to gather comprehensive data on the quality of life and treatment satisfaction of patients with DVT and/or PE. The GARFIELD-VTE registry is also planning to enrol ~10,000 patients, beginning in It will explore acute and long-term Thrombosis and Haemostasis 113.2/2015 Schattauer 2015

11 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE 241 Table 5: Summary of current guidelines for application of the NOACs for prevention and treatment of VTE. Executive summary: Antithrombotic therapy and prevention of thrombosis, 9th edition 2012: American College of Chest Physicians (ACCP) (97) Choice of anticoagulant regimen for long-term therapy for VTE In patients with DVT of the leg and no cancer, we suggest VKA therapy over LMWH for long-term therapy (Grade 2C). For patients with DVT and no cancer who are not treated with VKA therapy, we suggest LMWH over dabigatran or rivaroxaban for long-term therapy (Grade 2C). In patients with DVT of the leg and cancer, we suggest LMWH over VKA therapy (Grade 2B). In patients with DVT and cancer who are not treated with LMWH, we suggest VKA over dabigatran or rivaroxaban for long-term therapy (Grade 2B). In patients with PE and no cancer, we suggest VKA therapy over LMWH for long-term therapy (Grade 2C). For patients with PE and no cancer who are not treated with VKA therapy, we suggest LMWH over dabigatran or rivaroxaban for long-term therapy (Grade 2C). In patients with PE and cancer, we suggest LMWH over VKA therapy (Grade 2B). In patients with PE and cancer who are not treated with LMWH, we suggest VKA over dabigatran or rivaroxaban for long-term therapy (Grade 2C). Choice of anticoagulant regimen for preventative VTE therapy after major orthopaedic surgery In patients undergoing THR or TKR, we recommend use of one of the following for a minimum of 10 to 14 days rather than no antithrombotic prophylaxis: LMWH, fondaparinux, apixaban, dabigatran, rivaroxaban, low-dose UFH, adjusted-dose VKA, aspirin (all Grade 1B), or an intermittent pneumatic compression device (Grade 1C). In patients undergoing THR or TKR, irrespective of the concomitant use of an intermittent pneumatic compression device or length of treatment, we suggest the use of LMWH in preference to the other agents we have recommended as alternatives, fondaparinux, apixaban, dabigatran, rivaroxaban, low-dose UFH (all Grade 2B), adjusted-dose VKA, or aspirin (all Grade 2C). In patients undergoing major orthopaedic surgery and who decline or are uncooperative with injections or an intermittent pneumatic compression device, we recommend using apixaban or dabigatran (alternatively rivaroxaban or adjusted-dose VKA if apixaban of dabigatran are unavailable) rather than alternative forms of prophylaxis (all Grade 1B). National Institute for Health and Care Excellence (NICE) clinical guideline 2012: Venous thromboembolic diseases Choice of anticoagulant regimen for long-term therapy for VTE Rivaroxaban (Xarelto, Bayer) is indicated for the treatment of DVT, and prevention of recurrent DVT and PE following an acute DVT in adults. For the initial treatment of acute DVT, the recommended dosage of rivaroxaban is 15 mg twice daily for the first 21 days followed by 20 mg once daily for continued treatment and prevention of recurrence (98). Rivaroxaban is recommended as an option for treating PE and preventing recurrent DVT and PE in adults. For the initial treatment of acute PE, the recommended dosage of rivaroxaban is 15 mg twice daily for the first 21 days followed by 20 mg once daily for continued treatment and prevention of recurrent VTE (99). Choice of anticoagulant regimen for preventative VTE therapy after major orthopaedic surgery Rivaroxaban, within its marketing authorisation, is recommended as an option for the prevention of venous thromboembolism in adults having elective THR/ TKR surgery. The summary of product characteristics states that rivaroxaban should be taken orally once daily in 10 mg doses. The initial dose should be taken 610 hours after surgery, provided that haemostasis has been established (100). Apixaban is recommended as an option for the prevention of venous thromboembolism in adults after elective hip or knee replacement surgery. The recommended dosage of apixaban in the summary of product characteristics is 2.5 mg orally twice daily. The initial dose should be taken 1224 hours after surgery. The duration of treatment depends on the individual risk of the patient for VTE, which is determined by the type of orthopaedic surgery. Recommended treatment durations are 3238 days for patients having hip replacement surgery and 1014 days for patients having knee replacement surgery (101). Dabigatran etexilate, within its marketing authorisation, is recommended as an option for the primary prevention of venous thromboembolic events in adults who have undergone elective THR/TKR surgery. The summary of product characteristics states that dabigatran etexilate treatment should be started within 14 hours of surgery with a half dose of 110 mg. Thereafter, treatment is continued with a standard dose of 220 mg once daily for 10 days after knee replacement and for 2835 days after hip replacement (102). DVT, deep-vein thrombosis; LMWH, low-molecular-weight heparin; NOAC, non-vitamin K antagonist oral anticoagulant; PE, pulmonary embolism; THR, total hip replacement; TKR, total knee replacement; UFH, unfractionated heparin, VKA, vitamin K antagonist; VTE, venous thromboembolism. management, and outcomes in patients with symptomatic VTE treated in a real-world setting, including treatment duration, incidence of complications, practical aspects and healthcare resource utilisation. Considerations for NOACs for the acute and extended treatment of VTE As with studies of VTE prevention, differences in study design prohibit direct comparisons between the NOACs ( Table 3). AMPLIFY, RE-COVER and Hokusai-VTE were double-blind studies (53, 55, 56), and EINSTEIN DVT and EINSTEIN PE were open-label in design (57, 58). There are strengths and weaknesses associated with both types of design: open-label studies may be more reflective of clinical practice, but may have a risk of reporting bias of adverse events. Double-blind studies are complex, which could reduce patient compliance or result in high withdrawal rates (71). In terms of initial intensive treatment, benefits were seen with both apixaban (10 mg bid) given for seven days and rivaroxaban (15 mg bid) for 21 days (55, 57, 58). There is no optimal duration for this period of initial intensive anticoagulation, but seven days Schattauer 2015 Thrombosis and Haemostasis 113.2/2015

12 242 Beyer-Westendorf, Ageno: Benefitrisk profile of NOACs in VTE Table 6: Cost-effectiveness assessment of NOACs for VTE management. Indication VTE prevention in MOS VTE prevention in MOS b VTE prevention in MOS VTE prevention in MOS VTE prevention in MOS VTE prevention in MOS VTE prevention in MOS VTE prevention in MOS VTE acute treatment VTE secondary prevention VTE secondary prevention NOAC and comparator Dabigatran vs enoxaparin Dabigatran vs enoxaparin Dabigatran 220 mg once daily vs enoxaparin Dabigatran 150 mg once daily vs enoxaparin Dabigatran 220 mg once daily vs fondaparinux Dabigatran 150 mg once daily vs fondaparinux Apixaban vs enoxaparin Rivaroxaban vs enoxaparin Rivaroxaban vs enoxaparin Rivaroxaban vs enoxaparin Rivaroxaban vs enoxaparin + VKA Rivaroxaban vs warfarin Rivaroxaban vs placebo Country a a Country for which cost-effectiveness analysis has been conducted; b subgroup analysis in patients aged > 75 years or with moderate renal impairment (defined as CrCl 3049 ml/min). +, cost-effectiveness confirmed; -, cost-effectiveness not confirmed; MOS, major orthopaedic surgery; NOAC, non-vka oral anticoagulant; od, THR, total hip replacement; TKR, total knee replacement; VKA, vitamin K antagonist; VTE, venous thromboembolism. UK UK UK UK Canada USA Canada France, Italy and Spain USA USA USA Cost-effectiveness + THR + TKR + THR + TKR + THR + TKR + THR TKR THR TKR THR TKR + THR + TKR + THR + TKR + THR + TKR + THR + TKR Author and reference Wolowacz (87) Wolowacz (88) Holmes (89) Holmes (89) Revankar (90) Kwong (91) Diamantopoulos (92) Monreal (103) Lefebvre (94) Seaman (95) Coleman (96) is similar to the duration of standard LMWH therapy (510 days) (6), whereas treatment for 21 days may be supported by earlier studies showing an increased risk of recurrence in the first few weeks after the initial event (72, 73). A high proportion of patients in EINSTEIN DVT, EINSTEIN PE and AMPLIFY received parenteral anticoagulation with LMWH, heparin or fondaparinux prior to randomisation, but this was usually for less than 23 days (55, 57, 58). The clinical benefits of apixaban and rivaroxaban can, therefore, be attributed to this single-drug approach, and were unlikely to be augmented by initial parenteral anticoagulation. In RE-COVER and Hokusai-VTE, initial parenteral anticoagulation was administered for a median of nine days and seven days, respectively (53, 56). Guidelines now specify that dabigatran is indicated for the treatment of VTE in patients who have been treated with a parenteral anticoagulant for 510 days (104). Assessment of anatomical severity suggested that the populations were similar across the studies with the exception of AMPLIFY, which contained larger numbers of patients with unprovoked VTE owing to this specific inclusion criterion. Differences can be noted between the VTE treatment studies regarding the type of bleeding outcomes used, as well as the definition criteria used ( Table 3). The dabigatran studies (RE-COVER and RE-COVER II) did not specify a particular bleeding outcome as the principal safety outcome (53, 67). For all studies, the definitions of major and non-major clinically relevant bleeding were adapted from ISTH criteria (73). Major bleeding definitions were broadly consistent across studies, whereas some minor differences were noted among agents for non-major clinically relevant bleeding ( Table 3). Other characteristics of the studies, including treatment regimens, the type of index event, anatomical extent of PE and the proportion of patients with unprovoked VTE or renal impairment are presented in Table 3. AMPLIFY and Hokusai-VTE had stricter inclusion criteria than EINSTEIN DVT, EINSTEIN PE and RE-COVER, and did not include patients with cancer for whom long-term treatment with LMWH was planned (55, 56). Patients with cancer-associated VTE are a challenging population to treat. VKAs may not always be appropriate in patients with cancer because of unpredictable Thrombosis and Haemostasis 113.2/2015 Schattauer 2015