Article Text

Download PDFPDF

How I treat cancer-associated thrombosis
  1. Florian Moik1,
  2. Ingrid Pabinger1,
  3. Cihan Ay1,2
  1. 1Clinical Division of Haematology and Haemostaseology, Department of Medicine I, Comprehensive Cancer Center Vienna, Medical University of Vienna, Vienna, Austria
  2. 2I.M. Sechenov First Moscow State Medical University (Sechenov University), Moscow, Russia
  1. Correspondence to Dr Cihan Ay; cihan.ay{at}


Patients with cancer are at an increased risk of symptomatic venous thromboembolism (VTE). In addition, an increasing number of patients with incidental thromboembolic events have been recorded in clinical practice. Therapeutic anticoagulation is crucial to prevent thrombus progression and reduce risk of recurrence; however, this comes at the price of an increased bleeding risk, which necessitates a personalised approach to choose the most appropriate type of therapy. Over the last decade, low-molecular-weight heparin has been the preferred anticoagulant agent for patients with cancer-associated thrombosis due to better efficacy and similar safety profile compared with vitamin K antagonists. While direct oral anticoagulants (DOAC) have emerged as new option for treatment of VTE in a general population, only limited data have been available specifically for patients with cancer until recently. Randomised, controlled trials have now been published, establishing DOAC as an alternative for the treatment of cancer-associated thrombosis. However, the improvement in the therapeutic armamentarium is accompanied by a number of special considerations. For instance, risk of bleeding is elevated in patients with cancer-associated VTE receiving DOAC, especially in certain tumour types (eg, gastrointestinal), and no guidance exists regarding their use in patients with severe thrombocytopaenia. Furthermore, DOAC are prone to certain drug–drug interactions and their effect might be altered due to nausea and vomiting in patients receiving chemotherapy. Here, we provide guidance on how to treat cancer-associated VTE and how new evidence from randomised controlled trials can be implemented in clinical practice. There are still clinical scenarios where robust evidence is lacking and treatment recommendations are based on extrapolations from other populations or expert opinion only. Therefore, additional research in special subpopulations is needed to optimise management of patients in challenging clinical scenarios.

  • venous thromboembolism
  • thrombosis
  • anticoagulation
  • cancer
  • direct oral anticoagulants

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, any changes made are indicated, and the use is non-commercial. See:

View Full Text

Statistics from

Venous thromboembolism in patients with cancer

Venous thromboembolism (VTE), comprising deep vein thrombosis (DVT) and pulmonary embolism (PE), is a common complication in patients with cancer. The incidence of VTE in patients with cancer is elevated compared with the general population, with reported annual rates in a pooled analysis of 38 cohort studies between 0.5% and 20%, depending on specific cancer subpopulation, compared with an annual incidence rate of 0.1%–0.2% in patients without cancer.1 2 In addition, thrombosis at unusual sites, such as in the splanchnic veins, or related with a central venous catheter (CVC) is frequently observed in patients with cancer.

Cancer-associated thromboembolism (CAT) causes increased morbidity, sometimes delay of oncological treatment and an increase in healthcare expenses.3–5 Furthermore, VTE is among the leading causes of death in patients with cancer and the occurrence of thrombotic events is a negative prognostic factor beyond direct VTE-related mortality, underlining the complex interaction between the haemostatic system and malignancy.6–8

Therapeutic anticoagulation in patients with cancer-associated VTE requires carefully balancing risk and benefit. The management of patients with CAT is challenged by a higher risk of both recurrent VTE and bleeding events compared with patients with VTE without cancer, and oral anticoagulation can be further complicated by severe thrombocytopaenia, potential drug–drug interactions and nausea and vomiting.9 10

Here, we provide a concise overview on newly published randomised controlled trials, on how most recent evidence has been incorporated in updated guidelines for treatment of VTE in patients with cancer and our approach to patients with cancer-associated VTE. We also discuss several special issues and clinical scenarios, such as potential drug interactions of direct oral anticoagulants (DOAC), management of anticoagulation in patients with severe thrombocytopaenia, incidentally diagnosed asymptomatic VTE and catheter-related thrombosis (CRT).

Anticoagulation in patients with acute cancer-associated VTE

In the past two decades, the recommended treatment for patients with cancer and acute VTE in international guidelines was low-molecular-weight heparin (LMWH). This has been based on the pivotal CLOT trial, comparing LMWH (dalteparin) to vitamin K antagonists (VKA), that found lower rates of recurrent VTE at 6 months (9% vs 17%; HR: 0.48; 95% CI 0.30 to 0.77) and a similar risk of bleeding events (6% vs 4%, p=0.27) in patients treated with dalteparin.11 Treatment and secondary prevention of patients with VTE in a general population has been revolutionised by the development and introduction of DOAC in clinical practice. However, patients with cancer were underrepresented in clinical trials evaluating the efficacy and safety of these agents against VKA and details on their cancer status were not clearly defined.12–15 In a meta-analysis, including subgroups of patients with cancer from phase III trials comparing DOAC to VKA, efficacy and safety were comparable.16 However, as the comparative agent in these studies was VKA, which was not the preferred agent in patients with cancer according to guideline recommendations, only limited clinical implications could be drawn.

In 2018, the results of two phase III trials have been published, which specifically aimed at testing the efficacy and safety of DOAC against LMWH for the treatment of CAT. The Hokusai VTE cancer trial, a prospective, open label, blinded endpoint evaluation (PROBE), non-inferiority trial, compared edoxaban, a direct coagulation factor Xa inhibitor, to anticoagulation therapy with dalteparin for a treatment duration of 6–12 months. The primary outcome, a composite of recurrent VTE and major bleeding events (according to ISTH (International Society on Thrombosis and Haemostasis) definition) at 12 months, did not differ between edoxaban and dalteparin (edoxaban: 12.8%; LMWH: 13.5%; HR: 0.97; 95% CI 0.70 to 0.36, p=0.006 for non-inferiority). Patients treated with edoxaban experienced fewer VTE events (7.9% vs 11.3%; HR: 0.71; 95% CI 0.48 to 1.06), but rates of major bleeding were higher (6.9% vs 4.0%; HR: 1.77; 95% CI 1.03 to 3.04), mainly due to an excess of gastrointestinal (GI) bleeding in patients with GI malignancies.17 In SELECT-D, an open label, pilot study, rivaroxaban, also a direct factor Xa inhibitor, was compared with LMWH (dalteparin) for the treatment of cancer-associated VTE. The 6-month cumulative incidence of recurrent VTE was lower in patients treated with rivaroxaban (4% vs 11%, HR: 0.43; 95% CI 0.19 to 0.99) and the incidence of major bleeding was higher (6% vs 4%, HR: 1.83; 95% CI 0.68 to 4.96), as was the rate of clinically relevant non-major bleeding (CRNMB) (13% vs 4%, HR: 3.76; 95% CI 1.63 to 8.69), comprising mostly of upper GI and urothelial bleeding.18 Also in the SELECT-D study, GI bleeding tended to be higher in patients with gastro-oesophageal and colorectal cancer. Recruitment of patients with upper GI malignancy (oesophageal cancer and cancer of the gastro-oesophageal junction) was terminated during the ongoing SELECT-D trial, due to a non-significant rise in major bleeding events.18 Importantly, risk of intracranial bleeding in both the HOKUSAI VTE cancer and SELECT-D study in patients receiving a DOAC was not increased. Fatal bleeding occurred in one patient each in patients receiving rivaroxaban and dalteparin in SELECT-D, no patient in the edoxaban group and two patients in the dalteparin group in the Hokusai VTE cancer trial.17 18

In 2019, results from the ADAM-VTE trial have been published, comparing the oral factor Xa inhibitor apixaban to dalteparin for the treatment of cancer-associated VTE, including upper extremity and splanchnic vein thrombosis (SVT). The primary outcome event of major bleeding (according to ISTH definition) occurred in 0% of 145 patients in the apixaban group and 1.4% of 142 patients in the dalteparin group. Secondary outcomes included recurrent VTE, with rates significantly lower in the apixaban arm compared with the dalteparin arm (0.7% vs 6.3%, HR: 0.099; 95% CI 0.013 to 0.780).19 The frequency of CRNMB was 6% in both groups. Interestingly, the rate of major bleeding, recurrent VTE and mortality in the LMWH arm in the ADAM-VTE trial was lower compared with the LMWH arm in the Hokusai VTE cancer and SELECT-D studies, suggesting that a lower risk population was included in the ADAM-VTE study. A large randomised controlled non-inferiority trial comparing apixaban to LMWH (dalteparin) for treatment of acute cancer-associated VTE, the CARAVAGGIO study, which is ongoing, will provide further evidence for efficacy and safety of apixaban for treatment of cancer-associated VTE.20

Based on these studies, anticoagulation with either edoxaban or rivaroxaban has been incorporated in updated guidelines for management and treatment of cancer-associated VTE as alterative to LMWH.21–23 At the time of publication of these guidelines, the full publication of the ADAM VTE trial was not available to expand the recommendation also to apixaban. However, several factors have to be considered prior to deciding on treating a patient with a DOAC. In figure 1, we provide a treatment algorithm, which is based on evidence from interventional trials but also on expert opinion because of limited available data. The suggestions should help to guide the clinical decisions in daily practice, with particular value on patients’ safety. Furthermore, VKA might still be a treatment option for patients with no access or contraindications to DOAC and LMWH.22 As for these reasons, VKA are still frequently used in many parts of the world and knowledge and skills in VKA management, in particular dose adjustment and monitoring, are still important.

Figure 1

Treatment algorithm for patients with cancer and acute VTE; 1: VTE including symptomatic and incidentally detected events (excluding isolated, asymptomatic subsegmental pulmonary embolism); 2: high risk: symptomatic, proximal PE, proximal DVT or history of previous thrombotic events in acute phase of VTE (<30 days since diagnosis of the event); 3: low risk: distal DVT, incidental subsegmental PE, CRT and/or patients in subacute phase of VTE (>30 days since event); 4: target platelet count for transfusion: 40–50 G/L; 5: luminal GI tumours (intact primary tumour), tumours at risk of bleeding from GU tract, nephrostomy tubes, active GI mucosal abnormalities (gastric/duodenal ulcer, gastritis, colitis, esophagitis, etc); 6: strong interaction with CYP3A4 or P-gp suspected; 7: DOAC: edoxaban: 60 mg once daily after LMWH lead-in (reduced dose of 30 mg once daily: creatinine clearance below 50 mL/min, body weight of <60 kg or concomitant use of a potent P-glycoprotein (P-gp) inhibitor); rivaroxaban: 15 mg twice daily for 3 weeks, then 20 mg once daily; apixaban: 10 mg twice daily for 7 days, followed by 5 mg twice daily; 8: dalteparin: 200 IU/kg daily during the first month, then 150 IU/kg daily; 9: rivaroxaban absorption reduced when not taken with food; 10: for patients treated with LMWH: switch to DOAC or escalate dose; 11: retrievable IVC filters represent an option until anticoagulation is possible only in very selected cases depending on thrombus location (ie, proximal DVT) and timing since VTE diagnosis according to guidance from the ISTH SSC.25 Abbreviations: DVT, deep vein thrombosis; DOAC, direct oral anticoagulant; IVC, inferior vena cava; GI, gastrointestinal; GU, genito-urinary; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin; VTE, venous thromboembolism; VKA, vitamin K antagonist.

The advantage of DOAC lies in its oral administration and easy, standardised dosing and thereby might enhance patient compliance. LMWH is favourable due to experience in use in this setting and its applicability in patients with contraindications to DOAC. VKA can be administered if both LMWH and DOAC are contraindicated and is an economical, globally more easily available alternative for anticoagulation in patients with CAT.

Table 1 summarises the phase III and observational studies on the safety and efficacy of different anticoagulation strategies in patients with cancer and acute VTE.

Table 1

Summary of randomised controlled trials on anticoagulation in patients with cancer-associated VTE

Practical considerations of DOAC use in patients with cancer

Tumour types with high bleeding risk

In patients with CAT treated with a DOAC, more bleeding events in certain subpopulations were observed. Therefore, an individual risk–benefit evaluation is needed and their uncritical use should be avoided. The risk of bleeding seems to be increased in colorectal and gastro-oesophageal cancers and possibly also genito-urinary cancers. Therefore, in patients with luminal GI cancer with the primary tumour in place, tumours at risk of bleeding from the genito-urinary (GU) tract, nephrostomy tubes or active GI mucosal abnormalities (eg, gastric/duodenal ulcer, gastritis, colitis, esophagitis) DOAC are not suggested as the first-line treatment option.21 22

Frail patients

SELECT-D, Hokusai VTE cancer and ADAM-VTE have included patients with Eastern Cooperative Oncology Group (ECOG) performance status of ≤2 and therefore, data derived from these studies might not be generalisable to frail patients or those with unfavourable performance status.17 18


Another common factor influencing the individual risk of bleeding is the presence of thrombocytopaenia, which complicates treatment of CAT. Despite the substantial risk of bleeding in patients with cancer and thrombocytopaenia, rates of recurrent VTE are not reduced. For patients with cancer, acute VTE and platelet counts above 50 G/L, full-dose anticoagulation, either with LMWH or DOAC, can be administered. As suggested in a guidance document by an international expert group, for patients with cancer-associated VTE with platelet counts between 25 and 50 G/L, two possible strategies can be followed, depending on the underlying risk of VTE progression: (1) for patients with high-risk VTE features such as symptomatic, proximal PE, proximal DVT or history of previous thrombotic events who are in the acute phase of VTE (<30 days since diagnosis of the event), full-dose anticoagulation with LMWH or unfractionated heparin (UFH) in combination with platelet transfusion to keep thrombocyte counts above 40–50 G/L is suggested. (2) For patients with low risk of VTE progression (distal DVT, incidental subsegmental PE, CRT and/or patients in the subacute phase of VTE (>30 days since event), anticoagulation with half-therapeutic or prophylactic dose of LMWH is suggested. Anticoagulation should be withheld temporarily for patients with platelet counts below 25 G/L and resumed in full dose above 50 G/L. There is no clear guidance for use of DOAC in patients with severe thrombocytopaenia. As supported by the Hokusai VTE cancer study, DOAC can be used in patients with a platelet count >50 G/L and should be paused when platelets drop <50 G/L, and switching to LMWH should be considered on a case-by-case basis.24 25 In patients with acute VTE and absolute contraindications for anticoagulation, including active haemorrhage and severe, prolonged, transfusion resistant thrombocytopaenia, retrievable inferior vena cava (IVC) filters might be considered in very selected cases until anticoagulation can safely be initiated, depending on type of VTE event (ie, proximal DVT) and time since diagnosis, with increasing interval arguing against its usage.25 The authors suggest a conservative approach towards the application of IVC filters in this setting.

Drug–drug interactions

Risk of bleeding and efficacy can further be influenced by drug–drug interactions. DOAC are substrates of the P-glycoprotein (P-gp) and are in part metabolised through CYP3A4. Therefore, their plasma levels can be altered by a variety of drugs, especially some anti-cancer drugs that interfere with CYP3A4 or P-gp. LMWH is the suggested alternative in patients with suspected strong interactions, as patients with concomitant use of certain drugs that strongly influence CYP3A4 or P-gp have been excluded from inclusion to Hokusai VTE cancer and SELECT-D.26 Drugs that might influence efficacy and safety of DOAC comprise both antineoplastic agents (eg, paclitaxel, certain tyrosine kinase inhibitors, bicalutamide, enzalutamide, abiraterone) and supportive care medication (eg, dexamethasone, prednisone, azol antifungals, neurokinin-1 antagonists). However, the clinical impact of these suspected interactions is unknown as pharmacodynamics, pharmacokinetic and clinical outcome studies are not available.27

Nausea and vomiting

A common problem in patients with cancer that might also contribute to alterations in DOAC bioavailability is nausea and vomiting, influencing uptake of orally administered drugs. It is suggested to optimise antiemetic therapy and to temporarily switch anticoagulation to parenteral routes (eg, LMWH) in the case of both prolonged vomiting episodes (>24 hours) or acute episodes (<24 hours) and to resume DOACs after the episodes have resolved.28 This scheme might ensure consistent anticoagulation in a patients with VTE and nausea and vomiting.

Altered GI anatomy

Patients with cancer often undergo surgical procedures which result in anatomic changes and might affect GI absorption of DOAC, such as total/partial gastrectomy, Roux-en-Y gastric bypass (RYB) or colectomy.29 Because of the location of absorption of rivaroxaban, which is predominantly in the stomach, total gastrectomy might affect its bioavailability.30 The effect of partial gastrectomy and RYB are uncertain. Apixaban is mainly absorbed in the proximal small bowel.31 Its bioavailability is probably less affected by gastric resection and possibly reduced by colectomy. Levels of edoxaban, on the other hand, are possibly reduced in patients with a history of gastric resection and RYB and unlikely affected in patients who underwent colectomy.29 32 In the absence of dedicated studies assessing the efficacy of DOAC after GI resection or bypass in patients with cancer, most statements are based on extrapolations from data on physiology of DOAC absorption, case reports and case series of patients mainly in bariatric surgery.29 33

Duration of anticoagulation

Patients with cancer-associated VTE and indication for anticoagulation should preferably be treated for 6 months.22 Duration of anticoagulation should be extended in patients with ongoing active malignancy, for example in metastatic disease, and/or ongoing antineoplastic therapy and should be reassessed periodically for its risk–benefit ratio.21–23 34

Special clinical events

In addition to balancing risk of bleeding and recurrent VTE according to underlying risk factors such as type of malignancy, low platelet counts and potential drug–drug interactions, some special clinical scenarios of cancer-associated thrombosis need special consideration prior to deciding on a specific therapeutic approach.

Catheter-related thrombosis

Patients with cancer often receive antineoplastic therapy via CVCs, which pose a significant risk of CRT. However, evidence on therapeutic management of these events is limited. A retrospective multi-centre analysis of therapeutic strategies in CRT revealed that treatment approaches in clinical practice are heterogeneous, with a significant percentage of patients (16%) who do not receive anticoagulation but rather are being treated with removal of the catheter alone.35 According to guidelines, patients with CRT are suggested to undergo anticoagulation for a minimum of 3 months and as long as the CVC remains in situ.22 Preferably, LMWH should be used, as only very limited data exist on the use of DOACs in this setting. Based on expert opinion, the catheter may remain in place in the absence of infection, malfunction and incorrect positioning, and if it is planned to continue to use the catheter. Removal of the CVC without the administration of anticoagulation can be done in patients with high risk of or active bleeding.36

Incidental PE

Due to regular restaging procedures in patients with cancer, incidentally detected, asymptomatic PE is a dilemma to manage. Despite its asymptomatic presentation, cohort studies found a similar risk of recurrence in patients with incidental VTE compared with symptomatic events.37 Therefore, anticoagulation should be initiated in patients with (i) clinical features of VTE who then cannot be classified as truly ‘asymptomatic’, (ii) PE involving the segmental or more proximal pulmonary arterial branches or multifocal subsegmental branches and (iii) PE of a single subsegmental branch who are found to have accompanying asymptomatic proximal DVT. For patients with isolated, subsegmental, asymptomatic, independently confirmed PE with or without distal DVT, a case-by-case decision (anticoagulation or close clinical monitoring) can be followed.38

Splanchnic vein thrombosis

Thrombosis of the portal, mesenteric or splenic veins, summarised as SVT, either symptomatic or incidentally discovered on imaging procedures, can complicate management in patients with cancer, with the highest risk in patients with hepatocellular cancer, cholangiocarcinoma or pancreatic cancer.39 Furthermore, direct tumour ingrowth into splanchnic vessels, also known as tumour thrombus, consisting solely on antineoplastic proliferation, most frequently observed in renal cell carcinoma, might mimic ‘real’ SVT.40 In patients with confirmed symptomatic SVT, excluding tumour thrombus by imaging, continuous anticoagulation with LMWH over VKA is recommended for a minimum of 3 months or longer for patients with persistent thrombotic risk.41 42 For patients with asymptomatic SVT, anticoagulation remains controversial, with only very limited data to support either strategy. Therefore, a case-by-case decision considering individual risk of bleeding and recurrence has to be made.23 41 42


Clinical management of cancer-associated thrombosis is a challenging, multilayered issue. The therapeutic armamentarium has broadened by emerging evidence on the safety and efficacy of DOAC (figure 1). In general, we treat patients with active cancer and newly diagnosed symptomatic or incidental VTE with DOAC, if they (i) are not at an increased risk of bleeding from the GI or GU tract and do not have high-grade thrombocytopenia (<50 G/L) and (ii) do not currently or are not suspected in the near future to be treated with medication that strongly interact with DOAC. As an alternative, we use LMWH. However, VKA might be a reasonable alternative, if DOAC and LMWH are contraindicated or unavailable. All treatment decisions should be made in the sense of informed shared decision-making after educating the patient. Treatment should regularly be clinically monitored to optimise therapy adherence, and toevaluate issues that could affect bioavailability, for example, due to nausea and vomiting or newly administered drugs that might have a potential for drug–drug interactions. We treat patients with cancer-associated DVT and PE preferably for 6 months, and continue anticoagulation for secondary prevention of VTE, if cancer is not in complete remission and/or anti-cancer therapies are ongoing.


View Abstract


  • Contributors FM and CA have written the manuscript. FM, IP and CA have revised the manuscript and approved the final version for publication.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.