Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Review - Intervention

Oral direct thrombin inhibitors or oral factor Xa inhibitors versus conventional anticoagulants for the treatment of pulmonary embolism

Authors' declarations of interest

Version published: 14 April 2023 Version history

https://doi.org/10.1002/14651858.CD010957.pub3
Collapse all Expand all

Background

Pulmonary embolism (PE) is a potentially life‐threatening condition in which a clot can migrate from the deep veins, most commonly in the leg, to the lungs. Conventional treatment of PE used unfractionated heparin (UFH), low molecular weight heparin (LMWH), fondaparinux, and vitamin K antagonists (VKAs). Recently, two forms of direct oral anticoagulants (DOACs) have been developed: oral direct thrombin inhibitors (DTIs) and oral factor Xa inhibitors. DOACs have characteristics that may be favourable to conventional treatment, including oral administration, a predictable effect, no need for frequent monitoring or re‐dosing, and few known drug interactions. This review reports the efficacy and safety of these drugs in the long‐term treatment of PE (minimum duration of three months). This is an update of a Cochrane Review first published in 2015. 

Objectives

To assess the efficacy and safety of oral DTIs and oral factor Xa inhibitors versus conventional anticoagulants for the long‐term treatment of PE.

Search methods

The Cochrane Vascular Information Specialist searched the Cochrane Vascular Specialised Register, CENTRAL, MEDLINE, Embase and CINAHL databases, the World Health Organization International Clinical Trials Registry Platform and the ClinicalTrials.gov trials registers to 2 March 2022. We checked the reference lists of relevant articles for additional studies.

Selection criteria

We included randomised controlled trials (RCTs) in which people with a PE confirmed by standard imaging techniques were allocated to receive an oral DTI or an oral factor Xa inhibitor compared with a conventional anticoagulant or compared with each other for the long‐term treatment of PE (minimum duration three months).

Data collection and analysis

We used standard Cochrane methods. Our primary outcomes were recurrent PE, recurrent venous thromboembolism (VTE), and deep vein thrombosis (DVT). Secondary outcomes were all‐cause mortality, major bleeding, and health‐related quality of life. We used GRADE to assess the certainty of evidence for each outcome.

Main results

We identified five additional RCTs with 1484 participants for this update. Together with the previously included trials, we have included ten RCTs with a total of 13,073 participants. Two studies investigated an oral DTI (dabigatran) and eight studies investigated oral factor Xa inhibitors (three rivaroxaban, three apixaban, and two edoxaban). The studies were of good methodological quality overall. 

Meta‐analysis showed no clear difference in the efficacy and safety of oral DTI compared with conventional anticoagulation in preventing recurrent PE (odds ratio (OR) 1.02, 95% confidence interval (CI) 0.50 to 2.04; 2 studies, 1602 participants; moderate‐certainty evidence), recurrent VTE (OR 0.93, 95% CI 0.52 to 1.66; 2 studies, 1602 participants; moderate‐certainty evidence), DVT (OR 0.79, 95% CI 0.29 to 2.13; 2 studies, 1602 participants; moderate‐certainty evidence), and major bleeding (OR 0.50, 95% CI 0.15 to 1.68; 2 studies, 1527 participants; moderate‐certainty evidence). We downgraded the certainty of evidence by one level for imprecision due to the low number of events.

There was also no clear difference between the oral factor Xa inhibitors and conventional anticoagulation in the prevention of recurrent PE (OR 0.92, 95% CI 0.66 to 1.29; 3 studies, 8186 participants; moderate‐certainty evidence), recurrent VTE (OR 0.83, 95% CI 0.66 to 1.03; 8 studies, 11,416 participants; moderate‐certainty evidence), DVT (OR 0.77, 95% CI 0.48 to 1.25; 2 studies, 8151 participants; moderate‐certainty evidence), all‐cause mortality (OR 1.16, 95% CI 0.79 to 1.70; 1 study, 4817 participants; moderate‐certainty evidence) and major bleeding (OR 0.71, 95% CI 0.36 to 1.41; 8 studies, 11,447 participants; low‐certainty evidence); the heterogeneity for major bleeding was significant (I2 = 79%). We downgraded the certainty of the evidence to moderate and low because of imprecision due to the low number of events and inconsistency due to clinical heterogeneity. None of the included studies measured health‐related quality of life.

Authors' conclusions

Available evidence shows there is probably little or no difference between DOACs and conventional anticoagulation in the prevention of recurrent PE, recurrent VTE, DVT, all‐cause mortality, and major bleeding. The certainty of evidence was moderate or low. Future large clinical trials are required to identify if individual drugs differ in effectiveness and bleeding risk, and to explore effect differences in subgroups, including people with cancer and obesity.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Are direct oral anticoagulants (a type of 'blood thinner') better than traditional anticoagulants for treating a pulmonary embolism (a blood clot in the lung)?

What is a pulmonary embolism?

A pulmonary embolism occurs when a piece of blood clot breaks off from a clot somewhere else in the body and travels in the blood to the lungs. This can be life‐threatening and occurs in approximately 4 to 12 per 10,000 people. The chances of getting a pulmonary embolism can increase with risk factors, including previous clots, prolonged periods of immobility (such as travelling on aeroplanes or taking bed rest), cancer, exposure to oestrogens (pregnancy, oral contraceptives, or hormone replacement therapy), blood disorders (thrombophilia), and trauma. 

How is a pulmonary embolism treated?

Until recently, the standard treatment for a pulmonary embolism was an anticoagulant: a medicine that either treats or prevents blood clots, often called a 'blood thinner'. Conventional anticoagulants include heparin, fondaparinux, and vitamin K antagonists. However, these drugs can cause side effects and have limitations. 

Two types of anticoagulant have been developed: direct thrombin inhibitors (DTIs) and factor Xa inhibitors. These anticoagulants are given orally (that is, by mouth, in the form of a pill), have a predictable effect, do not require frequent monitoring or re‐dosing (taking multiple doses), and have few known interactions with other medicines. For these reasons, direct oral anticoagulants have become the medicines of choice for treating DVT.

What did we want to find out?

We wanted to find out if direct oral anticoagulants are useful and safe for treating people with a pulmonary embolism, compared with conventional anticoagulants. We looked at whether 3 months' treatment or longer prevented further blood clots (recurrent deep vein thrombosis (DVT), when a clot forms in a deep vein, usually in the leg), recurrent pulmonary embolism, and pulmonary embolisms. The main safety outcomes included death and unwanted, harmful adverse events, such as major bleeding.

What did we do?

We searched for studies in which people with a pulmonary embolism confirmed by standard imaging techniques were randomly allocated to one of two treatment groups. These types of studies give the most reliable evidence about treatment effects. People in the experimental groups received an oral DTI or an oral factor Xa inhibitor, and their results were compared to the results of people given conventional anticoagulation. All participants were given long‐term treatment of pulmonary embolism (a minimum duration of 3 months). 

What did we find?

After searching for relevant studies, we included 10 studies with a combined total of 13,073 participants. Studies compared oral DTIs and factor Xa inhibitors with conventional anticoagulation. We combined the data from the studies and found that there was no clear difference in the incidence of:

‐ recurrent pulmonary embolism;
‐ recurrent deep vein thrombosis (DVT: when a blood clot forms, usually in a deep vein of the leg or pelvis);
‐ recurrent venous thromboembolism (when DVT and pulmonary embolism occur together);
‐ death; 
‐ major bleeding

This review showed that there was no clear difference between the direct oral anticoagulants and conventional treatment in preventing recurrent PE, recurrent VTE, DVT, mortality, and major bleeding. No study measured health‐related quality of life.

What are the limitations of the evidence?

We are moderately confident in this evidence. This was because the number of events involved in the studies was small and there were differences in how individual studies were carried out.

How up to date is this evidence?

This review updates a previous Cochrane Review. The evidence is up to date to March 2022.

Key messages

Current evidence shows there is probably little or no difference between direct oral anticoagulants and conventional anticoagulation for preventing recurrent pulmonary embolism, recurrent venous thromboembolism (VTE), deep vein thrombosis (DVT), all‐cause mortality, and major bleeding in people who are being treated for a pulmonary embolism.

Authors' conclusions

Implications for practice

Available evidence shows there is little or no difference between direct oral anticoagulants (DOACs) and conventional anticoagulation in the prevention of recurrent pulmonary embolism (PE), recurrent venous thromboembolism (VTE), deep vein thrombosis (DVT), all‐cause mortality, and major bleeding. According to GRADE criteria, the certainty of evidence was moderate to low. DOACs, such as direct thrombin inhibitors (DTIs) and factor Xa inhibitors, may therefore be an alternative to conventional anticoagulation treatment for acute pulmonary embolism. One potential practical advantage of DOACs over conventional anticoagulants is their ease of use due to fixed doses and no need for routine monitoring with blood tests.

Implications for research

There is some evidence of wide inter‐individual variation in anticoagulant effect from the fixed doses of DOACs as currently prescribed. This may be of clinical importance, and further research is needed to compare DOACs directly with one another to see which one is most effective and safe, specifically in relation to bleeding risk in various subgroups, such as in people with cancer or obesity. Such studies would need to be very large and may not be considered financially viable to sponsoring drug companies. Any impact on the decision to use extended phase anticoagulation and interruption of procedures with DOAC use should also be investigated. Further research is also required to establish other factors associated with the use of DOACs, such as adherence, quality of life, cost‐effectiveness, and tolerability. Finally, research is required in categories of venous thrombosis not specifically examined in the studies included here, such as in people with malignancy or a thrombophilic abnormality, as well as travel‐associated venous thrombosis.

Summary of findings

Open in table viewer
Summary of findings 1. Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism

Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism

Patient or population: people with a pulmonary embolism, confirmed by standard imaging techniques
Setting: hospital
Intervention: oral DTIs
Comparison: conventional anticoagulation

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with conventional anticoagulation

Risk with oral DTI

Recurrent PEa

 

Follow‐up: 

6 months

Study population

OR 1.02
(0.50 to 2.04)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

20 per 1000

20 per 1000
(10 to 40)

Recurrent VTEc

 

Follow‐up: 

6 months

Study population

OR 0.93
(0.52 to 1.66)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

31 per 1000

29 per 1000
(16 to 50)

DVTd

 

Follow‐up: 

6 months

Study population

OR 0.79
(0.29 to 2.13)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

11 per 1000

9 per 1000
(3 to 23)

All‐cause mortality

See comment

See comment

See comment

‐

RE‐COVER 2009 and RE‐COVER II 2014 did not report on all‐cause mortality.

Major bleedinge

 

Follow‐up: 

6 months

Study population

OR 0.50
(0.15 to 1.68)

1527
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

10 per 1000

5 per 1000
(2 to 17)

Health‐related quality of life

See comment

See comment

See comment

‐

RE‐COVER 2009 and RE‐COVER II 2014 did not measure health‐related quality of life.

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: confidence interval; CTPA: computed tomographic pulmonary angiography; DVT: deep vein thrombosis; ISTH: International Society on Thrombosis and Haemostasis; OR: odds ratio; PE: pulmonary embolism; RCT: randomised controlled trial; V/Q: ventilation/perfusion; VTE: venous thromboembolism

GRADE Working Group grades of evidence
High certainty: we are very confident that the true effect lies close to that of the estimate of the effect
Moderate certainty: we are moderately confident in the effect estimate, the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low certainty: our confidence in the effect estimate is limited, the true effect may be substantially different from the estimate of the effect
Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

aConfirmed by V/Q lung scanning, pulmonary angiography, or CTPA
bWe downgraded one level for imprecision due to the low number of events. The possibility of publication bias is not excluded but we did not consider it sufficient to downgrade the certainty of evidence.
cVTE includes clinically overt DVT and PE. Clinically overt DVT, confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound); or clinically overt PE, confirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
dClinically overt DVT confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound). 
eAs defined by the ISTH (Schulman 2005): 1. Fatal bleeding, and/or 2. symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra‐articular or pericardial, or intramuscular with compartment syndrome, and/or 3. bleeding causing a fall in haemoglobin level of 20 g/L (1.24 mmol/L) or more, or leading to transfusion of two or more units of whole blood or packed red cells.

Open in table viewer
Summary of findings 2. Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism

Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism

Patient or population: people with a pulmonary embolism, confirmed by standard imaging techniques
Setting: hospital
Intervention: oral factor Xa inhibitors
Comparison: conventional anticoagulation

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with conventional  anticoagulation

Risk with oral factor Xa inhibitors

Recurrent PEa

 

Follow‐up: 

0 to 12 months

Study population

OR 0.92
(0.66 to 1.29)

8186
(3 RCTs)

⊕⊕⊕⊝

Moderateb 

‐

18 per 1000

16 per 1000
(12 to 23)

Recurrent VTEc

 

Follow‐up: 

0 to 12 months

Study population

OR 0.83
(0.66 to 1.03)

11,416
(8 RCTs)

⊕⊕⊕⊝
Moderateb

 

2 of 8 studies reported no events

32 per 1000

26 per 1000
(21 to 33)

DVTd

 

Follow‐up: 

5 days to 12 months

Study population

OR 0.77
(0.48 to 1.25)

8151
(2 RCTs)

⊕⊕⊕⊝
Moderateb

‐

10 per 1000

7 per 1000
(5 to 12)

All‐cause mortality

 

Follow‐up: 

0 to 12 months

Study population

OR 1.16
(0.79 to 1.70)

4817
(1 RCT)

⊕⊕⊕⊝
Moderateb

‐

21 per 1000

24 per 1000
(16 to 35)

Major bleedinge

 

Follow‐up: 

0 to 12 months

Study population

OR 0.71
(0.36 to 1.41) 

11,447
(8 RCTs)

⊕⊕⊝⊝
Lowb,f

2 of 8 studies reported no events

23 per 1000

16 per 1000
(8 to 32)

Health‐related quality of life

See comment

See comment

See comment

‐

The studies did not measure health‐related quality of life.

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: confidence interval; CTPA: computed tomographic pulmonary angiography; DVT: deep vein thrombosis; ISTH: International Society on Thrombosis and Haemostasis; OR: odds ratio; PE: pulmonary embolism; RCT: randomised controlled trial; V/Q: ventilation/perfusion; VTE: venous thromboembolism

GRADE Working Group grades of evidence
High certainty: we are very confident that the true effect lies close to that of the estimate of the effect
Moderate certainty: we are moderately confident in the effect estimate, the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low certainty: our confidence in the effect estimate is limited, the true effect may be substantially different from the estimate of the effect
Very low certainty: we have very little confidence in the effect estimate, the true effect is likely to be substantially different from the estimate of effect

aConfirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
bWe downgraded one level for imprecision due to the low number of events. The possibility of publication bias is not excluded but we did not consider it sufficient to downgrade the certainty of evidence.
cVTE includes clinically overt DVT and PE. Clinically overt DVT, confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound); or clinically overt PE, confirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
dClinically overt DVT confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound). 
eAs defined by the ISTH (Schulman 2005): 1. Fatal bleeding, and/or 2. symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra‐articular or pericardial, or intramuscular with compartment syndrome, and/or 3. bleeding causing a fall in haemoglobin level of 20 g/L (1.24 mmol/L) or more, or leading to transfusion of two or more units of whole blood or packed red cells.
fWe downgraded one level for serious inconsistency (I2 = 79% due to clinical heterogeneity).

Background

Description of the condition

Pulmonary embolism (PE) is a potentially life‐threatening condition in which a blood clot blocks the supply of blood to the lungs. PE is often a consequence of a thrombus in the deep veins of the legs (deep vein thrombosis, DVT) that dislodges and migrates in the blood to the pulmonary arteries. The incidence of PE has been estimated as 4 to 12 per 10,000 people annually (Keller 2020; Konstantinides 2020; Wendelboe 2016), and longitudinal studies have revealed that the annual PE prevalence rates tend to rise over time (Dentali 2016; de Miguel‐Diez 2014; Keller 2020; Lehnert 2018). DVT is present in approximately 70% to 80% of people with PE, yet only 15% of PE cases have symptoms of DVT (Huerta 2007). One long‐term complication of PE is chronic thromboembolic pulmonary hypertension (CTPH). CTPH occurs when the clot obstructs the pulmonary arteries, causing excessive pressure in the pulmonary artery and stress to the right ventricle. CTPH is uncommon but it can result in heart failure (NICE 2020). It is estimated that up to a quarter of all PE patients present with sudden death (Heit 2015). In the USA, PE is one of the leading causes of cardiovascular mortality, contributing to nearly 300,000 deaths annually (Wendelboe 2016).

Risk factors for PE are similar to those for DVT and are classified as provoked or unprovoked (Kearon 2012). For unprovoked PE, no clear precipitating risk factor can be identified; risk factors are either hereditary or more often acquired. For provoked PE, risk factors include cancer, acute medical illness, surgery, trauma, immobility (often in hospital and lasting at least three days), obesity, inflammatory diseases/infection, hormone therapy (oestrogen‐containing), pregnancy (particularly the postpartum period), long‐distance travel, recent hospitalisation, and antiphospholipid syndrome (APS) (Kakkos 2021).

People presenting with signs or symptoms of PE, such as chest pain, shortness of breath, or coughing up blood, will have their general medical history assessed, undergo a physical examination, and may be offered a chest X‑ray to exclude other causes (NICE 2020). However, it can be particularly challenging to diagnose PE as the symptoms (dyspnoea, pleuritic chest pain, retrosternal chest pain, cough, and haemoptysis) are not specific (NICE 2020). In severe cases, right ventricle failure leads to dizziness, syncope, tachypnoea, tachycardia, hypoxia, elevated jugular venous pressure, systemic hypotension, and cardiogenic shock (NICE 2020). The UK National Institute for Health and Care Excellence (NICE) recommends that people presenting with suspected PE should be assessed using a two‐level PE Wells score (NICE 2020; Wells 2000). Points are awarded for the presence of clinical features, including clinical signs and symptoms of DVT (minimum of leg swelling and pain with palpation of the deep veins), heart rate greater than 100 beats per minute, immobilisation for more than three days or surgery in the previous four weeks, previous DVT/PE, haemoptysis, and malignancy (on treatment, treated in the last six months, or palliative) (NICE 2020). For people with a low pre‐test probability, the use of a D‐dimer assay combined with a clinical prediction rule has a high negative predictive value and avoids the need for unnecessary imaging (Qaseem 2007). However, for people who have an intermediate or high pre‐test probability of PE, imaging is essential. People with a score of greater than four are judged to be likely to have had a PE and should undergo immediate diagnostic imaging. If this cannot be performed immediately, patients should be given immediate interim parenteral anticoagulant therapy until the imaging test is done. Those with a negative diagnosis in whom a DVT is likely should be given a proximal leg vein ultrasound scan (NICE 2020).

There are three types of imaging techniques used to diagnose PE: computed tomography pulmonary angiogram (CTPA), ventilation‐perfusion (V/Q) scan, and pulmonary angiography.

CTPA involves injecting a contrast agent intravenously and performing a computed tomography (CT) scan of the chest to visualise the pulmonary arteries and detect any thrombi in the pulmonary arteries down to the subsegmental branches. The procedure has over 90% specificity and sensitivity in diagnosing PE in the main, lobar, and segmental pulmonary arteries (Riedel 2004). However, the radiation dose administered to the individual is much larger than that of a V/Q scan, and thus people who have a CTPA may be at an increased lifetime risk of cancer (Anderson 2009). CTPA use is limited in people with iodine allergy, hyperthyroidism, those who are pregnant and breastfeeding, and is contraindicated in people with severe renal failure (Konstantinides 2020).

A V/Q scan comprises two parts: the ventilation part, where the patient breathes in a radioisotope (in the form of a gas or an aerosol) and the perfusion part, where the patient is given an intravenous injection of the isotope. A gamma camera is used to detect where the isotopes are in the lungs and the images show which areas of the lungs are ventilated but not perfused (NICE 2020). Another version of this test, the V/Q single photon emission computed tomography (V/Q SPECT) has been developed. The camera is rotated around the patient, thus generating three‐dimensional images and leading to a more accurate diagnosis (Laurence 2012).

Pulmonary angiography is an imaging method that uses x‐rays and a special dye to see how blood flows through the lung. The procedure is invasive, with related risks of mortality and complications, especially for people with haemodynamic compromise or respiratory failure (Stein 1992). For several decades, pulmonary angiography was the "gold standard" for diagnosing or ruling out acute PE, but it is now rarely used because less‐invasive CTPA provides comparable diagnostic accuracy (Konstantinides 2020).

Description of the intervention

Conventional treatment of PE used unfractionated heparin (UFH), low molecular weight heparin (LMWH), fondaparinux, or vitamin K antagonists (VKAs). These drugs block the action of thrombin either by "activating naturally occurring thrombin inhibitors or by inhibiting specific factors in the coagulation system that subsequently impact on thrombin generation or activity" (Weitz 2003). Although heparins and VKAs are effective anticoagulants, there are limitations associated with each. LMWH must be administered parenterally and may be associated with an increased risk of bleeding, loss of bone strength, elevated liver enzymes, and heparin‐induced thrombocytopenia (Konstantinides 2020). Meanwhile, VKAs have a narrow therapeutic window, require frequent monitoring and dosage adjustments, and can have multiple interactions with other drugs (Ageno 2012).

Two further classes of oral anticoagulants have been developed: direct thrombin inhibitors (DTIs) and factor Xa inhibitors. DTIs and factor Xa inhibitors have characteristics that may be favourable to heparin and VKAs, including ease of oral administration, a predictable effect, lack of frequent monitoring or re‐dosing, and fewer known drug interactions (compared with VKA) (Fox 2012).  

The latest American College of Chest Physicians (ACCP) guidelines recommended apixaban, dabigatran, edoxaban, or rivaroxaban over VKAs as treatment‐phase (first three months) anticoagulant therapy for people with VTE (DVT of the leg or PE); and recommended an oral Xa inhibitor (apixaban, edoxaban, rivaroxaban) over LMWH for the initiation and treatment phases of therapy for acute VTE associated with cancer (Kearon 2016; Stevens 2021). Similarly, the 2019 European Society of Cardiology (ESC) guidelines recommended DOACs in preference to VKAs in eligible individuals ready for an oral anticoagulant (Konstantinides 2020). The NICE 2020 guidelines recommend offering either apixaban or rivaroxaban as initial choices, and suggested other regimens only for people suitable for neither: consider a DOAC for people with active cancer and confirmed proximal DVT or PE, and consider other strategies when DOAC is unsuitable. 

According to research by Lutsey and colleagues, the use of DOACs (especially rivaroxaban and apixaban) to treat VTE has increased dramatically in the USA since the US Food and Drug Administration (FDA) approved them for this application (Lutsey 2019). Drawing on individual health insurance data for 2012 to 2017, they found that DOACs accounted for less than 2% of the prescriptions for VTE treatment at the beginning of 2012 and increased to more than 80% by the fourth quarter of 2017 (Lutsey 2019).

How the intervention might work

Oral direct thrombin inhibitors

DTIs work by binding directly to the enzyme thrombin without the need for a co‐factor, such as antithrombin. Unlike heparins and VKAs, DTIs can inhibit both soluble thrombin and fibrin‐bound thrombin (Kam 2005). Other advantages include a more predictable anticoagulant effect because of their lack of binding to other proteins, lack of an antiplatelet effect, and no suspected concern of heparin‐induced thrombocytopenia (HIT) (Lee 2011). There are several types of DTIs.

Dabigatran

Dabigatran etexilate is a reversible oral DTI that is metabolised to its active ingredient, dabigatran, in the gastrointestinal tract (Ageno 2012). It does not require anticoagulation monitoring, is excreted by the kidneys, and has a half‐life of 12 to 17 hours. As well as a treatment for venous thrombosis, this drug has been involved in many large randomised studies of atrial fibrillation (Connolly 2009), acute coronary syndromes (Oldgren 2011), prevention of thrombosis following orthopaedic surgery (Eriksson 2007), and in people with mechanical heart valves (Van de Werf 2012). In common with the other DOACs, dabigatran is associated with a lower incidence of intracranial haemorrhage (compared with VKAs). However, again compared with VKAs, dabigatran showed a higher incidence of indigestion and heartburn and a higher incidence of gastrointestinal bleeding (Baetz 2008). For the treatment of PE in people with creatinine clearance (CrCL) of more than 30 mL/min, the recommended dosage of dabigatran is 150 mg twice daily following at least five days of initial therapy with a parenteral anticoagulant. For people aged 80 years or older and for people having verapamil, the recommended dose is 110 mg twice daily. In people aged 75 to 80 years, people with moderately reduced kidney function, people with gastritis, oesophagitis or gastro‐oesophageal reflux, and people at increased risk of bleeding, either dose (300 mg or 220 mg) can be given, based on an individual assessment. Dabigatran is not recommended in people with CrCL of less than 30 mL/min (Konstantinides 2020; NICE 2020). 

Ximelagatran

Ximelagatran is a prodrug that is metabolised to melagatran as it is better absorbed from the gastrointestinal tract (Kam 2005). It has a plasma half‐life of three hours, has a predictable response after oral administration, and does not require coagulation monitoring. Ximelagatran was found to be effective in the treatment of venous thromboembolism but caused unacceptable liver toxicity (Boudes 2006), and, therefore, was never licensed. For the treatment of PE, the usual dosing of ximelagatran for adults is 48 mg twice daily (Ho 2006).

Oral factor Xa inhibitors

Factor Xa inhibitors bind directly to the active site of factor Xa, thus blocking the activity of the clotting factor. Unlike indirect factor Xa inhibitors such as fondaparinux, direct factor Xa inhibitors "inactivate free FXa and FXa incorporated with the prothrombinase complex equally well" and do not require interaction with the inhibitor antithrombin (Eriksson 2009). They have been shown to be non‐inferior to VKAs but without the need for regular blood test monitoring. They appear to have fewer drug interactions (compared with VKAs) and no food or alcohol interactions.

Rivaroxaban

Rivaroxaban is a reversible direct factor Xa inhibitor. The plasma half‐life is estimated to be eight to 10 hours if renal function is normal (Spyropoulos 2012). For the treatment of PE, the recommended dosage of rivaroxaban is 15 mg twice daily for the first 21 days followed by 15 mg once daily in people with moderate (CrCL 30 mL/min to 49 mL/min) or severe (CrCL 15 mL/min to 29 ml/min) renal impairment, if their risk of bleeding outweighs the risk of recurrent DVT or PE (NICE 2020). 

Apixaban

Apixaban is an oral, small molecule, reversible inhibitor of factor Xa with a plasma half‐life of eight to 15 hours. For the treatment of PE, the recommended dosage of apixaban is 10 mg twice daily for the first seven days followed by 5 mg twice daily for continued treatment (NICE 2020). For people with severe renal impairment (CrCL 15 mL/min to 29 mL/min), apixaban should be used with caution (Eriksson 2009). 

Betrixaban

Betrixaban is an orally administered direct factor Xa inhibitor. It has a half‐life of 19 to 27 hours (Palladino 2013). For the prophylaxis of PE, the recommended dose of betrixaban is an initial single dose of 160 mg starting on day 1, followed by 80 mg once daily, taken for 35 to 42 days at the same time each day with food. For people with severe renal impairment (CrCL 15 mL/min to 30 mL/min computed by Cockcroft‐Gault using actual body weight), the recommended dose of betrixaban is an initial single dose of 80 mg followed by 40 mg once daily (FDA 2017). 

Edoxaban

Edoxaban is an oral direct inhibitor of activated factor X that is rapidly absorbed with a half‐life of nine to 11 hours. Edoxaban has a dual mechanism of elimination with one‐third eliminated via the kidneys and the remainder excreted in the faeces (Eikelboom 2010). For the treatment of PE, the recommended dose of edoxaban is 60 mg taken orally once daily following at least five days of initial therapy with a parenteral anticoagulant. The edoxaban dose should be reduced to 30 mg once daily in people with CrCL 15 mL/min to 50 mL/min, those who weigh less than or equal to 60 kg, or those who are taking certain concomitant P‐glycoprotein inhibitor medications (NICE 2020).

Why it is important to do this review

The efficacy and safety of oral DTIs and oral factor Xa inhibitors for the treatment of VTE has been examined in several randomised controlled trials (the EINSTEIN‐PE study (EINSTEIN‐PE 2012), the ODIXa‐DVT study (Agnelli 2007), the Botticelli study (Botticelli Investigators 2008), the AMPLIFY study (Agnelli 2013), the RE‐COVER II study (Schulman 2011), and the THRIVE studies (Eriksson 2003)). Several non‐Cochrane systematic reviews have examined the benefits and harms of DTIs and factor Xa inhibitors versus VKAs in the treatment of acute VTE (Fox 2012; Samaranayake 2022; Song 2021; Wu 2022). However, specific data on people with PE were not reported, and appropriate subgroup analyses that considered some clinical factors (such as the history of VTE, age, and active cancer) were lacking. It is important to assess the efficacy and safety of oral DTIs and oral factor Xa inhibitors for the treatment of PE to help both healthcare professionals and people with PE choose the most appropriate treatment. Since the first version of this review was completed (Robertson 2015a), several new RCTs have been published (AMPLIFY‐J 2015; Caravaggio 2020; Hokusai VTE Cancer 2018; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018). Thus, it is necessary to update the review and present the most up‐to‐date evidence to aid decision‐making in clinical practice. For this update, due to the limited number of included studies, subgroup analyses were only possible for active cancer and the different types of factor Xa inhibitors.

Objectives

To assess the efficacy and safety of oral DTIs and oral factor Xa inhibitors versus conventional anticoagulants for the long‐term treatment of PE.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) in which participants with a confirmed pulmonary embolism (PE) were allocated to receive an oral direct thrombin inhibitor (DTI) or an oral factor Xa inhibitor for the treatment of PE. We included studies published and in progress if preliminary results were available. We placed no restrictions on publication status and non‐English studies were eligible for inclusion in the review. We excluded DTIs and factor Xa inhibitors that were not given by the oral route.

Types of participants

We included participants with a PE, confirmed by standard imaging techniques (CTPA, V/Q scan, or pulmonary angiography).

Types of interventions

We included the following interventions:

  • oral DTIs (e.g. dabigatran, ximelagatran) (although ximelagatran was withdrawn from the market in 2006 due to safety issues, we included it in the review to make the results as comprehensive as possible);

  • oral factor Xa inhibitors (e.g. rivaroxaban, apixaban, betrixaban, edoxaban);

  • other anticoagulants (e.g. LMWH, UFH, VKAs).

We included the following comparisons:

  • one oral DTI versus another oral DTI;

  • one oral factor Xa inhibitor versus another oral factor Xa inhibitor;

  • oral DTI versus oral factor Xa inhibitor;

  • oral DTI or oral factor Xa inhibitor versus another anticoagulant.

Treatment had to be for a minimum duration of three months as this is conventional anticoagulation practice for a PE.

Types of outcome measures

Primary outcomes

  • Recurrent PE, confirmed by standard imaging techniques (CTPA, V/Q scan, or pulmonary angiography)

  • Recurrent venous thromboembolism (VTE, clinically overt DVT, confirmed by standard imaging techniques, including venography, impedance plethysmography, whole‐leg compression ultrasound, or proximal compression ultrasound; or clinically overt PE, confirmed by CTPA, V/Q scan, or pulmonary angiography)

  • Clinically overt DVT, confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, or proximal compression ultrasound).

Secondary outcomes

  • All‐cause mortality

  • Adverse effects of treatment, including major bleeding (as defined by the International Society on Thrombosis and Haemostasis (ISTH); Schulman 2005):

    • fatal bleeding;

    • symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra‐articular or pericardial, or intramuscular with compartment syndrome;

    • bleeding causing a fall in haemoglobin level of 20 g/L (1.24 mmol/L) or more, or leading to transfusion of two or more units of whole blood or packed red cells;

    • any combination of the three points above.

  •  Health‐related quality of life, as reported in included studies

Search methods for identification of studies

Electronic searches

The Cochrane Vascular Information Specialist conducted systematic searches of the following databases, from inception to 2 March 2022, for RCTs and controlled clinical trials without language, publication year, or publication status restrictions:

  • the Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web);

  • the Cochrane Central Register of Controlled Trials (CENTRAL; 2022 via the Cochrane Register of Studies Online (CRSO);

  • MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE);

  • Embase Ovid;

  • CINAHL (Cumulative Index to Nursing and Allied Health Literature) Ebsco.

We developed search strategies for other databases from the search strategy designed for MEDLINE. Where appropriate, they were combined with adaptations of the highly sensitive search strategy designed by Cochrane for identifying RCTs and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions Chapter 4, Lefebvre 2021). Search strategies for major databases are provided in Appendix 1.

We searched the following trials registries:

  • the World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch);

  • ClinicalTrials.gov (clinicaltrials.gov).

The most recent searches were carried out on 2 March 2022.

Searching other resources

We reviewed the reference lists of relevant articles for potential additional citations.

Data collection and analysis

Selection of studies

We used Covidence to screen retrieved records (Covidence). Seven review authors (ML, JL, XH, LY, XW, QW, SX), working in pairs, independently screened all titles and abstracts based on the specified inclusion and exclusion criteria, and then reviewed full‐text records of all potentially eligible articles. We linked together multiple reports of the same study and used the most comprehensive report as the primary reference. We resolved any disagreements by discussion. We illustrated the study selection process in a PRISMA diagram (Page 2021). We listed studies excluded after full‐text assessment in the Characteristics of excluded studies table, and provided the reasons for exclusion.

Data extraction and management

Six review authors (ML, JL, XW, XH, QW, SX) independently extracted the data from the included studies, using a modified version of the Cochrane Vascular standard data extraction form. We resolved any disagreements in data extraction and management through discussion and we sought the opinion of a third author (LY or KY), when necessary. We collected the following data: 

  • study details (e.g. trial name, year, country, authors);

  • methods (e.g. study design, number of participants and study centres, withdrawals and drop‐outs, treatment duration, follow‐up time);

  • participant characteristics (e.g. country, setting, age, sex, inclusion and exclusion criteria, history of VTE, previous major surgery, active cancer, pregnancy, recent period of immobility, hereditary or acquired thrombophilia);

  • interventions (e.g. oral DTIs, oral factor Xa inhibitors);

  • comparisons (e.g. other anticoagulants, i.e. LMWH, UFH, VKAs);

  • outcomes (e.g. recurrent PE, recurrent VTE, clinically overt DVT, all‐cause mortality, major bleeding, health‐related quality of life);

  • funding source;

  • declarations of interest of the study authors.

Assessment of risk of bias in included studies

Pairs of reviewers (ML, JL, XW, XH) independently assessed the risk of bias using the Cochrane risk of bias tool, as described in section 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). The tool provides a protocol for judgements on sequence generation, allocation methods, blinding, incomplete outcome data, selective outcome reporting, and any other relevant biases. We judged each of these domains as either high, low, or unclear risk of bias according to Higgins 2017 and provided support for each judgement. We presented the conclusions in a Risk of bias in included studies table. We resolved any disagreements through discussions with a third author (LY or KY).

Measures of treatment effect

For dichotomous outcomes, we used odds ratios (ORs) as the effect measure, with 95% confidence intervals (CIs). For continuous data, we planned to calculate mean differences (MDs) with 95% CIs. If similar outcomes were measured using different scales, we planned to calculate the standardised mean difference (SMD).

Unit of analysis issues

The unit of analysis in this review was each participant recruited into each included study.

Dealing with missing data

We based the analysis on intention‐to‐treat (ITT) data from the individual clinical trials whenever possible.  We adopted the 'available‐case analysis' if the primary studies did not use ITT analysis. We contacted the study authors for additional information where there were missing or unavailable data. One study author provided missing outcome data when requested (Hokusai‐VTE 2013). In addition, some studies only reported overall data about VTE, with no specific data on PE. We contacted these study authors for detailed information. When we did not get a response within one month, we excluded these studies for the reason "unable to obtain specific outcome data for people with a pulmonary embolism" (see Characteristics of excluded studies table). 

Assessment of heterogeneity

We assessed clinical and methodological heterogeneity by checking the study details. We assessed statistical heterogeneity between the trials: by visually examining the forest plots to check for overlapping CIs; with the Chi2 test for homogeneity with a 10% level of significance; and by using the I2 statistic to measure the degree of inconsistency between the studies (Higgins 2022). We interpreted I² as described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022):

  • 0% to 40%: might not be important;

  • 30% to 60%: may represent moderate heterogeneity;

  • 50% to 90%: may represent substantial heterogeneity;

  • 75% to 100%: may represent considerable heterogeneity.

We planned to investigate the reasons for significant heterogeneity where I2 was greater than 50% by examining the characteristics of included articles, including participants, interventions, comparisons, outcomes, and study design.

Assessment of reporting biases

We planned to assess publication bias by funnel plots if a sufficient number of studies (10 or more) were available in the meta‐analyses. In order to investigate any potential selective reporting bias, we compared the full‐text reports of the included studies with the registration information and protocols, and we also searched for eligible studies that had been registered and should have been completed, but were without available published data.

Data synthesis

We used Review Manager Web (RevMan Web) to synthesise the data with a meta‐analysis, following the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022; RevMan Web 2022). One review author (ML) entered the data into RevMan Web, and the second review author (JL) cross‐checked the data entry. We resolved any discrepancies by consulting the source publication. We used a random‐effects model to synthesise the data, even with a small I², as we expected clinical heterogeneity across studies to exist, due to different oral factor Xa inhibitors (e.g. apixaban, rivaroxaban, edoxaban), different conventional thrombin inhibitors in the control group (e.g. warfarin, dalteparin), different treatment durations (e.g. three, six, and 12 months). If meta‐analysis was not appropriate or possible, we planned to report the results using a narrative synthesis (McKenzie 2002).

Subgroup analysis and investigation of heterogeneity

We planned to conduct subgroup analysis for the factors below if there were two or more studies in each subgroup. We planned to analyse the subgroup differences using the Chi2 test, set at a P value of 0.05.

  • History of VTE.

  • Age.

  • Active cancer (treatment within the last six months or palliative).

  • Pregnancy.

  • Major surgery requiring general or regional anaesthesia in the previous 12 weeks.

  • Recent period of immobility (bedridden for three or more days in the previous 12 weeks).

  • Thrombophilia (hereditary or acquired).

  • Treatment duration (three months or more than three months).

  • Types of factor Xa inhibitors.

For this update, due to the limited number of included studies, subgroup analyses were only possible for active cancer and the types of factor Xa inhibitors.

Sensitivity analysis

We planned to perform sensitivity analyses to explore the robustness of the results. We intended to exclude studies that we judged to be at high risk of bias (defined as high risk in any domain). We also planned to perform sensitivity analyses with and without studies in ximelagatran, given that this drug is no longer available. However, we found no studies that tested ximelagatran in people with PE.

Summary of findings and assessment of the certainty of the evidence

We developed summary of findings tables, using GRADEpro GDT software, to present the results of this review (GRADEpro GDT; Schünemann 2022a). We created one table each for our two comparisons: 'Oral DTIs versus conventional anticoagulation for the treatment of PE' and 'Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of PE'. We assessed the certainty of the body of evidence for each outcome as high, moderate, low, or very low by considering the risk of bias, inconsistency, indirectness, imprecision, and publication bias according to the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2022b). We assessed the certainty of evidence for the following outcomes: recurrent PE, recurrent VTE, DVT, all‐cause mortality, major bleeding, and health‐related quality of life, as described in the Types of outcome measures section. 

Results

Description of studies

See Figure 1.

Figure 1
Study flow diagram

Study flow diagram

Results of the search

For this update, the searches identified 13,412 records, leaving 10,797 records after deduplication. We assessed 10,611 records as not relevant, based on the title and abstract screening. We assessed 186 potentially relevant records by full text. We identified five new included studies (24 reports) and 39 additional reports to already included studies. We excluded 29 studies (90 reports) with reasons, and discarded 20 further reports as not relevant. We identified one study (one report) as awaiting classification, and 10 ongoing studies (12 reports). The previous version of this review included five studies (22 reports). Thus, a total of 10 studies (85 reports) are now included in this review. If an article reported on two studies, we listed the article as an additional publication under each study, which makes both the counts of total reports in included and excluded studies 91. See Figure 1, Characteristics of included studies, Characteristics of excluded studies, and Characteristics of ongoing studies.

Included studies

Ten RCTs involving 13,073 adults met the inclusion criteria for this review (AMPLIFY 2013; AMPLIFY‐J 2015; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018; RE‐COVER 2009; RE‐COVER II 2014). See Characteristics of included studies.

AMPLIFY 2013 was a double‐blind study in which 5395 participants with a deep vein thrombosis (DVT) or pulmonary embolism (PE) were randomised to receive oral apixaban 10 mg twice daily for the first seven days, followed by 5 mg twice daily for six months or enoxaparin 1 mg/kg of body weight every 12 hours for at least five days and warfarin concomitantly for six months. Participants were followed up for six months. Outcomes included a composite of recurrent symptomatic venous thromboembolism (VTE) (fatal or non‐fatal PE and DVT), mortality related to VTE, major bleeding, and clinically relevant non‐major bleeding.

AMPLIFY‐J 2015 was a randomised, active‐controlled, open‐label study in which 80 participants with a DVT or PE were randomised to receive oral apixaban 10 mg for seven days as an initial therapy, followed by 5 mg for 23 weeks (n = 40) or continuous infusion of unfractionated heparin (UFH) for at least five days and warfarin concomitantly for 24 weeks (n = 40). Participants were followed up at zero, two, 12, 24, and 28 weeks. Outcomes included major bleeding, clinically relevant non‐major bleeding, all bleeding events, adjudicated recurrent symptomatic VTE (non‐fatal DVT or non‐fatal PE) or VTE‐related death during 24 weeks, and thrombotic burden deterioration at two, 12, and 24 weeks.

Caravaggio 2020 was a multinational, randomised, controlled, investigator‐initiated, open‐label, non‐inferiority trial in which 1155 people with a cancer‐associated DVT or PE were randomised to receive oral apixaban 10 mg twice daily for the first seven days, followed by 5 mg twice daily for six months (n = 576) or dalteparin 200 IU/kg of body weight once daily for the first month, followed by 150 IU/kg of body weight daily for six months (n = 579). Participants were followed up at enrolment and at four weeks, then at three, six, and seven months after randomisation. Outcomes included recurrent VTE, recurrent DVT, recurrent PE, fatal PE, major bleeding, major gastrointestinal bleeding, major non‐gastrointestinal bleeding, clinically relevant non‐major bleeding, death from any cause, and event‐free survival.

EINSTEIN‐PE 2012 was an open‐label study in which 4832 participants were randomised to receive oral rivaroxaban 15 mg twice daily for the first three weeks, followed by 20 mg once daily (n = 2419), or enoxaparin 1.0 mg/kg of body weight twice daily and either warfarin or acenocoumarol, started within 48 hours of randomisation (n = 2413). Participants were followed up at three, six, and 12 months. Outcomes included recurrent PE, recurrent DVT, major bleeding, and all‐cause mortality.

Hokusai VTE Cancer 2018 was a multinational, prospective, randomised, open‐label, blinded endpoint, non‐inferiority study in which 1046 participants with a cancer‐associated DVT or PE were randomised to receive oral edoxaban at a fixed dose of 60 mg or 30 mg once daily for six to 12 months (n = 522) or dalteparin 200 IU/kg of body weight once daily for 30 days, with a maximum daily dose of 18,000 IU, followed by dalteparin 150 IU/kg of body weight once daily for six to 12 months (n = 524). Participants were followed up on day 31 after randomisation and at three, six, nine, and 12 months. Outcomes included a composite of recurrent VTE, major bleeding, clinically relevant non‐major bleeding, event‐free survival, VTE‐related death, mortality from all causes, recurrent DVT, and recurrent PE.

Hokusai‐VTE 2013 was a double‐blind study in which 3319 participants were randomised to receive 60 mg oral edoxaban once daily (n = 1650) or dose‐adjusted warfarin therapy and edoxaban‐like placebo (n = 1669). Outcomes were measured monthly for one year. Results were presented for all participants with VTE but specific outcome data for the subset of participants with PE were obtained through communication with the author.

J‐EINSTEIN DVT and PE 2015 was an open‐label, randomised trial in which 40 participants with a DVT or PE were randomised to receive oral rivaroxaban 15 mg twice daily for a total of three weeks in a double‐blind fashion, followed by open‐label rivaroxaban 15 mg once daily for three, six, or 12 months (n = 33) or UFH for at least five days, overlapping with and followed by an international normalised ratio (INR) (range 1.5 to 2.5) titrated warfarin for three, six, or 12 months (n = 7). Participants were followed up on day 22 and at the end of the three, six, or 12 months' intended treatment period. Outcomes included the occurrence of symptomatic recurrent VTE or asymptomatic deterioration, venous ultrasound and spiral CT, major bleeding, and clinically relevant non‐major bleeding.

MERCURY PE 2018 was a randomised, open‐label, parallel‐group, multicenter study in which 114 participants with PE were randomised to receive rivaroxaban with food, 15 mg twice daily for 21 days and then 20 mg once daily to study completion for three months (n = 51), or standard of care (any Food and Drug Administration (FDA)‐approved anticoagulant strategy) for three months (n = 63). Participants were followed up at on day 30 and day 90. Outcomes included the total amount of time spent in the hospital, VTE, bleeding events, major bleeding, recurrent VTE, VTE‐related death, satisfaction with anticlot treatment, total costs of care, and clinically relevant non‐major bleeding.

RE‐COVER 2009 was a phase III, non‐inferiority, double‐blind, double‐dummy trial in which people with VTE (n = 2539) were given 150 mg dabigatran twice daily or warfarin. In addition, initial treatment with an approved parenteral anticoagulant (UFH administered intravenously or low molecular weight heparin (LMWH) administered subcutaneously) was started before participants were randomised. Treatment was for a period of six months and included sham monitoring of INR and sham titration of warfarin in the control group. To gain regulatory approval, the study was repeated with an identical design (RE‐COVER II 2014). RE‐COVER 2009 and RE‐COVER II 2014 only reported outcome data of VTE, not separate data for PE, so we took data from a pooled analysis published in an additional report (Goldhaber 2016). 

Excluded studies

See Characteristics of excluded studies.

We excluded 29 studies (ADAM VTE trial 2020; AMPLIFY Extended 2013; Borsi 2021; Botticelli DVT 2008; CASTA DIVA Trial 2022; COBRRA pilot feasibility study 2017; CONKO‐011 2015; de Athayde Soares 2019; DIVERSITY trial 2021; EINSTEIN‐CHOICE trial 2017; Einstein‐DVT Dose 2008; Einstein DVT 2013; EINSTEIN Extension 2007; EINSTEIN‐Jr Trial 2020; Farhan 2019; IRIVASC‐Trial 2022; Mokadem 2021; ODIXa‐DVT 2007; Ohmori 2018; Piazza 2014; PRAIS trial 2019; PRIORITY 2022; REMEDY 2013; RE‐SONATE 2013; SELECT‐D 2018; Sukovatykh 2017; THRIVE 2005; THRIVE I 2003; THRIVE III 2003). We excluded 11 studies as participants had DVT only (Botticelli DVT 2008; de Athayde Soares 2019; Einstein‐DVT Dose 2008; Einstein DVT 2013; Farhan 2019; Mokadem 2021; ODIXa‐DVT 2007; Ohmori 2018; Piazza 2014; PRAIS trial 2019; Sukovatykh 2017). We excluded 11 studies as although all participants had a VTE, specific data on the subgroup with PE were not published (ADAM VTE trial 2020; Borsi 2021; CASTA DIVA Trial 2022; COBRRA pilot feasibility study 2017; CONKO‐011 2015; DIVERSITY trial 2021; EINSTEIN‐Jr Trial 2020; IRIVASC‐Trial 2022; PRIORITY 2022; SELECT‐D 2018; THRIVE 2005). We made attempts to contact the authors for these data but were unsuccessful. We excluded three studies as they were extended studies testing the effectiveness of DOACs as prophylaxis rather than the treatment of PE (AMPLIFY Extended 2013; EINSTEIN Extension 2007; REMEDY 2013). We excluded the THRIVE I 2003 study as treatment was for less than three months, and the THRIVE III 2003 study as the control arm was a placebo. We excluded EINSTEIN‐CHOICE trial 2017 as the control arm was aspirin. Finally, we excluded the REMEDY 2013 study from this review as participants were already included in the RE‐COVER 2009 and RE‐COVER II 2014 studies. 

Studies awaiting classification

One trial record is awaiting classification as there are currently insufficient details to assess its eligibility for inclusion (NCT01780987).

Ongoing studies

Ten trials are ongoing and there are currently no suitable data available for including in this review (EudraCT 2014‐002606‐20; NCT02464969; NCT02664155; NCT02744092; NCT02798471; NCT03129555; NCT03266783; NCT05171049; Pettit 2018; UMIN000020069). See Characteristics of ongoing studies. 

Risk of bias in included studies

See Figure 2 and Figure 3.

Figure 2
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Figure 3
Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Allocation

All 10 included studies stated that they used a computerised system to generate the random sequence, so we deemed the risk of selection bias relating to random sequence generation to be low. All 10 included studies reported that treatment allocation was concealed with the use of a computerised system, so we judged them to be at low risk of selection bias for allocation concealment (AMPLIFY 2013; AMPLIFY‐J 2015; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018; RE‐COVER 2009; RE‐COVER II 2014).

Blinding

The AMPLIFY‐J 2015, Caravaggio 2020, EINSTEIN‐PE 2012, Hokusai VTE Cancer 2018, J‐EINSTEIN DVT and PE 2015, and MERCURY PE 2018 studies were open‐label: blinding of participants and personnel was not conducted. However, we judged that the lack of blinding in the control group was unlikely to have affected the outcome, and we thus judged these studies to have a low risk of performance bias. The AMPLIFY 2013, RE‐COVER 2009, RE‐COVER II 2014, and Hokusai‐VTE 2013 studies were double‐blind and therefore we judged them to be at low risk of performance bias.

All studies blinded outcome assessors to treatment; therefore, we judged them to be at low risk of detection bias. 

Incomplete outcome data

Nine studies sufficiently reported that fewer than 20% of participants dropped out or withdrew, and these numbers were balanced across treatment groups. Therefore, we judged them to be at low risk of attrition bias (AMPLIFY‐J 2015; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018; RE‐COVER 2009; RE‐COVER II 2014). The AMPLIFY 2013 study inappropriately excluded a number of randomised participants from the ITT analysis. Furthermore, a large number of participants within each treatment group were classified as discontinuing the study for "other reasons" with no explanations given, and we therefore deemed the risk of attrition bias to be unclear.

Selective reporting

Eight studies clearly pre‐specified the study outcomes and data on the expected outcomes were presented (AMPLIFY‐J 2015; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018; RE‐COVER 2009; RE‐COVER II 2014); we judged these studies to be at low risk of reporting bias. The AMPLIFY 2013 study pre‐defined minor bleeding as a secondary outcome but data were not reported in the paper, and therefore we deemed the risk of reporting bias in this study to be high. The Caravaggio 2020 study pre‐defined quality of life as a secondary outcome in the protocol but the data on this outcome were not reported in the paper; therefore we deemed the risk of reporting bias to be high.

Other potential sources of bias

We did not find any methodological issues that might directly lead to a risk of bias; therefore, we deemed the risk of other bias to be low in all 10 included studies. 

Effects of interventions

See: Summary of findings 1 Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism; Summary of findings 2 Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism

We identified two studies that compared an oral DTI versus conventional anticoagulation with warfarin (RE‐COVER 2009; RE‐COVER II 2014), and eight studies that compared an oral factor Xa inhibitor versus conventional anticoagulation with warfarin or dalteparin (AMPLIFY‐J 2015; AMPLIFY 2013; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018). We did not find any studies comparing one DTI with another DTI, one factor Xa inhibitor with another factor Xa inhibitor, or an oral DTI with a factor Xa inhibitor. We used a random‐effects model for all analyses due to clinical heterogeneity across studies (different oral factor Xa inhibitors, different conventional thrombin inhibitors in the control group, and different treatment durations).

Oral direct thrombin inhibitor versus conventional anticoagulation

In the meta‐analysis of oral DTIs versus conventional anticoagulation, we used data from Goldhaber 2016, which reported combined data from the RE‐COVER 2009 and RE‐COVER II 2014 studies. This is reflected in the data analysis tables and summary of findings Table 1 by showing only one study for this comparison.

Recurrent pulmonary embolism

Two studies with a combined total of 1602 participants measured recurrent pulmonary embolism (PE) (RE‐COVER 2009; RE‐COVER II 2014). There was no clear difference in the rate of recurrent PE between participants treated with dabigatran (16 events/795 participants) and those treated with conventional anticoagulation (16 events/807 participants) leading to an odds ratio (OR) of 1.02 (95% confidence interval (CI) 0.50 to 2.04; 2 studies, 1602 participants; moderate‐certainty evidence; Analysis 1.1).

Recurrent venous thromboembolism

Two studies with a combined total of 1602 participants measured recurrent venous thromboembolism (VTE) (RE‐COVER 2009; RE‐COVER II 2014). There was no clear difference in the rate of recurrent VTE between participants treated with dabigatran (23 events/795 participants) and those treated with conventional anticoagulation (25 events/807 participants) leading to an OR of 0.93 (95% CI 0.52 to 1.66; 2 studies, 1602 participants; moderate‐certainty evidence; Analysis 1.2).

Deep vein thrombosis

Two studies with a combined total of 1602 participants measured deep vein thrombosis (DVT) (RE‐COVER 2009; RE‐COVER II 2014). There was no clear difference in the rate of DVT between participants treated with dabigatran (7 events/795 participants) and those treated with conventional anticoagulation (9 events/807 participants) leading to an OR of 0.79 (95% CI 0.29 to 2.13; 2 studies, 1602 participants; moderate‐certainty evidence; Analysis 1.3).

All‐cause mortality

Neither study presented results on all‐cause mortality for the specific group of participants with PE.

Adverse effects of treatment

Both RE‐COVER 2009 and RE‐COVER II 2014 measured major bleeding (as defined by the International Society on Thrombosis and Haemostasis (ISTH); Schulman 2005). There was no clear difference in the rate of major bleeding between participants treated with oral DTIs (4 events/759 participants) and those treated with conventional anticoagulation (8 events/768 participants) leading to an OR of 0.50 (95% CI 0.15 to 1.68; 2 studies, 1602 participants; moderate‐certainty evidence; Analysis 1.4).

Health‐related quality of life

Health‐related quality of life was not a reported outcome in either RE‐COVER 2009 or RE‐COVER II 2014.

Oral factor Xa inhibitor versus conventional anticoagulation

See summary of findings Table 2.

Recurrent pulmonary embolism

Data on recurrent PE was available in three studies with 8186 participants (AMPLIFY‐J 2015; EINSTEIN‐PE 2012; Hokusai‐VTE 2013). The duration of treatment in all three studies was longer than three months. Meta‐analysis showed no clear difference in the rate of recurrent PE between participants treated with oral factor Xa inhibitors (67 events/4087 participants) and those treated with conventional anticoagulation (73 events/4099 participants), leading to an OR of 0.92 (95% CI 0.66 to 1.29; I2 = 0; 3 studies, 8186 participants; moderate‐certainty evidence; Analysis 2.1). 

Recurrent venous thromboembolism

We included eight studies with a combined total of 11,416 participants in a meta‐analysis (AMPLIFY‐J 2015; AMPLIFY 2013; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018). The treatment duration of MERCURY PE 2018 was three months, and the events in both groups were 0, so it was not estimated in the meta‐analysis. The treatment duration of the other seven studies was longer than three months. Meta‐analysis showed no clear difference in the rate of recurrent VTE between participants treated with oral factor Xa inhibitors (150 events/5698 participants) and those treated with conventional anticoagulation (183 events/5718 participants), leading to an OR of 0.83 (95% CI 0.66 to 1.03; I2 = 0%; 8 studies, 11,416 participants; moderate‐certainty evidence; Analysis 2.2). 

Deep vein thrombosis

We included two studies with a combined total of 8151 participants in a meta‐analysis (EINSTEIN‐PE 2012; Hokusai‐VTE 2013). There was no clear difference in the rate of recurrent DVT between participants treated with oral factor Xa inhibitors (30 events/4069 participants) and those treated with conventional anticoagulation (39 events/4082 participants), leading to an OR of 0.77 (95% CI 0.48 to 1.25; I2 = 0; 2 studies, 8151 participants; moderate‐certainty evidence; Analysis 2.3). The AMPLIFY 2013 study did not present DVT data for the subgroup of participants with a PE and therefore we did not include it in this meta‐analysis.

All‐cause mortality

Three studies measured all‐cause mortality (AMPLIFY‐J 2015; EINSTEIN‐PE 2012; MERCURY PE 2018). In the AMPLIFY‐J 2015 and MERCURY PE 2018 studies, the events in both groups were 0, so they were not estimated in the meta‐analysis. There was no clear difference in the rate of all‐cause mortality between participants treated with the oral factor Xa inhibitor rivaroxaban (2.40%; 58 events/2412 participants) and those treated with conventional anticoagulation (50 events/2405 participants), leading to an OR of 1.16 (95% CI 0.79 to 1.70; 1 study, 4817 participants; moderate‐certainty evidence; Analysis 2.4). 

Adverse effects of treatment

Eight studies measured major bleeding (as defined by ISTH, Schulman 2005) (AMPLIFY‐J 2015; AMPLIFY 2013; Caravaggio 2020; EINSTEIN‐PE 2012; Hokusai VTE Cancer 2018; Hokusai‐VTE 2013; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018). There was no clear difference in the rate of major bleeding between participants treated with oral factor Xa inhibitors (95 events/5721 participants) and those treated with conventional anticoagulation (130 events/5726 participants) leading to an OR of 0.71 (95% CI 0.36 to 1.41; 8 studies, 11,447 participants; low‐certainty evidence; Analysis 2.5). Considerable heterogeneity was detected between the studies (I2 = 79%), likely due to clinical heterogeneity.

Health‐related quality of life

Health‐related quality of life was not reported in any of the included studies.

Subgroup analysis

For this update, due to the limited number of included studies, subgroup analyses were only possible for participants who received oral factor Xa inhibitors. We performed subgroup analysis by active cancer and different types of factor Xa inhibitors for the outcome of recurrent VTE and major bleeding. 

Results of subgroup analyses showed no clear difference in the incidence of recurrent VTE in participants without cancer (OR 0.89, 95% CI 0.68 to 1.17; 6 studies, 9898 participants; moderate‐certainty evidence; Analysis 2.2) compared to participants with cancer (OR 0.65, 95% CI 0.42 to 1.01; 3 studies, 1518 participants; high‐certainty evidence; Analysis 2.2). The test for subgroup differences did not indicate a difference (P = 0.23). For the subgroup analysis based on the different types of factor Xa inhibitors, there was also no clear difference in the incidence of recurrent VTE when participants received treatment with apixaban (OR 0.86, 95% CI 0.54 to 1.35; 3 studies, 2459 participants; moderate‐certainty evidence; Analysis 2.6), rivaroxaban (OR 1.14, 95% CI 0.75 to 1.71; 3 studies, 4981 participants; moderate‐certainty evidence; Analysis 2.6), and edoxaban (OR 0.66, 95% CI 0.48 to 0.92; 2 studies, 3976 participants; moderate‐certainty evidence; Analysis 2.6) compared with conventional anticoagulation. The test for subgroup differences did not indicate a difference (P = 0.13). 

Subgroup analysis of the incidence of major bleeding in participants without cancer (OR 0.45, 95% CI 0.19 to 1.06; 6 studies, 10,152 participants; low‐certainty evidence; Analysis 2.5) versus participants with cancer (OR 1.51, 95% CI 0.84 to 2.71; 2 studies, 1295 participants; low‐certainty evidence; Analysis 2.5) suggested a possible subgroup difference (test for subgroup differences: P = 0.02). Due to the limited number of studies in this analysis, and as CIs overlap, this should be interpreted with caution. When analysed according to different types of factor Xa inhibitors, there was no clear difference in the incidence of major bleeding between apixaban and conventional anticoagulation (OR 0.36, 95% CI 0.07 to 1.91; 3 studies, 2503 participants; low‐certainty evidence; Analysis 2.7), or between edoxaban and conventional anticoagulation (OR 1.44, 95% CI 0.80 to 2.58; 2 studies, 3976 participants; low‐certainty evidence; Analysis 2.7). However, when treated with rivaroxaban, participants had a lower incidence of major bleeding compared to those treated with conventional anticoagulation (OR 0.49, 95% CI 0.31 to 0.79; 3 studies, 4968 participants; low‐certainty evidence; Analysis 2.7). The test for subgroup differences indicated a possible difference (P = 0.01). Due to the limited number of studies in this analysis, and as CIs overlap, this should be interpreted with caution. 

Sensitivity analysis

We performed sensitivity analysis by excluding studies we judged to be at high risk of bias (AMPLIFY 2013; Caravaggio 2020). Based on the available data, only sensitivity analyses of oral factor Xa inhibitors for PE in the incidence of recurrent VTE and major bleeding were possible. The results showed that oral factor Xa inhibitors versus conventional anticoagulation made no clear difference in reducing the incidence of recurrent VTE (OR 0.78, 95% CI 0.54 to 1.14; 6 studies, 8992 participants; moderate‐certainty evidence; Analysis 2.8) and major bleeding (OR 0.91, 95% CI 0.42 to 1.97; 6 studies, 8979 participants; low‐certainty evidence; Analysis 2.9). The results of the sensitivity analyses are consistent with the results of Analysis 2.2 and  Analysis 2.5, which indicates that the results are robust.

Discussion

We included an additional five studies for this update, bringing the total to 10 included studies involving 13,073 participants. Two studies investigated an oral DTI (dabigatran) and eight studies investigated oral factor Xa inhibitors (three rivaroxaban, three apixaban, and two edoxaban). Overall the studies were of good methodological quality. The certainty of the evidence was moderate or low.

Summary of main results

Recurrent pulmonary embolism

Moderate‐certainty evidence showed no clear difference between both oral DTIs and factor Xa inhibitors compared to conventional anticoagulation in preventing recurrent pulmonary embolism (PE). We downgraded the certainty of the evidence by one level for imprecision. 

Recurrent venous thromboembolism

Moderate‐certainty evidence showed no clear difference between oral DTIs and conventional anticoagulation in the prevention of recurrent venous thromboembolism (VTE) during treatment, indicating that neither was more nor less effective. We downgraded the certainty of the evidence by one level for imprecision.

Similarly, for factor Xa inhibitors, moderate‐certainty evidence showed no clear difference between factor Xa inhibitors and conventional anticoagulation in preventing recurrent VTE, indicating that neither was more nor less effective. We downgraded the certainty of the evidence by one level for imprecision. 

Subgroup analyses showed no clear difference in the incidence of recurrent VTE between participants with and without cancer, and between participants who received different types of factor Xa inhibitors. Given the limited number of studies providing data in each subgroup, this may reflect a lack of information. More included studies may change the evidence in the future.

Deep vein thrombosis

Moderate‐certainty evidence showed no clear difference between both oral DTIs and factor Xa inhibitors compared to conventional anticoagulation in preventing deep vein thrombosis (DVT). We downgraded the certainty of the evidence by one level for imprecision.

All‐cause mortality

No study measured all‐cause mortality in participants treated with oral DTIs. Moderate‐certainty evidence showed no clear difference between oral factor Xa inhibitors and conventional therapy in preventing all‐cause mortality. We downgraded the certainty of the evidence by one level for imprecision.

Major bleeding

Moderate‐certainty evidence indicated no clear difference between oral DTIs and conventional anticoagulation in major bleeding. We downgraded the certainty of the evidence by one level for imprecision. For factor Xa inhibitors, low‐certainty evidence showed no clear difference between factor Xa inhibitors and conventional anticoagulation in preventing major bleeding. We downgraded the certainty of the evidence by one level for serious inconsistency due to clinical heterogeneity and by one level for imprecision. 

Subgroup analyses suggested possible subgroup differences in the incidence of major bleeding between participants without cancer versus those with cancer, and between participants receiving different types of factor Xa inhibitors. Due to the limited number of studies in the subgroups, and as CIs overlap, this should be interpreted with caution. More included studies may change the evidence in the future. 

Health‐related quality of life was not reported in the included studies.

Overall completeness and applicability of evidence

This review assessed whether long‐term treatment with oral anticoagulants – DTIs and factor Xa inhibitors – reduced the rate of recurrent PE, recurrent VTE, DVT, all‐cause mortality, and major bleeding in people with PE. Two studies tested DTIs and eight studies tested factor Xa inhibitors within similar study populations. The included studies analysed and reported all of our outcomes of interest, except for health‐related quality of life. Heterogeneity was high for major bleeding in the studies testing factor Xa inhibitors. As expected, this is likely due to clinical heterogeneity, because included studies varied in factor Xa inhibitors (three apixaban, three rivaroxaban, two edoxaban), comparators (six warfarin, two dalteparin), and treatment duration (three months (one study), five and a half months (one study), six months (two studies), 12 months (four studies)). 

For this update, we performed subgroup analyses for active cancer and different types of factor Xa inhibitors. Limited data indicated no clear difference in the incidence of recurrent VTE between participants with and without cancer, and between participants who received different types of factor Xa inhibitors. For major bleeding, both subgroup analyses (i.e. of people experiencing PE with versus without cancer, and subgroups of different types of factor Xa inhibitors) suggested possible subgroup differences. Bleeding risk is an important consideration when using DOACs for cancer patients, particularly in patients with gastrointestinal malignancies (Thapa 2019). A systematic review and meta‐analysis of the Hokusai VTE Cancer 2018 and SELECT‐D 2018 studies suggested that the risk of major bleeding in cancer patients was approximately doubled when using DOACs compared with LMWH (Li 2019). For factor Xa inhibitors, while efficacy with apixaban, edoxaban, and rivaroxaban versus dalteparin has been consistent in the treatment of cancer‐associated VTE, one study showed heterogeneity is evident with respect to major gastrointestinal bleeding, with an increased risk with edoxaban and rivaroxaban but not apixaban (Athanazio 2022). All of these results are based on limited evidence; more evidence is needed to confirm these conclusions in the future.

We were unable to perform subgroup analyses for other factors because of the lack of participant‐level data. More evidence is needed, as these analyses might be important to guide clinical management in people with different risk factors for PE.

Although many consider DVT and PE to be manifestations of the same disorder, we elected to study these two conditions separately as there is evidence of clinically significant differences between them. First, the majority of recurrent events occur at the same site as the original thrombosis (in other words, in people presenting with a PE, a recurrent event after treatment is much more likely to be another PE). Second, both oral contraceptive use and Factor V Leiden mutation are more likely to be associated with DVT than PE. Third, lung disease is much more likely to be associated with PE. Consequently, a review on the effectiveness of oral DTIs and factor Xa inhibitors for the long‐term treatment of DVT was published at the same time as the previous version of this review (Robertson 2015b). It has also now been updated (Wang 2023). 

We did not find any studies comparing:

  • one oral DTI versus another oral DTI;

  • one oral factor Xa inhibitor versus another oral factor Xa inhibitor;

  • oral DTI versus oral factor Xa inhibitor.

We identified no additional studies investigating DTIs for this update; therefore, this review cannot provide more evidence for this intervention. Given the recent clinical thinking that DOACs have a class effect, we will pool the results of all DTIs and factor Xa inhibitors in future updates. This may increase the power of the studies, reduce type II errors, and allow any effects to be determined clearly.

In 2020, the UK's National Institute for Health and Care Excellence (NICE) measured the cost‐effectiveness of DOACs versus conventional anticoagulation for the treatment of DVT and PE (NICE 2020). Assuming that people remain on the same drug in the initial and extended phases of treatment, apixaban was highly cost‐effective both in people with a DVT and people with a PE. Rivaroxaban had the next most favourable effect on major bleeding and generated the second‐highest total quality‐adjusted life years (QALYs). The cost of the two drugs was similar and the difference in total costs was mainly led by differences in the number of bleeding events. 

Certainty of the evidence

We assessed the overall risk of bias as low in eight of 10 included studies, reflecting good methodological quality. We judged two studies to have a high risk of reporting bias. Six of the 10 included studies were open‐label because of the complexity of monitoring international normalised ratio (INR) in the conventional anticoagulation arm. However, all outcomes were assessed by observers who were blinded to the treatment and all safety outcomes were adjudicated by a central independent committee in each included study. We could not investigate publication bias because we could not assess asymmetry in a funnel plot with the limited number of studies included in the meta‐analysis. 

For the comparison of oral DTIs versus conventional anticoagulation, we graded the certainty of the evidence as moderate, downgrading by one level for imprecision due to the low number of events. For oral factor Xa inhibitors versus conventional anticoagulation, we downgraded the evidence for recurrent PE, recurrent VTE, DVT, and all‐cause mortality to moderate due to the low number of events. For this comparison, we downgraded the evidence for major bleeding by two levels to low, due to the low number of events and substantial clinical heterogeneity. See summary of findings Table 1; summary of findings Table 2.

Potential biases in the review process

The search was as comprehensive as possible, and we are confident that we have included all relevant studies. However, the possibility remains that some relevant trials, particularly in the 'grey' literature (for example, conference proceedings), have been missed. Eight review authors independently performed study selection and data extraction in order to minimise bias in the review process. We strictly adhered to the inclusion and exclusion criteria set out in the protocol in order to limit subjectivity. We performed data collection according to the process suggested by Cochrane. We contacted study authors in an attempt to obtain PE‐specific data. We also followed Cochrane processes as described by Higgins 2017 for assessing the risk of bias. Two of the included studies, RE‐COVER 2009 and RE‐COVER II 2014, only reported outcome data of VTE, not separate data for PE, so we took data from a pooled analysis published in an additional report (Goldhaber 2016). This was the best available evidence.  

Agreements and disagreements with other studies or reviews

This review assesses the efficacy and safety of oral DTIs and oral factor Xa inhibitors for the long‐term treatment of PE. A similar review by Dentali 2015 evaluated the difference in the safety and efficacy of DOACs compared to conventional treatment in participants presenting with DVT and with PE. The Dentali 2015 review included six studies (27,023 participants) and indicated that DOACs appear to have similar efficacy and safety profiles compared to VKAs for people with DVT and PE. Dentali 2015 focused on both PE and DVT, but did not include some more recent RCTs that are included in this update (AMPLIFY‐J 2015; Caravaggio 2020; Hokusai VTE Cancer 2018; J‐EINSTEIN DVT and PE 2015; MERCURY PE 2018).

Fourteen other systematic reviews have assessed the same oral anticoagulants (Antoniazzi 2013; Castellucci 2013; Cohen 2015; Fox 2012; Gómez‐Outes 2014; Gómez‐Outes 2015; Hirschl 2014; Kang 2014; Moik 2020; Samaranayake 2022; Sardar 2014; Song 2021; Van der Huille 2014; Wu 2022). However, these reviews focused on people with a VTE, and did not report specific data on the subgroup with a PE. Four reviews indicated that DOACs can decrease the risk of recurrent VTE compared to conventional treatment (Moik 2020; Samaranayake 2022; Song 2021; Wu 2022); seven reviews found that DOACs are associated with less bleeding than conventional treatment (Antoniazzi 2013; Fox 2012; Gómez‐Outes 2014; Gómez‐Outes 2015; Hirschl 2014; Van der Huille 2014; Wu 2022); one review found that DOACs are associated with lower all‐cause mortality compared with those receiving conventional anticoagulants in people with venous thromboembolism (Wu 2022).

The Fox 2012 review performed meta‐analysis by brand rather than class of drug and found no difference in recurrent VTE between the two treatment groups. Rivaroxaban was the only drug found to be significantly associated with fewer major bleeding episodes (odds ratio (OR) 0.57, 95% confidence interval (CI) 0.39 to 0.84). All‐cause mortality did not differ between the two treatment groups.

The Van der Huille 2014 review showed no difference between the two treatment groups in terms of recurrent VTE, fatal PE, and all‐cause mortality. However, DOACs were associated with a significantly reduced risk of major bleeding (relative risk (RR) 0.60, 95% CI 0.41 to 0.88) and fatal bleeding (RR 0.36, 95% CI 0.15 to 0.87).

Hirschl 2014 found no differences between DOACs and conventional treatment regarding recurrent VTE and death. However, bleeding was reduced by rivaroxaban (RR 0.55, 95% CI 0.38 to 0.81), apixaban (RR 0.31, 95% CI 0.17 to 0.55), and edoxaban (RR 0.81, 95% CI 0.71 to 0.93).

The Gómez‐Outes 2014 review found no clear difference in the risk of recurrent VTE between the two treatment groups (RR 0.91, 95% CI 0.79 to 1.06), but DOACs were associated with reduced major bleeding (absolute risk difference of ‐0.6%, 95% CI ‐1.0% to ‐0.3%).

A network meta‐analysis by Kang 2014 found that DOACs did not differ in the risk of mortality or recurrent VTE. However, dabigatran was associated with increased major bleeding compared to apixaban (RR 2.69, 95% CI 1.19 to 6.07), and edoxaban also had a higher bleeding rate compared to apixaban (RR 2.74, 95% CI 1.40 to 5.39).

The Wu 2022 review included 65 real‐world evidence studies and found that, in people with VTE, rivaroxaban and apixaban are associated with a lower risk of recurrent VTE (rivaroxaban: HR 0.68, 95% CI 0.60 to 0.76; apixaban: HR 0.83, 95% CI 0.73 to 0.93), major bleeding events (rivaroxaban: HR 0.73, 95% CI 0.65 to 0.81; apixaban: HR 0.76, 95% CI 0.68 to 0.85), and all‐cause mortality (HR 0.56, 95% CI 0.39 to 0.80) compared with those receiving conventional anticoagulants.

The Song 2021 review included four RCTs and 14 retrospective studies and found that DOACs decreased the risk of recurrent VTE (RCTs: OR 0.60, 95% CI 0.45 to 0.80; retrospective studies: OR 0.73, 95% CI 0.59 to 0.90) and recurrent DVT (RCTs: OR 0.54, 95% CI 0.36 to 0.80; retrospective studies: OR 0.20, 95% CI 0.06 to 0.63), but not PE recurrence (OR 0.75, 95% CI 0.53 to 1.05), fatal PE (OR 0.87, 95% CI 0.33 to 2.26), and major bleeding events (OR 1.04, 95% CI 0.84 to 1.28) compared with LMWHs in people with cancer. 

The Samaranayake 2022 review conducted a network meta‐analysis including four RCTs with 2907 participants with venous thromboembolism. They found that the overall risk of recurrent VTE was lower in the DOACs group compared to the dalteparin group (RR 0.63, 95% CI 0.44 to 0.91). There was no statistically significant difference in risk of major bleeding (RR 1.31, 95% CI 0.83 to 2.07) and all‐cause mortality (RR 1.0, 95% CI 0.84 to 1.18) at six months' follow‐up between DOACs and LMWH. 

A review by Moik 2020 that included four RCTs (with 2894 participants with cancer‐associated VTE) found that DOACs significantly reduced recurrent VTE compared to LMWHs (RR 0.62, 95% CI 0.43 to 0.91), but were associated with a non‐significant increase in major bleeding (RR 1.31, 95% CI 0.83 to 2.08). Mortality risks were comparable between groups (RR 0.99, 95% CI 0.83 to 1.18). 

The Gómez‐Outes 2015 review found that DOACs were associated with lower major bleeding (RR 0.62, 95% CI 0.45 to 0.85), while there was no clear difference in recurrent VTE rate in people receiving DOACs or standard therapy (RR 0.91, 95% CI 0.79 to 1.05).

A review by Cohen 2015 conducted a network meta‐analysis to compare the efficacy and safety of DOACs for the initial and long‐term treatment of VTE. They found that apixaban was associated with a statistically significantly reduced risk of major bleeding compared with rivaroxaban (RR 0.47, 95% CI 0.36 to 0.61), dabigatran (RR 0.69, 95% CI 0.51 to 0.94), and edoxaban (RR 0.54, 95% CI 0.41 to 0.69). Dabigatran was associated with a significantly lower risk of major bleeding compared with rivaroxaban (RR 0.68, 95% CI 0.53 to 0.87) and edoxaban (RR 0.77, 95% CI 0.60 to 0.99). There were no clear differences between the DOACs with regard to the risk of VTE and related death.

The Antoniazzi 2013 review included participants with VTE and atrial fibrillation. Eight studies were included and results showed that the risk of major bleeding was lower in participants treated with dabigatran (RR 0.83, 95% CI 0.78 to 0.97) compared with the control group.

The reviews by Castellucci 2013, Cohen 2016, Sardar 2014, and Wang 2018 compared oral anticoagulants and antiplatelet drugs, but the focus was on the secondary prevention of venous thromboembolism rather than treatment. 

Study flow diagram

Figures and Tables -
Figure 1

Study flow diagram

Navigate to figure in ReviewOpen in new tab

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Figures and Tables -
Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Navigate to figure in ReviewOpen in new tab

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Figures and Tables -
Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Navigate to figure in ReviewOpen in new tab

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 1: Recurrent pulmonary embolism

Figures and Tables -
Analysis 1.1

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 1: Recurrent pulmonary embolism

Navigate to figure in ReviewOpen in new tab

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 2: Recurrent venous thromboembolism

Figures and Tables -
Analysis 1.2

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 2: Recurrent venous thromboembolism

Navigate to figure in ReviewOpen in new tab

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 3: Deep vein thrombosis

Figures and Tables -
Analysis 1.3

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 3: Deep vein thrombosis

Navigate to figure in ReviewOpen in new tab

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 4: Major bleeding

Figures and Tables -
Analysis 1.4

Comparison 1: Oral DTIs versus conventional anticoagulation, Outcome 4: Major bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 1: Recurrent pulmonary embolism

Figures and Tables -
Analysis 2.1

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 1: Recurrent pulmonary embolism

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 2: Recurrent venous thromboembolism

Figures and Tables -
Analysis 2.2

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 2: Recurrent venous thromboembolism

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 3: Deep vein thrombosis

Figures and Tables -
Analysis 2.3

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 3: Deep vein thrombosis

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 4: All‐cause mortality

Figures and Tables -
Analysis 2.4

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 4: All‐cause mortality

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 5: Major bleeding

Figures and Tables -
Analysis 2.5

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 5: Major bleeding

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 6: Recurrent venous thromboembolism (subgroup analysis based on different types of factor Xa inhibitors)

Figures and Tables -
Analysis 2.6

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 6: Recurrent venous thromboembolism (subgroup analysis based on different types of factor Xa inhibitors)

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 7: Major bleeding (subgroup analysis based on different types of factor Xa inhibitors)

Figures and Tables -
Analysis 2.7

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 7: Major bleeding (subgroup analysis based on different types of factor Xa inhibitors)

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 8: Recurrent venous thromboembolism (sensitivity analysis by including only studies at low risk of bias)

Figures and Tables -
Analysis 2.8

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 8: Recurrent venous thromboembolism (sensitivity analysis by including only studies at low risk of bias)

Navigate to figure in ReviewOpen in new tab

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 9: Major bleeding (sensitivity analysis by including only studies at low risk of bias)

Figures and Tables -
Analysis 2.9

Comparison 2: Oral factor Xa inhibitors versus conventional anticoagulation, Outcome 9: Major bleeding (sensitivity analysis by including only studies at low risk of bias)

Navigate to figure in ReviewOpen in new tab
Summary of findings 1. Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism

Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism

Patient or population: people with a pulmonary embolism, confirmed by standard imaging techniques
Setting: hospital
Intervention: oral DTIs
Comparison: conventional anticoagulation

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with conventional anticoagulation

Risk with oral DTI

Recurrent PEa

 

Follow‐up: 

6 months

Study population

OR 1.02
(0.50 to 2.04)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

20 per 1000

20 per 1000
(10 to 40)

Recurrent VTEc

 

Follow‐up: 

6 months

Study population

OR 0.93
(0.52 to 1.66)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

31 per 1000

29 per 1000
(16 to 50)

DVTd

 

Follow‐up: 

6 months

Study population

OR 0.79
(0.29 to 2.13)

1602
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

11 per 1000

9 per 1000
(3 to 23)

All‐cause mortality

See comment

See comment

See comment

‐

RE‐COVER 2009 and RE‐COVER II 2014 did not report on all‐cause mortality.

Major bleedinge

 

Follow‐up: 

6 months

Study population

OR 0.50
(0.15 to 1.68)

1527
(1 RCT)

⊕⊕⊕⊝
Moderateb

The data from RE‐COVER 2009 and RE‐COVER II 2014 were taken from one pooled analysis and are therefore shown as one study in our analyses.

10 per 1000

5 per 1000
(2 to 17)

Health‐related quality of life

See comment

See comment

See comment

‐

RE‐COVER 2009 and RE‐COVER II 2014 did not measure health‐related quality of life.

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: confidence interval; CTPA: computed tomographic pulmonary angiography; DVT: deep vein thrombosis; ISTH: International Society on Thrombosis and Haemostasis; OR: odds ratio; PE: pulmonary embolism; RCT: randomised controlled trial; V/Q: ventilation/perfusion; VTE: venous thromboembolism

GRADE Working Group grades of evidence
High certainty: we are very confident that the true effect lies close to that of the estimate of the effect
Moderate certainty: we are moderately confident in the effect estimate, the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low certainty: our confidence in the effect estimate is limited, the true effect may be substantially different from the estimate of the effect
Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect

aConfirmed by V/Q lung scanning, pulmonary angiography, or CTPA
bWe downgraded one level for imprecision due to the low number of events. The possibility of publication bias is not excluded but we did not consider it sufficient to downgrade the certainty of evidence.
cVTE includes clinically overt DVT and PE. Clinically overt DVT, confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound); or clinically overt PE, confirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
dClinically overt DVT confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound). 
eAs defined by the ISTH (Schulman 2005): 1. Fatal bleeding, and/or 2. symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra‐articular or pericardial, or intramuscular with compartment syndrome, and/or 3. bleeding causing a fall in haemoglobin level of 20 g/L (1.24 mmol/L) or more, or leading to transfusion of two or more units of whole blood or packed red cells.

Figures and Tables -
Summary of findings 1. Oral direct thrombin inhibitors (DTIs) versus conventional anticoagulation for the treatment of pulmonary embolism
Navigate to table in Review
Summary of findings 2. Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism

Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism

Patient or population: people with a pulmonary embolism, confirmed by standard imaging techniques
Setting: hospital
Intervention: oral factor Xa inhibitors
Comparison: conventional anticoagulation

Outcomes

Anticipated absolute effects* (95% CI)

Relative effect
(95% CI)

№ of participants
(studies)

Certainty of the evidence
(GRADE)

Comments

Risk with conventional  anticoagulation

Risk with oral factor Xa inhibitors

Recurrent PEa

 

Follow‐up: 

0 to 12 months

Study population

OR 0.92
(0.66 to 1.29)

8186
(3 RCTs)

⊕⊕⊕⊝

Moderateb 

‐

18 per 1000

16 per 1000
(12 to 23)

Recurrent VTEc

 

Follow‐up: 

0 to 12 months

Study population

OR 0.83
(0.66 to 1.03)

11,416
(8 RCTs)

⊕⊕⊕⊝
Moderateb

 

2 of 8 studies reported no events

32 per 1000

26 per 1000
(21 to 33)

DVTd

 

Follow‐up: 

5 days to 12 months

Study population

OR 0.77
(0.48 to 1.25)

8151
(2 RCTs)

⊕⊕⊕⊝
Moderateb

‐

10 per 1000

7 per 1000
(5 to 12)

All‐cause mortality

 

Follow‐up: 

0 to 12 months

Study population

OR 1.16
(0.79 to 1.70)

4817
(1 RCT)

⊕⊕⊕⊝
Moderateb

‐

21 per 1000

24 per 1000
(16 to 35)

Major bleedinge

 

Follow‐up: 

0 to 12 months

Study population

OR 0.71
(0.36 to 1.41) 

11,447
(8 RCTs)

⊕⊕⊝⊝
Lowb,f

2 of 8 studies reported no events

23 per 1000

16 per 1000
(8 to 32)

Health‐related quality of life

See comment

See comment

See comment

‐

The studies did not measure health‐related quality of life.

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: confidence interval; CTPA: computed tomographic pulmonary angiography; DVT: deep vein thrombosis; ISTH: International Society on Thrombosis and Haemostasis; OR: odds ratio; PE: pulmonary embolism; RCT: randomised controlled trial; V/Q: ventilation/perfusion; VTE: venous thromboembolism

GRADE Working Group grades of evidence
High certainty: we are very confident that the true effect lies close to that of the estimate of the effect
Moderate certainty: we are moderately confident in the effect estimate, the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low certainty: our confidence in the effect estimate is limited, the true effect may be substantially different from the estimate of the effect
Very low certainty: we have very little confidence in the effect estimate, the true effect is likely to be substantially different from the estimate of effect

aConfirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
bWe downgraded one level for imprecision due to the low number of events. The possibility of publication bias is not excluded but we did not consider it sufficient to downgrade the certainty of evidence.
cVTE includes clinically overt DVT and PE. Clinically overt DVT, confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound); or clinically overt PE, confirmed by V/Q lung scanning, pulmonary angiography, or CTPA.
dClinically overt DVT confirmed by standard imaging techniques (venography, impedance plethysmography, whole‐leg compression ultrasound, proximal compression ultrasound). 
eAs defined by the ISTH (Schulman 2005): 1. Fatal bleeding, and/or 2. symptomatic bleeding in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, intra‐articular or pericardial, or intramuscular with compartment syndrome, and/or 3. bleeding causing a fall in haemoglobin level of 20 g/L (1.24 mmol/L) or more, or leading to transfusion of two or more units of whole blood or packed red cells.
fWe downgraded one level for serious inconsistency (I2 = 79% due to clinical heterogeneity).

Figures and Tables -
Summary of findings 2. Oral factor Xa inhibitors versus conventional anticoagulation for the treatment of pulmonary embolism
Navigate to table in Review
Comparison 1. Oral DTIs versus conventional anticoagulation

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1.1 Recurrent pulmonary embolism Show forest plot

1

Odds Ratio (M‐H, Random, 95% CI)

Totals not selected

1.2 Recurrent venous thromboembolism Show forest plot

1

Odds Ratio (M‐H, Random, 95% CI)

Totals not selected

1.3 Deep vein thrombosis Show forest plot

1

Odds Ratio (M‐H, Random, 95% CI)

Totals not selected

1.4 Major bleeding Show forest plot

1

Odds Ratio (M‐H, Random, 95% CI)

Totals not selected
Figures and Tables -
Comparison 1. Oral DTIs versus conventional anticoagulation
Navigate to table in Review
Comparison 2. Oral factor Xa inhibitors versus conventional anticoagulation

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

2.1 Recurrent pulmonary embolism Show forest plot

3

8186

Odds Ratio (M‐H, Random, 95% CI)

0.92 [0.66, 1.29]

2.2 Recurrent venous thromboembolism Show forest plot

8

11416

Odds Ratio (M‐H, Random, 95% CI)

0.83 [0.66, 1.03]

2.2.1 Non‐cancer associated pulmonary embolism

6

9898

Odds Ratio (M‐H, Random, 95% CI)

0.89 [0.68, 1.17]

2.2.2 Cancer associated pulmonary embolism

3

1518

Odds Ratio (M‐H, Random, 95% CI)

0.65 [0.42, 1.01]

2.3 Deep vein thrombosis Show forest plot

2

8151

Odds Ratio (M‐H, Random, 95% CI)

0.77 [0.48, 1.25]

2.4 All‐cause mortality Show forest plot

3

Odds Ratio (M‐H, Random, 95% CI)

Totals not selected

2.5 Major bleeding Show forest plot

8

11447

Odds Ratio (M‐H, Random, 95% CI)

0.71 [0.36, 1.41]

2.5.1 Non‐cancer associated pulmonary embolism

6

10152

Odds Ratio (M‐H, Random, 95% CI)

0.45 [0.19, 1.06]

2.5.2 Cancer associated pulmonary embolism

2

1295

Odds Ratio (M‐H, Random, 95% CI)

1.51 [0.84, 2.71]

2.6 Recurrent venous thromboembolism (subgroup analysis based on different types of factor Xa inhibitors) Show forest plot

8

11416

Odds Ratio (M‐H, Random, 95% CI)

0.82 [0.66, 1.04]

2.6.1 Apixaban

3

2459

Odds Ratio (M‐H, Random, 95% CI)

0.86 [0.54, 1.35]

2.6.2 Rivaroxaban

3

4981

Odds Ratio (M‐H, Random, 95% CI)

1.14 [0.75, 1.71]

2.6.3 Edoxaban

2

3976

Odds Ratio (M‐H, Random, 95% CI)

0.66 [0.48, 0.92]

2.7 Major bleeding (subgroup analysis based on different types of factor Xa inhibitors) Show forest plot

8

11447

Odds Ratio (M‐H, Random, 95% CI)

0.71 [0.36, 1.41]

2.7.1 Apixaban

3

2503

Odds Ratio (M‐H, Random, 95% CI)

0.36 [0.07, 1.91]

2.7.2 Rivaroxaban

3

4968

Odds Ratio (M‐H, Random, 95% CI)

0.49 [0.31, 0.79]

2.7.3 Edoxaban

2

3976

Odds Ratio (M‐H, Random, 95% CI)

1.44 [0.80, 2.58]

2.8 Recurrent venous thromboembolism (sensitivity analysis by including only studies at low risk of bias) Show forest plot

6

8992

Odds Ratio (M‐H, Random, 95% CI)

0.78 [0.54, 1.14]

2.9 Major bleeding (sensitivity analysis by including only studies at low risk of bias) Show forest plot

6

8979

Odds Ratio (M‐H, Random, 95% CI)

0.91 [0.42, 1.97]
Figures and Tables -
Comparison 2. Oral factor Xa inhibitors versus conventional anticoagulation
Navigate to table in Review