It’s All Relative: Rivaroxaban for Extended Secondary Prevention of VTE


Written By

Reviewed By


Weitz JI, Lensing AWA, Prins MH, et al. Rivaroxaban or aspirin for extended treatment of venous thromboembolism. N Engl J Med 2017;376(13):1211-22.

Whether – or not - to continue anticoagulation therapy to prevent venous thromboembolism (VTE) after the initial acute treatment period requires shared decision making.  The risks and benefits are patient and situation-specific.  Numerous factors need to be considered. And the risks and benefits change over time. The EINSTEIN CHOICE study provides some important new insights to help inform this decision.1


Tens of thousands die from venous thromboembolism (VTE) every year.2 Prevention is the key to reducing morbidity and mortality. Current guidelines recommend full-dose anticoagulation for at least three months for the acute treatment of VTE.  Acute treatment can be followed by extended secondary prophylaxis based on the risk for recurrent VTE and bleeding.3  Among patients with a first unprovoked VTE, 10% will have a recurrent event within one year after completing anticoagulation therapy and 30% within five years.  The duration of prior therapy does not influence the risk of VTE recurrence once anticoagulation therapy is stopped.4,5


Extended therapy with warfarin, dabigatran, rivaroxaban, or apixaban at full therapeutic doses are all highly effective in preventing VTE recurrence (Table 1) in patients at high risk of recurrent VTE.3,6-9  Aspirin is also effective but appears to be less effective than full-dose anticoagulation therapy.10  Would low-dose therapy with a direct oral anticoagulant (DOAC) be as effective as standard treatment doses for the extended secondary prevention of recurrent VTE? Is low-dose DOAC therapy more effective than aspirin? Is low-dose DOAC therapy safer?  Does it improve the benefit:risk ratio? These are the questions the EINSTEIN CHOICE study attempted to answer.1


Table 1:  Summary of Trials for Extended Anticoagulaton for Secondary Prevention of VTE6-9 




Primary Efficacy Outcome
(Active vs. Placebo)

Recurrence Rate for VTE (Active vs. Placebo)
HR (95% CI)

Major Bleeding
(Active vs. Placebo)



Rivaroxaban 20 mg (n=602) vs. placebo (n=594) daily x 6-12 months after completion of 6 or 12 months of VTE treatment

Recurrent VTE:
1.3% vs. 7.1%;

0.18 (0.09-0.39);
p < 0.001

0.7% vs. 0%; p=0.11



Apixaban 5 mg (n=813) vs. 2.5 mg (n=840) BID for up to 12 months vs. placebo (n=829) after completion of 6 or 12 months of VTE treatment

Recurrent VTE or death from any cause:

5 mg: 4.2%
2.5 mg: 3.8%
Placebo: 11.6%

2.5 mg vs. placebo:
0.33 (0.22-0.48);
p < 0.01

5 mg vs. placebo:
0.36 (0.25-0.53);
p < 0.001

5 mg: 0.1%
2.5 mg: 0.2%
Placebo: 0.5%

P values not reported



Dabigatran 150 mg (n=681) vs. placebo (n=662) twice daily x 6 months after 6-18 months of VTE treatment

Recurrent or fatal VTE or unexplained death: 0.4% vs. 5.6%

0.08 (0.02-0.25);
p < 0.001

0.3% vs. 0%
P value not reported



Low-intensity warfarin (target INR 1.5-2.0; n=255) vs. placebo (n=253) after 3 to 12 months of VTE treatment

Recurrent VTE: 2.6 vs. 7.2 per 100 person-years

0.36 (0.19-0.67);
p < 0.001

0.9 vs. 0.4 per 100 person-years (p=0.25)

BID: twice daily; CI: confidence interval; HR: Hazard Ratio; INR: international normalized ratio; VTE: venous thromboembolism; HR: hazard ratio


The EINSTEIN CHOICE study enrolled 3,365 patients and randomized them to receive rivaroxaban 20 mg, rivaroxaban 10 mg, or aspirin 100 mg daily. Patients who had a clear indication for extended anticoagulation were excluded along with patients with active bleeding or patients at a high risk for bleeding. 2  The patient population was heterogeneous. The average age of patients was 59 years old, 77% were Caucasian, and 45% were female.


The primary outcome was a composite of recurrent symptomatic or fatal VTE or unexplained death.  Other efficacy outcomes included myocardial infarction, ischemic stroke, and death from any cause. After nearly 1 year of follow-up,  there was a statistically significant difference between both doses of rivaroxaban versus aspirin – but no statistical difference between rivaroxaban doses – in the primary outcome (See Table 2). Patients treated with rivaroxaban 20 mg or 10 mg daily at any point in the study were 66% and 74%, respectively, less likely to experience the primary outcome compared to patients treated with aspirin (absolute risk reduction [ARR] 2.9% and 3.2%, respectively). A Kaplan-Meier curve was used to illustrate the difference of cumulative primary outcome events over time between the three treatment groups.  Clear differences between the two rivaroxaban groups and aspirin could be seen as early as 30 days after randomization.


Table 2: Major Results from EINSTEIN CHOICE1



Primary Efficacy Outcomea

n (%)


Recurrence Rate for VTE (Active vs. Comparator)

HR (95% CI)



Incidence of Major Bleeding

n (%)

Major Bleeding

(Active vs. Comparator)

HR (95% CI)

Rivaroxaban 20 mg (N=1107)

17 (1.5)

0.34 (0.20-0.59);
p < 0.001



6 (0.5)

2.01 (0.5-8.04); p=0.32

Rivaroxaban 10 mg (N=1127)

13 (1.2)

0.26 (0.14-0.47);
p < 0.001



5 (0.4)

1.64 (0.39-6.84); p=0.50

Aspirin (N=1131)

50 (4.4)




3 (0.3)


ARR: absolute risk reduction; CI: confidence interval; HR: Hazard Ratio; NNT: number needed to treat; VTE: venous thromboembolism

a Recurrent venous thromboembolism

b Number of patients who would need to be treated for up to 12 months with rivaroxaban instead of aspirin to prevent one episode of fatal or nonfatal recurrent VTE without increasing the risk of bleeding


Major bleeding and clinically non-relevant nonmajor bleeding were also evaluated.  There was no difference in major bleeding between the three treatment groups. Bleeding events were defined by the International Society on Thrombosis and Hemostasis bleeding scale.11 Major bleeding was defined as clinically overt and associated with a fall in hemoglobin ≥ 2 g/dL, leading to a transfusion of ≥ 2 units packed red blood cells or whole blood, bleeding in a critical site (e.g., intracranial), and contributing to death.


With a clear reduction in recurrent VTE but without an increase in major bleeding risk, the results of this study show that extended treatment beyond the recommended duration of 6-12 months with rivaroxaban at therapeutic and prophylactic doses has a clinically important impact on patient outcomes.


Currently, rivaroxaban 10 mg daily is only approved for VTE prophylaxis following knee and hip replacement.3 Rivaroxaban 20 mg daily is indicated for VTE treatment (following an induction regimen of 15 mg twice daily for 21 days) and stroke prevention in patients with atrial fibrillation.2 Currently, there were no guideline recommendations nor FDA approval for the use of rivaroxaban for extended VTE prophylaxis beyond the initial treatment period, so it was reasonable to study both treatment and prophylactic doses. Comparing rivaroxaban to aspirin 100 mg daily is reasonable as aspirin has previously been studied for prevention of recurrent VTE. However, guidelines do not recommend the use of aspirin over anticoagulation therapy unless the patient refuses anticoagulation.2,


In patients with a previous VTE treated for at least 3 months, extended anticoagulation remains a topic of debate requiring a careful assessment of patient specific characteristics in order to weigh the risk of a recurrent thrombotic event versus bleeding.3 Due to the increased risk of bleeding with therapeutic anticoagulation, this study aimed to evaluate the efficacy and safety of a prophylactic dose of rivaroxaban to prevent recurrent VTE. The results of this study are consistent with findings of the AMPLIFY-EXT trial, which found extended anticoagulation with apixaban at either a treatment or a prophylactic dose resulted in a significant reduction in the risk of recurrent VTE and a similar safety profile compared to placebo (Table 1).6 Unlike the AMPLIFY-EXT trial, the EINSTEIN CHOICE trial used an active control, aspirin, rather than placebo. The EINSTEIN CHOICE trial also reinforces the concept that aspirin therapy is less effective than anticoagulant therapy and supports the CHEST guideline recommendation against the use of aspirin in place of anticoagulation for the prevention of recurrent VTE.3 The decision to stop anticoagulation therapy and start aspirin places the patient at a higher risk of recurrent VTE but is not safer. Since the EINSTEIN CHOICE study found that both the 20 mg and 10 mg doses of rivaroxaban effectively reduced the risk of recurrent VTE with a similar bleeding risk when compared to aspirin, rivaroxaban is a more favorable option compared to low-dose aspirin for extended anticoagulation. It should be noted, the EINSTEIN CHOICE study was designed to compare rivaroxaban 10 mg and 20 mg with low-dose aspirin (100 mg) and not designed to determine if the 10mg dose is non-inferiority to rivaroxaban 20 mg. None-the-less, the event rates and NNT are very similar.


Now that low-dose apixaban and rivaroxaban both have data to support their use for extended treatment of VTE in patients at equipoise regarding the continuation of anticoagulant therapy, which treatment should be use? There are no head-to-head trials comparing apixaban to rivaroxaban in either treatment or prophylactic doses. One of the benefits of rivaroxaban is its once daily dosing, whereas apixaban is dosed twice daily even when used in a prophylactic dose. Once daily dosing may be preferred by some patients for ease of administration. However, because apixaban has data to support its use in patients with renal impairment, it may be preferred over rivaroxaban in patients with impaired or declining renal function. The clinical trials evaluating rivaroxaban for the treatment of VTE excluded patients with a CrCl < 30 mL/min. The bleeding profile of each drug should also be considered, in addition to patient age, comorbidities, and drug interactions. The choice of anticoagulant will undoubtedly be limited for many patients by drug formularies and insurance coverage.


Rivaroxaban 10 mg once daily offers a favorable efficacy and safety profile for patients who choose extended VTE therapy.  Once daily dosing makes it more convenient.  Should rivaroxaban be the preferred option, even in patients who have been initially treated with warfarin? Or another DOAC?  What do you think?




Can you help me understand why 60% of the patients had provoked VTE? I'm not sure how those patients would have clinical equipoise to continue therapy since guidelines for those patients have a defined duration of treatment. Only about 7% had a known thrombophilia, only about 3% had active cancer... This is compared to AMPLIFY-EXT where >90% were unprovoked DVTs. Thanks in advance for your thoughts, I am not an anticoag guru!

Great question! You're right - patients with an unprovoked VTE are more likely to benefit from extended AC therapy based on risk of recurrence after stopping therapy (10% at 1 year & 30% at 5 years). Those with a provoked VTE are also at risk of recurrence, albeit less so (1-5% at 1 year and 3-15% at 5 years). The authors of EINSTEIN CHOICE acknowledge that the "risk of recurrence among those in whom [VTE] was provoked by minor transient or persistent risk factors is less certain." EINSTEIN CHOICE doesn't provide details on the patients with provoked VTEs, unfortunately. However, whether patients had a provoked or unprovoked VTE, rivaroxaban reduced the risk of recurrence by up to 70%. Given all of this, the benefit of extended AC therapy is more clear with an unprovoked VTE vs. provoked, but both groups may benefit. In each case, the risk of bleeding should be assessed prior to extending therapy and ongoing benefit should be assessed an readdressed at least annually.