People leaving hospital after medical illness do not benefit from extended clot reducing treatment

Taking rivaroxaban after discharge from hospital does not significantly reduce the risk of venous thromboembolism, either blood clots in large veins or of dying from clots travelling to the lungs.

People who are hospitalised with conditions such as heart failure and stroke are at an increased risk of blood clots. This risk is further increased by factors such as age, lack of mobility and previous illness or clots. These patients will usually be given anticoagulant treatment to reduce blood clotting while in hospital, but it isn't certain whether continuing an anticoagulant is beneficial.

This large multinational trial sought to find out whether using the anticoagulant rivaroxaban for 45 days after discharge was effective. Continued use offered a very small reduction in the risk of a venous thromboembolism, but also slightly increased the risk of major bleeding. Neither result reached statistical significance.  The trial was terminated because of the lower than expected number of clots.

Why was this study needed?

As many as 1 in 20 people will be affected at some point in their lives by a blood clot in a large vein. If unsuspected or left untreated, part of the clot can travel to the lungs (a pulmonary embolism).

Half of all cases are linked to hospitalisation for medical illness or surgery. People at risk are usually given treatment to reduce blood clotting while they are in hospital.

Currently, it is uncertain whether this treatment should be continued after discharge because, although the chance of a clot forming is reduced, the risk of bleeding is increased.

What did this study do?

This randomised controlled trial included 12,019 patients from 671 centres in 36 countries.

Their average age was 70, and all had been hospitalised for a medical illness such as heart failure, chronic obstructive pulmonary disease or stroke.  All were on heparin in hospital and risk of venous thromboembolism (VTE) was assessed using the IMPROVE score. On discharge, they were randomly assigned to take placebo or rivaroxaban for 45 days.

People were excluded if they were already taking an anticoagulant (such as warfarin or dabigatran) or dual antiplatelet therapy (aspirin plus dipyridamol) or if they had active cancer, recent bleeding or a high risk of bleeding.

The inclusion criteria may have led to a greater proportion of patients at a lower than intended risk thereby diminishing the benefit of anticoagulant prophylaxis. In addition, half were taking aspirin which may also have reduced their risk of embolism.

What did it find?

  • There was no difference in the proportion of people who had a symptomatic or fatal VTE. Either occurred in 0.83% (50/6,007) of the rivaroxaban group and 1.10% (66/6,012) of the placebo group (hazard ratio [HR] 0.76, 95% confidence interval [CI] 0.52 to 1.09). The difference in risk between the intervention and placebo group was minimal at 0.27% (95% CI, -0.61 to 0.08).
  • When looking separately at symptomatic non-fatal VTE, there were slightly fewer cases, 0.18% (11/6,007) in the rivaroxaban group compared to 0.42% (25/6,012) of the placebo group (HR 0.44, 95% CI 0.22 to 0.89).
  • Deaths due to VTE were similar at 0.72% (43/6,007) in the rivaroxaban group compared to 0.77% (46/6,012) in the placebo group.
  • Safety-wise, major bleeding occurred in 0.28% (17/5,982) patients in the rivaroxaban group and 0.15% (9/5,980) patients in the placebo group (HR 1.88, 95% CI 0.84 to 4.23). The difference in risk between the intervention and placebo group was 0.13% (95% CI, -0.03 to 0.30).

What does current guidance say on this issue?

The NICE 2018 guideline - venous thromboembolism in over 16s: reducing the risk of hospital-acquired deep vein thrombosis or pulmonary embolism - does not specifically focus on treatment post hospital discharge. However, it does state that those who are discharged with anticoagulants or other prophylaxis should be confident using it, know the symptoms and who to contact should they arise.

What are the implications?

In this select group of patients, rivaroxaban did not significantly reduce the incidence of fatal or symptomatic venous thromboembolism.

There were fewer than predicted events in both groups, so it may be that rivaroxaban could still be beneficial for people who are at high risk.

The rate of venous thromboembolism was lower than expected, and there remains some uncertainty in management of people at higher risk.

Citation and Funding

Spyropoulos AC, Ageno W, Albers GW et al; MARINER Investigators. Rivaroxaban for thromboprophylaxis after hospitalization for medical illness. N Engl J Med. 2018;379(12):1118-27.

Funded by Janssen Research and Development; MARINER ClinicalTrials.gov number, NCT02111564.



E-Learning for Health. E-VTE. London: NHS Health Education England; 2014.

Gibson CM, Spyropoulos AC, Cohen AT et al. The IMPROVEDD VTE risk score: incorporation of D-dimer into the IMPROVE score to improve venous thromboembolism risk stratification. TH Open. 2017;1(1): e56-65.

NHS website. Deep vein thrombosis. London: Department of Health and Social Care; updated 2016.

NICE. Venous thromboembolism in over 16s: reducing the risk of hospital-acquired deep vein thrombosis or pulmonary embolism.NG89. London: National Institute for Health and Care Excellence; 2018.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre



Venous thromboembolism  risk was assessed according to their score on the modified International Medical Prevention Registry on Venous Thromboembolism (IMPROVE).



Expert commentary

It is routine in the UK to use low dose heparinoids or mechanical methods to prevent venous thromboembolism in acutely ill medical patients, but there is debate about how long to continue this treatment, especially for those with long stays in hospital or other care facilities.

This trial does not directly answer this question but sheds some light on it, by testing the use of a different anti-thrombotic approach using rivaroxaban for seven weeks after a 3 to 10-day acute hospital stay.

The largely neutral results show that the actual rates of venous thromboembolism and significant bleeding are both low and that any benefits of thromboprophylaxis at this point are tiny. Prolonged thromboprophylaxis seems to be relatively unimportant compared to its value in the acute setting.

JRF Gladman, Professor of the Medicine of Older People, Division of Rehabilitation and Ageing, School of Medicine

The commentator declares no conflicting interests


Expert commentary

The role for thromboprophylaxis in medical patients after discharge remains controversial. Newer trials have sought to recruit larger numbers, focus on patients at highest risk and assess symptomatic events.

The MARINER study identified increased risk patients by using the IMPROVE risk score combined with D-dimer. Conventional thromboprophylaxis was given during admission and patients randomized at discharge to 45 days of rivaroxaban or placebo.

This large and rigorously designed medical thromboprophylaxis study showed low event rates and failed to demonstrate a benefit suggesting a limited utility for such an approach in medical patients in the future.

Roopen Arya, Professor of Thrombosis and Haemostasis, King’s College Hospital

The commentator declares previous support from the manufacturer of rivaroxaban