Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Is long-term Rivaroxaban superior to Warfarin in pulmonary embolism at 6 months?

Three Part Question

In [patients with proven pulmonary embolism (PE)] is [Rivaroxaban better than warfarin] at [reducing mortality and recurrence of venous thromboembolism at 6 months]?

Clinical Scenario

A 52-year-old lady has presented to the Emergency Department with a suspected PE. This is confirmed by CT pulmonary angiography (CTPA). Consequently she requires anticoagulation. Hospital guidelines suggest the use of a low molecular weight heparin (LMWH) followed by 6 months of Warfarin therapy. You wonder whether Rivaroxaban, a novel oral anticoagulant (NOAC), would be a better treatment option for her and patients like her.

Search Strategy

(lung embolism/OR pulmonary embol$.mp. OR lung infarction/OR pulmonary infarct$.mp. OR thromboembol$.mp. OR lung embol$.mp. OR AND (exp blood clotting factor 10a inhibitor/ OR exp rivaroxaban/ OR OR AND (exp heparin/ OR exp low molecular weight heparin/ OR OR exp tinzaparin/ OR exp dalteparin/ OR exp enoxaparin/ OR exp nadroparin/ OR exp warfarin/ OR OR exp coumarin/ OR OR vitamin k OR exp antivitamin K/) AND (exp cardiovascular mortality/ OR exp mortality/ OR recurrent or exp recurrent disease/ OR recurrent OR recurrent

LIMIT to Human and English Language.

Search Outcome

Embase (1974 to 2014 June 8): 1419 papers.

Ovid Medline (1946 to June Week 2 2014): 55 papers.

Cochrane Central Register of Controlled Trials (May 2014) & Cochrane Database of Systematic Reviews (2005 to May 2014): 30 papers.

3 Relevant papers were found; 2 were only available as Conference Abstracts but have been included for completeness. These are presented in the table

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
The Einstein-PE Investigators,
The Netherlands
4817 patients with an acute, symptomatic PE (objectively diagnosed), with or without DVT were included—2412 patients were treated with Rivaroxaban and 2405 received standard therapy. Exclusion criteria: LMWH/unfractionated heparin administered >48 h or >1 dose of warfarin prior to randomisation; received other treatment (eg, thrombolysis, embolectomy); creatinine clearance <30 mL/min; clinically significant liver disease or an ALT >3×upper limit of normal; bacterial endocarditis; contraindication to anticoagulation; systolic blood pressure >180 mm Hg or diastolic >110 mm Hg; pregnancy, breast feeding or child-bearing age without use of contraception; use of a strong inhibitor or inducer of cytochrome P-450 3A4; participation in another pharmacotherapeutic programme with 30 days; life expectancy <3 months. Randomised, open-label trial with an event-driven non-inferiority design. Patients from 263 sites across 38 countries. Patients either received: 1. Rivaroxaban (15 mg twice daily for 3 weeks followed by 20 mg once daily) or 2. Standard therapy. This involved enoxaparin at 1 mg/kg body weight twice daily until the INR was between 2 and 3 and they had received at least 5 days of Enoxaparin; Warfarin or Acenocoumarol were started within 48 h of randomisation and adjusted accordingly to maintain an INR of 2.0–3.0. Intended length of treatment was decided by the clinician before randomisation; this was also the duration of followed up. A ‘pre-study’ phase was used to confirm the adequacy of the Rivaroxban dose since the regimen was taken from a population with DVTs, not PEs Symptomatic recurrent VTE (fatal/non-fatal DVT or PE)50 (2.1%) on Rivaroxaban vs 44 (1.8%) receiving standard therapy. HR 1.12; CI 0.75 to 1.68; p=0.003 (p=0.57 for superiority).Open design potentially resulted in a degree of bias against the Rivaroxaban group since there was high clinical suspicion for VTE recurrence in this group and yet similar rates after these patients underwent objective diagnostic assessment. Exclusion criteria were extensive and may have significantly changed the general patient cohort who are at risk of PE ie, does excluding patients who have contraindications to anticoagulation treatment exclude all patients with cancer? Further, these patients generally receive long-term LMWH rather than Warfarin anyway. The regimen for the standard therapy (described earlier) is slightly different to current NICE guidelines in use. Outcomes are less reliable after 6 months since not all patients (257; 10.7%) completed their target treatment durations due to discontinuation of the study. Mortality figures include deaths in which PE could not be ruled out; how reliable is this? The mean age is relatively young (58) with few patients over the age of 80; this is not particularly representative of the UK population that is at increased risk of PE
Clinically relevant bleeding (major or clinically relevant non-major bleeding)249 (10.3%) on Rivaroxaban vs 274 (11.4%) receiving standard therapy (HR 0.90; CI 0.76 to 1.07; p=0.23) Major bleeding alone occurred in 26 (1.1%) vs 52 (2.2%) patients on Rivaroxaban or standard therapy, respectively (HR 0.49; CI 0.31 to 0.79; p=0.003).
Net Clinical Benefit83 (3.4%) on Rivaroxaban vs 96 (4.0%) receiving standard therapy (HR 0.85; CI 0.62–1.14; p=0.28).
Mortality58 (2.4%) on Rivaroxaban vs 50 (2.1%) receiving standard therapy (HR 1.13; CI 0.77 to 1.65; p=0.53). Due to PE or PE not ruled out: 11 vs 7 on Rivaroxaban vs standard therapy.
Wasserlauf et al.
5 RCTs including 23,063 patients. 11 536 patients were assigned to Rivaroxaban and 11 527 to a Vitamin K Antagonist (VKA). 4832 of these patients were treated for acute, symptomatic pulmonary emboli Systematic review and meta-analysis of RCTs, with predominantly safety outcomes. Inclusion restricted to treatment durations of >30 days Fatal bleedingSignificantly decreased in Rivaroxaban group. Relative risk (RR): 0.48; CI: 0.31-0.75; Number needed to treat (NNT) to prevent one major bleed: 500 Conference Abstract presented at the American Heart Association 2012 Scientific Sessions and Resuscitation Science Symposium. Results presented are for all patients studied, not specifically for those presenting with acute, symptomatic PE. Primary outcome involves major/fatal bleeding with no mention of recurrent venous thromboembolism (VTE) as an outcome. All-cause mortality is not particularly helpful since this is low in patients with PE and is therefore difficult to statistically analyse; further to this it is unclear whether any mortality represented here is attributable to PE or to other comorbidities. Although only studies with outcomes >30 days have been included, their actual duration may still be too limited to fully assess efficacy and safety outcomes for patients being anticoagulated for 3–6 months
All-cause mortalityNo increase seen in Rivaroxaban group (RR 0.89; CI 0.73 to 1.09).
Kohler et al,
2249 patients enrolled onto a regional registry, of which 72 were treated with Rivaroxaban for an acute PE. Average age 67.3 years; 55.6% female; 23.6% treated for recurrent VTE. Treatment discontinuation rates were approximately 5% Conference Abstract only; presented at the 55th Annual Meeting of the American Society of Haematology, ASH 2013. Prospectively collected data to assess efficacy and safety outcomes of Rivaroxaban in unselected patients with acute PE Recurrent VTE1/72 (1.4%) patients. Equivalent to 1.7/100 patient years.Prospectively collected data with potentially pre-selected patients being assigned to the registry in which all patients receive Rivaroxaban. No comparative group in this unselected patient cohort. Only small patient numbers are being treated specifically for acute PE. It states that 23.6% of patients in the registry were being treated for recurrent VTE. It is unclear whether some of these patients were also included in the 72 patients with acute PE. Such patients would have had previous anticoagulation, potentially affecting the clinical outcomes. No mention of the Rivaroxaban regimen used. Further details of distribution of PE, other treatments received, underlying conditions etc. would be required to fully analyse the results
Mortality3/72 (4.2%) patients; none attributable to VTE. 5 events/100 patient years.
Major bleeding2/72 (2.8%) patients; both non-fatal vaginal bleeds. 3.3 events/100 patient years.
Minor bleeding23/72 (31.9%) patients. 38.3 events/100 patient years.


The anticoagulation regimen used in the treatment of acute, symptomatic PE has been established for the best part of half a century and continues to be used in guidelines around the world. The recent introduction of NOACs may potentially transform the way in which we treat such patients. This would likely be appreciated by patients and clinicians given the inconvenient, time-consuming and often unreliable schedule that currently has to be followed when patients are prescribed a Vitamin K Antagonist (VKA) such as Warfarin.

There has only been one randomised controlled trial to date that has assessed clinical efficacy and safety outcomes when directly comparing Rivaroxaban with the standard therapy of our current practice. This trial found that with Rivaroxaban there were similar rates of recurrent VTE and death but significantly fewer episodes of major bleeding than with Warfarin. There are, of course, differences between the design of this trial and standard practice in the UK; most of these have been addressed by the National Institute for Health and Clinical Excellence (NICE) in their recommendations.4 Therefore, accounting for the study weaknesses I have discussed in the table above, the appraising committee believes the results of this trial to be generalisable to the UK population, excluding the relatively low average age in the trial. They also highlighted the need for further studies to perform sub-group analyses of patients with underlying conditions that constitute a significant portion of patients presenting with an acute PE. For example, patients with cancer and severe renal disease are potentially at greater risk of both recurrent VTE and haemorrhagic complications and therefore whether Rivaroxaban sustains its promising efficacy and safety profile in these patients is up for debate.

Finally, a key thing to consider for any potential treatment is cost-effectiveness. A conference abstract, published in July 2013, discussed this in relation to the National Health Service (NHS).5 The authors concluded that cost savings of £396, £213 and £133 for 3-, 6- and 9-month courses of treatment respectively could be achieved by using Rivaroxaban instead of LMWH/VKA. For patients requiring life-long anticoagulation an incremental cost-effectiveness ratio (ICER) of £7072/quality-adjusted-life-years (QALY) was calculated; this was again deemed cost-effective.

Editor Comment

ALT, alanine transaminase; DVT, deep vein thrombosis; INR, International normalised ratio; LMWH, low-molecular weight heparin; NICE, National Institute for Health and Clinical Excellence; PE, pulmonary embolism; RCT, randomised controlled trial; VTE, venous thromboembolism.

Clinical Bottom Line

There is a severely limited amount of evidence surrounding the efficacy and safety of Rivaroxaban, especially concerning its use in the general patient population in all except those with an absolute contraindication to its use. Despite this, the Einstein-PE study published promising outcome data in relation to Rivaroxaban and its use in patients with acute, symptomatic PE. NICE have also appraised their work and as a result created their own recommendations. With a similar efficacy and safety profile to low molecular weight heparin (LMWH)/VKA, as well as proving to be a cost-effective alternative, I have no doubt that with additional evidence Rivaroxaban or a similar NOAC has the ability to modernise anticoagulation treatment as we know it.


  1. The Einstein-PE Investigators. Oral Rivaroxaban for the treatment of symptomatic pulmonary embolism. N Engl J Med 2012;366:1287–1297.
  2. Wasserlauf G, Grandi SM, Filion KB, et al . Safety of Rivaroxaban versus vitamin K antagonists: A systematic review and meta-analysis. Circulation 2012;126:A16558.
  3. Kohler C, Werth S, Tittl L, et al. Acute treatment of pulmonary embolism with Rivaroxaban – Real life data from the prospective dresden NOAC registry. Blood 2013;122:2380.
  4. Adan J, Squire I, Ades AE, et al . Rivaroxaban for treating pulmonary embolism and preventing recurrent venous thromboembolism.[NICE ONLINE] NICE technology appraisal guidance [TA287] June 2013 (accessed 6 Jul 2014).
  5. McLeod E, Guillermin A, Hudson R, et al . Cost-effectiveness of Rivaroxaban for the treatment of pulmonary embolism and secondary prevention of venous thromboembolism – a UK perspective. J Thromb Haemost 2013;11(Suppl. 2):OC 02.6.