Intended for healthcare professionals


New oral anticoagulants for preventing venous thromboembolism

BMJ 2012; 344 doi: (Published 14 June 2012) Cite this as: BMJ 2012;344:e3820
  1. Elliott R Haut, associate professor of surgery, anaesthesiology/critical care medicine, and emergency medicine12,
  2. Brandyn D Lau, medical informatician, department of medicine1,
  3. Michael B Streiff, associate professor of medicine and pathology12
  1. 1Johns Hopkins Medical Institutions, Baltimore, MD 21287, USA
  2. 2Armstrong Institute for Patient Safety and Quality, Johns Hopkins Medicine, Baltimore, MD, USA
  1. ehaut1{at}

Are we at the point of diminishing returns?

Orthopaedic surgery is known to be associated with a high risk of venous thromboembolism, and prophylaxis for orthopaedic patients is vital. In the linked systematic review and meta-analysis (doi:10.1136/bmj.e3675), Gómez-Outes and colleagues present a comparative effectiveness review that examines newer anticoagulant agents (dabigatran, rivaroxaban, and apixaban) in the prevention of venous thromboembolism after hip or knee replacement surgery.1

This review used randomised controlled trials that had directly compared one of the newer agents with enoxaparin to indirectly compare the effects of these drugs on venous thromboembolism outcomes and clinically relevant bleeding. The review therefore gives us an idea of what the results of a head to head trial would be like. This type of network meta-analysis in which multiple treatments are compared using both direct comparisons of interventions within randomised controlled trials and indirect comparisons across trials on the basis of a common comparator is on the methodological cutting edge of systematic reviews, but it can be fraught with peril if not done with scientific rigour.2 It found that rivaroxaban led to significantly lower rates of symptomatic venous thromboembolism than enoxaparin but at the cost of significantly increased bleeding. Both apixaban and dabigatran were as effective as enoxaparin in preventing venous thromboembolism, but apixaban had a lower risk of bleeding.

The four agents showed no difference in efficacy with regard to the “net clinical endpoint”—a composite of symptomatic venous thromboembolism, major bleeding, and all cause mortality.1 Readers should take away four important conclusions.

Firstly, we are reaching the point of diminishing returns with newer anticoagulants for preventing venous thromboembolism. Pushing thrombosis rates lower (as was done with rivaroxaban) causes more bleeding. Unless new antithrombotic agents are developed that target pathological thrombus formation without disrupting normal postsurgical haemostasis, further reductions in venous thromboembolism will come at the cost of increased bleeding.

Secondly, patient preferences should be considered when making decisions about prophylaxis against venous thromboembolism. The four agents examined had different degrees of risk for venous thromboembolism and clinically relevant bleeding, yet all had similar overall outcomes as measured by the author’s composite net clinical end point.1 The authors weighted equally the negative consequences of a thromboembolism and those of a clinically relevant bleeding event. However, patients may not necessarily weight them the same. Some patients might prefer to receive a two unit blood transfusion rather than develop deep vein thrombosis, which requires months of anticoagulation and can result in post-thrombotic syndrome or pulmonary embolism. Others might prefer a symptomatic deep vein thrombosis, which can be identified and treated, rather than a devastating haemorrhagic stroke. The specific definition of “major bleeding” is crucial. The importance of including patient preferences and their impact on decision making have been emphasised by the National Institute for Health and Clinical Excellence (NICE),3 the American College of Chest Physicians,4 and the newly formed Patient-Centered Outcomes Research Institute (PCORI; When pooling different outcomes to determine the net clinical benefit of a given intervention for meta-analysis, it is important to appropriately weigh and balance the harms and benefits associated with the intervention.

Thirdly, in a real world setting effectiveness may differ greatly because adherence to drug regimens may vary as a result of differences in the route of administration (oral versus subcutaneous injections), frequency, duration of treatment, and side effects.

Fourthly, the authors re-emphasised the fact that venous thromboembolism still occurs, even with current best practice prophylactic regimens in highly selected patient populations in clinical trials.5 A zero rate of venous thromboembolism is an unattainable goal. Policy makers should not therefore designate it a “never event” or use venous thromboembolism rates alone for pay for performance or quality metrics, as proposed by the NHS Quality and Outcomes Framework in the United Kingdom and the Centers of Medicare and Medicaid Services in the United States.5 6 Instead, a more reasonable and attainable goal is to eliminate preventable harm—defined as venous thromboembolism associated with suboptimal prophylaxis.7

Throughout the world many organisations are creating guidelines for best practice in prophylaxis against venous thromboembolism in patients undergoing orthopaedic surgery, including NICE,2 the American College of Chest Physicians,4 and the American Association of Orthopaedic Surgeons.8 These organisations rely on synthesised evidence, such as the accompanying article and a recent systematic review sponsored by the Agency for Healthcare Research and Quality,9 to guide clinicians. However, even with published guidelines, shockingly few patients get the optimal prophylaxis they deserve. In one US study only 42% of patients with deep vein thrombosis had received appropriate prophylaxis during a recent hospital admission.10 The ENDORSE study, encompassing more than 68 000 patients in 32 countries, found similarly low adherence to prophylaxis guidelines (39.5% of medical patients and 58.5% of surgical patients).11 A 2012 study reported that only 40% of Austrian patients in intensive care received prophylaxis in line with guidelines.12

Clinical research and evidence based medicine have focused on generating and synthesising new evidence, yet it is equally, if not more, important to ensure that all patients receive optimal care. This knowledge translation aspect of medical care is a substantial gap in the medical literature. Pronovost and colleagues offered the “translating evidence into practice” (TRIP) scientific framework on which to base quality improvement interventions.13 Using this structure, our multidisciplinary collaborative has successfully implemented interventions that have improved prophylaxis against venous thromboembolism at our institution, and we urge clinicians elsewhere to concentrate on implementing evidence based measures.14


Cite this as: BMJ 2012;344:e3820


  • Research, doi:10.1136/bmj.e3675
  • Competing interests: All authors have completed the ICMJE uniform disclosure form at (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; ERH is the primary investigator of a mentored clinician scientist development award K08 1K08HS017952-01 from the Agency for Healthcare Research and Quality entitled “Does screening variability make DVT an unreliable quality measure of trauma care?”, receives royalties from Lippincott, Williams & Wilkins for a book Avoiding Common ICU Errors, and has given expert witness testimony in various medical malpractice cases. MBS has received research funding from Sanofi-Aventis and Bristol Myers Squibb and honorariums for CME lectures from Sanofi-Aventis and Ortho-McNeil; has consulted for Sanofi-Aventis, Eisai, Daiichi-Sankyo, and Janssen HealthCare, and has given expert witness testimony in various medical malpractice cases; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract