Bruyère Reports
New oral anticoagulants for venous thromboembolism prophylaxis. A Bruyère Rapid Review
Report authors: Elizabeth Ghogomu, Shalini Sani, Vivian Welch, Tim Veregin, Jean Chouinard
Executive Summary
In this rapid review we sought to find out whether new oral
anticoagulants (NOACs) could be used in place of low molecular weight heparins
(LMWHs) in all patient populations across Bruyère Health and whether the use of NOACs would achieve cost-savings
for Bruyère over the use of dalteparin, a LMWH. We compared the effectiveness,
cost-effectiveness and safety of NOACs versus LMWHs.
Venous thromboembolism (VTE) is the formation of a blood clot in a vein. The clot may get detached and travel in the blood (embolus) to other parts of the body. The incidence for VTE is 1 per 1000 person- years in the community and 96 per 1000 person-years in hospitalized patients.
Many cases of VTE are preventable with anticoagulants alone, or in combination with general methods (e.g. mobilization and leg exercises), and mechanical methods (e.g. graduated compression stockings). Dalteparin is currently the treatment of choice for venous thromboembolic prophylaxis at Bruyère Health as recommended by clinical practice guidelines. The new oral anticoagulants are increasingly being used for VTE prophylaxis in patients with atrial fibrillation and to a lesser extent for VTE prevention after knee or hip replacement surgery in geriatric rehabilitation.
We searched for systematic reviews, health technology assessments or economic evaluations and guidelines and found and screened 2000 potentially relevant articles. Forty met our inclusion criteria: 22 reviews focused on the effectiveness and safety of NOACs versus LMWHs; 10 economic evaluations assessed the cost-effectiveness of NOACs versus LMWHs with three done in Canada; and eight guidelines addressed VTE prophylaxis with NOACs.
Based on our findings we suggest the following:
- Patient risk assessment for VTE risk and risk of bleeding should be done before deciding whether or not VTE prophylaxis should be used, and which type. For VTE risk assessment, additional risks such as the clinical condition or reason for hospitalization should also be taken into account. When assessing patients for risk of bleeding, a balance between actual and perceived risk should be considered as well as contraindications for prophylaxis.
- Various guideline groups recommend thromboprophylaxis with either NOACs or LMWH in patients including elderly with hip or knee joint replacement surgery, provided patients have no contraindications. For patients undergoing orthopedic surgery who refuse injections, NOACs is recommended. NOACs are more cost-effective than LMWH for patients with hip or knee joint replacement even when risk of major bleeds is considered.
- For the medically ill, there is higher risk of bleeds (4 per 1000, from 1-7 more per 1000; high certainty of evidence) with NOACs and guideline groups recommend LMWH or unfractionated heparin for this population.
- For palliative care, we found no systematic reviews but two guidelines recommend the use of LMWH for thromboprophylaxis.
The full report is available for download.
Table of Contents
Key messages
New oral anticoagulants (NOACs) also known as Direct-acting oral anticoagulants (DOACs) are cost-effective and easier to administer than low molecular weight heparins (LMWHs). Antidotes for reversing the anticoagulant effect in case of severe bleeding exist for LMWHs. Three antidotes for NOACs are under development and one (idarucizumab (PRAXBIND) has recently been approved for dabigatran. The guideline recommendations were based on studies conducted before the availability of an antidote.
- The decision and choice of venous thromboembolism (VTE) prophylaxis should be based on patient risk assessments of VTE risk and risk of bleeding.
- New oral anticoagulants are recommended for VTE prophylaxis in patients with hip or knee joint replacement surgery provided they have no contraindications.
- There was insufficient evidence to support the use of NOACs instead of LMWHs in other patient populations.
- Although NOACs are cost-effective, the choice of VTE prophylaxis should be patient-centred, considering each patient’s needs, preferences, and values.
Background
The issue:
Venous thromboembolism (VTE) is the formation of a blood clot in a vein. The clot may get detached and travel in the blood (embolism) to other parts of the body. The incidence for venous thromboembolism is 1 per 1000 person- years in the community and 96 per 1000 person-years in hospitalized patients.Common risk factors are: increasing age, active or occult malignancy, some forms of cancer chemotherapy, previous VTE, varicose veins, obesity, prolonged severe immobility (prolonged bed rest, immobilization in a plaster cast or brace or prolonged travel resulting in limited movement and subsequent venous stasis), use of oestrogen-containing hormone replacement therapy or oral contraceptives in women, inherited or acquired thrombophilia, heart failure, myocardial infarction, stroke with immobility, acute inflammatory bowel disease, severe acute infection, nephrotic syndrome, pregnancy and the puerperium, trauma, anesthesia and surgery.
Many cases of VTE are preventable with anticoagulants alone, or in combination with general methods (e.g. mobilization and leg exercises), and mechanical methods (e.g. graduated compression stockings). Anticoagulant prophylaxis is recommended in patients with no contraindications such as active bleeding, previous major bleeding, known untreated bleeding disorder, severe renal or hepatic disorder, and thrombocytopenia. The standard anticoagulant prophylaxis is with the indirect thrombin inhibitor: unfractionated heparin, and low molecular weight heparin (LMWH, such as dalteparin, enoxaparin, nadroparin, and tinzaparin), fondaparinux, or vitamin K antagonists (VKAs) such as warfarin.
Direct-acting oral anticoagulants (DOACs) also known as new oral anticoagulants (NOACs) have been approved in recent years for VTE prophylaxis in Canada. These new oral anticoagulants (NOACs) include the direct thrombin inhibitor dabigatran, and the direct factor Xa inhibitors rivaroxaban and apixaban and have been studied in certain patient populations only. They may have the advantage of easier administration, orally instead of injections and requiring no dose adjustment and monitoring, but reversing their anticoagulant effects in case of major bleeding is a concern. Three antidotes are under development and one has recently been approved for dabigatran. LMWHs are the most often used type of VTE prophylaxis in Canada. These data show a slight decrease in LMWH usage from 2013 to 2014, and this trend may continue with an increase in the usage of other anticoagulants (apixaban, rivaroxaban, dabigatran). They may cause major bleeding as well but their effects are generally reversible.
Context:
Dalteparin, a low molecular weight heparin, is currently the treatment of choice for venous thromboembolic prophylaxis at Bruyère Health as recommended by clinical practice guidelines. The new oral anticoagulants are increasingly being used at Bruyère for VTE prophylaxis in patients with atrial fibrillation mostly, and also for VTE prevention after knee or hip replacement surgery in geriatric rehabilitation.
Dalteparin is administered parenterally and is more expensive per dose (considering prophylaxis-related drug costs) than new oral anticoagulants. It is the drug with the highest expenditure across inpatient programs in Bruyère.
Serious bleeding may occur with anticoagulant prophylaxis. This side effect is considered reversible with traditional anticoagulants such as warfarin, unfractionated heparin, and low molecular weight heparin. An antidote has recently been approved for dabigatran but there is none yet for rivaroxaban or apixaban. However, all episodes of serious bleeding at Bruyère would need to be transferred to acute care for urgent management. Bruyère Health Saint-Vincent Hospital estimates that there are fewer than 5 cases per year transferred to acute care from Bruyère Health Saint-Vincent Hospital.
It is unclear if new oral anticoagulants could be recommended for VTE prophylaxis in all patient populations at Bruyère.
Objectives
By comparing the effectiveness (including cost-effectiveness) and safety of new oral anticoagulants versus low molecular weight heparins for preventing venous thromboembolism (VTE) in adult patients in subacute care, this review will address the following questions:
- Will the use of NOACs for VTE prophylaxis across all patient populations at Bruyère have a financial advantage over the use of dalteparin?
- Does the lack of antidotes to address the risk of serious bleeding risk with NOACs preclude their use across all patient populations at Bruyère?
Methods
Eligibility and selection criteria:
We used the PICO (population, intervention, comparison, and outcome) framework to define the eligibility criteria.
- Population: subacute care patients 18 years or older – palliative care, geriatric and stroke rehabilitation, and complex continuing care (mixed population)
- Interventions: Direct-acting Oral Anticoagulants (DOACs) or New Oral anticoagulants (NOACs) such as dabigatran, rivaroxaban, apixaban for prophylaxis of VTE. Treatment with NOACs for stroke prevention was not included.
- Comparison: Low molecular weight heparin such as dalteparin, enoxaparin, tinzaparin for prophylaxis of VTE. We excluded articles that compared NOACs with other anticoagulants such as warfarin, antiplatelet drugs.
- Outcomes: morbidity e.g. venous thromboembolism events (DVT, PE), bleeding events, mortality, cost-effectiveness (hospital perspective), patient preference related to inconvenience of injections.
Bleeding events include major bleeding as well as clinically-relevant non-major bleeding events (such as nose bleed, gastrointestinal bleed, bleeding gums, hematuria, spontaneous skin hematoma, bleeding leading to hospitalization or surgery).
Major bleeding is defined in some of the included articles as a fall in hemoglobin of at least 20 g/L or transfusion of at least two units of red cells, or symptomatic bleeding into a critical area or organ, such as intracranial, intraspinal, intra-ocular, retroperitoneal, intra-articular, pericardial or intramuscular with compartment syndrome, or bleeding leading to death.
Clinically-relevant non-major bleeding is defined as any sign or symptom of hemorrhage that does not fit the criteria for the ISTH definition of major bleeding but does meet at least one of the following criteria: requires medical intervention by a healthcare professional, or leads to hospitalization or increased level of care, or prompts a face to face evaluation.
We excluded systematic reviews and clinical guidelines if they focused on treatment of acute venous thromboembolism, acute settings (e.g. emergency), outpatients, children or pregnant women. We also excluded articles that compared NOACs with other anticoagulants such as warfarin, antiplatelet drugs.
Literature search:
We searched the Trip Database on February 9 2016 and retrieved 383 articles. We also searched for relevant systematic reviews, health technology assessments, economic evaluations and clinical practice guidelines in PubMed, the Cochrane Library (DARE and HTA) up to March 14 2016 and retrieved 1617 articles.
Relevance assessment:
We screened the search results and reference lists of eligible articles in duplicate. Disagreements were resolved by consensus. We only considered articles in English or French and identified 40 articles (32 systematic reviews, health technology assessments (HTAs) and economic evaluations, and eight guidelines) that met our inclusion criteria.
Quality assessment and grading of evidence:
We assessed the quality of the included reviews and guidelines using AMSTAR and AGREE II respectively (see Appendix 2). The quality of the included reviews ranged from low to high with an AMSTAR score of 2 to 10 out of 11. The AGREE II score for the guidelines was good with scores ranging from 125 to 148 out of 168.
We also graded the quality of the evidence using GRADE. These ranged from low to high.
Evidence review
Evidence from systematic reviews and HTAs or economic evalutaion:
We identified 32 systematic reviews and HTAs or economic evaluations on new oral anticoagulants for the prevention of venous thromboembolism in hospitalized adult patients. We considered the three new anticoagulants approved in Canada (dabigatran, apixaban, and rivaroxaban) compared with low molecular weight heparins (LMWHs such as dalteparin, enoxaparin, or tinzaparin).
Of the 22 systematic reviews that focused on the effectiveness of NOACs, nine reviews assessed direct factor Xa inhibitors (rivaroxaban or apixaban), two assessed direct thrombin inhibitor (dabigatran) and 11 assessed both types of NOACs compared with LMWHs. Four recent systematic reviews considered only the recommended prophylactic doses of the NOACs whereas others included studies assessing other doses as well. Clinical categories rather than hospital settings were considered. Seventeen were in patients who had orthopedic surgery (hip/knee joint replacement), two in elderly patients older than 65 years, one in cancer patients, one in patients with renal impairment and one in a mixed population (surgery and medically ill).
Ten articles focused on cost-effectiveness of NOACs compared with LMWHs; three in Canada, two each in Norway and the UK; and one each in Australia, Ireland, and the US. We decided to focus on the three cost-effectiveness analyses done in Canada. Two assessed the cost-effectiveness of rivaroxaban compared to enoxaparin or dalteparin based on Ontario data and the other assessed apixaban compared to enoxaparin based on Quebec data. Both were in patients who had hip/knee joint replacement surgery. Cost-effectiveness evaluations have not been done in other patient populations.
Synthesis of findings
Clinical categories rather than hospital settings were considered. We focused on the findings from the reviews with the highest quality, the most recent search date, the recommended prophylactic doses (10 mg daily for rivaroxaban, 2.5 mg twice daily for apixaban, and 220 mg daily for dabigatran) and the outcomes of interest. Older systematic reviews included studies assessing other doses than the current recommended prophylactic doses.
No article assessed inconvenience of injections as an outcome. Bleeding events and VTE events were classified inconsistently across the articles. For example, most articles presented overall bleeding risk including both major bleeding and clinically relevant non-major bleeding events. Some articles also presented VTE events and mortality as a composite outcome. In all the included systematic reviews effectiveness was assessed by VTE events and all-cause mortality and safety by bleeding. One review also considered arterial thrombosis and assessed myocardial infarction and ischemic stroke as primary outcomes.
Findings from systematic reviews and HTAs or economic evaluations
Results are grouped by the type of patients: 1) orthopedic surgery patients, 2) elderly patients with hip or knee joint replacement surgery, 3) patients with cancer, 4) patients with renal impairment, 5) mixed population of patients (medically ill with infectious disease, cardiovascular disease and inflammatory disease).
Orthopedic surgery patients:
For VTE events and all-cause mortality, there were no important differences between NOACs and LMWH (0 fewer events per 1000 patients, high certainty evidence). For bleeding (including major bleeding and clinically relevant non-major bleeding events), there were no important differences between NOACs and LMWH (1 more event with NOACs per 1000 patients, moderate certainty evidence).
There are no head-to-head direct comparisons of specific NOACs. However, a network meta-analysis used indirect evidence to compare different NOACs to each other and enoxaparin(46). These analyses showed that rivaroxaban is more effective at preventing VTE (56 fewer (70 fewer to 34 fewer) events per 1000 patients compared to enoxaparin).
Cost-effectiveness in the Ontario setting was assessed in two studies using data from the provincial government and hospital perspective. Rivaroxaban was more cost-effective compared to enoxaparin after hip and knee joint replacement surgery. Rivaroxaban was associated with an overall cost savings of C$296.95 per patient who had hip replacement surgery, compared with enoxaparin. The cost savings per patient who had knee replacement surgery was up to C$150.44. Factors contributing to the cost-effectiveness include fewer symptomatic VTE events with rivaroxaban leading to a higher number of QALYs gained; the reduction of treatment-related monitoring needs and the reduction in long term complications that would impact upon healthcare resources.When rivaroxaban was compared to dalteparin in a sensitivity analysis, similar results were found with cost savings of C$374.17 in patients who had hip replacement surgery and C$180.83 in patients who had knee replacement surgery.
Diamantopoulos found similar results comparing rivaroxaban to enoxaparin and dalteparin(22). For rivaroxaban versus enoxaparin, cost savings of C$300 per patient who had hip replacement surgery and C$129 per patient who had total knee replacement surgery were found. When rivaroxaban was compared to dalteparin in a sensitivity analysis, similar results were found with cost savings of C$360 in patients who had hip replacement surgery and C$153 in patients who had knee replacement surgery.
In the Quebec setting, apixaban was equally found to be more cost-effective compared to enoxaparin with cost savings of C$277 in patients who had hip joint replacement surgery and C$181 in patients who had knee joint replacement surgery.
Elderly patients:
In elderly patients, >65 years, who had hip/knee joint replacement surgery, there were no important differences between NOACs and LMWH for VTE events including VTE-related deaths (6 fewer events per 1000 patients on NOACs, moderate certainty evidence), and risk of major bleeding (4 more events per 1000 patients on NOACs, moderate certainty evidence).
In elderly patients, ≥75 years old, VTE events including VTE-related deaths were similar but major bleeding was significantly lower in NOACs compared with LMWH.
Patients with cancer:
In a subgroup of 405 hospitalized patients with cancer, there was no difference between rivaroxaban compared to enoxaparin on VTE and VTE-related deaths (RR 1.34, 0.71-2.54; with 25 more events per 1000 patients, moderate certainty evidence), but rivaroxaban was associated with a higher risk of major bleeding (37 more per 1000). No studies assessed apixaban or dabigatran in cancer patients.
Patients with renal impairment:
Patients hospitalized for renal impairment, were assessed in one review. Dabigatran was compared to enoxaparin in 159 hospitalized patients who had moderate renal dysfunction (defined as creatinine clearance between 30 and 49 mL/min). The rates of VTE events were not significantly different for dabigatran and enoxaparin (43 per 1000 compared to 90 per 1000) but enoxaparin had higher rates of major bleeding than dabigatran (47 compared to 5 per 1000). Rivaroxaban and apixaban have not been studies in this patient population.
Mixed population (medically ill):
In a mixed population of patients with infectious disease (excluding septic shock), congestive heart failure, respiratory failure, ischemic stroke, acute rheumatic disorder, inflammatory bowel disease, there was a higher rate of major bleeding with NOACs than with enoxaparin (4 more events per 1000 patients). Other outcomes were not assessed in the systematic review.
Recommendations from clinical practice guidelines
Dabigatran, rivaroxaban, apixaban or LMWH were all recommended as first line preventive therapy for thromboembolism in patients with hip/knee joint replacement surgery in four guidelines although one preferred LMWH over NOACs. The ACCP guidelines recommended the use of dabigatran in patients undergoing major surgery (hip or knee replacement surgery or hip fracture surgery) who decline injections. Rivaroxaban and apixaban could be used if dabigatran was unavailable. In the ACCP guidelines, dabigatran alone is recommended in patients with a history of ischemic stroke or TIA and atrial fibrillation, including paroxysmal atrial fibrillation. However, it is contraindicated in patients with severe renal impairment. UFH is the preferred anticoagulant for VTE prophylaxis in patients with renal impairment and LMWHs in all other patient populations with no contraindications. Cost effectiveness analyses were done in four guidelines.
Patient preferences
We did not find any systematic reviews comparing NOACs to LMWHs which reported patient experience or preferences related to injections required by LMWH. We did not search for patient values about the outcomes of VTE or bleeding.
NICE guidelines reported about patient adherence to LMWH injections. In a study comparing dalteparin and enoxaparin in patients with spinal cord injury, adherence for subcutaneous LMWH injection during hospitalisation reached more than 99%, both for once and twice daily injections. In another study of LMWHs in out-patients with a knee plaster cast, 12% of 148 participants discontinued treatment due to discomfort or refusal to self-inject.
All the included guidelines except one recommended that the choice of thromboprophylactic agents should be based on availability, and individual patients’ risk characteristics and preferences. Three guidelines also considered cost and one, compliance. The ACCP clinical practice guideline group found that patient values and preferences for treatment choices vary widely, and made a recommendation that NOACs could be considered for patients who disliked or refused daily injections of LMWH.
VTE can result in complications such as post-thrombotic limb syndrome, pulmonary hypertension, stroke, heart failure and even death. VTE prophylaxis also carries some known risks e.g. bleeding which can be extremely frightening and uncomfortable for patients, and the consequences will depend on the site (e.g. intracranial bleeding) and severity particularly if the bleeding is difficult to stop due to anticoagulation effect. It is therefore important for clinicians to discuss potential risks and benefits of VTE prophylaxis with the patient so that they can make informed choices and develop an adequate treatment plan, taking into account the patient’s needs and preferences.
Discussion
Applicability of evidence/implementation:
LMWHs have been the treatment of choice for VTE prophylaxis in adult patients in subacute care at BCC. With the recently approved NOACs we sought to find out whether NOACs could be used in place of LMWHs. We compared the effectiveness and safety of NOACs versus LMWHs. Different outcomes were assessed in different patient populations. VTE events and mortality were assessed separately in some reviews and combined as a composite outcome in others. Some reviews focused on one outcome e.g. major bleeding. No review assessed patient values or preference related to the inconvenience of injections. We found three economic evaluations comparing rivaroxaban or apixaban with LMWHs in patients with hip and knee replacement surgery in Canada. NOACs have been assessed mostly in patients with hip and knee joint replacement surgery.
All 19 reviews that assessed NOACs versus enoxaparin in patients (including the elderly) with hip or knee replacement surgery found that NOACs had a marginal or superior effect in preventing VTE events and had similar or increased risk of bleeding than enoxaparin. Four of these reviews considered only the approved doses and found that NOACs prevented more VTE events and mortality (0 to 6 fewer VTE/deaths per 1000 patients) and had an increased risk of bleeding than enoxaparin (1 to 4 more bleeds per 1000 patients).
NOACs are easier to administer and more cost-effective than LMWH for hip or knee surgery patient population, however, LMWHs remain the preferred anticoagulant prophylactic drugs as recommended by seven international guideline groups. These guidelines were developed before the approval of an antidote for dabigatran therefore the reversal of bleeding risk of NOACs was still a challenge. Antidotes for rivaroxaban and apixaban are still under development.Data in other patient populations such as cancer, renal impairment and mixed population of medically ill are limited (few studies with few participants for quite a rare outcome).
Guideline recommendations have been limited to the patient populations in whom NOACs have been studied. Four guidelines out of eight recommended NOACs in patients with hip and knee joint replacement surgery. The recommendations are in line with the findings of the included reviews that assessed NOACs in patients with joint replacement surgery. The AAOS guidelines(9) suggest that NOACs should be considered only in patients who are not at elevated risk beyond that of the surgery itself for venous thromboembolism or bleeding. The ACCP guidelines recommend NOACs in patients undergoing major surgery who decline injections. Other guidelines suggest that patient values and preferences should be considered in the choice of venous thromboprophylaxis but no clear recommendations were made based on patient’s preferences and values as in the ACCP guidelines.
NOACs are recommended in patients with atrial fibrillation and risk of stroke based on studies comparing NOACs with warfarin, the most common thromboprophylactic treatment in this patient population. These were excluded from our synthesis. Also, the recommendation for dabigatran in patients with ischemic stroke is based on studies comparing it to antiplatelet drugs (also excluded from our synthesis).
Strengths and limitations:
The objective of this review was to identify evidence about the effectiveness, cost-effectiveness and safety of NOACs compared with LMWH for venous thromboprophylaxis in adult patients in subacute care. The LMWH commonly used at BCC is dalteparin. We found no review comparing NOACs to dalteparin except for two economic evaluations comparing rivaroxaban to dalteparin in sensitivity analyses. However, NOACs were compared to enoxaparin which is similar to dalteparin. All 10 cost-effectiveness evaluations showed that NOACs were more cost-effective than LMWHs for hip or knee surgery patient population.
Similar trends of effectiveness and safety effects were found in the systematic reviews and HTAs that assessed NOACs versus LMWHs. The inconvenience of injections was not assessed in any review or HTA but the administration cost of injections of LMWHs compared to oral administration of NOACs was considered in cost-effectiveness evaluations. Guideline recommendations were limited to patient populations that have been studied and were based on effectiveness and safety data. Guidelines also recommended that patient preferences and values should be considered in the choice of treatment.
Overall recommendations
Overall, hospitals should consider approaches that will likely increase provider compliance and patient adherence as well as improve patient outcomes.
Based on our findings, we suggest the following:
- Patient risk assessment for VTE risk and risk of bleeding should be done before deciding whether or not VTE prophylaxis should be used, and which type. For VTE risk assessment, additional risks such as the clinical condition or reason for hospitalization should also be taken into account. When assessing patients for risk of bleeding, a balance between actual and perceived risk should be considered as well as contraindications for prophylaxis.
- Various guideline groups recommend thromboprophylaxis with either NOACs or LMWH in patients including elderly with hip or knee joint replacement surgery, provided patients have no contraindications.
- For patients undergoing orthopedic surgery who refuse injections, NOACs is recommended.
- NOACs are more cost-effective than LMWH for patients with hip or knee joint replacement even when risk of major bleeds is considered.
- For the medically ill, there is higher risk of bleeds (4 per 1000, from 1-7 more per 1000; high certainty of evidence) and guideline groups recommend LMWH or or unfractionated heparin or unfractionated heparin for this population.
- For palliative care, we found no systematic reviews but two guidelines recommend the use of LMWH for thromboprophylaxis.