NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Adam SS, McDuffie JR, Lachiewicz PF, et al. Comparative Effectiveness of Newer Oral Anticoagulants and Standard Anticoagulant Regimens for Thromboprophylaxis in Patients Undergoing Total Hip or Knee Replacement [Internet]. Washington (DC): Department of Veterans Affairs (US); 2012 Dec.

Cover of Comparative Effectiveness of Newer Oral Anticoagulants and Standard Anticoagulant Regimens for Thromboprophylaxis in Patients Undergoing Total Hip or Knee Replacement

Comparative Effectiveness of Newer Oral Anticoagulants and Standard Anticoagulant Regimens for Thromboprophylaxis in Patients Undergoing Total Hip or Knee Replacement [Internet].

Show details


We identified six good-quality SRs that evaluated thromboprophylaxis using newer oral anticoagulants versus LMWH. One SR evaluated additional drug classes, including unfractionated heparin, aspirin, and vitamin K antagonists. Although we identified no direct comparisons of newer oral anticoagulants, two good-quality SRs indirectly compared one newer oral anticoagulant with another through common comparison to enoxaparin.21,23 Only one SR compared combined pharmacologic and mechanical thromboprophylaxis to either method alone.9 FXa inhibitors have been studied more extensively than DTIs. In the absence of head-to-head comparisons between newer oral anticoagulants, it is difficult to draw strong conclusions on inter- or intra-drug class differences. The main findings and strength of evidence from our literature synthesis are summarized by key question in the section that follows.


KQ 1. Newer Oral Anticoagulants Versus Standard Treatments

FXa inhibitors. Rivaroxaban and apixaban are the most commonly studied FXa inhibitors. The risk of symptomatic DVT was reduced with FXa inhibitors thromboprophylaxis compared with LMWH, while the risk of nonfatal PE and mortality was not significantly different (all high strength of evidence). The estimated absolute risk difference was 4 fewer symptomatic DVT events for each 1000 patients receiving thromboprophylaxis with FXa inhibitors over 5 weeks compared with LMWH. However, these benefits were offset by an increase in major bleeding (moderate strength of evidence). The absolute risk difference was 2 more major bleeding events per 1000 patients on FXa thromboprophylaxis over a period of 5 weeks. Higher doses of FXa inhibitors but not intermediate or low doses were associated with increased major bleeding.20 Subgroup analysis by specific drug and type of surgery showed a reduced risk of bleeding with apixaban compared with LMWH in TKR but not in THR; risk of major bleeding with rivaroxaban did not differ significantly for either surgery.35 No reviews identified trials comparing oral FXa inhibitors with warfarin, UFH, or oral antiplatelet agents.

Direct thrombin inhibitors. Dabigatran is the only FDA-approved oral DTI, and the only DTI evaluated in existing SRs. Compared with LMWH, dabigatran was not associated with significant differences for any outcome examined. The strength of evidence was low for most outcomes due to few events and imprecise estimates of effect; also, effects on mortality varied substantially across studies. In addition to the major outcomes, a subgroup analysis in one SR found no significant difference between both treatment groups on bleeding requiring rehospitalization.9 No reviews identified trials comparing oral DTIs with warfarin, unfractionated heparin, or oral antiplatelet agents.

Table 7 summarizes the findings and strength of evidence for the effects of newer oral anticoagulant drug classes compared with enoxaparin in patients undergoing THR or TKR surgery.

Table 7. Summary of the strength of evidence for KQ 1.

Table 7

Summary of the strength of evidence for KQ 1.

KQ 2. Combined Pharmacological and Mechanical Prophylaxis

No reviews identified trials comparing newer oral anticoagulants combined with mechanical prophylaxis to pharmacological prophylaxis alone. Even when considering standard treatments, very little data are available comparing combined-modality thromboprophylaxis and pharmacologic prophylaxis only. One SR found moderate strength of evidence that combined-modality thromboprophylaxis was associated with a decreased risk of overall DVT (including asymptomatic events) compared with pharmacologic prophylaxis only.9 The evidence was insufficient for all other outcomes.

KQ 3. Comparisons of Individual Newer Oral Anticoagulants

Only indirect comparisons of rivaroxaban, apixaban, and dabigatran were performed through common comparison with LMWH.23,35 These comparisons were made for only two of our major outcomes—symptomatic VTE (DVT or PE) and major bleeding. There were no significant differences in treatment effect for symptomatic VTE or major bleeding. Because these indirect comparisons are subject to confounding and the treatment effects were imprecise, we considered the strength of evidence low. Other outcomes reported included clinically relevant bleeding and net clinical endpoints. Rivaroxaban was found to be associated with an increased risk of clinically relevant bleeding, but there was no significant difference in net clinical endpoints (symptomatic VTE, major bleeding and death). Table 8 summarizes the findings and strength of evidence for between-drug comparisons of newer oral anticoagulants in patients undergoing THR or TKR.

Table 8. Summary of the strength of evidence for KQ 3.

Table 8

Summary of the strength of evidence for KQ 3.


Patients undergoing total knee replacement or total hip replacement are at a significant risk for VTE. The precise risk for VTE in contemporary orthopedic surgery is difficult to estimate because of changes in surgical management, the paucity of recent trials comparing thromboprophylaxis to placebo, and the high frequency of thromboprophylaxis in routine practice, making observational studies of natural history difficult to conduct. A recent, careful analysis estimated the prevalence of symptomatic VTE without thromboprophylaxis at 2.8 percent for the initial 14 days and 4.3 percent at 35 days following major orthopedic surgery.1 Appropriate use of perioperative thromboprophylaxis significantly reduces the risk of postoperative proximal VTE, but the evidence is much more limited for effects on symptomatic DVT, PE, and mortality.1,9

Our evidence synthesis primarily addresses the comparative effectiveness of newer oral anticoagulants compared with standard antithrombotic agents for VTE prophylaxis. These newer drugs have been compared only to LMWH and show similar effects on most major clinical outcomes, although the strength of evidence varies by drug class and specific drug. In evaluating whether to add these newer agents to the VA formulary and whether to promote a specific thromboprophylaxis strategy, consideration should be given to the evidence of effectiveness and the importance and variability of patient values and preferences, costs, and health care system resources for successfully implementing competing strategies. In the following section, we summarize recommendations from the two major U.S. clinical guideline panels that have addressed this issue.


Both the American College of Chest Physicians (ACCP) and the American Academy of Orthopaedic Surgeons (AAOS) have recently issued guidelines on thromboprophylaxis in patients undergoing TKR or THR.1,2,24 The ACCP recommends antithrombotic prophylaxis over no prophylaxis for patients undergoing TKR or THR. The AAOS guidelines suggest individual assessment of patients for thromboprophylaxis. For patients at average risk, the guidelines do not include a recommendation for a specific thromboprophylactic strategy, considering the evidence for comparative effectiveness to be inconclusive. In contrast, the ACCP guidelines make recommendations for specific strategies and include the following options: LMWH, fondaparinux, apixaban, dabigatran, rivaroxaban, low-dose unfractionated heparin, adjusted-dose vitamin K antagonists, aspirin, or an intermittent pneumatic compression device. However, in the absence of elevated bleeding risk, LMWH is recommended in preference to other agents. Factors identified as increasing the risk of bleeding include previous major bleeding, severe renal failure, concomitant antiplatelet use, and a history of or difficult-to-control surgical bleeding during the current operative procedure, extensive surgical dissection, and revision surgery.1 For patients with increased bleeding risk, ACCP recommends intermittent pneumatic compression device or no prophylaxis.

In making recommendations, both guideline panels considered benefits and potential harms but had different approaches to considering costs. The AAOS did not conduct cost analyses and instructed guideline members to consider costs only when the impact was likely to be substantial. The ACCP process considered costs when it was plausible that resource use might change the direction or strength of recommendation and when high-quality economic analyses were available; it is not stated whether costs were considered in the recommendations pertaining to major orthopedic surgery. Guideline recommendations regarding choice of thromboprophylaxis are summarized in Table 9. Finally, we note that these guidelines address other clinical management issues, including duration and timing of thromboprophylaxis and use of routine DVT screening, which are not summarized here since they are not directly germane to our key questions.

Table 9. U.S. guideline recommendations related to specific thromboprophylaxis strategies.

Table 9

U.S. guideline recommendations related to specific thromboprophylaxis strategies.


We think these results are likely to apply to Veteran populations, but recommend some caution when applying trial data to standard clinical practice. There were strict exclusion criteria for patients enrolled in these studies, including severe renal or hepatic impairment or high risk of bleeding. Patients enrolled in trials are often more adherent to treatment plans and are monitored more closely than patients in routine clinical care. As a result, treatment effects in standard clinical practice may differ from those observed in clinical trials. Furthermore, the definition of bleeding was not consistent across studies, and it did not always include surgical site bleeding, which can lead to infection, dehiscence, and reoperation. Another limitation is the treatments compared. Newer anticoagulants were compared exclusively with LMWH, an appropriate and widely used comparator, but there were no direct comparisons to other treatment options recommended by guideline panels. Finally, none of these studies included patients from the VA health care system. Compared to patients with private sector insurance, VA patients on average have a greater burden of chronic disease, which would likely increase bleeding risk. If the comparative treatment effects vary by the presence of certain comorbid conditions, these results may not be reproducible in VA settings.


Our study has a number of strengths, including a protocol-driven review, a comprehensive search, and careful quality assessment. Another strength is the opportunity for meta-synthesis from existing systematic reviews and the opportunity to carefully evaluate reasons for different findings or conclusions across published reviews. Limitations include the lack of head-to-head comparisons of the newer oral anticoagulants, which precludes strong conclusions on their comparative effectiveness. Further, the length of experience with these new anticoagulants is too short to allow identification of longer term adverse events that may only emerge with more widespread use.


We used the framework recommended by Robinson et al.42 to identify gaps in evidence and classify why these gaps exist (Table 10).

Table 10. Evidence gaps and future research.

Table 10

Evidence gaps and future research.

This approach considers PICOTS (population, intervention, comparator, outcomes, timing, and setting) to identify gaps and classifies them as due to: (1) insufficient or imprecise information, (2) biased information, (3) inconsistency or unknown consistency, and (4) not the right information. VA and other health care systems should consider their clinical and policy needs when deciding whether to invest in research to address gaps in evidence. Specific research questions can be evaluated quantitatively, using value-of-information analysis, which uses Bayesian methods to estimate the potential benefits of gathering further information through research.43


For THR or TKR, the 35-day rate of symptomatic VTE without thromboprophylaxis is estimated to be 4.3 percent. Pharmacological thromboprophylaxis decreases VTE by approximately 50 percent but with the tradeoff of increased bleeding. Newer oral anticoagulants have a more convenient route of administration compared with LMWH, and unlike adjusted dose warfarin, they do not require regular laboratory monitoring. Compared with LMWH, FXa inhibitors are associated with a reduced risk of symptomatic DVT, but mortality and nonfatal PE are not significantly different, and the risk of major bleeding episodes is increased.

There are no available studies on head-to-head comparisons of these novel anticoagulants. Longer clinical experience and direct drug-drug comparisons are needed to better assess the risk-to-benefit ratio of newer oral anticoagulants for surgical thromboprophylaxis. Based on current evidence, newer anticoagulants—particularly FXa inhibitors—are a reasonable option for thromboprophylaxis in patients undergoing total hip replacement or total knee replacement.


Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...