Venous thromboembolism (VTE) is a common complication

Transcription

1 Enoxaparin Versus Dabigatran or Rivaroxaban for Thromboprophylaxis After Hip or Knee Arthroplasty Results of Separate Pooled Analyses of Phase III Multicenter Randomized Trials Menno V. Huisman, MD, PhD*; Daniel J. Quinlan, MBBS*; Ola E. Dahl, MD; Sam Schulman, MD Downloaded from by guest on July 19, 2018 Background Dabigatran and rivaroxaban are novel oral anticoagulants approved for prevention of venous thromboembolism after hip or knee arthroplasty. However, information assessing clinically important efficacy and bleeding outcomes of these 2 new agents versus low-molecular-weight heparin (enoxaparin) is lacking. Methods and Results We separately pooled efficacy and safety data from 6 phase III randomized trials ( participants) comparing equivalent durations of treatment with enoxaparin (40 mg once daily [od] or 30 mg twice daily) versus dabigatran (220 mg od) or versus rivaroxaban (10 mg od) after hip or knee arthroplasty. Odds ratios (OR) for individual outcomes were calculated for each trial and were pooled using the Mantel-Haenszel method. Compared with dabigatran, enoxaparin had a similar risk of symptomatic venous thromboembolism plus all-cause mortality (0.9% versus 1.1%; OR, 0.76; 95% confidence interval [CI], 0.44 to 1.31; I 2 76%) and bleeding (5.0% versus 5.6%; OR, 0.90; 95% CI, 0.71 to 1.15; I 2 0%). Compared with rivaroxaban, enoxaparin had a 2-fold higher risk of symptomatic venous thromboembolism plus all-cause mortality (1.2% versus 0.6%; OR, 2.04; 95% CI, 1.32 to 3.17; P 0.001; number needed to treat, 167; I 2 0%) but demonstrated a significant lower risk of bleeding (2.5% versus 3.1%; OR, 0.79; 95% CI, 0.62 to 0.99; P 0.049; number needed to harm, 167; I 2 0%). Conclusions In patients undergoing hip or knee arthroplasty, enoxaparin and dabigatran showed similar rates of efficacy and bleeding. Enoxaparin was less effective than rivaroxaban but had a lower risk of bleeding. These results may have important implications for the choice of prophylactic agent in major joint arthroplasty. (Circ Cardiovasc Qual Outcomes. 2010;3:00-00.) Key Words: anticoagulant dabigatran enoxaparin rivaroxaban venous thromboembolism Venous thromboembolism (VTE) is a common complication after hip or knee arthroplasty. 1,2 Low-molecularweight heparin reduces the risk of VTE by approximately one half compared with placebo, but it is inconvenient for out-of-hospital use because of the requirement for daily subcutaneous injections. 3 Recent research has focused on the development of new oral anticoagulants that can be administered in fixed dosages with the expectation that they will provide safe and effective alternatives to existing therapies. Two novel agents are the direct thrombin inhibitor dabigatran etexilate (referred to hereafter as dabigatran) and the selective factor Xa inhibitor rivaroxaban, both of which have been approved for use in several countries for prevention of VTE in patients undergoing hip or knee arthroplasty. Enoxaparin was used as the comparator through the development programs of both agents. With the availability of these new oral anticoagulants, clinicians will need guidance on which one to use. When evaluating an anticoagulant, a reduction in risk of VTE is balanced against a risk of serious bleeding, particularly in joint arthroplasty because bleeding into the replaced joint can have a detrimental effect on the surgical outcome. 4 To evaluate the merits of the new agents, we performed separate pooled analyses of phase III randomized trials that compared enoxaparin versus dabigatran or versus rivaroxaban, assessing clinically important efficacy and bleeding outcomes. In each analysis, we compared trials using equivalent durations of prophylaxis and assessed efficacy and bleeding using similar definitions in all trials. Received March 23, 2010; accepted July 22, From the Department of General Internal Medicine-Endocrinology (M.V.H.), Leiden University Medical Center, Leiden, The Netherlands; the Department of Radiology (D.J.Q.), King s College Hospital, London, United Kingdom; the Department of Orthopedics (O.E.D.), Elverum Hospital, Oslo, Norway; Thrombosis Research Institute (O.E.D.), London, United Kingdom; and the Department of Medicine (S.S.), McMaster University, Hamilton, Ontario, Canada. *The first 2 authors contributed equally to this work. The online-only Data Supplement is available at Correspondence to Daniel J. Quinlan, MBBS, Department of Radiology, King s College Hospital, Denmark Hill, London, SE5 9RS, UK. dan.quinlan@consultoberon.com 2010 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at DOI: /CIRCOUTCOMES

2 2 Circ Cardiovasc Qual Outcomes November 2010 Downloaded from by guest on July 19, 2018 WHAT IS KNOWN Currently available anticoagulants are effective for prophylaxis of venous thromboembolism (VTE) after elective hip or knee arthroplasty, but either requires parenteral administration (eg, enoxaparin) or are difficult to manage because they require laboratory monitoring to adjust the dose (vitamin K antagonists). New oral anticoagulants, such as the direct thrombin inhibitor dabigatran and the selective factor Xa inhibitor rivaroxaban, target specific components of the coagulation cascade and can be administered in fixed dosages without laboratory monitoring. WHAT THE STUDY ADDS The study presents separate pooled analyses of efficacy and safety data from 6 phase III randomized trials comparing equivalent durations of treatment with enoxaparin (40 mg once daily or 30 mg twice daily) versus dabigatran (220 mg once daily) or rivaroxaban (10 mg once daily) after elective hip or knee arthroplasty. Enoxaparin had similar rates of efficacy and risk of bleeding compared with dabigatran. When compared with rivaroxaban, however, enoxaparin was associated with a significant 2-fold increase in the risk of symptomatic VTE plus all-cause mortality, equivalent to an excess of 6 events per 1000 patients treated; there was also a significantly lower risk of major or clinically relevant nonmajor bleeding, equivalent to 6 fewer events per 1000 patients treated. The findings may have important implications for the choice of prophylactic antithrombotic agent in patients undergoing hip or knee arthroplasty. Methods A protocol was prospectively developed, detailing the objectives, criteria for study selection, primary and secondary outcomes, and statistical methods. Each study included in this analysis had previously been approved by relevant local research ethics board, and all patients provided informed consent. Trial Selection Trial selection was confined to phase III randomized trials that compared equivalent durations of treatment for elective hip arthroplasty or elective knee arthroplasty, respectively, with enoxaparin (40 mg once daily or 30 mg twice daily) versus approved doses of dabigatran (220 mg once daily with a dose of 110 mg on the day of surgery) or versus approved doses of rivaroxaban (10 mg once daily) for the prophylaxis of VTE after elective hip or knee arthroplasty. Although 2 doses of dabigatran (220 and 150 mg once daily) were compared with enoxaparin, we restricted our primary assessment to the approved 220 mg once daily dose. Six randomized trials, involving a combined total of patients, were available for inclusion in the separate pooled analyses (8185 comparing enoxaparin versus dabigatran 5 7 and comparing enoxaparin versus rivaroxaban 8 10 ) (Table). One trial (REgulation of coagulation in ORthopedic surgery to prevent Deep venous thrombosis and pulmonary embolism [RECORD] 2) did not meet our inclusion criteria 11 because in that study different durations of treatment for rivaroxaban (31 to 39 days) and enoxaparin (10 to 14 days) were compared, and complete data were not available for all of the efficacy and safety end points with equivalent durations of treatment. All 6 trials provided information on the methods used to generate the randomized treatment allocation with adequate concealment of treatment allocation. All studies were double-blind. Outcomes and Data Extraction The primary efficacy outcome for each analysis was the composite of symptomatic VTE, which includes deep-vein thrombosis and pulmonary embolism and all-cause mortality during the treatment period. The primary bleeding outcome was the composite of major bleeding (the primary safety end point in each trial) and clinically relevant nonmajor bleeding during treatment, as defined in the individual trials. This bleeding end point definition was used because it was reported in all trials and accounted for differences in the definition of major bleeding used in the 2 trial programs (See supplemental Table 1). Secondary outcomes included individual components of the primary efficacy and bleeding outcomes; total VTE and all-cause mortality at end of treatment (the primary end point in each trial); major VTE plus VTE-related mortality at end of treatment; and adverse events during treatment including hepatic enzyme elevations and cardiovascular events. The latter also was assessed in the follow-up periods. Total VTE included symptomatic or asymptomatic deep-vein thrombosis documented by lower-limb venography and nonfatal pulmonary embolism; major VTE included proximal deep-vein thrombosis and nonfatal pulmonary embolism. Two investigators (D.J.Q., S.S.) independently extracted data on study design, study quality, and efficacy and safety outcomes during the treatment period and follow-up periods. We accepted the authors definitions for clinical outcomes and did not attempt to reclassify them retrospectively. The data abstracted for each trial were confirmed by reviewer consensus and then sent to the first or corresponding author for verification. Missing data were requested from the authors or sponsoring pharmaceutical company or obtained from Sponsor and Regulatory Agency websites. 12,13 Statistical Analysis We used a fixed-effects model based on the Mantel-Haenszel method for combining results from the individual trials. 14 We calculated the odds ratios (OR) and 95% confidence intervals (CI). All statistical calculations were performed using Review Manager (RevMan, version Copenhagen, The Cochrane Collaboration 2008) and Comprehensive Meta-Analysis (version 2.0, Biostat, Englewood, NJ). Heterogeneity was assessed using Cochran Q and the I 2 statistic: A probability value 0.10 was considered to denote statistically significant heterogeneity; where I 2 was 50%, heterogeneity was considered substantial. 15,16 We removed one study at a time to assess the source of the heterogeneity (where present). We performed the following subgroup analyses: enoxaparin versus dabigatran 150 mg once daily, which is recommended for specific patient populations, including those patients over 75 years of age or those with moderate renal impairment; and inclusion of primary efficacy and bleeding data comparing equivalent durations of treatment (at 14 days) from the RECORD2 trial 11 in the comparison of enoxaparin versus rivaroxaban. Sensitivity analyses were conducted to further explore the robustness of our results. To identify any study that may have exerted a disproportionate influence on the summary treatment effect, we deleted studies one at a time. We explored the potential for publication bias using funnel plots of effect size versus standard error. 17 We also compared results obtained using a fixed-effects model with those obtained using a randomeffects model. Results Enoxaparin Versus Dabigatran (3 Trials, n 8185) Two studies included patients undergoing elective knee arthroplasty 5,6 and 1 study included patients undergoing

3 Huisman et al Enoxaparin Versus Dabigatran or Rivaroxaban 3 Table. Summary of Trials Enoxaparin Versus Dabigatran Enoxaparin Versus Rivaroxaban RE-MOBILIZE (n 2615) 5 REMODEL (n 2076) 6 RE-NOVATE (n 3494) 7 RECORD1 (n 4541) 8 RECORD3 (n 2531) 9 RECORD4 (n 3148) 10 Downloaded from by guest on July 19, 2018 Eligibility TKA TKA THA TKA TKA TKA Age, mean, y Female, % Enoxaparin arm 30 mg bid starting h after surgery for d 40 mg od starting before surgery for 6 10 d 40 mg od starting before surgery for d 40 mg od starting before surgery for d 40 mg od starting before surgery for d 30 mg bid starting h after surgery for d Investigational arm 150 or 220 mg od starting 6 12 h after surgery for 150 or 220 mg od starting 1 4 h after surgery for 150 or 220 mg od starting 1 4 h after surgery for 10 mg od starting 6 8 h after surgery for d 10 mg od starting 6 8 h after surgery for d 10 mg od starting 6 8 h after surgery for d d* 6 10 d* d* Timing of randomization Postop Preop Preop Preop Preop Preop Treatment period First dose to 3 d after last dose study drug First dose to 3 d after last dose study drug First dose to 3 d after last dose study drug First dose to 2 d after last dose study drug First dose to 2 d after last dose study drug First dose to 2 d after last dose study drug Primary outcome Total VTE and all-cause mortality Total VTE and all-cause mortality Total VTE and all-cause mortality Total VTE and all-cause mortality Total VTE and all-cause mortality Total VTE and all-cause mortality Primary safety outcome Major bleeding Major bleeding Major bleeding Major bleeding Major bleeding Major bleeding Follow-up 3 mo 3 mo 3 mo THA indicates total hip arthroplasty; TKA, total knee arthroplasty; bid, twice daily; od, once daily; n, No. randomized; Preop, preoperative; and Postop, postoperative. *Half-dose was given on the day of surgery. Wound closure or adequate hemostasis. During the treatment period. After surgery. After the last dose of study medication. elective hip arthroplasty. 7 Enoxaparin was given at a dose of 40 mg once daily starting before surgery in 2 trials 6,7 and 30 mg twice daily starting after surgery in 1 trial. 5 Dabigatran was given at a dose of 150 or 220 mg once daily starting after surgery and continued for 6 to 14 days in the knee arthroplasty studies 5,6 and for 28 to 35 days in the hip arthroplasty study. 7 Follow-up was 3 months after surgery for all studies. The number of patients lost to follow-up at 3 months was not reported in any of the primary publications, but, based on unpublished data supplied by the sponsor, follow-up was 99% in all 3 trials. There was a similar risk of symptomatic VTE plus allcause mortality among patients treated with enoxaparin compared with dabigatran (0.9% versus 1.1%; OR, 0.76; 95% CI, 0.44 to 1.31), but there was significant heterogeneity (P 0.02, I 2 76%) for this outcome (Figure 1A). Statistical heterogeneity was reduced when the RE-MODEL study was removed from the analysis (OR, 0.49; 95% CI, 0.26 to 0.49; P for heterogeneity 0.19; I 2 41%). Secondary efficacy outcomes, including total VTE plus all-cause mortality and major VTE plus VTE-related mortality, showed no significant differences between the 2 treatment groups (heterogeneity tests: P 0.09; I 2 58% and P 0.22, I 2 35%, respectively) (data not shown). The composite of major and clinically relevant nonmajor bleeding was similar in patients treated with enoxaparin versus dabigatran (5.0% versus 5.6%; OR, 0.90; 95% CI, 0.71 to 1.15) without evidence for heterogeneity (Figure 1B). The absolute risk of bleeding in each of the 2 treatment groups was lower in the RE-MOBILIZE trial 5 (3.8% for enoxaparin, 3.3% for dabigatran) compared with the RE-NOVATE 7 (5.0% and 6.2%, respectively) and RE-MODEL 6 (6.6% and 7.4%, respectively) trials. Secondary bleeding outcomes, including major bleeding and its individual components, showed no significant differences between the 2 treatment groups (Figure 2). The majority of the major bleeding events were caused by bleeding leading to transfusion of 2 Uof blood (33 of 39 events with enoxaparin and 33 of 38 events with dabigatran) and bleeding associated with hemoglobin drop of 20 g/l compared with prerandomization level (32 of 39 events with enoxaparin and 29 of 38 events with dabigatran). With the exception of a moderate increase in the concentration of alanine aminotransferase (more than 3 times the upper limit of the normal range) at any time after baseline with enoxaparin (3.6%) compared with dabigatran (2.2%; P 0.003), there were no statistically significant differences between the treatments regarding any safety outcomes, including adverse events leading to discontinuation, cardiovascular events (during treatment or after discontinuation), and wound infections (Figure 2). Enoxaparin Versus Rivaroxaban (3 Trials; n ) Two studies included patients undergoing elective knee arthroplasty 9,10 and 1 study included patients undergoing elective hip arthroplasty. 8 Enoxaparin was given at a dose of 40 mg once daily starting before surgery in 2 trials 8,9 and 30 mg twice daily starting after surgery in 1 trial. 10 Rivaroxaban was given at a dose of 10 mg once daily starting after surgery in all studies and continued for 10 to 15 days in the knee arthroplasty studies 9,10 and for 28 to 35 days in the hip arthroplasty study. 10 Follow-up ranged from 1.5 to 2 months after surgery for all studies. The number of patients lost to follow-up at 3 months was not reported in any of the primary publications, but, based on data reported by the sponsor in a submission to the regulatory authorities, 13 follow-up was 99% in all 3 trials. There was a significant 2-fold higher risk of symptomatic VTE plus all-cause mortality (1.2% versus 0.6%; OR, 2.04;

4 4 Circ Cardiovasc Qual Outcomes November 2010 A n/n (%) Trial Enoxaparin Dabigatran OR (95% CI) RE-MOBILIZE 10/868 (1.2) 14/857 (1.6) 0.70 ( ) RE-MODEL 9/685 (1.3) 2/675 (0.3) 4.48 ( ) RE-NOVATE 4/1142 (0.4) 14/1137 (1.2) 0.28 ( ) Total 23/2695 (0.9) 30/2669 (1.1) 0.76 ( ) Test for heterogeneity: I 2 = 76%, P =.02 Test for overall effect: P = Favors Enoxaparin Favors Dabigatran B n/n (%) Trial Enoxaparin Dabigatran OR (95% CI) Downloaded from by guest on July 19, 2018 RE-MOBILIZE 33/868 (3.8) 28/857 (3.3) 1.17 ( ) RE-MODEL 46/694 (6.6) 50/679 (7.4) 0.89 ( ) RE-NOVATE 58/1154 (5.0) 71/1146 (6.2) 0.80 ( ) Total 137/2716 (5.0) 149/2682 (5.6) 0.90 ( ) Test for heterogeneity: I 2 = 0%, P =.49 Test for overall effect: P =.40 95% CI, 1.32 to 3.17) among patients treated with enoxaparin compared with rivaroxaban without evidence for heterogeneity (Figure 3A). A similar treatment effect was observed for the secondary efficacy outcomes of total VTE plus all-cause mortality and major VTE plus VTE-related mortality (heterogeneity tests: P 0.03; I 2 72% and P 0.04, I 2 69%, respectively) (data not shown). Statistical heterogeneity for both of these outcomes was reduced when the RECORD1 study was removed from the analysis (data not shown). The composite of major and clinically relevant nonmajor bleeding was significantly lower in patients treated with enoxaparin versus rivaroxaban (2.5% versus 3.1%; OR, 0.79; 95% CI, 0.62 to 0.99; P 0.049) without evidence for heterogeneity (Figure 3B). Secondary bleeding outcomes, including major bleeding and its individual components, showed similar but nonsignificant differences between the 2 treatment groups (Figure 4). The majority of the major bleeding events were caused by bleeding leading to reoperation (7 of 12 events with enoxaparin and 12 of 23 events with rivaroxaban). Enoxaparin compared with rivaroxaban was associated with a nonsignificant 75% reduction in a hemoglobin drop of 20 g/l compared with postoperative level (0.0% versus 0.1%; OR, 0.25; 95% CI, 0.05 to 1.18) as well as 75% reduction in bleeding associated with transfusion of 2 U of whole blood or packed cells (0.0% versus 0.1%; OR, 0.25; 95% CI, 0.05 to 1.18). Enoxaparin was associated with significant increases in adverse events leading to discontinuation (224 events versus 184 events; P 0.047), reported serious adverse events (397 events versus 316 events; P 0.002), and concentrations of alanine aminotransferase ( 3 times the upper limit of the Favors Enoxaparin Favors Dabigatran Figure 1. Enoxaparin versus dabigatran: A, Symptomatic VTE plus all-cause mortality; B, composite of major plus clinically relevant nonmajor bleeding. Vertical line indicates an OR of 1.0, equivalent to no difference between treatment groups. Size of the data markers is proportional to the study weight. normal range) at any time after baseline (2.4%) compared with rivaroxaban (1.7%; P 0.02). The higher number of serious adverse events and adverse leading to discontinuation with enoxaparin were mainly due to deep vein thrombosis and pulmonary embolism events. There were no statistically significant differences between the treatments regarding cardiovascular events (during treatment or after discontinuation) or wound infections. Subgroup and Sensitivity Analyses Efficacy and bleeding outcomes were similar with enoxaparin versus dabigatran 150 mg once daily (See supplemental Table 1). Inclusion of 14-day comparative data from RECORD2 in the analysis of enoxaparin versus rivaroxaban from published 18 and unpublished sources 13 showed a similar risk of symptomatic VTE plus all-cause mortality (1.3% versus 0.7%; OR, 2.07; 95% CI, 1.36 to 3.17) and major plus clinically relevant nonmajor bleeding (2.5% versus 3.1%; OR, 0.81; 95% CI, 0.66 to 1.01). Deleting individual studies yielded pooled results that were not significantly different from the overall pooled estimates, although in the comparison of enoxaparin versus dabigatran there was a statistically significant reduction in symptomatic VTE plus all-cause mortality after removal of the RE- MODEL trial. A funnel plot of effect size versus study precision was fairly symmetrical, with a similar number of studies on either side of the summary treatment effect for symptomatic VTE plus all-cause mortality and the composite of major plus clinically relevant nonmajor bleeding (figure not shown). There were no important differences between

5 Huisman et al Enoxaparin Versus Dabigatran or Rivaroxaban 5 Downloaded from by guest on July 19, 2018 Efficacy Symptomatic VTE 23/2695 (0.9) 25/2669 (0.9) 0.91 ( ).74 DVT 14/2695 (0.5) 14/2669 (0.5) 0.99 ( ).98 PE 9/2695 (0.3) 11/2669 (0.4) 0.82 ( ).65 Death 1/2695 (0.0) 5/2669 (0.2) 0.33 ( ).17 Bleeding Major bleeding 39/2716 (1.4) 38/2682 (1.4) 1.01 ( ).95 Bleeding leading to a fall in Hb 20 g/l compared with pre-operative level Clinically relevant non-major bleeding Any on-treatment adverse event 2226/2716 (82.0) 2190/2682 (81.7) 1.02 ( ).77 Adverse events leading to discontinuation ALT 3 x ULN any time post-baseline ALT 3 x ULN and BR 2 x ULN within 30 days post-baseline 32/2716 (1.2) 29/2682 (1.1) 1.09 ( ).73 Bleeding leading to re-operation 5/2716 (0.2) 5/2682 (0.2) 0.99 ( ).98 Bleeding warranting treatment cessation Bleeding leading to transfusion of 2 units of blood Adverse events Fatal bleeding 0/2716 (0.0) 1/2682 (0.0) 0.33 ( ).50 Bleeding into a critical organ 0/2716 (0.0) 1/2682 (0.0) 0.33 ( ).50 Serious on-treatment adverse events Cardiovascular Events 2/2716 (0.1) 3/2682 (0.1) 0.74 ( ).69 33/2716 (1.2) 33/2682 (1.2) 0.99 ( ).96 98/2716 (3.6) 111/2682 (4.1) 0.87 ( ) /2716 (5.6) 148/2682 (5.5) 95/2639 (3.6) 58/2613 (2.2) 1.65 ( ).003 2/2639 (0.1) 3/2613 (0.1) 0.71 ( ) Lower with Enoxaparin Lower with Dabigatran 1.02 ( ) /2716 (6.3) 179/2682 (6.7) 0.93 ( ).54 During treatment 17/2716 (0.6) 9/2682 (0.3) 1.87 ( ).13 After treatment discontinuation 7/2639 (0.3) 1/2611 (0.0) 6.94 ( ).07 Wound Complications Enoxaparin n/n (%) Dabigatran OR (95% Cl) Wound Infections 37/2716 (1.4) 30/2682 (1.1) 1.22 ( ).42 Figure 2. Secondary outcome analyses for enoxaparin versus dabigatran. ALT indicates alanine aminotransferase; BR, bilirubin; Hb, hemoglobin; and ULN, upper limit of normal. P value for overall effect is shown. Significant heterogeneity was seen for the outcomes: major bleeding leading to a fall in hemoglobin (P 0.07, I 2 63%) and major bleeding leading to transfusion (P 0.08, I 2 60%). P value outcomes or heterogeneity results obtained using the fixedversus a random-effects model. Discussion In the present report, separate pooled analyses of phase III trials of enoxaparin versus dabigatran or enoxaparin versus rivaroxaban after hip or knee arthroplasty indicated different risk-benefit profiles. Enoxaparin had similar rates of efficacy and risk of bleeding compared with dabigatran. When compared with rivaroxaban, however, enoxaparin was associated with a significant 2-fold increase in the risk of symptomatic VTE plus all-cause mortality, equivalent to an excess of 6 events per 1000 patients treated, although there was also a significantly lower risk of major or clinically relevant nonmajor bleeding, equivalent to 6 fewer events per 1000 patients treated. There was a discrepancy in the results of the primary trial outcome (total VTE plus all-cause mortality) in the individual trials comparing enoxaparin and dabigatran. Whereas the trials comparing enoxaparin 40 mg once daily with dabigatran showed similar efficacy outcomes, 6,7 the trial comparing enoxaparin 30 mg twice daily with dabigatran showed that enoxaparin was superior to dabigatran. 5 There were, however, no statistically significant differences in the risk of any category of bleeding between dabigatran and enoxaparin in any of the trials. In contrast, the individual trials comparing enoxaparin with rivaroxaban showed superior efficacy of rivaroxaban with regard to total VTE plus all-cause mortality, 8 10 major VTE 8,9 and symptomatic VTE, 9 with no significant difference in the risk of any category of bleeding with rivaroxaban compared with enoxaparin. Significant heterogeneity was demonstrated when comparing dabigatran with enoxaparin for the pooled outcome of symptomatic VTE plus all-cause mortality. This was probably due to the imprecise estimates of treatment effect observed in the RE-MODEL trial compared with the other 2

6 6 Circ Cardiovasc Qual Outcomes November 2010 A n/n (%) Trial Enoxaparin Rivaroxaban OR (95% CI) RECORD1 14/2206 (0.6) 10/2193 (0.5) 1.39 ( ) RECORD3 26/1217 (2.1) 8/1201 (0.7) 3.26 ( ) RECORD4 21/1508 (1.4) 12/1526 (0.8) 1.78 ( ) Total 61/4931 (1.2) 30/4920 (0.6) 2.04 ( ) Test for heterogeneity: I 2 = 13%, P =.32 Test for overall effect: P = Favors Enoxaparin Favors Rivaroxaban B n/n (%) Trial Enoxaparin Rivaroxaban OR (95% CI) Downloaded from by guest on July 19, 2018 RECORD1 56/2224 (2.5) 70/2209 (3.2) 0.79 ( ) RECORD3 34/1239 (2.7) 40/1220 (3.3) 0.83 ( ) RECORD4 34/1508 (2.3) 46/1526 (3.0) 0.74 ( ) Total 124/4971 (2.5) 156/4955 (3.1) 0.79 ( ) Test for heterogeneity: I 2 = 0%, P =.94 Test for overall effect: P =.049 trials. However, other secondary efficacy outcomes, including symptomatic pulmonary embolism and major VTE plus VTE-related mortality, showed no significant heterogeneity, and we believe that the heterogeneity for symptomatic VTE plus all-cause mortality is likely due to the play of chance. The ongoing hip arthroplasty trial RE-NOVATE II should provide further clarification. 19 In contrast, rivaroxaban demonstrated significantly better efficacy, not only for the primary efficacy outcome but also for the outcome of symptomatic deep-vein thrombosis, without significant heterogeneity. However, this was balanced by a significant increase in the risk of major and clinically relevant nonmajor bleeding. Secondary bleeding outcomes including major bleeding events, bleeding leading to transfusion of at least 2 U of whole blood or packed cells, and bleeding leading to a fall in hemoglobin were also more common in patients treated with rivaroxaban, although there was a lack of statistical power to detect clinically important differences caused by the small number of events. Subgroup analysis that included RECORD2 yielded similar results for the primary efficacy and bleeding outcomes; rivaroxaban showed better efficacy compared with enoxaparin but had a higher risk of bleeding. Although the clinical development programs for rivaroxaban and dabigatran for the prevention of VTE in major joint arthroplasty have some common methodological features type of surgery, comparator drug, use of extended-duration of prophylaxis after hip arthroplasty, and geographical distribution there are also important differences. In each of the individual trials, the primary outcome was the incidence of total VTE plus all-cause mortality. However, for patients treated with enoxaparin in the dabigatran trials, the rate of this Favors Enoxaparin Favors Rivaroxaban Figure 3. Enoxaparin versus rivaroxaban: A, Symptomatic VTE plus all-cause mortality; B, composite of major plus clinically relevant nonmajor bleeding. primary outcome was more than 2-fold higher than that observed in the rivaroxaban trials (mean, 20.1% versus 9.5%). In part, this may be due to recognized differences in the methods used by venogram adjudication committees to assess the incidence of asymptomatic deep-vein thrombosis. 20 Additionally, there was a 7-fold proportional difference in the rate of major bleeding for patients receiving enoxaparin in trials versus dabigatran than versus rivaroxaban (mean, 1.4% versus 0.2%). This is almost certainly related to recognized differences in the definition of major bleeding. 21 In the trials comparing enoxaparin with rivaroxaban, bleeding at the surgical site, which usually accounts for 80% to 90% of bleeding events, 6,7,22,23 was excluded from the definition of major bleeding, except if the event resulted in reoperation or death. However, the main driver of differences in the absolute rate of major bleeding between the 2 analyses relates to the individual categories bleeding associated with a fall in hemoglobin of at least 20 g/l and bleeding requiring transfusion of 2 or more units of blood. In the enoxaparin versus dabigatran trials, enoxaparin-treated patients had 32 and 33 events, respectively, in these 2 categories. In the trials versus rivaroxaban, only 1 event in each category was reported in patients receiving enoxaparin. This substantial difference not only relates to exclusion of surgical-site bleeding in the rivaroxaban trials but also to how the fall in hemoglobin after surgery was calculated. In the dabigatran trials, the reference hemoglobin value was the preoperative baseline level, whereas in the rivaroxaban trials it was the value at the time of the first rivaroxaban dose, 6 to 8 hours after surgery. 8 Similarities between the 2 trial programs enhance our decision to present 2 separate analyses side-by-side. In view

7 Huisman et al Enoxaparin Versus Dabigatran or Rivaroxaban 7 n/n (%) Enoxaparin Rivaroxaban OR (95% Cl) P value Efficacy Symptomatic VTE 53/4931 (1.1) 25/4920 (0.5) 2.13 ( ).002 DVT 39/4931 (0.8) 17/4920 (0.3) 2.30 ( ).004 PE 14/4931 (0.3) 9/4920 (0.2) 1.53 ( ).31 Death 9/4931 (0.2) 6/4920 (0.1) 1.46 ( ).46 Bleeding Major bleeding 12/4971 (0.2) 23/4955 (0.5) 0.52 ( ).07 Fatal bleeding 0/4971 (0.0) 2/4955 (0.0) 0.33 ( ).34 Bleeding into a critical organ 4/4971 (0.1) 3/4955 (0.1) 1.28 ( ).73 Clinically overt extrasurgicalsite bleeding leading to a fall in Hb 20 g/l compared with post-operative level 1/4971 (0.0) 7/4955 (0.1) 0.25 ( ).08 Bleeding leading to re-operation 7/4971 (0.1) 12/4955 (0.2) 0.58 ( ).25 Downloaded from by guest on July 19, 2018 Bleeding warranting treatment cessation Clinically overt extrasurgicalsite bleeding leading to transfusion of 2 units of blood Clinically relevant non-major bleeding Adverse events Adverse events leading to discontinuation ALT 3 x ULN and BR 2 x ULN during treatment Cardiovascular Events Wound Complications /4971 (2.3) 137/4955 (2.8) 0.81 ( ).10 Any on-treatment adverse event 3499/4971 (70.4) 3411/4955 (68.8) 1.08 ( ).08 Serious on-treatment adverse events 1/4971 (0.0) 7/4955 (0.1) 0.25 ( ) /4971 (4.5) 184/4955 (3.7) 1.22 ( ) /4971 (8.0) 316/4955 (6.4) 1.27 ( ).002 ALT 3 x ULN during treatment 115/4736 (2.4) 82/4749 (1.7) 1.42 ( ).02 4/4749 (0.1) 4/4759 (0.1) 1.00 ( ) 1.00 During treatment 20/4971 (0.4) 11/4955 (0.2) 1.81( ).11 After treatment discontinuation 10/4971 (0.2) 12/4955 (0.2) 0.84 ( ).67 Wound Infections 22/4971 (0.4) 19/4955 (0.4) 1.15 ( ) Lower with Enoxaparin Lower with Rivaroxaban Figure 4. Secondary outcome analyses for enoxaparin versus rivaroxaban. ALT indicates alanine aminotransferase; BR, bilirubin; Hb, hemoglobin; and ULN, upper limit of normal. Probability value for overall effect is shown. Significant heterogeneity was seen for cardiovascular events after treatment discontinuation (P 0.04, I 2 69%). of our aim to assess clinically important efficacy and bleeding outcomes, the primary outcomes we used in this analysis differed from those employed in each of the individual trials. Each of these outcomes was largely adjudicated by the same independent committees in each of the 6 trials, thereby reducing potential bias in interobserver interpretation. The outcome of symptomatic VTE plus all-cause mortality is a relatively robust outcome and not as sensitive to variability as for the assessment of asymptomatic venographic deep-vein thrombosis. 20 Because screening for asymptomatic deep-vein thrombosis is not done in clinical practice, the most relevant benefit for the practicing physician regarding VTE prophylaxis after hip and knee arthroplasty is the prevention of symptomatic events. Indeed, the event rates for enoxaparin for comparison with dabigatran or rivaroxaban were reasonably similar (0.9% and 1.2%). For the assessment of bleeding, we used the composite end point of major plus clinically relevant nonmajor bleeding rather than the primary safety end point of major bleeding used in each of the trial publications. This latter outcome, which comprised mainly excessive wound hematoma and surgical-site bleeding (data not shown), was largely defined in the same way in each trial program (See supplemental Table 2) and contributed toward most events reported for the composite outcome (209 of 286 events [73.1%] for enoxaparin versus dabigatran, and 249 of 280 events [88.9%] for enoxaparin versus rivaroxaban), thus accounting for the effect of marked differences in the definition of major bleeding within each trial program. Despite this, there was still a 2-fold difference in the composite bleeding event rate for enoxaparin in the dabigatran trials compared with rivaroxaban trials (5.0% versus 2.5%), which again precludes direct comparisons between the 2 trial programs.

8 8 Circ Cardiovasc Qual Outcomes November 2010 Downloaded from by guest on July 19, 2018 The safety profiles of dabigatran and rivaroxaban during treatment were similar to that of enoxaparin with respect to liver enzyme elevations and acute cardiovascular events, although the incidence of moderate transaminase elevations was significantly lower with either agent compared with enoxaparin. In addition, there was no significant increase in the rate of adverse events or wound infections with either compound compared with enoxaparin. The risk of death appeared to be lower for patients treated with enoxaparin versus dabigatran (1 event [0.1%] versus 5 events [0.2%]; OR, 0.33; 95% CI, 0.07 to 1.64) but higher for enoxaparin versus rivaroxaban (9 events [0.2%] versus 6 events [0.1%]; OR, 1.46; 95% CI, 0.54 to 3.96). The difference in risk is due to the difference in number of deaths in the enoxaparin groups (1 versus 9). A review of the causes of death did not show obvious differences between the different treatment groups (see supplemental Tables 3 and 4), and the overall mortality rate in the trials was generally low. Therefore, this is likely to be a chance finding because of the small number of events and the wide confidence intervals for the estimates of treatment effect. Although the large data set presented for safety and adverse event data with dabigatran and rivaroxaban are reassuring that there does not appear to be any major differences between these new anticoagulants and enoxaparin, because of the rarity of many of the events, much larger postmarketing data sets need to accrue before any definitive statements can be made. We acknowledge some limitations in our analyses. First, as explained, our results do not permit a direct comparison of dabigatran with rivaroxaban. The aim of our analysis was to present data for the 2 analyses side-by-side using similar end points, which would allow clinicians to make informed and objective assessment of the 2 new drug treatments compared with enoxaparin, a parenteral anticoagulant with a wellestablished efficacy and safety profile. Second, our analysis used study-level data rather than patient-level data. Although separate analyses based on individual-patient data would have been beneficial to assess the data with time-to-event methodology and explore baseline patient characteristics that might predict important clinical outcomes, the logistic challenges with obtaining multiple data sets meant that is was not possible to perform this type of analysis. Third, with only 3 trials within each meta-analysis, we were not able to fully explore individual subgroups according to the dose of enoxaparin (40 mg once daily versus 30 mg twice daily) or type of arthroplasty (hip versus knee), in which clinically important benefits might have been observed. Despite the relatively small number of trials included in our meta-analyses (3 in each comparison), each of the individual trials were substantially larger than previous randomized, controlled trials conducted in this surgical setting. 3 Other than the RE-NOVATE II study, 19 it is unlikely that any further randomized trials will be conducted with these agents. Therefore, data from this meta-analysis represents the best estimates of the efficacy and safety of enoxaparin compared with dabigatran or rivaroxaban in patients undergoing hip or knee arthroplasty. Forth, by nature of the trial design used in each study, patients with impaired renal function and the very elderly were excluded and thus general conclusions drawn from these analyses only pertain to the highly selected individuals and caution is needed with extrapolation to the general population. Finally, meta-analysis remains retrospective research that is subject to the methodological deficiencies of the included studies. However, we minimized the likelihood of bias by developing a detailed protocol before commencing this study, performing a meticulous and exhaustive search for both published and unpublished trial data, and using explicit methodology for study selection, data extraction, and data analysis. Our findings may have important implications for the choice of prophylactic agent in major joint arthroplasty. Balancing the risks of bleeding with the risks of thromboembolic events is commonly performed by surgeons performing lower-limb joint arthroplasty, and choice of prophylactic therapy is often determined on an individual basis. Because both dabigatran and rivaroxaban are administered orally and therefore are more convenient than parenterally administered anticoagulants, particularly for out-of-hospital prophylaxis, either agent may be an attractive alternative for patients undergoing total hip or knee arthroplasty. Acknowledgments We thank the authors of the studies included in our meta-analysis (B.I. Eriksson, J.S. Ginsberg, and B.L. Davidson) as well as S. Hantel from Boehringer Ingelheim for their assistance in verifying the accuracy of the data and providing missing data. Disclosures Dr Huisman received research grant support from AstraZeneca, Bayer, Boehringer Ingelheim, GlaxoSmithKline, and Pfizer and has been an investigator in trials for these companies. Dr Quinlan acted as a consultant for Boehringer Ingelheim. Dr Dahl acted as a consultant to Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Pfizer, Sanofi-Aventis, and Wyeth. Prof Schulman is an investigator for Boehringer Ingelheim and member of the safety committee for Bayer. References 1. Warwick D, Friedman RJ, Agnelli G, Gil-Garay E, Johnson K, FitzGerald G, Turibio FM. Insufficient duration of venous thromboembolism prophylaxis after total hip or knee replacement when compared with the time course of thromboembolic events: findings from the Global Orthopaedic Registry. J Bone Joint Surg Br. 2007;89: Sweetland S, Green J, Liu B, Berrington de GA, Canonico M, Reeves G, Beral V. Duration and magnitude of the postoperative risk of venous thromboembolism in middle aged women: prospective cohort study. BMJ. 2009;339:b4583 [Epub 3 Dec 2009]. 3. Geerts WH, Bergqvist D, Pineo GF, Heit JA, Samama CM, Lassen MR, Colwell CW. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008;133(Suppl):381S 453S. 4. Novicoff WM, Brown TE, Cui Q, Mihalko WM, Slone HS, Saleh KJ. Mandated venous thromboembolism prophylaxis: possible adverse outcomes. J Arthroplasty. 2008;23(Suppl 1): RE-MOBILIZE Writing Committee, Ginsberg JS, Davidson BL, Comp PC, Francis CW, Friedman RJ, Huo MH, Lieberman JR, Muntz JE, Raskob GE, Clements ML, Hantel S, Schnee JM, Caprini JA. Oral thrombin inhibitor dabigatran etexilate vs North American enoxaparin regimen for prevention of venous thromboembolism after knee arthroplasty surgery. J Arthroplasty. 2009;24: Eriksson BI, Dahl OE, Rosencher N, Kurth AA, van Dijk CN, Frostick SP, Kälebo P, Christiansen AV, Hantel S, Hettiarachchi R, Schnee J, Büller HR. Oral dabigatran etexilate vs subcutaneous enoxaparin for the prevention of venous thromboembolism after total knee replacement: the RE-MODEL randomized trial. J Thromb Haemost. 2007;5: Eriksson BI, Dahl OE, Rosencher N, Kurth AA, van Dijk CN, Frostick SP, Prins MH, Hettiarachchi R, Hantel S, Schnee J, Büller HR.

9 Huisman et al Enoxaparin Versus Dabigatran or Rivaroxaban 9 Downloaded from by guest on July 19, 2018 Dabigatran etexilate versus enoxaparin for prevention of venous thromboembolism after total hip replacement: a randomised, double-blind, non-inferiority trial Lancet. 2007;370: Eriksson BI, Borris LC, Friedman RJ, Haas S, Huisman MV, Kakkar AK, Bandel TJ, Beckmann H, Muehlhofer E, Misselwitz F, Geerts W. Rivaroxaban versus enoxaparin for thromboprophylaxis after hip arthroplasty. N Engl J Med. 2008;358: Lassen MR, Ageno W, Borris LC, Lieberman JR, Rosencher N, Bandel TJ, Misselwitz F, Turpie AG. Rivaroxaban versus enoxaparin for thromboprophylaxis after total knee arthroplasty. N Engl J Med. 2008;358: Turpie AG, Lassen MR, Davidson BL, Bauer KA, Gent M, Kwong LM, Cushner FD, Lotke PA, Berkowitz SD, Bandel TJ, Benson A, Misselwitz F, Fisher WD. Rivaroxaban versus enoxaparin for thromboprophylaxis after total knee arthroplasty (RECORD4): a randomised trial. Lancet. 2009;373: Kakkar AK, Brenner B, Dahl OE, Eriksson BI, Mouret P, Muntz J, Soglian AG, Pap AF, Misselwitz F, Haas S. Extended duration rivaroxaban versus short-term enoxaparin for the prevention of venous thromboembolism after total hip arthroplasty: a double-blind, randomised controlled trial. Lancet. 2008;372: Division of Medical Imaging and Hematology Products/Office of Oncology Drug Products/Office of New Drugs. FDA Advisory Committee Briefing Document. New Drug Application (NDA) : Rivaroxaban oral tablets (10 milligrams), Johnson & Johnson Pharmaceutical Research & Development, LLC, for the prophylaxis of deep vein thrombosis and pulmonary embolism in patients undergoing hip replacement or knee replacement surgery. Available at: Accessed December 1, Johnson & Johnson Pharmaceutical Research & Development LLC. Advisory Committee Briefing Book. Rivaroxaban for the prophylaxis of deep vein thrombosis (DVT) and pulmonary embolism (PE) in patients undergoing hip or knee replacement surgery. JNJ Available at: Accessed December 1, Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22: Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327: Deeks JJ, Higgins JP, Altman DG, on behalf of the Cochrane Statistical Methods Group. Chapter 9: Analysing data and undertaking meta-analyses. In: Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions. Version [updated September 2009]. The Cochrane Collaboration, Available at: Accessed December 1, Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315: Eriksson BI, Kakkar AK, Turpie AG, Gent M, Bandel TJ, Homering M, Misselwitz F, Lassen MR. Oral rivaroxaban for the prevention of symptomatic venous thromboembolism after elective hip and knee replacement. J Bone Joint Surg Br. 2009;91: Boehringer Ingelheim. A phase III randomised, parallel group, doubleblind, active controlled study to investigate the efficacy and safety of orally administered 220 mg dabigatran etexilate capsules (110 mg administered on the day of surgery followed by 220 mg once daily) compared to subcutaneous 40 mg enoxaparin once daily for days, in prevention of venous thromboembolism in patients with primary elective total hip arthroplasty surgery (RE-NOVATE II). ClinicalTrials.gov, Available at: dabigatran&rank 9. Accessed December 1, Quinlan DJ, Eikelboom JW, Dahl OE, Eriksson BI, Sidhu PS, Hirsh J. Association between asymptomatic deep vein thrombosis detected by venography and symptomatic venous thromboembolism in patients undergoing elective hip or knee surgery. J Thromb Haemost. 2007;5: Hull RD, Yusen R, Bergqvist D. Assessing the safety profiles of new anticoagulants for major orthopedic surgery thromboprophylaxis. Clin Appl Thromb Hemost. 2009;15: Lassen MR, Bauer KA, Eriksson BI, Turpie AG. Postoperative fondaparinux versus preoperative enoxaparin for prevention of venous thromboembolism in elective hip-replacement surgery: a randomised double-blind comparison. Lancet. 2002;359: Eriksson BI, Borris LC, Dahl OE, Haas S, Huisman MV, Kakkar AK, Muehlhofer E, Dierig C, Misselwitz F, Kälebo P. A once-daily, oral, direct Factor Xa inhibitor, rivaroxaban (BAY ), for thromboprophylaxis after total hip replacement. Circulation. 2006;114:

10 Enoxaparin Versus Dabigatran or Rivaroxaban for Thromboprophylaxis After Hip or Knee Arthroplasty: Results of Separate Pooled Analyses of Phase III Multicenter Randomized Trials Menno V. Huisman, Daniel J. Quinlan, Ola E. Dahl and Sam Schulman Downloaded from by guest on July 19, 2018 Circ Cardiovasc Qual Outcomes. published online October 5, 2010; Circulation: Cardiovascular Quality and Outcomes is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2010 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation: Cardiovascular Quality and Outcomes can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation: Cardiovascular Quality and Outcomes is online at:

11 SUPPLEMENTAL MATERIAL

12 Supplemental Tables Supplemental Table 1: Efficacy, bleeding and adverse events during treatment in randomized trials comparing enoxaparin vs. dabigatran 150 mg once-daily Outcome Total N Enoxaparin, n/n (%) Dabigatran, n/n (%) OR (95% CI)* P-value Efficacy Symptomatic VTE + all-cause mortality 5,418 23/2695 (0.9) 24/2723 (0.9) 0.97 ( ) 0.91 Symptomatic VTE 5,418 23/2695 (0.9) 20/2723 (0.7) 1.16 ( ) DVT 5,418 14/2695 (0.5) 18/2723 (0.7) 0.79 ( ) PE 5,418 9/2695 (0.3) 2/2723 (0.1) 3.85 ( ) 0.06 Death 5,418 1/2695 (0.1) 5/2723 (0.2) 0.34 ( ) 0.18 Major VTE plus VTE-related mortality 4,167 69/2096 (3.3) 78/2071 (3.8) 0.87 ( ) 0.40 Bleeding Major + clinically relevant non-major bleeding 5, /2716 (5.0) 154/2737 (5.6) 0.89 ( ) 0.34 Major bleeding 5,453 39/2716 (1.4) 29/2737 (1.1) 1.36 ( ) Fatal bleeding 5,453 0/2716 (0.0) 1/2737 (0.04) 0.34 ( ) Bleeding into a critical organ 5,453 0/2716 (0.0) 1/2737 (0.04) 0.34 ( ) Bleeding assoc with Hb drop 20 g/l compared with pre-op 5,453 32/2716 (1.2) 24/2737 (0.9) 1.35 ( ) Bleeding leading to re-operation 5,453 5/2716 (0.2) 4/2737 (0.2) 1.23 ( ) Bleeding warranting treatment cessation 5,453 2/2716 (0.1) 1/2737 (0.04) 1.68 ( ) Bleeding leading to transfusion of 2 U of blood 5,453 33/2716 (1.2) 17/2737(0.6) 1.97 ( ) 0.02 Clinically relevant non-major bleeding 5,453 98/2716 (3.6) 125/2737 (4.6) 0.78 ( ) 0.08