This analysis evaluates-from a US payer's perspective-the economic impact of using rivaroxaban versus enoxaparin for thromboprophylaxis following total hip or knee replacement surgery.
Venous thromboembolism (VTE) comprises deepvein thrombosis (DVT) and pulmonary embolism (PE) and represents a signifi cant health burden in the United States. Venous thromboembolism occurs frequently after major orthopedic surgery, and patients undergoing total hip replacement (THR) or total knee replacement (TKR) are among those with the highest risk for VTE.1 The economic impact of this complication is substantial because patients who develop VTE after major orthopedic surgery are more likely to be rehospitalized and/or remain hospitalized for a longer period.2 Furthermore, the number of total joint replacement surgeries in the United States is expected to rise as the population ages, making appropriate thromboprophylaxis strategies important for reducing the incidence and recurrence of VTE.3
Pharmacologic thromboprophylaxis in patients undergoing THR and TKR signifi cantly reduces the risk of postoperative VTE. The 9th American College of Chest Physicians evidence-based guidelines suggest at least 10 and up to 35 days of thromboprophylaxis with an anticoagulant after THR/TKR.1 However, continuing anticoagulant therapy after patients have been discharged is complicated. Information from the Multinational Global Orthopaedic Registry indicates that in real-world practice, the standards for prophylaxis recommended by the American College of Chest Physicians are largely not followed.4 This may reflect the myriad inconveniences associated with the current standard of care (eg, frequent monitoring and dose adjustments, daily self administered subcutaneous injections, narrow therapeutic window, numerous food and drug interactions) and bolsters the rationale for use of an oral, fixed-dose anticoagulant that can be administered without the practical management challenges associated with traditional anticoagulants.
Rivaroxaban is an oral, once-daily anticoagulant that directly inhibits Factor Xa activity, resulting in the inhibition of thrombin generation by both intrinsic and tissue factor pathways.5 The pivotal phase 3 RECORD trials evaluated thromboprophylaxis with rivaroxaban in more than 12,500 patients undergoing THR (RECORD trials 1 and 2) and TKR (RECORD trials 3 and 4) and provided the foundation for its approval in Canada and the United States.6-11
The goal of this analysis was to evaluate, from a payer’s perspective, the economic impact of using rivaroxaban versus enoxaparin, a low-molecular-weight heparin, for thromboprophylaxis after THR or TKR surgery.
METHODSClinical Data Sources for Model
The RECORD trials were randomized, multinational, double-blind, phase 3 trials conducted to assess the efficacy and safety of rivaroxaban compared with enoxaparin in patients undergoing THR or TKR. Data cited in the Xarelto (rivaroxaban) US package insert from the first, second, and third RECORD clinical trials were analyzed using an Excel-based model (Microsoft Corp, Redmond, Washington).6 The duration of thromboprophylaxis, dosing regimens, and rationale for data inclusion in the current study are summarized in
.6,12-15 Identical efficacy and safety end points were assessed in all the clinical trials. All patients were evaluated for asymptomatic DVT using ascending bilateral venography the day after the last dose of the study drug, or earlier if symptoms arose.8-11 Presence of PE was assessed by spiral computed tomography, perfusion-ventilation lung scintigraphy, or pulmonary angiography.8,10 All end points were analyzed by a central blinded adjudication committee up to day 42 for the RECORD 1 and 2 studies,8,9 and up to day 17 in the RECORD 3 study.10
For the THR analysis, the rivaroxaban arms from RECORD trials 1 and 2 were combined because the dosages of rivaroxaban and the durations of rivaroxaban administration in these 2 trials were similar (10 mg once daily for 35 days in RECORD trial 1 and 31 to 39 days in RECORD trial 2). The THR data for enoxaparin from only RECORD trial 1 were used for comparison because the duration of prophylaxis was the same as that for rivaroxaban (ie, 35 days). For TKR, data from RECORD trial 3 were used for both the rivaroxaban (10 mg once daily) and enoxaparin (40 mg once daily) arms (both for 12 days).
The model incorporated 4 clinical end points: total VTE, major VTE, symptomatic VTE, and major bleeding. The primary end point evaluated in the clinical trials was total VTE, which was defi ned as a composite of any DVT, nonfatal PE, or death from any cause up to 36 days postoperatively. Other end points assessed included major VTE—comprising proximal DVT, nonfatal PE, and VTErelated death—and symptomatic VTE that occurred during the prophylaxis period or within 30 to 35 days after the last dose of either drug.8,11 The primary safety outcome for all trials was the incidence of major bleeding events, which was defi ned as bleeding that was fatal, occurred in a critical organ, or required surgical intervention, or any clinically overt bleeding that occurred at an extra-surgical site that resulted in a drop in hemoglobin of 2 or more g/dL (20 g/L) or that required a transfusion of 2 or more units of blood products.8,11
Drug costs were based on wholesale acquisition costs for rivaroxaban and generic enoxaparin (lowest wholesale acquisition cost for generic enoxaparin as of July 2012),16 and calculations were made based on the dosing schedule and duration used in the RECORD trials. Only the costs of symptomatic events were considered for the event cost analysis (symptomatic VTEs and major bleeding events). Differences in total and major VTEs are presented, but they were not used in the cost analysis. Symptomatic VTE costs were derived from the study by Encinosa and Hellinger,17 who found that the cost of care for patients being treated in a hospital setting for postoperative VTE within 90 days of surgery was approximately $50,911, whereas patients who did not have any complicating factors in the same postoperative time period incurred costs of approximately $18,284. Once adjustments for age, sex, concurrent medical conditions, and other factors were made between the 2 groups, the cost difference was approximately $7838 (2002 dollars), which was adjusted to a value of $10,546 in 2010 dollars.17 In addition, costs associated with postoperative major bleeding events were derived from a model presented at the 2008 American Society of Hematology Meeting by Kwong and colleagues.12 The analysis used major bleeding event costs of $3828 in 2007, which were converted to $4195 in 2010 dollars.12
Applying the RECORD event data described previously, an economic model was created using Excel to compare the projected difference in clinical events and projected cost of thromboprophylaxis for a sample population of 1000 patients undergoing THR or TKR surgery. Results of the analysis were reported in terms of differences in event rates and associated costs of VTE, major VTE, symptomatic VTE, and major bleeding events or the separate THR and TKR cohorts of patients receiving rivaroxaban versus enoxaparin. Sensitivity analyses were undertaken to explore the impact of varying generic enoxaparin and event costs by ±30%.
Clinical Event and Cost Comparison Model After Total Hip Replacement
In the RECORD 1 and 2 trials, rivaroxaban 10 mg once daily was administered for 35 days. At a cost of $7.72 per day (for a 35-day study duration), drug acquisition costs were $270.20 per patient (total for 1000 patients was $270,200). The enoxaparin drug acquisition costs per patient for the same 35-day study duration were $770.00 based on a per day cost of $22.00 for a 40-mg generic (total for 1000 patients was $770,000). This resulted in cost savings of approximately $499,800, or $499.80 per patient over the course of thromboprophylaxis therapy with rivaroxaban.
Based on the events data from RECORD trials 1 and 2 for rivaroxaban and from the RECORD trial 1 data for enoxaparin, 1000 THR patients treated with rivaroxaban instead of enoxaparin would have 24.2 fewer total VTEs, 17.2 fewer major VTEs, 2.8 fewer symptomatic VTEs, and 1.2 additional major bleeding events. Monetizing the symptomatic events, $29,032 was saved because of the symptomatic VTEs avoided. After including the additional cost of prophylaxis-related bleeding events of $4991, there was an overall projected cost savings from clinical events avoided of $24,041. Once drug acquisition cost savings were included, the model projected potential cost savings of $523,841 associated with the use of rivaroxaban compared with prophylaxis with enoxaparin in this patient population (or a savings of ~$523.84 per patient). These projections are summarized in Table 2.
Clinical Event and Cost Comparison Model After Total Knee Replacement
In this 1000-patient example, the duration of thromboprophylaxis for both anticoagulants was 12 days on average, based on RECORD trial 3 data. Drug costs for rivaroxaban 10 mg at $7.72 per day were $92,640 over the course of the study period for 1000 patients. The prophylactic dose for enoxaparin was based on the RECORD trial 3 dosing regimen of 40 mg once daily. Daily drug costs were $22.00 per patient, or $264,000 for all patients receiving enoxaparin. Switching the 1000-patient cohort from generic enoxaparin to rivaroxaban would result in potential drug cost savings of $171,360 over the course of thromboprophylaxis.
Based on the event data from the RECORD 3 trial for rivaroxaban and enoxaparin, the 1000-patient TKR cohort would experience 91.1 fewer total VTEs, 15.0 fewer major VTEs, and 12.9 fewer symptomatic VTEs, with 0.9 additional major bleeding events using rivaroxaban instead of enoxaparin. Thus, the use of rivaroxaban in this patient population would result in potential event savings of $136,490 in symptomatic VTE events avoided. When the cost of possible major bleeding events of approximately $3819 was accounted for, there was a projected cost benefi t in adverse events avoided of $132,671. Overall total cost savings, including drug acquisition costs, were $304,031 (or $304.03 per patient) if rivaroxaban was used for thromboprophylaxis in the TKR population. These projections are summarized in Table 2.
The overall findings of rivaroxaban cost savings remained after performing sensitivity analyses. Decreasing generic enoxaparin prices by 30% reduced the magnitude of rivaroxaban cost savings in THR ($292,840) and TKR ($224,831), while increasing enoxaparin prices by 30% increased the cost savings associated with rivaroxaban (THR $754,841; TKR $383,231). Increasing and decreasing costs of symptomatic events (symptomatic VTEs and major bleeding) did not alter the results significantly (data not shown).
Using data from 3 large published clinical trials from the rivaroxaban US package insert,6 this economic model demonstrated significant cost savings and improvement in patient outcomes when rivaroxaban was used for thromboprophylaxis after total joint replacement surgery. For a cohort of 1000 patients over a 35-day prophylaxis regimen following THR, savings of nearly $500 per patient were projected when using rivaroxaban for VTE prophylaxis. In the TKR population, there were projected savings of approximately $304 per patient over the course of therapy when rivaroxaban was used. This analysis was based on data within a 90-day postoperative period; however, it is possible that the concurrent minimization of the risk of subsequent VTE and complications arising from post-thrombotic syndrome and/or chronic pulmonary hypertension might translate into even greater cost savings.1,2,13-15,18 It is important to note that these event-related costs only included symptomatic events.
Asymptomatic events could also lead to increased costs and negative clinical consequences, but these costs were not included in this economic model because only shortterm costs were assessed. It must be noted, however, that a number of these initially asymptomatic events may become symptomatic or, as a worse consequence, only present as PE at death.2,14,15 In a retrospective study (1979- 1983) done in the United Kingdom, it was found that DVT was particularly common after THR surgery, with approximately 50% of individuals undergoing hip replacement found to have a DVT by venography.13 These investigators found that, based on case records of 195 autopsies in which the cause of death was attributed to PE, no symptoms of DVT were reported in 77% (124/162) of the cases that reported fi nding DVT at autopsy. In addition, the percentage of patients with PE postmortem was shown to be 52% in individuals undergoing THR surgery.13
While these findings are intriguing, it is important to consider some of the limitations of this assessment. The definition of major bleeding may differ from the definition typically used in other studies.19 In the RECORD trials, the definition of major bleeding did not include bleeding at the surgical site unless the event required surgical intervention or was fatal. It is possible that this may have resulted in fewer major bleeding events being reported. However, the definition was discussed with regulatory authorities and agreed upon in advance of the analyses to allow for the assessment of clinically relevant bleeding events.9,11 In addition, the costs associated with training patients on how to use enoxaparin were not included. Furthermore, this model was not intended to be used for comparison with other agents, and any changes made in dosing or duration will likely alter the results. Finally, the prescribed dose of generic enoxaparin for TKR in the United States is often 30 mg twice daily ($33.00/day vs $22.00/day for 40 mg once daily), serving to increase the cost difference between generic enoxaparin and rivaroxaban and increasing the cost savings realized by rivaroxaban administration.
Rivaroxaban, a direct Factor Xa inhibitor, provided a safe and effective alternative to enoxaparin for prevention of VTE in patients undergoing THR or TKR. When analyzing these data from a payer’s perspective, there was a clear and signifi cant economic impact associated with VTE, which makes the choice of appropriate thromboprophylaxis an important one. Based on data from the rivaroxaban US package insert and 3 large head-to-head clinical trials, the use of rivaroxaban over enoxaparin not only decreased the number of thrombotic events, but also generated significant potential cost savings in both drug acquisition costs and event-related costs.