Rivaroxaban for Prevention of Venous Thromboembolism after Total Knee Replacement: Impact on Healthcare Costs Based on the RECORD4 Study.

Introduction: Rivaroxaban is a novel, oral, direct Factor Xa inhibitor in development for the prevention of venous thromboembolism (VTE) after major orthopaedic surgery. Two large, phase III, randomized controlled trials compared rivaroxaban with subcutaneous (sc) enoxaparin following total knee replacement (TKR). In the recently published RECORD3 trial, rivaroxaban, 10 mg once daily (od), showed superior efficacy to an enoxaparin regimen of 40 mg od. However, enoxaparin 30 mg twice daily (bid) is the more widely used regimen in North America for VTE prophylaxis following TKR. The RECORD4 trial compared rivaroxaban 10 mg od versus enoxaparin 30 mg bid (both for 12 ± 2 days), for prevention of VTE following TKA. In the modified intent to treat population, the primary outcome (deep vein thrombosis [DVT], pulmonary embolism [PE], and all-cause mortality) occurred in 6.9% of rivaroxaban patients and 10.1% of enoxaparin patients (RRR 32%; p<0.012). Symptomatic VTE was observed in 0.7% of rivaroxaban patients and 1.2% of enoxaparin patients, although this difference was not statistically significant. There was no difference in major bleeding between rivaroxaban and enoxaparin. This analysis was designed to show the impact of the efficacy improvement with rivaroxaban and of oral vs. subcutaneous administration on US healthcare costs from a payer’s perspective.

Methods: The impact of rivaroxaban was assessed using the efficacy data from RECORD4, and the reduced administration cost associated with oral therapy. The treatment cost of symptomatic VTE and major bleeding were taken from published managed care data in the US. For costing purposes, the duration of hospitalization for TKR (4 days) was obtained from a published US orthopaedic registry. It was assumed that asymptomatic events had no impact on healthcare costs, and conservatively assumed that nurses spend three minutes per day administering enoxaparin and training patients to self-inject for outpatient use while patients are in the hospital. This is likely to be an underestimate in light of other studies and clinical experience of VTE post TKR in the US clinical practice, so a sensitivity analysis included incremental costs associated with home nurse visits to administer sc injections. The duration of prophylaxis was assumed to be for 14 days. As rivaroxaban is not yet approved and available in the US, the analysis assumed similar drug acquisition costs to enoxaparin 40 mg, which is less expensive than enoxaparin 30mg bid.

Results: The total cost associated with healthcare resource use in the US for the duration of treatment was $469 per patient with enoxaparin 30 mg bid injection and $307 with oral rivaroxaban 10 mg od. This implies a saving of $162 per patient. This improvement was driven primarily by the reduced drug acquisition relative to enoxaparin 30 mg bid, and monitoring costs. In addition, published data suggest home nurse visits during post-hospital prophylaxis period cost an average of $100 for enoxaparin; suggesting that potential savings with rivaroxaban could rise to $262 per patient. This analysis excludes any treatment benefits from a reduction in symptomatic events or the impact of post-thrombotic syndrome. However, pooled data from all phase III rivaroxaban trials (RECORD1, 2, 3, 4) vs. enoxaparin showed a significant reduction in symptomatic VTE and death until day 12±2 (P<0.001). Estimated US costs for the treatment for symptomatic VTE range from $9,805 to $14,146 per event. Given the reduced incidence of symptomatic events with rivaroxaban, there may be even further savings in healthcare costs associated with rivaroxaban.

Conclusions: The improved efficacy of rivaroxaban over enoxaparin 30 mg BID for VTE prevention after TKR and reduced administration and acquisition costs are expected to result in savings to healthcare resources. With more than 300,000 US patients having TKR annually, these potential cost savings are significant.