Summary

• The economic value proposition of XARELTO is driven by the relative risk reduction of key clinical events (reduction in stroke and systemic embolism) as observed in the ROCKET AF study for stroke prevention in atrial fibrillation (AF).
• Economic benefits - stroke risk reduction in nonvalvular AF (NVAF):
  • A cost-effectiveness model of XARELTO and warfarin for the reduction of stroke in AF patients found that XARELTO appears to be a cost-effective alternative to warfarin, given the assumptions and limitations of the model. XARELTO therapy is associated with higher drug acquisition costs than warfarin, but these are partly offset by decreased drug administration and event treatment costs.
  • Sensitivity analyses found that the key drivers of cost-effectiveness of XARELTO were treatment efficacy rates, frequency and cost of monitoring visits to the anticoagulation clinic for warfarin, and the baseline myocardial infarction (MI) rate.
  • A cost-effectiveness analysis compared the use of XARELTO vs adjusted-dose warfarin for the prevention of strokes in patients with AF and demonstrated similar results.¹ The authors concluded that XARELTO has a high probability of being a cost-effective alternative to warfarin.
  • A Markov model-based study compared the cost-effectiveness of apixaban with that of XARELTO, warfarin, or dabigatran for stroke prevention in patients with NVAF in the Chinese population. In the base case analysis, the quality-adjusted life years (QALYs) gained with XARELTO, apixaban, and dabigatran 150 mg BID were 5.870, 6.017, and 6.022, respectively.²
  • A systematic review and meta-analysis comparing the cost-effectiveness of DOACs vs warfarin for stroke prevention in AF by pooling incremental net benefits (INBs) demonstrated that XARELTO was cost-effective compared with warfarin in high-income countries from a third-party payer perspective but not from a societal perspective.³
• Additional citations identified during a literature search have been included in the REFERENCES section for your review.^4-9

THE XARELTO COST-EFFECTIVENESS MODEL-REDUCTION OF RISK OF STROKE AND SYSTEMIC EMBOLISM IN AF

Overview-The Economic Value Proposition

The economic value proposition of XARELTO is driven by relative risk reduction of key clinical events (reduction in stroke and systemic embolism) as observed in the ROCKET AF study. XARELTO is cost-effective compared to warfarin due to risk reduction in stroke, intracranial hemorrhage, MI, and vascular death without a significant increase in major bleeding. XARELTO provides additional economic benefits by eliminating the need for coagulation monitoring and associated lab or office visits. These results work together to strengthen the risk-benefit aspect of the value proposition.

Results from an economic model comparing XARELTO with warfarin in stroke prevention in AF, based mainly on population and clinical events from the ROCKET study, suggest that XARELTO is a cost-effective alternative to warfarin for stroke prevention in AF. Although treatment with XARELTO is associated with higher drug acquisition costs than warfarin, these are partly offset by decreased drug administration and event treatment costs. The following sections describe the model. The model has also been summarized in 2 published manuscripts.^10,11

Structure of the Economic Model

A Markov model was developed to assess the cost-effectiveness of XARELTO relative to warfarin for stroke prevention in AF. Patients enter the model with stable uncomplicated nonvalvular AF and are being treated/initiated with oral anticoagulants (therapy initiation). The major complications considered in the model are:

• Stroke (ischemic and hemorrhagic)
• Systemic embolism
• MI
• Major bleeds (major extracranial bleeds and intracranial bleeds, excluding hemorrhagic strokes)

In the ROCKET trial, the composite primary endpoint was ischemic stroke, hemorrhagic stroke, and systemic embolism. Therefore, the model considers them together as 1 health state. Major bleed events in the model exclude hemorrhagic strokes as this is part of the primary efficacy endpoint. Patients progress between the health states every 3 months (cycle length), according to transition probabilities derived from ROCKET or reported in the literature. Each health state has a cost and a utility weight describing the quality of life associated with it. The sum of the costs and utilities in the cohort progression are used to calculate the ICERs for cost per life year (LY) gained as the main model output.

The base case population is modeled on that of the ROCKET study using the prespecified safety on treatment analysis with age of the cohort being 65 years. The primary analysis is based on the prespecified analysis and is focused on the on-treatment population. The distribution of risk of stroke uses the CHADS₂ (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke [double weight]) system: 0% - low risk, 13% - moderate risk and 87% - high risk, where low risk is defined by CHADS₂ risk score of 0-1, medium risk is defined as CHADS₂ risk score of 2 and high risk is defined as a score of 3 or higher.

The interventions assumed in the model are:

• XARELTO 20 mg QD
• Warfarin 4.5 mg QD (target INR 2.5)

The model assumes that if patients discontinue from XARELTO or warfarin, they will switch to aspirin or no treatment.

The model has a lifetime horizon to fully incorporate the costs and consequences of AF. The base case analysis takes the perspective of the United States (US) health care payer, capturing the 2 management models in the US: clinic- and primary care-based. Costs are reported in 2013 US dollars, updated from the original 2010 model (Table: Resource Use and Costs).

Resource Use and Health Care Costs

The resource use and cost inputs considered in the model are drug acquisition, drug administration & monitoring, and event treatment (Table: Resource Use and Costs). The drug acquisition costs used in the base case analysis are wholesale acquisition costs. The frequency of monitoring visits in the maintenance phase in both the anticoagulation clinic and general practitioner office/usual care settings is based on a 2008 US study by Nichol et al. The cost of each anticoagulation monitoring visit for warfarin patients is based on the resource use described in a study by Lafata et al¹², with unit component costs from Medicare fees for appropriate current procedural terminology codes for 2013. Acute events costs were obtained either from the published literature (acute intracranial hemorrhage and acute MI, Freeman 2011¹³), or, where published payer perspective values were not available, acute costs were derived from the 2011 Medicare costs on the HCUPNet website (HCUPNet 2011)¹⁴, and inflated to 2013 equivalent costs. Where several related diagnosis related group codes were applicable for a particular event (e.g. with or without major complications), costs were weighted based on their relative 2011 total discharges, as listed on HCUPNet.

The values of clinical inputs for the model and corresponding ranges used for sensitivity analyses are shown in Table: Clinical Inputs. Most of the inputs were estimated from values observed in the ROCKET study. For the sensitivity analyses, the lower and upper values of these rates were calculated as the 95% confidence interval (CI) of a beta probability distribution based on the data from the ROCKET trial. In summary, the clinical inputs considered are:

• Baseline risk of the clinical events
• Treatment efficacy - measured by relative risk of an event following treatment
• Treatment safety - also measured by relative risk for developing major bleeds as a result of treatment
• Case fatality rate - used to determine mortality due to the event in the model cycle in which it occurred; this was assumed to be the same regardless of treatment regimen This also applies to mortality observed in the postevent states
• Discontinuation & persistence rate (as described in the previous section)

Outcome Measure

The primary outcome measure in the model was the cost per LY gained. Cost per QALY gained was a secondary outcome.

Sensitivity Analysis

One-way sensitivity analyses were performed to determine the key drivers of cost-effectiveness in the model. Key parameters were varied to low and high values within plausible ranges. For parameters such as clinical efficacy values which are based on robust studies or reviews, the reported 95% CI was used. For parameters for which there are still uncertainties surrounding the source data, ranges reported in the literature were used. For the 2-way sensitivity analysis, 2 parameters were selected from the top drivers of cost-effectiveness, as determined by the 1-way sensitivity analyses. The 2 parameters selected were the frequency and unit cost of warfarin monitoring visits because these were the parameters judged most likely to vary in US clinical practice. The frequency of warfarin monitoring visits was varied within the range used for the 1-way sensitivity analysis, while the unit cost of warfarin monitoring visits was varied within the arbitrary range used for the 1-way sensitivity analysis.

Results

Base Case Analysis

The base case results including incremental values are presented in Table: Base Case Results (Per Patient Per Lifetime). XARELTO has an ICER of $48,545 per LY gained, and $43,276 per QALY gained, when compared to warfarin. XARELTO treatment is associated with higher overall cost than warfarin treatment. This is due mainly to drug acquisition costs given that drug administration and event treatment costs are both lower for patients treated with XARELTO. Although XARELTO treatment results in a higher incidence of major nonfatal bleeds (2 additional events per 100 patients), it is associated with a lower incidence of strokes than warfarin treatment (incrementally 3 stroke events avoided per 100 patients), resulting in more LYs gained for XARELTO patients.

One-Way Sensitivity Analysis

Key results of the 1-way sensitivity analysis are depicted in a tornado diagram in Figure: Tornado Diagram of Cost-Drivers of ICER (Cost/LY Gained). This shows the parameters that cause the cost-effectiveness ratios to change by $15,000 or more.

The top key drivers of cost-effectiveness are:

• Relative risks of XARELTO vs warfarin in preventing the primary efficacy endpoint (stroke/systemic embolism) and MI
• Frequency and cost of monitoring visits to the anticoagulation clinic for warfarin
• Baseline MI rate

Varying the relative risk for the primary endpoint to the lower and upper limits of its 95% CI (0.65 to 0.95) produced ICERs (cost/LY) of $19,965 and $155,691, respectively. The same trend of results applies when the relative risk of MI with XARELTO is varied ($38,795 and $76,641/LY gained).

Reducing the baseline MI rate from 0.28% to 0.03% increases the ICER of XARELTO to $31,398/LY gained, while increasing the rate to 1.22% leads to an improvement in the ICER ($60,535/LY). The unit cost and frequency of visits for anticoagulation monitoring are likewise sensitive parameters. A 25% increase in the unit cost of this monitoring improves the cost-effectiveness of XARELTO (from $48,525/LY to $34,316/LY gained) and vice-versa (a 25% decrease in cost leads to $62,733/LY). The frequency of visits to the anticoagulation clinic for warfarin monitoring was inversely related to the cost-effectiveness of XARELTO. Varying the visits by 25% in either direction produced ICERs of $36,003/LY gained and $61,046/LY, respectively. An alternate scenario in which the only 31% of patients were monitored in an anticoagulation clinic with the remainder seen in a usual care clinic as in Nichol’s study, produced a lower ICER of $46,562/LY.

Two-Way Sensitivity Analysis

Figures: Two-Way Sensitivity Analysis - ICER (Cost/LY Gained) and Two-Way Sensitivity Analysis - Annual Warfarin Monitoring Cost illustrate the different ICERs (cost/LY gained) and annual warfarin monitoring costs achieved for each combination of values tested for the 2 parameters (annual warfarin monitoring frequency and unit cost of anticoagulation visit. Red and green squares represent combinations resulting in ICERs of greater than $50,000 or less than $50,000 per LY gained, respectively. The square with a bold outline represents the combinations of values used in the base case analysis.

Subgroup Analyses

The results shown in Table: Subgroup Analysis are based on the safety on treatment in the ROCKET AF population. The subgroup in which XARELTO had the best cost-effectiveness ratio (LY gained) is the vitamin K antagonist (VKA)-difficult - high resource use and poor INR control subgroup, followed by the high resource use subgroup, then the poor INR control subgroup. The ICER for the prior stroke subgroup is higher than that of the high risk of stroke subgroup. This can be partly explained by the fact that a better relative risk reduction of the primary efficacy endpoint was observed in the ROCKET AF study with the high-risk subgroup compared to the prior stroke subgroup.

Conclusions

The results of the model predict that XARELTO is a cost-effective alternative to warfarin for the prevention of stroke in US AF patients. XARELTO treatment is associated with higher drug acquisition costs than warfarin, but these are partly offset by decreased drug administration and event treatment costs.

Cost-Effectiveness of XARELTO Compared to adjusted-dose Warfarin for Stroke Prophylaxis in AF - BASE CASE ANALYSIS

A cost-effectiveness analysis compared the use of XARELTO (20 mg QD) vs adjusted-dose warfarin for the prevention of strokes in patients with AF and demonstrated similar results (Lee 2012).¹ The base case of this Markov model used a hypothetical cohort of patients age 65 years or older with a CHADS₂ score of ≥3 and no contraindications to anticoagulation. All patients began the simulation in “well with AF” state and progressed through the model based on probabilities and rates published in the ROCKET AF trial as well as other published anticoagulation literature (Table 1). The model found a high probability that XARELTO is a cost-effective alternative to warfarin for the prevention of stroke. It estimated total lifetime costs for 65-year-old AF patients of $94,456 per patient on XARELTO and $88,544 per patient on adjusted-dose warfarin, with a quality-adjusted life expectancy of 10.03 QALYs per patient on XARELTO, and 9.81 QALYs per patient on adjusted-dose warfarin. With these results, the authors concluded that XARELTO has a high probability of being a cost-effective strategy given its ICER of $26,873/QALY.

These results are similar to the previously presented cost effectiveness model with a similar ICER. See Table: Base Case Model Variables and Ranges Used in Sensitivity Analysis.

Cost-effectiveness of XARELTO compared to other DOACs - Stroke prevention in patients with NVAF

A cost effectiveness analysis compared the costs and clinical outcomes of apixaban 5 mg (twice daily [BID]) with XARELTO 20 mg (once daily [QD]), warfarin (adjusted target international normalized ratio [INR]), dabigatran 110 mg (BID), or dabigatran 150 mg (BID) for stroke prevention in patients with NVAF in the Chinese population. Using the Clinical Data Analysis and Reporting System (CDARS) database, patient profiles, costs, and part of transition probabilities were collected retrospectively for incident patients with NVAF.²

Results

• A total of 71,705 patients with NVAF were included; of these, 49.7% were male and 50.3% were female, with mean age of 73.2 and 78.9 years, respectively.

Base Case Analysis

• In the base case, the QALYs gained with XARELTO, apixaban, and dabigatran 150 mg BID were 5.870, 6.017, and 6.022, respectively.
• Compared with XARELTO, which had low cost ($8088), apixaban was associated with improved QALY gained (0.147 QALYs) at the cost of $152, leading to an incremental cost-effectiveness ratio (ICER) of $1034/QALY gained below the threshold of willingness to pay (WTP).

Cost-Effective Acceptability

• Using the Monte Carlo simulation (WTP threshold of $46,091/QALY gained in the probabilistic sensitivity analysis (PSA), including 2000 iterations), the probability of XARELTO 20 mg, warfarin, dabigatran 110 mg, dabigatran 150 mg, and apixaban 5 mg being cost-effective was 0%, 0%, 29.4%, 33.2%, and 37.4%, respectively.

Cost-Effectiveness of XARELTO Compared to Warfarin for Stroke Prophylaxis in AF - pooled iNCREMENTAL NET BENEFITS

A systematic review and meta-analysis compared the cost-effectiveness of DOACs vs warfarin, including XARELTO vs warfarin (n=34), for stroke prevention in AF by pooling INBs (Noviyani 2022).³ The search strategy was comprised of data from inception to December 2019. A total of 100 studies (consisting of 224 comparisons) were included in the review, of which 86, 13, and 1 were conducted in high-income countries, upper-middle income countries, and low/middle income country, respectively. The primary outcome of interest, INB, was calculated as a difference of incremental effectiveness multiplied by WTP threshold minus the incremental cost. A positive INB indicated favoring treatment (ie, intervention is cost-effective), whereas a negative INB indicated favoring comparator (ie, intervention is not cost-effective). Heterogeneity was evaluated by using the Cochrane-Q test and the I² statistic.

In high-income countries (26 studies with 28 comparisons), the pooled INBs for XARELTO vs warfarin were $7664.58 (95% CI, $2979.79-$12,349.37; I²=0%) from a third-party payer perspective and $10,345.74 (95% CI, -$15,461.54 to $36,153.02; I²=30.7%) from a societal perspective. In upper-middle income countries (7 studies with 10 comparisons), the pooled INBs were -$27,567.34 (95% CI, -$170,185.85 to $115,051.17; I²=99.9%) from a third-party payer perspective and -$14,350.24 (95% CI, -$21,631.83 to -$7068.64; I²=68.3%) from a societal perspective.

XARELTO was cost-effective compared with warfarin in high-income countries with lifetime horizon from a third-party payer perspective but not from a societal perspective. XARELTO was significantly not cost-effective compared with warfarin in upper-middle income countries.

LITeRATURE SEARCH

A literature search of MEDLINE^®, EMBASE^®, BIOSIS Previews^®, DERWENT^® (and/or other resources, including internal/external databases) was conducted on 16 December 2023.