Cost-effectiveness of frontline eltrombopag with immunosuppressive therapy for severe aplastic anemia

Abstract

The randomized phase 3 RACE trial demonstrated eltrombopag used in conjunction with traditional immunosuppressive therapy (IST) to be an efficacious treatment for severe aplastic anemia. However, eltrombopag can pose a significant cost burden. Using clinical data from the RACE trial, we performed the first cost-effectiveness analysis of frontline eltrombopag plus IST in adults with newly diagnosed severe aplastic anemia across all accepted willingness-to-pay (WTP) thresholds from the United States modified societal perspective. The primary outcome was the incremental net monetary benefit across WTP thresholds of $50,000-$150,000/quality-adjusted life year (QALY) with base-case WTP threshold of $120,000/QALY. Frontline eltrombopag plus IST versus frontline IST alone accrued discounted costs of $593,000 and $561,000, and discounted QALYs of 11.53 and 10.73, respectively, with a per-patient incremental net monetary benefit of $64,000 [95% credible interval −$91,000 to $262,000] favoring frontline eltrombopag plus IST. Deterministic sensitivity analyses demonstrated that the 6-month overall response risk ratio (with versus without frontline eltrombopag) had the greatest effect on model results. In probabilistic sensitivity analysis, frontline eltrombopag with IST was favored in 78% of 10,000 second-order Monte Carlo iterations, as well as 58% and 82% of 10,000 iterations at WTPs of $50,000 and $150,000/QALY, respectively. At current pricing, frontline eltrombopag is the cost-effective therapeutic strategy in adults with severe aplastic anemia in the US. Longer-term clinical and patient utility data will likely help further clarify the cost-effectiveness of eltrombopag in this setting.

Introduction

Acquired aplastic anemia is a rare autoimmune bone marrow failure disorder (incidence of 2–3 cases per million per year) that is characterized by a bimodal age distribution with incidence peaks at 10–25 years and over 60 years and ultimately manifests as pancytopenia with a hypocellular bone marrow.^1–3 Disease severity is determined by absolute neutrophil count, platelet count, and reticulocyte count in the peripheral blood.⁴ Patients with severe aplastic anemia have a 70% mortality rate at 2 years without treatment.^4,5 Traditionally, in addition to supportive care including blood product transfusions and infection management, immunosuppressive therapy (IST) and allogeneic hematopoietic stem cell transplant (allo-HSCT) have comprised the cornerstones of treatment for severe aplastic anemia.⁶ Furthermore, aplastic anemia can be complicated by the development of paroxysmal nocturnal hemoglobinuria (PNH) and clonal evolution to myelodysplastic syndrome and acute myeloid leukemia.^7,8

For younger patients (<40 years of age) with severe aplastic anemia and an HLA-matched sibling donor available, proceeding to allo-HSCT is recommended.^1,2 However, for patients older than 40 years of age and younger patients without an HLA-matched sibling donor, IST consisting of horse antithymocyte globulin in combination with cyclosporin A and eltrombopag is the recommended standard of care. This recommendation is based on a phase 1–2 study conducted by the National Institutes of Health (NIH) in the United States (US) and the randomized phase 3 RACE trial conducted in Europe.^9,10

In the phase 3 RACE trial, eltrombopag (Novartis, Basel, Switzerland), an oral thrombopoietin receptor agonist, was confirmed to be an efficacious treatment in conjunction with traditional IST for the treatment of severe aplastic anemia in the frontline setting.⁹ Specifically, patients treated with high-dose eltrombopag and IST with horse antithymocyte globulin and cyclosporin A achieved a 68% overall response rate at 6 months, compared to 44% in patients treated with IST alone. These results corroborated the earlier NIH phase 1–2 trial that led to US Food and Drug Administration approval of eltrombopag for frontline treatment of severe aplastic anemia in 2018, following initial approval for second-line treatment after IST failure in 2014.^10–12 However, eltrombopag can pose a significant cost burden at approximately $157,000 for a 6-month course, and its cost-effectiveness relative to traditional therapy is presently unknown.¹³ Thus, we performed the first cost-effectiveness analysis of frontline eltrombopag plus IST versus IST alone in adults with newly diagnosed severe aplastic anemia in the US.

Methods

Patients and Intervention

We constructed our cohort model to reflect patients enrolled in the RACE trial; namely, patients had newly diagnosed severe or very severe aplastic anemia based upon the Camitta criteria and were ineligible for frontline allo-HSCT.⁹ Other eligibility criteria for the RACE trial included age 15 years or older without prior immunosuppressive therapy with antithymocyte globulin or other lymphocyte depleting agents. This multi-center, randomized, phase 3 trial examined the safety and efficacy of IST (horse antithymocyte globulin and cyclosporin A) with and without eltrombopag, comparing the complete hematologic response at 3 months between the two therapeutic strategies with a secondary end point of overall response at 6 months. The median age was 53 years, and 55% of patients were male. Sixty-six percent of patients had severe aplastic anemia; the remainder of patients had very severe aplastic anemia. Eighty-four percent of patients in the IST only arm in the trial who went on to receive second-line therapy received eltrombopag.

Model Construction

We employed a Markov cohort model design to compare IST alone with IST plus eltrombopag as the frontline therapies for patients with untreated severe aplastic anemia (Figure 1). In both arms, patients entered either a remission or persistent disease state after frontline therapy. Patients who achieved remission could remain in long-term remission or experience relapse. Those who had persistent disease or experienced relapse could receive additional treatments, such as allo-HSCT, eltrombopag, and danazol. Following these treatments, patients could transition to either long-term remission or stay in refractory disease over the lifetime. Patients could move to the death state at any point, due to both disease-specific mortality and sex- and age-dependent background mortality in the US.¹⁴

Figure 1. State transition diagram.

Patients with untreated severe aplastic anemia entered the model in either A) frontline eltrombopag with IST arm or B) IST alone arm. In both arms, patients transition to either a remission or persistent disease state following frontline therapy. Those who achieved remission could remain in long-term remission or experience relapse. Patients with persistent disease or relapse could receive additional treatments (e.g., HSCT, eltrombopag, and danazol). After these treatments, patients could transition to either long-term remission or stay in refractory disease throughout their lifetime. Patients could move to the death state at any point, due to both disease-specific mortality and sex- and age-dependent background mortality.

We selected a 6-month Markov cycle based on a prior cost-effectiveness analysis in patients with severe acquired aplastic anemia¹⁵ and parameterized our model in TreeAge Pro Healthcare 2025 (TreeAge Software, Williamstown, MA). We used a lifetime analytic time horizon with a 3% annual discount rate for health and cost outcomes.¹⁶ We performed our cost-effectiveness analysis from the US societal perspective across all accepted willingness-to-pay (WTP) thresholds ranging from $50,000 to $150,000 per quality-adjusted life year (QALY), with a base-case WTP of $120,000/QALY based upon the latest clinical practice guidelines.^17,18 The primary outcome was the incremental cost-effectiveness ratio (ICER) and its non-ratio reformulation, known as the incremental net monetary benefit (iNMB).¹⁸ We internally validated our model in accordance with the International Society for Pharmacoeconomics and Outcomes Research guidelines, counting the incidence of clinical events that include relapse, allo-HSCT utilization, second-line eltrombopag (or eltrombopag reintroduction) and PNH incidence, all with a comparison to RACE trial data.¹⁹ The Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline was implemented where applicable (Supplement).

Transition Probabilities

Base-case estimates for the transition probabilities utilized in our model are listed in Table 1. Probabilities for complete remission, relapse, and eltrombopag re-treatment or second-line treatment (in IST plus eltrombopag or IST alone arms, respectively) were all obtained from the RACE trial.⁹ Probabilities for allo-HSCT and danazol treatment as well as for death associated with disease were also obtained from the trial. US Life Tables were employed to inform sex- and age-specific background mortality.¹⁴ Patients with refractory disease were assumed to have an increased risk of death, modeled with a hazard ratio based on the long-term (6-year) results of a NIH phase 2 study in US patients with relapsed severe aplastic anemia relative to an age- and sex-matched general population cohort.²⁰

Table 1.

Model Clinical Parameters, Costs, and Utilities

Parameter	Base Case	Probability Distribution (Range)	Study or Data Source
Annual discount rate, %	3	Fixed	Sanders et al., 201616
Median starting age of population, years	53	Fixed	Latour et al., 20229
Mean weight, kg	70	Fixed	Patel et al., 202132
Probability of complete remission at 3 months (eltrombopag arm), %	22	Beta-PERT (17.6, 26.4)	Latour et al., 20229
Probability of overall response at 6 months (IST alone), %	41	Beta-PERT (32.8, 49.2)	Latour et al., 20229
Adjusted risk ratio of overall response at 6 months	1.71	Beta-PERT (1.33, 2.21)	Latour et al., 20229
Probability of relapse after initial response (eltrombopag arm), %	37.5	Beta-PERT (30, 45)	Latour et al., 20229
Probability of relapse after initial response (IST alone arm), %	49.5	Beta-PERT (39.6, 59.4)	Latour et al., 20229
Probability of eltrombopag retreatment as 2nd line treatment (eltrombopag arm)		Beta-PERT (−20%, +20%)	Latour et al., 20229
6–12 months	0.485
12–18 months	0.242
18–24 months	0.242
Probability of eltrombopag treatment as 2nd line treatment (IST alone arm)		Beta-PERT (−20%, +20%)	Latour et al., 20229
6–12 months	0.513
12–18 months	0.398
18–24 months	0.406
Probability of HSCT (eltrombopag arm)		Beta-PERT (−20%, 20%)	Latour et al., 20229
6–12 months	0.213
12–18 months	0.106
18–24 months	0.107
Probability of HSCT (IST alone arm)		Beta-PERT (−20%, +20%)	Latour et al., 20229
6–12 months	0.116
12–18 months	0.0901
18–24 months	0.0919
Probability of complete response after 2nd line eltrombopag, %	51.2	Beta-PERT (41, 61.4)	Ruan et al., 202045
Probability of complete response after HSCT, %	66	Beta-PERT (52.8, 79.2)	Bejanyan et al., 201946
Probability of PNH incidence in eltrombopag arm, %	1	Beta-PERT (0.8, 1.2)	Latour et al., 20229
Probability of PNH incidence in IST alone arm, %	7	Beta-PERT (5.6, 8.4)	Latour et al., 20229
Overall survival in eltrombopag arm, %			Latour et al., 20229
At 6 months	96.9	Beta-PERT (93.4, 100)
At 12 months	95.7	Beta-PERT (91.6, 99.8)
At 24 months	89.5	Beta-PERT (82.4, 96.6)
Overall survival in IST alone arm, %			Latour et al., 20229
At 6 months	93.1	Beta-PERT (88.1, 98.0)
At 12 months	88.9	Beta-PERT (82.8, 95.1)
At 24 months	85.0	Beta-PERT (77.7, 92.4)
Probability of background death, %	Varies by age	Fixed	US Life Tables (CDC)14
Mortality HR for refractory severe aplastic anemia	6.03	Beta-PERT (4.82, 7.24)	Aggarwal et al., 202320
Cost	Base Case Estimate (USD)	Probability Distribution (Range)	Study or Data Source
Eltrombopag, per month	39,300	Fixed	AWP13
Initial admission	15,222	Fixed	CMS23
Cyclosporine, per month	525	Fixed	AWP21
HSCT	323,990	Gamma (259,192, 388,788)	Hui et al., 201728
Danazol, total	1,261	Fixed	AWP22
AWP discount rate	0.28	Fixed	Abraham et al., 201447
Routine care, eltrombopag arm, per 6 months	44,458	Gamma (23,457, 67,644)	Tremblay et al., 201927; Peffault de Latour et al., 202024
Routine care, IST alone arm, per 6 months	67,328	Gamma (40,136, 97,427)	Tremblay et al., 201927; Peffault de Latour et al., 202024
Routine care, refractory disease, per 6 months	74,438	Gamma (52,728, 96,159)	Peffault de Latour et al., 202024
Routine care, remission, per 6 months		Fixed	Assumption
Year 1	187.6
Year 2–5	93.8
Year 6-	46.9
PNH treatment	82,714	Gamma (71,136, 99,703)	Cheng et al., 202129
End-of-life care for non-aplastic anemia-related death	55,620	Gamma (45,383, 64,939)	Hogan et al., 200130; Huntington et al., 201831
Hourly wage	32.66	Fixed	US Bureau of Labor Statistics33
Utilities/Disutilities	Base Case Estimate	Probability Distribution (Range)	Study or Data Source
Untreated severe aplastic anemia	0.604	Beta-PERT (0.48, 0.72)	Latour et al., 20229, EORTC QLQ-C3035
Transfusion dependent	0.6	Beta-PERT (0.52, 0.68)	Szende 200936
Transfusion independent	0.84	Beta-PERT (0.794, 0.886)	Szende 200936
Utility increment, eltrombopag arm		Beta-PERT (−20%, +20%)	Latour et al., 20229, EORTC QLQ-C3035
6 months	0.071
12 months	0.143
24 months	0.214
Utility increment, IST alone arm		Beta-PERT (−20%, +20%)	Latour et al., 20229, EORTC QLQ-C3035
6 months	0.143
12 months	0.214
24 months	0.214

Costs

The costs included in our model are summarized in Table 1. Oral drug costs, including those for eltrombopag, cyclosporine, and danazol, were based upon discounted average wholesale price, which was used to approximate Medicare average sales price-based reimbursement.^13,21,22 In particular, costs were included for both frontline eltrombopag (as well as eltrombopag re-introduction per trial supplement guidelines) as well as eltrombopag received in the second-line setting in the IST only arm. Inpatient costs associated with initial admission were based upon the Centers for Medicare & Medicaid Services inpatient prospective payment system.²³ Dosages were based upon the RACE trial.⁹ Specifically, antithymocyte globulin was administered at 40 mg/kg per day for four days, cyclosporine at 5 mg/kg per day for 12 months (then tapered for 12 months and discontinued at 24 months), and eltrombopag at 150 mg daily from day 14 through six months (or through three months for patients who had complete response at three months). Routine care costs, including those associated with hospitalizations, transfusion of blood products, and hematologist appointments, were calculated by using the healthcare resource utilization data for patients with severe aplastic anemia, derived from a retrospective chart review conducted in the US and France,²⁴ with the corresponding costs obtained from the Centers for Medicare & Medicaid Services physician fee schedule and hospital outpatient prospective payment systems.^25,26 Healthcare resource utilization data before and after eltrombopag initiation were applied to the IST only and IST with eltrombopag arms, respectively.^24,27 For patients in long-term remission, office visits were assumed to occur every 3 months for the first 2 years, every 6 months during years 3–5, and annually thereafter. Allo-HSCT and PNH treatment costs were based upon previously published literature.^27–29 The cost of end-of-life care for non-aplastic anemia-related death was based on prior work.^30,31 We assumed a mean weight of 70 kg for weight-based dosing.³² For societal costs, we accounted for productivity losses associated with healthcare resource utilization, which includes transfusions, outpatient office visits, emergency room visits, and hospitalizations.²⁴ The mean hourly wage was obtained from the US Bureau of Labor Statistics.³³ All costs were expressed in 2025 US dollars using the general Consumer Price Index.³⁴

Effectiveness

Effectiveness was measured in quality-adjusted life years (QALYs). Utility values at baseline were derived from quality-of-life data collected in the RACE trial by applying a linear transformation to patient-reported EORTC QLQ-C30 raw scores.^9,35 Utility for the remission health state was estimated using published EQ-5D values from patients with myelodysplastic syndrome, distinguishing between transfusion-dependent and transfusion-independent states.³⁶ As the RACE trial defined complete or partial response as an achievement of transfusion independence, we derived the utility for the remission health state by multiplicatively adjusting the baseline trial-specific utility to reflect the utility gain associated with transfusion independence. In a separate scenario analysis, we substituted utilities derived from EORTC QLQ-C30 patient-reported data in the RACE trial as reported at 6, 12, and 24 months. These utility values, representing a spectrum from death (0) to perfect health (1), were used to calculate QALYs, achieved by multiplying the utility value by the amount of time spent by a patient in any given health state.

Sensitivity and Scenario Analyses

To examine uncertainty in our model, we performed several sensitivity and scenario analyses. During one-way sensitivity analyses, parameters were varied 20% above and below their base case value, with all parameters affecting the iNMB at least +/− 10% presented in Figure 2. During probabilistic sensitivity analyses, we performed 10,000 second-order Monte Carlo simulations, each time randomly sampling from the distributions of model inputs (Table 1). Costs were parameterized with gamma distributions, while transition probabilities and utilities were parameterized with beta-PERT distributions. Additionally, we performed two scenario analyses. First, to see if PNH incidence—higher point estimate in the IST alone arm than with eltrombopag in the RACE trial—can impact our conclusion, we excluded the occurrence of PNH in both arms. The second scenario parameterized utilities using trial-specific utility increments, instead of accounting for transfusion-independence as defined in the RACE trial.

Figure 2. Tornado diagram.

This included parameters affecting incremental net monetary benefit (iNMB) greater than +/−10% when each parameter was varied across their full range in one-way sensitivity analysis. Blue denotes decreasing parameter value, and red denotes increasing parameter value.

Results

Model Validation

Table S1 shows the 2-year cumulative incidence of relapse, PNH, allo-HSCT, and second-line eltrombopag or eltrombopag reintroduction. Our model outputs are consistent with the RACE trial data.

Base Case, Sensitivity and Scenario Analyses

The frontline eltrombopag strategy accrued 11.53 discounted QALYs (14.43 unadjusted life years) at a discounted cost of $593,000, and the IST only strategy accrued 10.73 discounted QALYs (13.55 unadjusted life years) at a discounted cost of $561,000. This resulted in an ICER of $40,000/QALY and an iNMB expected gain of $64,000 per patient treated [95% credible interval −$91,000 to $262,000] at a WTP of $120,000/QALY (Table 2). In probabilistic sensitivity analyses, frontline eltrombopag was favored in 58%, 78%, and 82% of iterations at WTP thresholds of $50,000, $120,000, and $150,000/QALY, respectively (Figure 3). In scenario analysis excluding the increased PNH incidence point estimate with IST alone, the iNMB decreased from $64,000 to $60,000 [95% credible interval −$95,000 to $258,000], still in favor of frontline eltrombopag (i.e., $60,000 incremental net monetary benefit gained on average for every patient treated with eltrombopag). In scenario analysis employing trial-specific utilities instead of transfusion-independence, the frontline eltrombopag strategy accrued 11.56 discounted QALYs, whereas the IST alone strategy accrued 10.95 discounted QALYs, resulting in an ICER of $53,000/QALY and the iNMB of $42,000 [95% credible interval −$102,000 to $222,000] at a WTP of $120,000/QALY; again still in favor of frontline use of eltrombopag.

Table 2.

Base-Case Results

Base Case Model
Strategy	Cost (USD)	Incremental Cost (USD)	Effectiveness (QALY)	Incremental Effectiveness (QALY)	ICER (USD/QALY)	iNMB (USD) [95% credible interval]
IST alone	561,000	--	10.73	--	--	--
IST plus eltrombopag	593,000	32,000	11.53	0.80	40,000	64,000 [−91,000 to 262,000]

Abbreviations: IST, immunosuppressive therapy; ICER, incremental cost-effectiveness ratio; iNMB, incremental net monetary benefit; QALY, quality-adjusted life year; USD, US dollars

Figure 3. Cost-effectiveness acceptability curve.

At a willingness-to-pay threshold of $120,000/QALY, frontline eltrombopag was favored in 78% of the 10,000 second-order Monte Carlo iterations.

Discussion

Aplastic anemia is an acquired bone marrow failure syndrome with significant morbidity and mortality resulting in a substantial health economic burden despite the overall rarity of the disease. The recent, randomized RACE trial established the triplet combination of eltrombopag, cyclosporin A, and horse antithymocyte globulin as the new standard of care for patients with newly diagnosed severe aplastic anemia who are not planned for an upfront allo-HSCT, which constitutes the majority of patients with severe or very severe aplastic anemia. While the RACE trial did not formally compare first- versus second-line treatment with eltrombopag, it is notable that the majority of patients (>80%) in the standard IST arm received second-line treatment with eltrombopag.⁹ The use of eltrombopag in the second-line setting for patients with relapsed aplastic anemia who did not receive eltrombopag and are not eligible for allo-HSCT is also in line with current US and UK guidelines.^1,2 As such, our study addresses the cost-effectiveness of a clinically relevant scenario and is, to our knowledge, the first cost-effectiveness analysis of first- vs. second-line eltrombopag use in aplastic anemia. Our analysis demonstrated that the frontline use of eltrombopag in combination with IST may be a cost-effective strategy compared to second-line eltrombopag.

While clinically efficacious and typically only administered during the first 6 months of treatment, the addition of eltrombopag to frontline IST increases costs by $39,300 per month based on the current average wholesale price. Despite this significant increase in drug costs, triplet therapy still proved to be cost-effective in our model across all commonly used US WTP thresholds compared to a strategy that reserves eltrombopag for the second-line setting, favored in 58% to 82% of probabilistic sensitivity analysis iterations. Based upon our deterministic sensitivity analysis, a significant driver of this result, as well as of the uncertainty in our model, is the assumptions made surrounding the risk ratio for 6-month overall survival for eltrombopag plus IST versus IST alone based upon RACE trial parameters. This benefit in cost-effectiveness is also driven by the high costs of supportive care and management of complications for patients with refractory aplastic anemia as well as faster time to complete remission and lower rates of clonal evolution and development of PNH with triplet therapy. Although the latter (PNH) constitutes another very costly complication of aplastic anemia requiring long-term treatment with lifetime costs of several million US dollars, frontline eltrombopag treatment remained cost-effective even in a scenario without PNH development.³⁷ Furthermore, patients responding to frontline eltrombopag-based therapy are also able to benefit from decreasing red blood cell and platelet transfusions, minimizing lost time and improving productivity as well as from directly lowering associated healthcare utilization payer costs.^9,27

While a modest improvement in patient quality-of-life was detected in the RACE trial at 24 months in patients who received eltrombopag plus IST, longer-term patient outcome data would be helpful to clarify potentially increased health state utilities associated with the eltrombopag plus IST strategy several years after treatment initiation. With present quality-of-life data from the RACE trial, the total number of QALYs and unadjusted life years between the two strategies were similar, highlighting that the cost-effectiveness advantage with frontline eltrombopag plus IST is mostly driven by differences in supportive care costs rather than utilities. Longer-term data would also be beneficial to better characterize the impact of first-line eltrombopag treatment on the development of PNH and clonal evolution, which were only captured over a 24-month period in the RACE trial but can entail substantial costs to treat.³⁸

Important questions related to the use of first-line eltrombopag plus IST remain. Our study focused on adults with severe aplastic anemia who were ineligible for upfront allo-HSCT. Since patients <40 years of age with newly diagnosed severe aplastic anemia have excellent long-term outcomes after receipt of allo-HSCT from a matched sibling donor with long-term survival rates in young children and adolescents of >90% and >80%, respectively, upfront allo-HSCT remains the preferred option for eligible patients.^39–41 Conversely, even in the contemporary era allo-HSCT for patients >40 years of age remains associated with significant morbidity and mortality with 5-year overall survival rates of ~60% supporting the current practice of using eltrombopag plus IST in the frontline setting versus IST alone.^2,42 Whether the use of alternative graft sources such as haploidentical donors, which appear to be safe and effective and would substantially expand the donor pool, will meaningfully influence the decision-making regarding the use of allo-HSCT in aplastic anemia remains to be elucidated.⁴³ Currently, there are no studies comparing upfront allo-HSCT with eltrombopag plus IST for the treatment of aplastic anemia in any setting and as such, it is unknown whether an upfront allo-HSCT strategy or eltrombopag plus IST strategy would be the cost-effective approach for patients with severe aplastic anemia.

Our study has several strengths. First, the clinical parameters in our model are informed by data from the phase 3 RACE trial, the most recent and robust data for use of eltrombopag in adults with newly diagnosed severe aplastic anemia. Second, patient-reported outcome data were collected in the RACE trial, which we were able to incorporate into our model for health state utilities, in addition to utilities based on transfusion dependency derived from other patient populations. Both the base case and scenario analyses yielded consistent results, supporting the robustness of our findings. Third, our analysis incorporated costs associated with productivity loss informed by a published healthcare resource utilization study, enabling us to evaluate the cost-effectiveness of frontline eltrombopag from a societal perspective.²⁴ Fourth, we internally validated key clinical parameters across both treatment arms (receipt of HSCT, need for second-line eltrombopag or eltrombopag reintroduction, PNH incidence, and relapse) to accurately reflect clinical care received in the RACE trial.

Our study also has limitations. First, we drew upon data from the 197 patients enrolled in the RACE trial with median follow-up of 24 months; while a relatively small sample size and short time frame extrapolated to a lifetime horizon, this study represents the largest to date for the patient population of interest. Second, we focused our analysis on the US societal perspective while future studies will be required to understand the cost-effectiveness of eltrombopag in healthcare systems around the world. Third, we based utilities for transfusion-dependent and transfusion-independent health states off of published EQ-5D data from the myelodysplastic syndrome patient population based upon the expert opinion of the hematologists on our team; however, we conducted a scenario analysis using utility increments based upon EORTC QLQ-C30 data from the RACE trial in place of the transfusion-independent health state utility, which supported our base case analysis results. Fourth, the 24-month median follow-up period of the RACE trial is unlikely to fully capture late clonal evolution and other longer-term outcomes; thus, longer-term follow-up data will be essential to better refine cost and survival data that cane have implications for cost-effectiveness. Lastly, there is no well-established standard of care for patients with disease relapse after eltrombopag who are not eligible for allo-HSCT. For our model we included danazol as salvage treatment. Other treatment options including alemtuzumab are sometimes employed and are not included here.²⁰ However, only three patients in the RACE trial received alternative immunosuppressive agents, which limits the impact of these treatments on the overall results.⁹

In conclusion, at current pricing and based upon the RACE trial patient population, we found first-line eltrombopag plus IST to be a cost-effective strategy in US adults with newly diagnosed severe aplastic anemia who are not candidates for upfront transplantation. This finding expands the literature and helps inform the evidence to decision framework in upcoming American Society of Hematology guidelines on the management of people diagnosed with severe aplastic anemia.

Key points.

• Frontline eltrombopag plus immunosuppressive therapy (IST) is cost-effective compared with IST alone for adults with severe aplastic anemia

• Incremental net monetary benefit per patient treated with frontline eltrombopag is, on average, $64,000

Data sharing:

Data used in this study are from publicly sourced research, are outlined within the article, and available on request from the corresponding authors, Jan P. Bewersdorf (jan.bewersdorf@yale.edu) and George Goshua (george.goshua@yale.edu).