The cost‐effectiveness of isavuconazole compared to voriconazole, the standard of care in the treatment of patients with invasive mould diseases, prior to differential pathogen diagnosis in Spain

Abstract Background Invasive mould diseases are associated with high morbidity, mortality and economic impact. Its treatment is often started prior to differential pathogen diagnosis. Isavuconazole is approved for treatment of invasive aspergillosis (IA) and invasive mucormycosis (IM) when amphotericin‐B is not indicated. Objectives To estimate the cost‐effectiveness of isavuconazole vs voriconazole for the treatment of adult patients with possible IA prior to differential pathogen diagnosis, in Spain. Methods A decision tree analysis was performed using the Spanish Healthcare System perspective. Among all patients with possible IA, it was considered that 7.81% actually had IM. Costs for laboratory analysis, management of adverse events, hospitalisation and drugs per patient, deaths and long‐term effects in life years (LYs) and quality‐adjusted LYs (QALYs) were considered. Efficacy data were obtained from clinical trials and utilities from the literature. Deterministic and probabilistic sensitivity analyses (PSA) were conducted. Results In patients with possible IA and when compared to voricanozole, isavuconazole showed an incremental cost of 4758.53€, besides an incremental effectiveness of +0.49 LYs and +0.41 QALYs per patient. The Incremental Cost Effectiveness Ratio was 9622.52€ per LY gained and 11,734.79€ per QALY gained. The higher cost of isavuconazole was due to drug acquisition. Main parameters influencing results were mortality, treatment duration and hospitalisation days. The PSA results showed that isavuconazole has a probability of being cost‐effective of 67.34%, being dominant in 24.00% of cases. Conclusions Isavuconazole is a cost‐effective treatment compared to voriconazole for patients with possible IA for a willingness to pay threshold of 25,000€ per additional QALY.


| INTRODUC TI ON
Invasive mould diseases (IMD) are life-threatening infections, especially in immunocompromised patients such as those with haematologic diseases, those who have been subjected to a solid organ transplantation or haematopoietic stem cell transplantation (HSCT) and in critically ill patients with leukaemia or profound neutropaenia. 1,2 IMDs are caused by fungus of the genus Aspergillus (causing invasive aspergillosis, IA) or other filamentous fungi, for example, Mucorales (causing invasive mucormycosis, IM). These infections entail an important clinical burden to individuals who are already vulnerable as they are associated with high morbidity and mortality, with mortality rates ranging from 30% to 80% for IA and up to 97% when untreated IM. 3,4 Similarly, a delay in the establishment of an adequate therapy leads to increased mortality rates in certain type of patients and significant increase in treatment duration. 5 For instance, according to a study analysed by the FDA, a delay of 6 days in the initiation of treatment, increased mortality rates up to almost two times, from 48.6% to 82.9%. 6 IMDs have a high economic impact as well. 1,7,8 In line with available data in the website of the Spanish Ministry of Health RAE-CMBD (Registro Atención Especializada-Conjunto Mínimo Básico de Datos), 9 length of stay for IA in Spain in 2017 was 54.20 days and the mean associated costs were 18,235€; regarding IM the length of stay was 70 days and the mean costs were 24,020€.
Even though it is highly important to establish an early diagnosis for this pathology, the process is challenging because often non-specific symptoms are present. According to the recently published update of the consensus definition of invasive fungal disease (IFD) from the EORTC and the Mycoses Study Group Education and Research Consortium, probable IFD requires the presence of a host factor, a clinical feature and mycologic evidence; cases that meet the criteria for a host factor and a clinical feature but for which mycological evidence has not been found are considered possible IMD. 10 In contrast, obtaining a diagnosis of mucormycosis on histomorphological basis is challenging, and the most common cause for incorrect morphological diagnosis is the misidentification of Mucorales as Aspergillus spp. The application of immunohistochemistry or PCR techniques on either fresh or formalin-fixed paraffin-embedded tissue have been shown to be highly specific, although a variation in sensitivity has been reported. 11 The outcome and the management of the patient depend on a prompt and correct diagnosis and usually treatment is initiated before pathogen identification. 3 However, if the treatment turns out to be inappropriate, a delay in receiving the right treatment can increase the mortality rate, as reported by Chamilos et al where a 6-day delay was associated with a 2-fold increase in mortality rate. 12 Among the therapeutic options currently available to treat IA and IM, only isavuconazole and amphotericin-B (L-AMB) have approved indication for both types of infections. Voriconazole and posaconazole are exclusively indicated for the treatment of IA, particularly posaconazole is only indicated when the disease is refractory to L-AMB.
According to published therapeutic guidelines, isavuconazole is recommended as a rescue therapy in IM patients, as well as posaconazole, although the recommendation for posaconazole is off-label. 11 The SECURE pivotal trial 13 has shown isavuconazole to have non-inferior efficacy and survival to voriconazole, and better safety profile, in patients with IA. On the other hand, the VITAL trial has demonstrated the efficacy of isavuconazole as primary therapy, in refractory patients and salvage, in the treatment of IM. 14 Isavuconazole, on top of its broad spectrum, being active against Aspergillus and Mucorales, which is a clear advantage when treating patients with possible IA (thus, lacking pathogen confirmation), offers several additional advantages relative to voriconazole, that entail benefits to the patients and could carry economic consequences as well; it has lower drug-drug interactions; therefore, it can be more safely given to patients receiving additional medications; isavuconazole does not need routinely therapeutic drug monitoring, as it is the case for voriconazole; patients receiving it show fewer adverse events, and dose adjustment is not needed for patients with mild/ moderate hepatic impairment or renal impairment. 15,16 Several economic analyses of isavuconazole vs voriconazole for the treatment of IA have been conducted in different countries: in hospitalised patients with IA in the United States 17 and in patients with possible IA in Sweden 18 and the UK. 19 Cost-minimisation models of isavuconazole vs L-AMB followed by posaconazole for the treatment of IM have been also performed in Italy, 20 Germany 21 and the UK. 22 The clinical situation described in the UK cost-effectiveness model 19 clearly reflects the real-world treatment approach in Spain when facing patients with possible IA, in which is necessary to start treatment before getting the differential diagnosis between IA and IM. Although isavuconazole is more expensive than voriconazole, it is effective for both IA and IM and has a better safety profile. Therefore, the objective of the present study was to conduct a health economic analysis from the Spanish National Health Service perspective (NHS), by estimating the cost-effectiveness of isavuconazole vs voriconazole, the standard of care in Spain, for the initial treatment of IMD prior to differential diagnosis between IA and IM.

| Economic model
A cost-effectiveness and cost-utility analysis, with a lifetime horizon and from the perspective of the NHS, was adapted to Spain. 19 In order to reflect the short-term patient pathway from initial symptoms to eventual result after antifungal treatment (resolution of infection or cost-effectiveness, decision tree, invasive aspergillosis, isavuconazole, mucormycosis death), a decision tree originally developed with the collaboration of UK expert's panel was analysed by local experts (1 hospital pharmacist, 1 clinical pharmacologist and 1 haematologist, from three different autonomous communities of Spain) and considered a representative approach of Spanish patients' pathology pathways ( Figure 1). First, the model reflects that a population of patients with 'possible IA' can either be treated with isavuconazole or voriconazole. Secondly, patients are subdivided into having IA or IM, regardless of a clinical confirmation. In this step, a prevalence of 7.81% was used to estimate those patients incorrectly diagnosed with possible IA but actually having IM instead (tree branches with blue dots in Figure 1). Due to a lack of specific data from Spain, this prevalence was estimated from a UK study. 23 Then, depending on treatment response, discontinuations or di-

| Second-line treatment
The percentage of patients receiving second-line treatment was calculated using data from the SECURE and VITAL trials by considering the number of patients that discontinued first-line treatment due to no or insufficient response to treatment, adverse events or F I G U R E 1 Decision tree model structure. IA, Invasive aspergillosis; IM, Mucormycosis; L-AMB, Liposomal amphotericin-B intercurrent illness. 13,14 Since no statistically differences were found between isavuconazole and voriconazole, these percentages were equally applied for both therapies (Table 1). Accordingly, 27.13% of IA patients and 14.29% of IM patients treated with isavuconazole will switch to a second-line treatment.
In case of pathogen confirmation, 100% of IM patients treated with voriconazole will switch to a second-line treatment.

| All-cause mortality
All-cause mortality over 84-day from the SECURE trial 13 (Table 1) was used to approximate mortality rate in IA patients, regardless of the initial treatment (isavuconazole or voriconazole) since no statistical differences were reported in the trial. In order to avoid double counting in the number of deaths, this mortality was also assumed for patients with IA that switch to L-AMB.
The 84-day all-cause mortality in IM patients treated with isavuconazole was extracted from the VITAL study 14 (Table 1). According to published data, 12 an increased mortality probability due to a delay in the correct treatment was applied to IM patients treated with voriconazole that switch to L-AMB. On the other hand, patients lacking pathogen confirmation were assumed to have the same mortality as untreated patients 6 ( Table 1).

| Dosing
Dosing of all treatments in the model was aligned to their corresponding SmPC (refs FT) and, for L-AMB, the FungiScope ™ matched control study 14 to reflect real-world dosing ( Table 2) The model considers that, according to VITAL study data, 14 75% patients will start with IV treatment and subsequently step-down to oral therapy, while the remaining 25% will directly start with oral therapy.
Second-line treatment consists of IV L-AMB followed by oral voriconazole/posaconazole at a 50%/50% ratio, except for IM patients treated with voriconazole that switch to L-AMB: in this case, only oral posaconazole is considered. In IA, oral voriconazole was included as step-down therapy after L-AMB as fungal load is significantly reduced after therapy with L-AMB and thus voriconazole is efficacious again (expert panel input).

| Duration
Treatment durations are summarised in Figure 1 and Table 1. For IA patients, it was calculated by adjusting isavuconazole treatment duration in the SECURE study (47.0 days in total: 8.1 days IV, 38.9 days oral) 28 according to whether patients responded to, and remained on, first-line treatment or discontinued treatment and switched to second-line therapy. Since no statistical differences among therapies were found, a total treatment duration of 70.70 days for IA was assumed for both therapies. As previously published, the duration of second-line therapy was assumed to be equal to the duration of first-line therapy in patients who responded to treatment (70.70 days). 29 Pts moving to 2nd line 13,14 All-cause mortality 6,12,13 Adjusted treatment duration (d) 13

| Costs
Only direct costs were included in the model, namely: drug acquisition, hospitalisations, adverse events (AEs) and laboratory analysis.
All costs were expressed as 2020 Euros.

| Drug acquisition costs
Acquisition costs (Table 2) were estimated using the ex-factory price (EFP) 32 including the corresponding mandatory deduction outlined in Royal Decree 08/2010 33 and the reference price 34,35 established or the lowest EFP among the available.

| Hospitalisations
The cost per hospital day used in the model was 568.48€. This was calculated from Spanish health system reference costs as the mean of the latest published prices for hospital stay in every autonomous community. 36 Length of hospitalisation stay for IA patients was extracted from the SECURE study. 13,28 Since no statically differences were found between isavuconazole and voriconazole, isavuconazole's mean hospitalisation length stay, 18.60 days, was used for both therapies.
This duration was also assumed to second-line therapy.
IM patients with a full course treatment of either isavuconazole or voriconazole were assumed to have the same hospital stay length as reported in the VITAL study, 14

| Adverse events
The model includes adverse events reported in the SECURE study 15 for which statistically significant differences between isavuconazole and voriconazole treatments were found (see Table 3). For the treatment with L-AMB, unlike the UK original model, nephrotoxicity 37 (associated with 9 days of additional hospital treatment 29 ) and hypokalaemia 38 were also considered. All the costs of adverse events were calculated from Spanish health system reference costs as the mean of the latest published prices for related specialist outpatient visit in autonomous communities. 36 Due to a lack of specific information, hypokalaemia cost was assumed to correspond to a nephrology specialist visit cost (206€). 36 Importantly, the model adjusted the average calculated AE costs by the duration of treatment for every patient subgroup.

| Laboratory tests
Laboratory monitorisation tests were estimated taking into account length of treatment, adverse events and precautions included in the summary of product characteristics of each drug 26 (

| Utility and life expectancy
Utility and life expectancy were defined using the most frequent underlying condition of IMD in the SECURE and VITAL trials, which is acute myeloid leukaemia. Therefore, utility was assumed to be 0.82, as reported by Leunis A et al 40

| Cost-effectiveness outcomes
Results are presented through the main cost-effectiveness metric which is the incremental cost-effectiveness and cost-utility ratio (ICER/ICUR). To calculate it, both long-term effects in life years (LYs) and quality-adjusted life years (QALYs) were reported.
Costs and outcomes were discounted using a 3.00% discount rate. 42 Results of the analysis were checked using a willingness to pay (WTP) threshold of 25,000 Euros per additional QALY gained. 43 Thus, the evaluated strategy is considered cost-effective when the ICUR < 25,000 €/QALY.

| Sensitivity analysis
Sensitivity analyses were performed in order to evaluate the robustness of the model and analyse the uncertainty around the parameter's estimates and assumptions.  The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to.

| Probabilistic sensitivity analysis
No ethical approval was required as the research in this article related to an economic model based on data from published literature.

| Base case
In the situation considered in this economic model, which is the treatment of IMD prior to differential diagnosis between IA and IM,

| Deterministic sensitivity analysis
Values for 122 parameters were tested in the deterministic sensitivity analysis. As shown in Figure 2, main parameters influencing results were mortality, treatment duration and number of hospitalisation days. It was seen that results were especially sensitive to a reduction in mortality of IA patients treated with isavuconazole in a 20%, or an increase of this same input for voriconazole in the same percentage; the ICER was decreased to 2,600€/QALY. Additionally, treatment length of IA patients and length of hospital stay did have also important influence in the results. An increase in the number of hospital days for isavuconazole by 20% or a decrease for voriconazole, doubled the ICERs.

| Probabilistic sensitivity analysis
In the probabilistic sensitivity analysis, the average ICUR was compared with a WPT of 25, 000€ per QALY. 43 Results showed that 65.36% of simulations (that represent the differences in costs and effects between comparators) were in the top right quadrant, indicating that isavuconazole is more efficient and more expensive in most of the cases. Besides, 24.00% of the simulations were dominant meaning more health gains and less expensive (Figure 3).
Data were represented in a cost-effectiveness acceptability curve (CEAC) in order to estimate the probability of isavuconazole of being cost-effective for a range of WPT ranging from 0 to 100,000€/QALY (Figure 4). According to these results, isavuconazole has a probability of 67.34% of being cost-effective for a WPT of 25, 000€/QALY.

| DISCUSS ION
Invasive mould diseases are very severe infections that have a high morbidity and mortality, entailing high associated costs. The early differential diagnosis between IA and IM is very challenging, so in real life setting usually treatment is initiated before pathogen identification, because a delayed treatment initiation can severely influence the outcomes. 5 Under these circumstances, isavuconazole and voriconazole are two main therapeutic options that can be chosen by the physician, taking into account both clinical and economic factors, however, only isavuconazole is effective and indicated for treating both IA and IM. 13,14,26 Taking into consideration this context, this modelling study was undertaken to compare, from the Spanish NHS perspective, the cost-effectiveness of the use of isavuconazole vs the standard of care, voriconazole, when at the time of therapy initiation, the differential diagnosis between IA and IM has not been achieved. the decision tree and the numbers for mortality and treatments. The incidence of IM is low and there is a lack of data in the long-term consequences of antimycotic treatments, thus assumptions had to be made to overcome this lack of real data in the area.

F I G U R E 2
Another limitation of this cost-effective analysis would be that the clinical parameters considered for IA patients were the same for both treatment arms (isavuconazole and voriconazole). Thus, the differences in QALYs and LYs are mainly due to the fact that voriconazole is not effective in treating IM, resulting in high mortality rates for these patients because there is a delay in effective therapy initiation (at the time the pathogen is identified); together with an improvement in IA outcomes such as reduction in mortality, laboratory costs, adverse events and hospitalisations.

| CON CLUS IONS
In conclusion, isavuconazole is cost-effective over voriconazole for the treatment of patients with possible IA. Considering the demonstrated efficacy of isavuconazole against Aspergillus spp.
and Mucorales spp and its additional clinical benefits, these results strongly suggest that using isavuconazole instead of voriconazole could provide clinical and economic advantages for IA patients without pathogen identification at the moment of therapy initiation and thus being beneficial for patients and for the Spanish NHS.