Cost-effectiveness of everolimus plus reduced tacrolimus in de novo liver-recipients in the Italian setting

Abstract

Introduction: Long-term exposure to calcineurin inhibitor-based immunosuppressant (IS) therapy in liver transplant (LT) recipients is associated with renal complications. In the randomized trial H2304, everolimus + reduced-dose tacrolimus (EVR + rTAC) demonstrated equivalent efficacy and superior renal function compared to standard-dose tacrolimus.

Methods: To evaluate the cost-effectiveness of EVR + rTAC vs TAC, in de novo LT patients, a Markov model simulating both liver and kidney function was developed and estimated the long-term outcomes of IS following LT. The analysis used the Italian healthcare payer perspective.

Results: Patients treated with EVR + rTAC gained on average 1.92 years and 1.62 quality-adjusted life years (QALYs). The incremental cost-effectiveness ratios (ICER) were €35,851 and €42,567 for LY gained and QALY gained, respectively. For the hepatitis-c sub-population, the ICERs decreased to €22,519 and €30,658, respectively.

Conclusion: EVR + rTAC improves survival and quality-of-life and is a cost-effective alternative to calcineurin-inhibitor monotherapy for patients requiring LT.

Keywords:

• Cost-effectiveness
• Post-liver transplantation
• Everolimus
• Immunosuppressant therapies

Introduction

Liver transplantation (LT) is a widely accepted form of therapy for patients suffering from acute liver failure, end-stage liver disease, and hepatic malignancies¹. In Europe, the primary indications for LT have changed over time. In contrast to the 1980s when hepatic malignancies were a leading cause for transplantation (50%), liver cirrhosis caused by viral hepatitis and chronic alcoholism are currently the most common indications, accounting for 23% and 20% of the liver transplants, while primary hepatic malignancies represent only 16% of all liver transplantation cases². Worldwide, the number of patients receiving LT has also steadily increased over time. The European Liver Transplant Registry (ELTR; 2012) estimates that 5800 LTs are performed annually in Europe². In 2010, the rate of LT in Italy was ∼16.69 per million population³.

Post-transplantation graft and patient survival rates have improved significantly following advances in surgical procedures and with the introduction of calcineurin inhibitors (CNI). In Europe, 1-year patient survival rates increased from only 33% before 1985 to 76% in 1990–1994 and 85% after 2004². Despite these improvements in survival rates, liver transplant recipients are still at risk from a number of complications related to surgery, disease recurrence (e.g. hepatitis-c virus [HCV] progression), and long-term use of immunosuppression (IS) regimens^4–9.

CNIs have been the cornerstone of IS therapy over the last 30 years, improving post-LT patient and graft survival and reducing the incidence of acute rejection (AR) episodes. However, prolonged exposure to CNIs is associated with several complications such as nephrotoxicity that can lead to chronic kidney disease (CKD), neurotoxicity, malignancies, metabolic complications, and hypertension. Studies have shown CKD to affect cumulatively ∼25% of liver transplant recipients over the first 10 years of IS therapy^6–11.

Thus, the current focus of post-LT care involves developing treatment strategies that help reduce the long-term complications caused by immunosuppressant strategies while effectively preventing organ rejection and graft loss^9,10,12–15.

Certican (everolimus; EVR) is a proliferation signal inhibitor (also known as mammalian target of rapamycin (mTOR) inhibitor)) that provides effective immunosuppression while protecting against renal dysfunction when used in combination with reduced doses of CNI. In the early 2000s, EVR, in combination with cyclosporine and steroids, was approved in many European countries for both kidney and heart transplants.

Efficacy and safety of EVR in combination with reduced-dose tacrolimus (TAC) was compared to standard-exposure TAC in a 24-month, multi-center, open-label randomized, controlled study (H2304)^16–19. This trial included de novo liver transplant recipients and was designed to assess the maintenance of renal function by allowing for the reduction or elimination of TAC early post-transplant. The extension study (H2304E1)²⁰ to the aforementioned randomized controlled trial collected data up to 36 months post-transplant. These trials concluded that patients receiving EVR in combination with reduced-dose TAC (EVR + rTAC) demonstrated non-inferior efficacy and superior renal function compared to those who received the standard-dose TAC. Furthermore, the improvements in renal function were sustained at month 24. In the extension population a similar effect from month 24 to month 36 was observed.

Based on this evidence, the use of EVR in combination with reduced-dose TAC and steroids in adult LT patients was approved by the European health authorities in 2012 and the US Food and Drug Administration (FDA) in 2013^9,21.

The objective of this study is to explore the cost-effectiveness of EVR + reduced-dose TAC compared to standard-dose TAC alone in de novo liver-recipients in Italy. To our knowledge, there is little evidence on the long-term economic impact of IS therapy in post-LT patients.

Methods

Model structure

A Markov state-transition model was developed in Microsoft Excel 2007 (Microsoft Corporation, Redmond, WA) to estimate the long-term costs and effects associated with IS following LT. Based on the design of the pivotal EVR Phase III trial (study H2304), the target population consisted of a representative group of adults aged ≥18 years who received a primary allograft from a deceased donor.

The Markov model was designed to simulate the development of post-transplant complications along two pathways: (1) liver-related: leading to graft lost and (2) kidney-related: culminating in ESRD. All patients entered the model at day of transplant and, following a run-in period of 30 days, were randomized to receive EVR + reduced-dose TAC or standard-dose TAC. All patients followed both the liver and kidney pathways (Figure 1). This structure allowed all possible combinations of liver status and kidney function; which were assumed to be independent. The liver pathway was considered to be the main model pathway. Costs and utilities (or decrements) were assigned to each health state. For those experiencing renal complications, incremental costs and utility decrements were attributed and applied to those accrued in the liver pathway. The only connection between the pathways was the death status as patients dying from renal complications had to exit also the liver pathway and vice versa. Patients in need of a second LT following graft loss were absorbed into the post-LT health state and did not re-enter the model. All health states were associated with a risk of death.

Figure 1. Liver transplantation model structure.

The cycle length was annual, with the exception of the first year post-LT that was split into three cycles (0–3, 3–6, and 6–12 months) to allow the model to report the early complications post-LT. Half cycle correction was incorporated to compensate for the fact that in Markov models transitions are assumed to occur at the beginning or end of the cycle. In the clinical setting, patients could transition between health states at any given time, thus the half cycle correction was applied from year 2, as the three monthly cycles in year 1 were deemed sufficient to capture the transition of patients.

The analysis was based on the Italian healthcare payer perspective and included direct healthcare costs only. A lifetime horizon was considered; however, the model was built with the flexibility to allow for results over shorter time scales (i.e. 1, 2, 3, 5, 10, 20, and 30 years).

The model structure and assumptions were reviewed and validated by clinical experts.

Model inputs

Efficacy

Most efficacy data were obtained from the H2304 study and, wherever possible, estimates from published literature were used to populate the model. Data on patient response to the treatment of acute rejection (AR) by level of severity were obtained from consultations with clinical experts.

The risks of AR, change in renal function, and hepatitis C (HCV) progression were specific to treatment. For AR, odds ratios (OR) for each treatment regimen were calculated based on the analysis of the H2304 and the H2304E1 patient level data (PLD).

Post-LT, HCV recurrence was assumed to occur in the first 3 months and disease progression was characterized based on the following three transitions post-LT: (1) mild-to-moderate HCV, (2) moderate HCV to compensated cirrhosis, and (3) compensated cirrhosis to decompensated cirrhosis. Data on the potential effect of the EVR + rTAC regimen and the standard-dose TAC regimen on HCV progression were obtained from the H2304 clinical trial.

The H2304 PLD were used to derive the transition probabilities between CKD stages for EVR + rTAC and standard-dose TAC alone arms for the first 3 years. The ORs were then assessed for each model cycle.

Treatment-emergent adverse events (TEAEs), associated with the two IS regimens, were also incorporated in the model. The model assumed TEAEs (with the exception of renal dysfunction) did not occur after year 3 due to the lack of clinical data. In the absence of long-term data, the analysis assumed that death due to graft loss was not treatment-specific.

Transition probabilities

The early transition probabilities (TPs; i.e. for the first 3 years) associated with hepatic complications were derived using the H2304 and the H2304E1 PLD and supplemented with the published literature²². As data on the risk of AR among patients experiencing liver complications such as HCV fibrosis and cirrhosis were not available, conservative assumptions were made following consultations with the clinical experts.

In terms of the kidney pathway, TPs within the 36 months following liver transplantation were derived from H2304 and the H2304E1 PLD. For subsequent years of the model, TPs related to progressive kidney failure were derived from UNOS database analysis²³. The TP for EVR + reduced TAC were calculated using the OR derived from the trial data. However, as no patient experienced CKD stage 5 within the H2304 and the H2304E1 timeframe, the annual probability of patients requiring a renal transplant was assumed to be 0.02%.

Resource use and costs

A focused search of the literature was conducted to identify Italian costs associated with the liver and renal pathways. This was supplemented by local expert opinion.

Drug costs for all of the treatment strategies considered in the model were obtained from FARMADATI ITALIA (2013; Table 1).

Table 1. Unit cost per milligram.

	Price per mg (€)	Source
Everolimus	€12.65	Farmadati Italia 2013^26
Tacrolimus, CAP	€1.80	Farmadati Italia 2013^26

CAP, Capsule.

Costs associated with monitoring following transplant included outpatient and inpatient care, and laboratory tests. Resource use estimates were based on clinical opinion.

The costs associated with each complication (Table 2 [t]insert Table 2 near here[/t]) were obtained from the literature, and these costs were independent of the monitoring costs²⁴. All costs were adjusted to 2014 prices using the Italian ISTAT (The National Institute of Statistics) consumer price index for medical care.

Table 2. Estimated annual health states costs (EUR; €).

Health state	Annual cost estimate		Source
	Original	Inflated 2013
Liver pathway
Recurrent Hepatitis C	€78.55	€87.33	Ruggeri et al.^24
HCV on anti-viral therapy	€14,129.18	€14,957.59	Camma et al.^28
Hepatitis C (F1–F2)	€110.19	€110.19	Assumption: half the cost of F3–F4
Hepatitis C (F3–F4)	€220.37	€233.29	Camma et al.^28
Compensated cirrhosis (F5–F6)	€479.07	€507.16	Camma et al.^28
Decompensated cirrhosis	€4994.28	€5287.10	Camma et al.^28
Recurrent HCC	€6438.79	€6816.30	Camma et al.^28
Acute rejection
Mild	€5674.93	€6803.67	Assumption: 60% of the cost of moderate AR
Moderate	€9458.22	€11,339.45	Assumption: half the cost of severe AR
Severe	€18,916.43	€22,678.91	Rufat et al.^29; costs converted from GBP (£12 595) to Euros using an exchange rate of 1.5019
Graft loss leading to re-transplant	€90.733.60	€96,053.43	Camma et al.^28
Post-acute rejection	€50.00	€50.00	ASSUMPTION
Post-liver transplant	€5923.46	€6270.76	Camma et al.^28
Kidney pathway
CKD: stage 1	—	—	Assumption
CKD: stage 2	€450.00	€498.66	Cicchetti et al.^30; Stage 2
CKD: stage 3a	€890.00	€986.25	Cicchetti et al.^30; Stage 3
CKD: stage 3b	€2141.00	€2372.53	Assumption: Average between stage 3a/4
CKD: stage 4	€3392.00	€3758.82	Cicchetti et al.^30; Stage 4
CKD: stage 5	€52,830.00	€58,543.20	Cicchetti et al.^30; Stage 5
Renal transplantation	€47,173.89	€49,939.76	DRG tariff 2011 Accordo Interrergionale per lacompensazione della mobilita sanitaria
Post-renal transplant	€6283.00	€6962.46	Villa et al.^31

CKD, Chronic kidney disease; HCC, Hepatocellular carcinoma.

Utilities

The baseline utility for the study population is the value associated with post-LT without major complications. Quality-of-life data were not collected as part of H2304, thus utility values were based on baseline utility estimates of EQ-5D scores from Russell et al.²⁵. The estimates were derived through calibration of the model utilities to match the values reported in Russell et al. (early post-transplant =0.746; intermediate post-transplant =0.765; late post-transplant =0.817), as described in Table 3 ²⁵. Quality-of-life is affected each time patients experience complications (hepatic and/or renal). Thus, the utilities associated with the liver health states were combined with the baseline utilities using the multiplicative approach and each kidney complication and adverse event were then applied as decrements using a subtractive approach. These values were taken from the literature.

Table 3. Health state utilities post-liver transplantation.

Health state	Utility	Method	Source
Post-liver transplant without major complication			Multiplicative factor for post-liver transplant condition
Early post-transplant (<6 months)	0.83	NA	Assumption: parameter of calibration
Intermediate post-transplant (7–36 months)	0.93	NA	Assumption: parameter of calibration
Late post-transplant (>36 months)	0.92	NA	Assumption: parameter of calibration
Liver pathway
HCV (F0–F2)	0.82	HUI3	Sullivan et al.^32
HCV (F3–F4)	0.80	NA	Assumption: lower than F0-F2
HCV anti-viral therapy	0.77	NA	Assumption; decrement of 0.05 due to adverse events
Compensated cirrhosis	0.78	HUI3	Sullivan et al.^32
Decompensated cirrhosis	0.65	NA	Sullivan et al.^32
HCC	0.25	HUI3	Sullivan et al.^32
Re-transplantation	0.50	HUI3	Sullivan et al.^32
Post-2nd liver transplant	0.70	HUI3	Sullivan et al.^32
Acute rejection
Mild	0.90	Other	Das^33; mean estimate
Moderate	0.85	Other	Das^33; lower limit
Severe	0.70	Other	Das^33; upper limit
Post-acute rejection	0.95	NA	Assumption
Kidney pathway
Reduced KF - Stage 1 & 2	1.00	NA	Assumption: no decrement in utility
Reduced KF - Stage 3a	0.87	TTO	Gorodetskaya et al.^34
Reduced KF - Stage 3b	0.86	TTO	Assumption: utility value between stage 3 and stage 4
Reduced KF - Stage 4	0.85	TTO	Gorodetskaya et al.^34
Reduced KF - Stage 5 (with dialysis)	0.77	TTO	Gorodetskaya et al.^34
Renal transplant	0.70	NA	McLaughlin et al.^35
Post-renal transplant	0.74	TTO	Laupacis et al.^36

TTO, time trade off; NA, Not applicable.

Model outcomes

The main outcome of consideration was the incremental cost-effectiveness ratio (ICER). The analyses also considered other effectiveness outcomes such as life years (LYs) and quality-adjusted life years (QALY). The ICER is measured by comparing the value to a threshold level of willingness to pay (WTP). According to the AIES (Italian Association of Health Economics) a technology within the range of €25,000–€40,000 per QALY is considered to be valuable²⁷. In addition, clinical outcomes such as the number of AR and graft losses avoided were also examined, as well as the economic costs associated with renal and liver complications.

Sensitivity analyses

In order to assess the uncertainty in the results, deterministic and probabilistic sensitivity analyses were carried out. In the deterministic sensitivity analysis (DSA) the following inputs were varied: transition probabilities (± 25%), utilities (± 25%), health state costs (± 20%), drug costs (± 20%), AE costs (± 20%), and discounting (3%). In terms of the efficacy parameters, the proportion of patients treated with HCV was decreased from 50% in the base case to 25% in the DSA. For those patients who experienced a mild acute rejection, no treatment of these patients was also investigated. In the clinical study report two methods were used to estimate the OR associated HCV progression on reduced dose tacrolimus. In the base case a value of 0.37 was applied and the DSA examined an alternative value of 0.68.

In the probabilistic sensitivity analysis (PSA) TPs, utilities, health state costs, drug costs, and AE costs were varied. A beta distribution was applied to the TPs and health state utilities, a log-normal for AE utilities, and a gamma distribution for all cost parameters.

Sub-group analyses were also conducted. The following patient groups were investigated based on liver disease background and renal function at baseline: HCV, hepatocellular carcinoma (HCC), CKD stage 1–2, 3a–3b, and 4–5.

Results

Base case

The model predicts that, with a mean life expectancy of 18 years, patients treated with EVR + reduced TAC would gain 1.92 LY and 1.62 QALY vs TAC alone (Table 4 near here[/t]), with a cost per LY gained of €35,851 and the cost per QALY gained close to the upper bound of the AIES threshold (€42,567/QALY). Pharmacological costs represented the major share of the total costs for post-LT patients. The higher IS cost of EVR + reduced TAC was slightly offset by lower management costs. Costs associated with renal complications were more favorable in the EVR arm, accounting for 15% of the total costs compared to 21% in the TAC arm; this was explained by the lower number of cases of severe CKD (stage 3–5) in the EVR arm. EVR + rTAC was also associated with a considerable reduction in the frequency of acute rejections (20%), resulting in a slight cost offset in the EVR arm.

Table 4. Lifetime base case results.

	EVR + reduced TAC	TAC alone	Incremental
Outcomes
LY	18.16	16.24	1.92
QALY	14.74	13.12	1.62
Acute rejections	2.321	2912	−591
Costs
Total cost	€274,422	€205,432	€68,989
IS cost	€198,975	€129,031	€69,944
Renal complications	€41,938	€43,619	−€1681
Liver complications	€4460	€5513	−€1054
Monitoring and AE	€26,099	€24,244	€1855
ICER
Cost per LY gained			€35,851
Cost per QALY gained			€42,567

ICER, Incremental cost effectiveness ratio; LY, Life Years; QALY, Quality Adjusted Life Years; IS, Immunosuppressant.

The EVR regimen demonstrated that it was a favorable strategy in the sub-group of patients with liver complications at baseline. For the HCV sub-group, the incremental cost per LY and QALY gained decreased to €22,519 and €30,658, respectively (Figure 2).

Figure 2. Sub-group analyses results (lifetime horizon).

Sensitivity analyses

The tornado diagram in Figure 3 illustrates the difference from the base case in the lifetime cost per QALY observed for a series of sensitivity analyses performed on key parameters. The results of the DSA demonstrated that the most sensitive parameters were the price of EVR, the use of generic TAC, discount rate, the utilities associated with baseline and CKD stages (1–3a), as well as the effect of reduced-dose TAC on HCV progression.

Figure 3. Tornado diagram: cost per QALY gained.

Probabilistic sensitivity analysis showed that the EVR strategy remained costlier, but also more effective than TAC alone across the simulations. The cost-effectiveness acceptability curve (Figure 4) shows a 77% probability of ICERs remaining below the €50,000 threshold. The probability of not exceeding the threshold of €40,000 was 39%.

Figure 4. Cost-effectiveness acceptability curve.

Discussion

The long-term use of CNI-based IS regimens is associated with several complications that have a significant impact on patient and graft survival^2,4,6–10. Thus, there is a need to develop treatment strategies that can reduce CNI-induced complications while effectively preventing organ rejection and graft loss^9,10,12–15.

EVR is an alternative IS agent that was approved by the FDA and the European health authorities for LT^9,21. EVR being a proliferation signal inhibitor (mTOR inhibitor) has a different pharmacological profile to the CNIs and, hence, provides effective immunosuppression with less CNI-related adverse effects if combined with reduced exposure to CNIs. Phase III trials have shown that patients receiving EVR with reduced dose TAC experienced non-inferior efficacy and superior renal function compared to those who received the standard-dose TAC^16,17,20.

Furthermore, our cost-effectiveness analysis has shown that the EVR with reduced dose TAC regimen is associated with gains in LY (1.92) and QALY (1.62) years vs standard-dose TAC. Everolimus with reduced dose TAC was associated with substantially lower renal complications and ARs. Although the ICER is close to the upper bound of the AIES threshold (€42,567/QALY) for the overall population, it is worth noting that EVR with reduced dose TAC has a favorable cost-effectiveness profile compared to the standard-dose TAC regimen among patients with HCV who received a de novo liver transplantation. In Europe, these sub-groups account for ∼40% of all transplantation cases.

Our analysis has several strengths. First, the efficacy data vs standard-dose TAC is derived from a head to head clinical trial and is robust in sensitivity analyses considered in the model. Second, patient level data were also used, which increases the validity of the model. It also allowed for the exploration of sub-group analyses. The base case results were found to be robust in several sensitivity analyses that were conducted. Finally, to our knowledge this is the first cost-effectiveness model investigating post-liver transplantation and the impact of change in renal function on patient survival.

This analysis is not without limitations. Due to a lack of data in the literature or in the study some assumptions were applied for transition probabilities and mortality. At the time of writing only one published model existed measuring fibrosis progression post-liver transplant. Therefore, many assumptions were applied for acute rejection and the progression of HCV and HCC. As the transition probabilities within CKD stages were not available in the UNOS database after 5 years post-transplant, the model assumed these parameters remained constant from year 5 onwards. There was also limited published evidence on mortality rates relevant to certain health states (i.e. compensated cirrhosis, HCC, acute rejection, CKD stage 1 and death following renal transplant) considered in the model, and long-term TEAEs beyond the 3 years trial period.

Transparency

Declaration of funding

This study was funded by Novartis Pharma.

Declaration of financial/other relationships

JG and AR are employees of Novartis Pharma. RC, VD, and FB are employees of Mapi Group, who received funding from Novartis Pharma to conduct this study. PDS was the lead investigator for the randomized controlled trials of everolimus and received funding from Novartis Pharma. JFR has received compensation for consultancy work from Novartis Pharma to conduct this study. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgements

The authors take full responsibility for the content of this contribution and would like to thank Dr. William Irish of CTI Clinical Trial and Consulting Services for performing the UNOS database analysis.