Table 1Summary of the Manufacturer’s Economic Submission
View in own window
| Drug Product | Riociguat (Adempas) |
| Study Question | “What is the incremental cost per QALY gained, from a provincial government payer perspective for Adempas compared to placebo for the management of CTEPH in Canadian patients in WHO functional class II or III with inoperable CTEPH, or patients with persistent or recurrent pulmonary hypertension following pulmonary endarterectomy, over a patient lifetime horizon (maximum 20 years)?” |
| Type of Economic Evaluation |
|
| Target Population | Inoperable CTEPH patients or CTEPH patients with persistent or recurrent pulmonary hypertension after pulmonary endarterectomy |
| Treatment | 0.5 to 2.5 mg three times a day |
| Outcomes |
CUA: cost per LY gained, cost per QALY gained CCA: 6MWD, PVR, NT-proBNP, WHO FC, TTCW, Borg Dyspnea Index, EQ-5D, LPH, and SF-36
|
| Comparators |
|
| Perspective | Canadian ministry of health |
| Time Horizon | Lifetime (max 20 years) |
| Manufacturer’s Results (Base Case) |
Riociguat vs. placebo: $173,524 per QALY Riociguat vs. generic bosentan: $187,347 per QALY Riociguat vs. Tracleer: riociguat dominates brand-name bosentan
|
| Key Limitations and CDR Estimate(s) |
Riociguat is priced at $42.75 per tablet, or $128.25 daily. Generic and brand-name bosentan are priced at $44.93 daily and $128.36 daily, respectively. The utilization of brand-name vs. generic bosentan varies across drug plans (from 0% to 95% of claims are for brand-name bosentan).
Riociguat vs. Placebo
The long-term efficacy of riociguat is unclear. ICURs are robust in the CDR analyses, based on different assumptions on the relative efficacy of transition probabilities (± 20%). ICURs increase when a shorter time horizon is used, which indicates that much of the benefit occurs in the future. The potential benefit of riociguat on mortality may be double counted. ICUR increases to $350,519 per QALY in the CDR analyses when mortality is mediated only through FC health status.
Riociguat vs. Generic/Brand-Name Bosentan (Most Relevant Comparator)
The true relative efficacy is unknown. A manufacturer-funded ITC was performed, but the results were not used to inform transition through the FC health states. However, ICURs are robust (versus generic bosentan) in the CDR analyses, based on different assumptions on the relative efficacy of transition probabilities (± 20%). ICURs increase when a shorter time horizon is used (and indicate that much of the benefit occurs over a long time frame). The potential benefit of riociguat on mortality may be double counted and the ITC showed no statistically significant difference between riociguat and bosentan for this outcome. Where mortality risk is mediated only through FC health state status, riociguat is dominated by generic bosentan (more costly and less effective) and riociguat is less costly but less effective than Tracleer (CDR analysis). The true differences in mortality between riociguat and bosentan are unknown.
|
6MWD = six-minute walk distance; CCA = Cost-consequence analysis; CDR = Common Drug Review; CTEPH = chronic thromboembolic pulmonary hypertension; CUA = cost-utility analysis; EQ-5D = EuroQol 5-Dimensions Questionnaire; FC = functional class; ICUR = incremental cost-utility ratio; ITC = indirect treatment comparison; LPH = living with pulmonary hypertension questionnaire; LY = life-year; NT-proBNP = N-terminal prohormone of brain natriuretic peptide; PVR = pulmonary vascular resistance; QALY = quality-adjusted life-year; SF-36 = Short-Form 36 Health Survey; TTCW = time to clinical worsening; vs. = versus; WHO FC = World Health Organization functional class.
Background
Riociguat (Adempas) is being reviewed for the treatment of inoperable chronic thromboembolic pulmonary hypertension (CTEPH), or persistent or recurrent CTEPH after surgical treatment in adult patients (≥ 18 years of age) with World Health Organization functional class (WHO FC) II or III pulmonary hypertension. Riociguat is administered based on an individual dose titration of between 0.5 to 2.5 mg taken three times a day. The manufacturer submitted a price of $42.75 per tablet (0.5 mg, 1 mg, 1.5 mg, 2 mg, or 2.5 mg), or $128.25 daily.1
Summary of Economic Analysis
The manufacturer conducted a cost-utility analysis (CUA) from a Canadian public-payer perspective, over a 20-year time horizon, comparing riociguat with placebo and riociguat with generic and brand-name bosentan (Tracleer).
The 16-week cycle Markov model included the following health states: WHO FC II, WHO FC III, WHO FC IV, and death. WHO FC I was not included in the model since the CHEST-1 trial did not recruit patients in that health state. The clinical data from Phase III CHEST-1 and CHEST-2 trials were used to establish the characteristics of patients entering the economic model, transition probabilities between FC for placebo (CHEST-1) and riociguat (CHEST-2) for the first model cycle (16 weeks), the frequency of adverse events, and utility measurement. Comparison between riociguat and bosentan was performed with an indirect treatment comparison (ITC) using CHEST-1 and the BENEFiT trial.2 Within each Markov cycle, patients can remain in the same health state, improve by one FC, worsen by one FC, or die. After the first cycle, FC transitions were derived from the extrapolation of survival curves derived from statistical fitting of the trial data. Only liver toxicity and hypotension were included in the model as adverse events. Mortality data by FC were from a European chart review commissioned by the manufacturer.1
Utilities associated with FC status were collected from CHEST-1. Disutilities associated with adverse events were not considered in the model. Drug costs for riociguat were provided by the manufacturer, while costs for Tracleer and generic bosentan were obtained from the Quebec Formulary (June 2013). Treatment-specific one-off initiation costs were based on discussions with clinical experts. Supportive care use (such as supplemental oxygen, warfarin, and diuretics) was based on the European chart review, with the unit cost derived from Canadian sources such as the Ontario Drug Benefit (ODB) Formulary and a study on the cost of management of warfarin.3 Similarly, ongoing health care resource utilization associated with CTEPH (hospitalizations, specialists visits, and examination and diagnostic testing) were also based on the European chart review, with the unit costs estimated from Canadian sources (Ontario Schedule of Benefits, Ontario Case Costing Initiative (OCCI), and BC Medical Services Commission payment schedule).
Results of Manufacturer’s Analysis
Using the health-payer perspective, the manufacturer reports an incremental cost per quality-adjusted life-year (QALY) for riociguat compared with placebo of $173,524. The incremental cost per QALY for riociguat compared with generic bosentan is $187,347. Riociguat dominates Tracleer.
Interpretations and Key Limitations
Short Duration of Clinical Trials and Assumption of Long-Term Relative Efficacy
Given the duration of existing trials (16 weeks for CHEST-1 and approximately one year for CHEST-2), it is not established that long-term differences will occur in the clinically important outcome of FC (the major factor driving quality of life and disease costs). If the treatment effect is not durable or attenuates, the cost-effectiveness ratio will be greater.
Mortality and Relative Efficacy
In the model, mortality is assumed to increase by worsening FC (which, according to the clinical expert, is a reasonable assumption), but mortality is also impacted by treatment strategy, regardless of FC health state (informed by the ITC, but not stratified by FC status). This might lead to double counting of the potential mortality benefit of riociguat. True differences in mortality between riociguat and comparators are not known.
Transition Probabilities
The manufacturer states that it was not possible to derive the odds ratio (OR) from a formal indirect comparison between riociguat and bosentan for each FC health state, since individual patient data were not available. Therefore, the ORs for transition to FC health states in patients treated with bosentan were estimated from the BENEFiT study (bosentan group only) using a calibration approach. Of note, the ITC4 submitted by the manufacturer reported non-significantly increased odds of being in a better functional class (FC I or II versus FC III, IV, or death) at the study end point when treated with riociguat compared to bosentan (OR 1.15, 95% credible interval [Crl], 0.51 to 2.61). The true relative efficacy of riociguat and bosentan is not clear. This is a key limitation, as in Canada bosentan is currently used in the majority of patients who would be eligible for riociguat.
Titration Cost Not Included in the Cost-Utility Analysis
There were four nursing visits for treatment initiation with riociguat in the cost-consequence analysis (CCA), but only one visit in the CUA model. However, given the high drug cost, the impact of the titration cost on the incremental cost-utility ratio (ICUR) is minimal.
Results of CADTH Common Drug Review Analysis
Riociguat Versus Placebo
Uncertainty in long-term efficacy: Shortening the time horizon to five years, the incremental QALYs associated with riociguat compared with placebo decreased from 0.887 to 0.275, and the cost per QALY increased to $434,311 for riociguat versus placebo, highlighting that a majority of the incremental benefit accrued in the model is well beyond the timeframe of current randomized controlled trials (RCTs).
Mortality might be double counted. ICUR increases to $350,519 per QALY in the CADTH Common Drug Review (CDR) analyses when mortality risk by FC class only is considered.
Uncertainty in transition probabilities through FC. Exploring the ± 20% for the transition probabilities did not significantly alter results.
Riociguat Versus Generic Bosentan
Uncertainty in long-term efficacy: When shortening the time horizon to five years, the incremental QALYs decreased from 0.416 to 0.137 for riociguat compared with generic bosentan, and the cost per QALY increased to $492,361.
Mortality might be double counted. Riociguat is dominated by generic bosentan (more costly and less effective: 5.387 versus 5.839 QALYs, respectively) in the CDR analyses when mortality risk by FC class only is considered, although true differences in mortality are not known.
Uncertainty in transition probabilities through FC. Exploring the ± 20% for the transition probabilities did not significantly alter results.
Riociguat Versus Brand-Name Bosentan (Tracleer)
Uncertainty in long-term relative efficacy: Riociguat dominated Tracleer in all time horizons tested.
Mortality might be double counted. Riociguat is less costly but less effective than Tracleer when mortality risk by FC class only is considered in sensitivity analyses. The cost per QALY for Tracleer is $227,457 compared with riociguat.
Issues for Consideration
According to the clinical expert, the majority of eligible patients are currently treated with bosentan (through special authorization programs). There is wide variation in the proportion of patients treated with Tracleer versus generic bosentan (from 0 to approximately 95%), which has important implications for incremental cost. Riociguat and Tracleer have almost the same daily cost ($128.25 versus $128.36, respectively). Tracleer appears to be favoured by clinical experts due to industry-supported patient programs, concerns around the range of bioavailability of generics, and switching “stable” patients to generic. The true difference in efficacy and side effects of generic versus brand-name bosentan is not known. As the proportion of patients on Tracleer versus generics increases, incremental drug costs for riociguat become smaller.
According to the clinical expert, since there is currently no Health Canada–approved drug for this indicated patient group, riociguat will become first-line treatment if listed by drug plans.
This medication, if approved, is unlikely to modify consideration of surgical management which, if feasible, is considered optimal treatment.
Riociguat is also indicated for treatment in patients with pulmonary arterial hypertension (PAH), and is likely to be used in non-CTEPH PAH.
Conclusions
In the CDR reanalysis, eliminating the possible double counting of mortality benefit, the ICUR increases to $350,519 per QALY for riociguat versus placebo, and riociguat results in lower QALYs than bosentan (riociguat is dominated by generic bosentan; riociguat is less costly, but less effective than Tracleer). However, true differences in mortality between riociguat and bosentan are unclear. There is significant uncertainty in the ICUR given lack of head-to-head trials of riociguat versus bosentan, the approach to modelling relative efficacy, and lack of data on long-term outcomes. Several scenarios result in greater ICURs than the base case presented by the manufacturer.
If drug costs of riociguat versus bosentan only are examined, riociguat has similar costs in jurisdictions where all patients are receiving Tracleer, but incremental costs of riociguat versus bosentan rise as this proportion falls.
