Conclusions

The CADTH clinical review concluded ravulizumab is noninferior to eculizumab in transfusion avoidance, occurrence of breakthrough hemolysis (BTH), lactate dehydrogenase (LDH) normalization, and hemoglobin stabilization over 26 weeks of treatment in adult patients with paroxysmal nocturnal hemoglobinuria (PNH), with maintenance of efficacy up to 52 weeks of treatment.

Based on the findings of the CADTH clinical review, CADTH assumed equal efficacy of ravulizumab and eculizumab. Further, CADTH undertook reanalyses to address limitations which included removing treatment-specific utility differences within health states for ravulizumab and eculizumab and removing the utility increment due to frequency of health care visits for patients receiving ravulizumab. Based on the CADTH reanalysis, conclusions remain similar to the sponsor’s: ravulizumab is equally as effective and potentially less costly compared with eculizumab when patients are treated for a lifetime (up to 100 years of age).

These results are driven by the lower drug acquisition cost of ravulizumab maintenance doses, when considering the publicly available list price of eculizumab over a lifetime time horizon (Table 7 for cost comparison). Because loading dose costs are higher for ravulizumab than eculizumab, cost savings with ravulizumab are realized much further into the time horizon. Patients who are treatment-naive would need to receive ravulizumab for more than 26 years before cost savings are realized or 34 years for patients who are treatment-experienced.

There is further uncertainty associated with the potential cost savings with ravulizumab, as eculizumab was previously reviewed by CADTH for the same indication and received a “do not list” recommendation, in part because it was found to have an incremental cost-effectiveness ratio (ICER) of $2.4 million per quality-adjusted life-year (QALY) gained compared with supportive care and would require a substantial reduction in price to be considered cost-effective. While participating drug plans may list eculizumab, the actual price may be lower than the current list price. Based on a CADTH threshold analysis, should the actual price of eculizumab for the participating drug plans be 1% less than the current list price, ravulizumab would be more costly than eculizumab. Therefore, ravulizumab is unlikely to be cost-effective at its submitted price and a price reduction is likely required, the magnitude of which is unknown.

Further uncertainty remains which could not be addressed by CADTH: up-dosing with either treatment would add drug costs and have an uncertain effect on clinical outcomes; reduced infusion frequency associated with ravulizumab could be preferred by patients, the impact (utility) of which was not appropriately captured in the analysis; and, as best supportive care was not included in the analysis as a comparator, the cost-effectiveness of ravulizumab compared to no active comparator is unknown.

Economic Evaluation

Summary of Sponsor’s Economic Evaluation Results

All analyses were run probabilistically (1,000 iterations). The deterministic and probabilistic results were similar. The probabilistic findings are presented below.

Base-Case Results

Ravulizumab was associated with a QALY gain of 0.92 at a cost that was $42,858 less than eculizumab, resulting in ravulizumab dominating (i.e., being more effective and less costly) eculizumab (Table 3). Ravulizumab was dominant in 95% of iterations. At a willingness-to-pay threshold of $50,000 per QALY gained, there was a 97% probability of ravulizumab being cost-effective compared to eculizumab.

Drug costs accounted for more than 99% of total costs for both treatments. There was no life-year gain associated with ravulizumab. All of the QALY gain for ravulizumab compared with eculizumab was accrued in the no BTH health state. This occurred because 0 ravulizumab life-years were spent in any of the BTH-related health states; all ravulizumab life-years accrued in the no BTH and spontaneous remission states.

Table 3

Summary of the Sponsor’s Economic Evaluation Results.

Sensitivity and Scenario Analysis Results

The same conclusions made in the aggregate population were also made in the treatment-naive and treatment-experienced cohorts, with ravulizumab dominating eculizumab. A breakdown of incremental costs and QALYs by cohort and for the aggregate population is provided in Table 9.

The sponsor conducted a number of scenario analyses examining uncertainty. The only scenario whereby ravulizumab was not dominant occurred when the time horizon was shortened to 10 years. This occurred because it takes greater than 10 years to recuperate the additional costs of the ravulizumab loading dose for patients who are switching from eculizumab and because loading dose costs for ravulizumab are higher than those for eculizumab.

While cost-effectiveness conclusions associated with other scenarios remained unchanged from the sponsor’s base case, a few resulted in substantial changes to incremental costs and QALYs. The scenarios exploring up-dosing of eculizumab for incomplete inhibition of C5-related BTH events led to greater cost savings (meaning additional eculizumab costs relative to ravulizumab). Having an increased background mortality risk for patients with PNH led to fewer incremental costs and similar QALYs.

CADTH Appraisal of the Sponsor’s Economic Evaluation

CADTH identified several key limitations to the sponsor’s analysis that have notable implications on the economic analysis.

• The CADTH clinical review concluded that ravulizumab is noninferior to eculizumab and conclusions on superiority of ravulizumab to eculizumab could not be drawn. The data used to inform treatment efficacy parameters in the sponsor’s model were derived from 2 noninferiority phase III trials comparing the effectiveness of ravulizumab with eculizumab across a variety of outcomes. According to the CADTH clinical review, the noninferiority margins for ravulizumab were met for all outcomes; however, superiority was not reached. The CADTH clinical review report therefore concluded that ravulizumab was noninferior to eculizumab in transfusion avoidance, occurrence of BTH, LDH normalization, and hemoglobin stabilization. Given this, the sponsor’s base case, which used trial data to inform BTH-related transition probabilities, transfusion probabilities, and transfusion volumes, does not align with the CADTH clinical review conclusions, as the transition probabilities for ravulizumab were set to perform better than eculizumab across all outcomes, based on rates observed in the clinical trials. In particular, the model predicted no incomplete C5 inhibition-related events for patients receiving ravulizumab over the entire time horizon, which is uncertain given the conclusions of the clinical review on the short-term evidence, as well as a lack of long-term evidence in support. According to the clinical expert consulted for this review, there are anecdotal cases in jurisdictions where ravulizumab is available where incomplete C5 inhibition has led to a shorter dosing interval than the labelled 8-week interval to maintain C5 inhibition. It is therefore unreasonable to assume there will be no instances of incomplete C5 inhibition with ravulizumab for the entire model time horizon. While there is potential for ravulizumab to perform better than eculizumab based on its pharmacokinetic profile, CADTH concluded that the sponsor’s base-case transition probabilities were inappropriate and not supported by clinical evidence given the trial’s inability to demonstrate that ravulizumab is superior to eculizumab.
  In addition to the conclusions related to BTH events, ravulizumab did not demonstrate superiority to eculizumab for other efficacy outcomes parameterized in the model, including the probability of transfusion and volume required. According to the clinical expert consulted by CADTH for this review, outside of the context of BTH events, transfusions would not be expected to differ between treatment populations. As CADTH concluded noninferiority of ravulizumab to eculizumab, transfusions related to BTH events were also assumed to be similar across comparators. Finally, according to the clinical expert consulted for this review, among those requiring transfusions, the volume of pRBCs received is not expected to differ between treatments.
  • In CADTH reanalyses, to align with the finding of ravulizumab being noninferior to eculizumab, CADTH set efficacy for ravulizumab to be equal to eculizumab.
  • To explore uncertainty surrounding the comparative clinical effectiveness between treatments, CADTH conducted several scenarios: 1 where eculizumab efficacy was set to be equal to ravulizumab, and 1 using the sponsor’s efficacy assumptions which assume ravulizumab is more efficacious than eculizumab.
• Health states used in the model did not capture all aspects of the condition. To capture the costs and health-related quality of life associated with PNH disease progression and the impacts of treatment, the sponsor’s model was based on BTH events. BTH was defined as having at least 1 new or worsening symptom or sign of intravascular hemolysis (fatigue, hemoglobinuria, abdominal pain, shortness of breath [dyspnea], anemia [hemoglobin < 10 g/dL], major adverse vascular event [including thrombosis], dysphagia, or erectile dysfunction) in the presence of elevated LDH 2 or more × the upper limit of normal (ULN), after prior LDH reduction to less than 1.5 × ULN on therapy.^3,4 While prevention of BTH events is noted to be an important component of the treatment of PNH, both clinical expert feedback and patient input received noted thrombosis to be the most devastating consequence of disease, which was not explicitly modelled. Other symptoms noted to be important in the patient input received by CADTH included fatigue, difficulty swallowing, pulmonary hypertension, chronic kidney disease, and shortness of breath, none of which were explicitly modelled. While the clinical expert noted that some symptoms such as improvements in fatigue are related to BTH events, this was not explicitly modelled, and overall BTH it is a poor proxy for many other outcomes important to patients and which affect health system costs.
  • CADTH was unable to incorporate other important aspects of PNH in the model. Given that ravulizumab demonstrated noninferiority to eculizumab across most outcomes, the influence of not including all aspects of the condition in the model on cost-effectiveness results should be limited.
• Health state utility values and approach to their implementation w inappropriate. Several issues were identified with the approach to identifying and incorporating patient utility values in the submitted model. The sponsor conducted a regression analysis on data from Study 301 and Study 302 to estimate utility impacts associated with BTH events and transfusions. A treatment indicator was included to explore a difference in utility between ravulizumab and eculizumab, independent of BTH events and transfusions and led to treatment-specific utility values which were used in the sponsor’s base case.⁵ The sponsor derived ravulizumab base utility values by adding the treatment indicator for ravulizumab onto eculizumab utilities, meaning the utility for ravulizumab patients was fixed to be 0.01 and 0.02 better than for patients receiving eculizumab in Cohort 1 and Cohort 2, respectively.⁵ The use of treatment-specific utility values is contradictory to CADTH guidelines that specify that utilities be associated with health states.¹³ All outcomes associated with treatment, along with their impact on patient utility, should be explicitly modelled, rather than captured using a treatment-specific utility value. Including a treatment indicator to capture a difference in utility between treatments that has not been modelled is therefore inappropriate.
  Second, to capture the utility difference associated with reduced visit frequency due to less frequent treatment administration with ravulizumab, the sponsor added a utility benefit for ravulizumab of 0.02. The value was based on an assumption and informed by a study demonstrating a higher utility for cardiovascular patients who had 7 or less general practice visits compared to those who were seen more than 7 times annually.¹¹ First, whether a utility benefit captured in this patient population and visit setting would apply to patients with PNH receiving infusions is uncertain. Second, a more appropriate way of explicitly incorporating the quality of life benefits associated with infusions would have been to incorporate an administration disutility for all patients and apply it during cycles that they received an infusion. As patients receiving ravulizumab would receive fewer infusions, this would result in an overall higher QALY gain, but it is uncertain whether the difference in QALYs between treatments due to fewer treatment administrations would be accurately represented by the sponsor’s assumed utility increment for ravulizumab. Finally, whether reduced infusion frequency would improve health-related quality of life or quality of life in general is uncertain as mentioned by the clinical expert consulted by CADTH. From a health care system perspective, health-related quality of life benefits should be captured, whereas general quality of life benefits would be captured from a broader societal perspective which is not considered in the CADTH base case.
  Finally, the sponsor estimated utility values in the model by mapping clinical trial EORTC QLQ-C30 data to EQ-5D values using an algorithm published by Longworth at al.¹⁰ According to CADTH Guidelines for Economic Evaluation, mapping as a means of deriving health utilities is not recommended.¹³ Instead, CADTH prefers the use of a generic indirect utility measure to obtain utility scores for the economic model.¹³ This is because the utility values garnered through mapping can vary dramatically depending on instruments being mapped, the algorithm used for mapping, and the severity of the included health states. As noted in the guidelines and by the sponsor, there are several published algorithms available to map EORTC QLQ-C30 data to EQ-5D values.⁵ As demonstrated by the sponsor’s scenario analyses exploring the use of a different mapping algorithm by McKenzie et al.,¹⁴ not only do the utility values derived vary from those derived using the Longworth algorithm, so too do the resulting total QALYs estimated in the model. Whereas the sponsor’s base case led to 0.92 incremental QALYs between ravulizumab and eculizumab, when the McKenzie algorithm was used it led to 1.59 incremental QALYs, a 53% difference.⁵ This demonstrates the uncertainty that mapping introduces into the analysis and highlights its inappropriateness.
  • CADTH was unable to address the uncertainty associated with deriving utility estimates using mapping. In the CADTH reanalysis, both the treatment indicator and the utility increment for reduced treatment administration visits with ravulizumab were removed as they do not explicitly model outcomes associated with complement inhibitor treatment. Additionally, the value of the reduced visit increment is uncertain.
• Up-doses associated with BTH events are highly uncertain and may affect total treatment-related costs. The sponsor’s model allowed for the flexibility to consider an increased dose in response to incomplete C5 inhibition or CAC-related BTH events. No up-dosing was considered in the sponsor’s base case. According to the clinical expert consulted by CADTH for this review, some up-dosing for patients receiving eculizumab is occurring in Canadian clinical practice, with a proportion of patients receiving a maintenance dose of 1,200 mg rather than 900 mg to maintain C5 inhibition. However, in Study 301 and Study 302, only the 900 mg dose of eculizumab was allowed; therefore, the effect of up-dosing on BTH events is not captured by the trial. Consequently, while the option to up-dose in the model adds costs and removes the disutility associated with a BTH event, it does not influence other efficacy outcomes such as BTH transitions, transfusions, or thrombosis. Further, while Study 301 and Study 302 did not capture incomplete C5 inhibition for patients receiving ravulizumab, it is uncertain whether C5 inhibition would be sustained in the long-term. According to the clinical expert consulted for this review, there are anecdotal cases in jurisdictions where ravulizumab is available where incomplete C5 inhibition has led to a shorter dosing interval than the labelled 8-week interval. Taken together, due to the uncertainty regarding the form up-dosing would take should it occur (higher dose or reduced intervals), the proportion of patients receiving eculizumab and ravulizumab who would require 1-off or continuously higher levels of drug to maintain C5 inhibition and the influence of higher doses on drug efficacy, CADTH agrees with the sponsor’s base case which excludes the consideration of up-dosing for both treatments. If increased doses are required of either drug, it will add costs and have an uncertain influence on outcomes.
  • Due to the uncertainty regarding the proportion of patients on ravulizumab and eculizumab who would require 1-off or continuous up-dosing, how up-dosing would occur (e.g., higher dose given at the same administration intervals, shortening the time between administrations, whether ravulizumab patients would be up-dosed on ravulizumab or receive a dose of eculizumab), and the influence of higher doses on model efficacy parameters, CADTH did not incorporate up-dosing in the reanalysis. If only patients receiving eculizumab require a higher dose to maintain C5 inhibition for the model time horizon, ravulizumab will dominate as it will be associated with lower costs. If patients receiving ravulizumab also require a higher dose, the influence on the cost-effectiveness results are uncertain.
• The cost-effectiveness of ravulizumab compared with best supportive care is unknown. In the sponsor’s submission, ravulizumab was compared with eculizumab. Best supportive care was not included as a comparator. According to the clinical expert consulted by CADTH for this review, eculizumab is the most relevant comparator to ravulizumab. However, as eculizumab received a “do not list” recommendation during its review by CADTH for PNH, for participating plans that do not reimburse eculizumab, the cost-effectiveness of ravulizumab is unknown. The CADTH review of eculizumab estimated an ICER of $2.4 million per QALY gained with eculizumab compared with supportive care and noted that without a substantial reduction in price, it would not be considered cost-effective.¹⁵ This is noteworthy as participating drug plans that do reimburse eculizumab may have negotiated some price reduction. Based on a CADTH threshold analysis, should the price of eculizumab be 1% less than the current list price, ravulizumab will be more costly.
  • CADTH was unable to address this limitation. As such, the cost-effectiveness of ravulizumab relative to best supportive care is unknown.

The following limitations were identified but were not deemed key limitations:

• The rate of spontaneous remission incorporated in the model may be an overestimate. In the sponsor’s model, spontaneous remission was informed by a study by Hilleman et al. which followed patients with PNH for 30 years and found approximately 15% of patients had a spontaneous clinical recovery. As noted by the sponsor, the rate of spontaneous remission is uncertain, and, according to the clinical expert consulted by CADTH for this review, the rate may be closer to 5%. However, as spontaneous remission is assumed to be equal in both arms, it is not expected that the rate of spontaneous remission will drive results.
  • CADTH conducted a scenario exploring the impact of no spontaneous remission as a scenario analysis.
• The difference in infusion times between comparators was not explored. As noted in the CADTH clinical review report, eculizumab and ravulizumab administration differs in both frequency and duration of the infusion., Eculizumab is infused over 35 minutes,¹ whereas ravulizumab maintenance doses are infused over 120 to 140 minutes. The sponsor did not include administration costs as part of the submission as it was assumed that these would be covered for the duration of the model time horizon. It would have been more accurate to include the option to explore administration frequencies and durations as a model parameter to explore the scenario where the infusions differ in time should funding through the patient support program change in the future.
  • CADTH was unable to address this limitation. The cost per infusion would be expected to be higher for ravulizumab than eculizumab because of the duration; however, overall infusions costs would be lower for ravulizumab as there are fewer annual infusions.
• Mortality associated with PNH is uncertain. In the sponsor’s base case, background mortality was assumed to be equal to that of the general population and the sponsor assumed that there was no excess mortality risk associated with BTH events. According to the clinical expert consulted by CADTH for this review, it is expected that the mortality of patients treated with complement inhibitors will follow that of the general population for the first 10 years, and then will become higher than that of the general population due to factors such as bone marrow failure, which is not prevented by complement inhibitor treatment. Second, the clinical expert consulted for this review also noted that there is a mortality risk associated with experiencing a BTH event; however, the magnitude of the risk is uncertain.
  As background mortality was assumed to be equal across treatments, an increasing mortality risk over time for all patients would not be expected to influence incremental life-years between treatments. Additionally, as the CADTH base case has assumed the risk of experiencing a BTH event would be equal across treatments, mortality associated with BTH events will also not influence incremental life-years.
  • CADTH was unable to address this limitation; however, it is not expected to influence cost-effectiveness results.

Additionally, the following key assumptions were made by the sponsor and have been appraised by CADTH (Table 4).

Table 4

Key Assumptions of the Submitted Economic Evaluation (Not Noted as Limitations to the Submission).

CADTH Reanalyses of the Economic Evaluation

Base-Case Results

The CADTH base case was derived by making changes in model parameter values and assumptions, in consultation with clinical experts. CADTH reanalyses addressed several limitations within the economic model, summarized in Table 5. CADTH was unable to address limitations regarding health states not fully reflecting all key aspects of PNH, the approach to incorporating treatment frequency impact on utility and the magnitude of health-related quality of life benefit from administration frequency, the difference in infusion times between comparators, and uncertainty surrounding mortality assumptions.

Table 5

CADTH Revisions to the Submitted Economic Evaluation.

The results of CADTH’s stepped analysis are presented in Table 6. CADTH’s base-case reanalysis demonstrates that, compared with eculizumab, ravulizumab yields the same number of QALYs and is associated with lower costs (–$13,386), resulting in ravulizumab dominating eculizumab (Table 6). Ravulizumab was dominant in all steps of the analysis. Removing the utility increment related to fewer treatment administrations for ravulizumab resulted in the greatest decrease in total QALYs with ravulizumab. The majority (56%) of the QALYs for both comparators was accrued in the no BTH health state (Table 10). Similar to the sponsor’s base case, nearly all (99.7%) of total costs for both comparators were drug costs (Table 10). When considering drug costs alone, in the treatment-naive population, patients must remain on treatment for 26.75 years before overall drug costs with ravulizumab become lower than eculizumab. In the treatment-experienced population, it took 34.73 years of treatment for ravulizumab to become less costly than eculizumab.

Table 6

Summary of the Stepped Analysis of the CADTH Reanalysis Results.

Scenario Analysis Results

To address remaining uncertainty regarding parameterization of the model, CADTH conducted several scenario analyses. Full results are presented in Table 11. Ravulizumab remained dominant in all scenarios apart from a scenario exploring its cost-effectiveness in patients weighing more than 100 kg. In this scenario, eculizumab dominated ravulizumab, meaning that ravulizumab is not cost-effective compared to eculizumab in this patient population. Using the sponsor’s base-case efficacy parameters led to a 0.10 QALY gain for ravulizumab compared to eculizumab, highlighting that the improvements in BTH event frequencies and transfusions used in the sponsor’s analysis do not drive the cost-effectiveness of ravulizumab. Applying the sponsor’s utility increment associated with fewer visits for ravulizumab led to a QALY gain of 0.39 for ravulizumab. This scenario is uncertain as the value of the increment used was based on assumption, and whether reduced visit frequencies influence health-related quality of life rather than general quality of life is uncertain. CADTH conducted a threshold analysis using the CADTH base case to examine the price for eculizumab at which ravulizumab would be considered cost-effective. If the price per vial of eculizumab is $6,734 or more, ravulizumab results in cost savings over a lifetime time horizon; however, if the confidentially negotiated price of eculizumab is $6,733 per vial or less, ravulizumab will not be considered cost-effective compared to eculizumab and eculizumab will dominate ravulizumab.

Issues for Consideration

• Should eculizumab biosimilars become available, and should these biosimilars be considered equivalent to eculizumab, ravulizumab is unlikely to remain less costly than eculizumab biosimilars, and eculizumab biosimilars would be considered optimal.
• Administration fees for complement inhibitor therapies were not included in the sponsor’s model as they were assumed to be covered by the sponsor’s patient support program for the entirety of the model time horizon. If this were to change, it is expected that administration fees will be greater for eculizumab compared with ravulizumab due to the greater frequency of administration; however, chair time for a given administration will be longer for patients receiving ravulizumab.

Overall Conclusions

The CADTH clinical review found that ravulizumab is noninferior to eculizumab in transfusion avoidance, occurrence of BTH, LDH normalization, and hemoglobin stabilization over 26 weeks of treatment in adult patients with PNH, with maintenance of efficacy up to 52 weeks of treatment.

In the submitted model, ravulizumab was set to perform better across relevant outcomes for the entire model time horizon, which is uncertain in the long-term. Additional identified limitations include that the health states used in the model did not fully capture all aspects of PNH that may impact patient health-related quality of life, mortality, or health system costs. Several limitations were also identified with the utility values, including the use of treatment-specific utility values and the sponsor added a utility increment to ravulizumab associated with fewer administration visits, which was inappropriate because the utility increment value was uncertain, and it is uncertain if reduced visit frequency would be associated with improved health-related quality of life specifically. CADTH also noted that up-dosing practices in the real-world setting are highly uncertain but may affect the estimated cost-effectiveness of ravulizumab in comparison with eculizumab.

Given the findings of the CADTH clinical review, CADTH assumed equal efficacy for ravulizumab and eculizumab. Further, CADTH undertook reanalyses to address limitations which included removing treatment-specific utility differences within health states for ravulizumab and eculizumab and removing the utility increment due to frequency of health care visits for patients receiving ravulizumab. Based on the CADTH reanalysis, conclusions remain similar to the sponsor’s: ravulizumab is equally as effective and potentially less costly compared with eculizumab when patients are treated for a lifetime (up to 100 years of age).

These results are driven by the lower drug acquisition cost of ravulizumab maintenance doses, when considering the publicly available list price of eculizumab over a lifetime time horizon (Table 7 for cost comparison). Because loading dose costs are higher for ravulizumab than eculizumab, cost savings with ravulizumab are realized much further into the time horizon. Patients who are treatment-naive would need to receive ravulizumab for more than 26 years before cost savings are realized or 34 years for patients who are treatment-experienced.

There is further uncertainty associated with the potential cost savings with ravulizumab, as eculizumab was previously reviewed by CADTH for the same indication and received a “do not list” recommendation, in part because it was found to have an ICER of $2.4 million per QALY gained compared with supportive care and would require a substantial reduction in price to be considered cost-effective. While participating drug plans may list eculizumab, the actual price may be lower than the current list price. Based on a CADTH threshold analysis, should the actual price of eculizumab for the participating drug plans be 1% less than the current list price, ravulizumab would be more costly than eculizumab. Therefore, ravulizumab is unlikely to be cost-effective at its submitted price and a price reduction is likely required, the magnitude of which is unknown.

Further uncertainty remains which could not be addressed by CADTH: up-dosing with either treatment would add drug costs and have an uncertain effect on clinical outcomes; reduced infusion frequency associated with ravulizumab could be preferred by patients, the impact (utility) of which was not appropriately captured in the analysis; and, as best supportive care was not included in the analysis as a comparator, the cost-effectiveness of ravulizumab compared to no active comparator is unknown. Though the model did not capture outcomes identified as important based on patient and clinician input such as thrombosis, this is not expected to influence conclusions as ravulizumab was found to be noninferior to eculizumab for most outcomes.

Detailed Results of the Sponsor’s Base Case

Table 9

Summary of Results of the Sponsor’s Base Case by Treatment Status.

Detailed Results of CADTH Base Case

Table 10

Disaggregated Summary of CADTH’s Economic Evaluation Results.

Scenario Analyses

Table 11

CADTH Scenario Analyses.

Summary of Sponsor’s BIA

The sponsor submitted a BIA estimating the budget impact of introducing ravulizumab for the treatment of adult patients with PNH. The BIA base case was undertaken from a publicly funded drug plan perspective considering only drug costs over a 3-year time horizon. The sponsor’s patient support plan data were used to estimate the number of individuals being treated for PNH and discontinuation rates (Table 13).

The sponsor assumed that no new patients were expected to start complement inhibitor therapy over the 3-year time horizon.¹⁶

The sponsor compared a reference scenario, where only eculizumab was available to treat adult patients with PNH with a new drug scenario where ravulizumab is also funded according to current eculizumab reimbursement criteria (i.e., 1) granulocyte clone > 10%; 2) LDH greater than 1.5 × ULN; and 3) at least 1 of the following: thrombosis, transfusions, anemia, pulmonary insufficiency, renal insufficiency, and smooth muscle spasms). Eculizumab costs included in the analyses accounted for up-dosing by assuming that 10% of patients were maintained at a dose of 1,200 mg every 2 weeks, and 90% were maintained at 900 mg.¹⁷ Ravulizumab costs assumed that 91% of patient weights fell between 60 to less than 100 kg; 7% were between 40 to less than 60 kg, and 2% were greater than 100 kg, based on Canadian PNH registry data.¹⁸

Table 13

Summary of Key Model Parameters.

Summary of the Sponsor’s BIA Results

The sponsor estimated the net budget impact of introducing ravulizumab for the treatment of adult patients with PNH would be $601,724 in Year 1, $335,187 in Year 2 and $118,759 in Year 3 for a total budget impact $1,055,670 over 3 years.

CADTH Appraisal of the Sponsor’s BIA

CADTH identified several key limitations to the sponsor’s analysis that have notable implications on the results of the BIA:

• The BIA population is not aligned with the Health Canada indication. Rather than using an epidemiological based approach and applying filters to derive the size of the eligible population, the sponsor estimated the market size of ravulizumab based on their confidential patient support program for eculizumab. As there may be reimbursement criteria for eculizumab that are narrower than the Health Canada indication, this means that the population estimated by the sponsor is smaller than the total Health Canada indicated eligible population size. According to the clinical expert consulted by CADTH for this review, there is a proportion of patients in clinical practice who meet the Health Canada indication for eculizumab but do not meet jurisdictional reimbursement criteria. These patients would therefore also be eligible for ravulizumab but would not be captured in the sponsor’s estimates.
  Additionally, using confidential data from the patient support program meant that CADTH was unable to validate key input parameters including population size, the number of new patients per year and discontinuation rate. This means there is uncertainty regarding the size of the eligible population for ravulizumab.
  • CADTH was unable to estimate the total potential size of the eligible population for ravulizumab using the sponsor’s existing model structure. Should ravulizumab be reimbursed in a manner similar to eculizumab, this is expected to have minimal influence on the budget impact estimate. If ravulizumab is reimbursed according to its Health Canada indication, the budget impact will be greater than estimated by the sponsor and CADTH.
  • CADTH explored a scenario analysis that increased the total size of the eligible patient population by 5% to explore the impact of reimbursement for a broader population.
• The anticipated uptake of ravulizumab is not aligned with clinical expert expectations. In the sponsor’s base case it was assumed that 15% of patients current receiving eculizumab would switch to ravulizumab each year should ravulizumab become available. Based on the number of patients in the BIA, this works out to ravulizumab having a market share of 15% / 28% / 39% in Year 1, 2, and 3, respectively. According to the clinical expert consulted by CADTH for this review, approximately 50% of eligible patients would be expected to uptake ravulizumab upon availability and this was expected to increase to nearly 100% by Year 3. This is echoed by patient feedback received as part of this review, which indicated that they would expect a large proportion of patients to switch to ravulizumab should it be reimbursed by public payers.
  • In CADTH reanalysis, the proportion of eligible patients who will use ravulizumab in Year 1, 2, and 3 were changed to 50% / 75% / 95%, respectively.
• The number of new PNH patients eligible for complement inhibitor therapy was lower than clinical expert expectations. The sponsor estimated the number of new PNH patients who would start complement inhibitor therapy for the baseline year to be |||| based on projections of the number of new PNH patients historically from 2016 to 2020. However, during Years 1, 2, and 3 of the BIA, zero new patients are assumed to start complement inhibitor therapy, despite the patient support program data showing greater than |||| new patients initiating treatment with eculizumab over the past 4 years. Additionally, feedback from the clinical expert consulted by CADTH for this review indicated that while there may be a low incidence of PNH, if patients who are diagnosed meet the reimbursement criteria for complement inhibitor therapy, they will initiate therapy. Finally, the sponsor assumed that 92.6% of naive patients will initiate ravulizumab in the new drug scenario, but clinical expert feedback indicated nearly all new patients would initiate ravulizumab instead of eculizumab should both be available.
  • In CADTH reanalyses, the number of new patients each year was based on the average number of new patients in the sponsor’s patient support program between 2016 and 2020. Additionally, it was assumed that 97% of new patients would start ravulizumab.
• The implementation of patient discontinuation in the model was inappropriate. The sponsor’s analysis assumed that 2.6% of patients currently receiving eculizumab and ravulizumab would discontinue therapy based on their patient support program. When CADTH validated this rate with the clinical expert consulted for this review, discontinuation was deemed to be higher than that observed in clinical practice.
  In addition, the sponsor’s implementation of the discontinuation rate led to their being different numbers of eligible PNH patients between the reference and new drug scenario, which is inappropriate and not expected. This is because the number of patients who discontinue is calculated by multiplying the number of current patients receiving eculizumab or ravulizumab by the discontinuation rate and leads to a larger total number of patients discontinuing in the reference rather than the new drug scenario (8.68 vs 7.59, respectively).
  • To address this, CADTH corrected the sponsor’s base case by setting the sponsor’s discontinuation rate parameter to 0%. This led to the same number of patients in the reference and new drug scenario.
  • To reflect feedback that a proportion of patients will discontinue, CADTH used the discontinuation rate in the ravulizumab arm from Study 302. To incorporate this, CADTH calculated the total number of patients who would discontinue each year, and weighted that value by the comparator’s respective annual market share to ensure that the number of patients discontinuing was the same in the reference and new drug scenarios.
• The dosing for eculizumab is not aligned with the pharmacoeconomic analysis. In the sponsor’s BIA, 90% of patients receiving eculizumab received a 900 mg dose, and 10% were assumed to be up-dosed and receiving the 1,200 mg dose. According to the clinical expert consulted by CADTH for this review, there are proportion of patients receiving eculizumab, however it is uncertain whether a proportion of patients receiving ravulizumab will also require up-dosing to maintain C5 inhibition. As the sponsor’s and CADTH’s pharmacoeconomic report assumes no up-dosing across comparators, to align with this, CADTH has assumed no up-dosing will occur in the BIA.
  • In CADTH reanalyses, 100% of eculizumab patients receive the 900 mg dose. To explore the impact of eculizumab up-dosing, in a scenario analysis, the sponsor’s assumption of 10% of patients receiving the 1,200 mg dose was explored.
• The number of ravulizumab administrations in the first year of treatment was underestimated. To calculate ravulizumab costs for patients who initiate treatment, the sponsor assumed that patients would receive 6.25 of maintenance administrations in the first year of treatment. However, CADTH calculated that patients will receive a loading dose administration, followed by 7 maintenance administrations in the first year of treatment. The following year, patients would receive 6 administrations, followed by 7 administrations in their third year of treatment, which is reflected in the sponsor’s rate of 6.5 annual administrations in subsequent years of therapy.
  • CADTH’s reanalysis assigned ravulizumab patients to receive 7 maintenance administrations in their first year of treatment to accurately capture the costs of switching from eculizumab to ravulizumab.
• The distribution of patient weights in the BIA are uncertain. In the sponsor’s analysis, the distribution of patients across weight dosing categories was based on Canadian PNH registry data.¹⁶ However, the clinical expert consulted for this review felt that the proportion of patients greater than 100 kg in his clinical practice was greater than the proportion in the registry. As ravulizumab dosing is weight based and those greater than 100 kg require a higher dose that leads to greater subsequent year annual costs than patients who receive eculizumab, this could add costs to the budget impact of reimbursing ravulizumab.
  • As a scenario analysis, CADTH explored the impact of assuming10% of the patient population was greater than 100 kg.

CADTH Reanalyses of the BIA

CADTH revised the sponsor’s base case by increasing the uptake of ravulizumab, changing the number of new patients eligible for complement inhibitor therapy from 0 to 14, increasing the proportion of new patients who initiate ravulizumab, using the discontinuation rate for ravulizumab observed in Study 302, assuming no up-dosing with eculizumab, and assuming that ravulizumab patients receive 7 maintenance dose administrations in the first year of treatment. Table 14 notes the assumptions used by the sponsor in comparison to those used by CADTH in its reanalysis.

Table 14

CADTH Revisions to the Submitted Budget Impact Analysis.

Applying these changes increased the total 3-year budget impact to $13,180,849. The results of the CADTH step-wise reanalyses are presented in summary format in Table 15 and a more detailed breakdown is presented in Table 16.

Table 15

Summary of the CADTH Reanalyses of the BIA.

CADTH also conducted additional scenario analyses to address remaining uncertainty:

1. Increase the total eligible population size by 5% to account for the proportion of patients who may be eligible for complement inhibitor therapy according to the Health Canada indication for complement inhibitors but not meet jurisdictional reimbursement criteria for eculizumab.
2. Assuming 10% of eculizumab patients receive the 1,200 mg dose of eculizumab.
3. Assuming 10% of patients are greater than 100 kg.

Table 16

Detailed Breakdown of the CADTH Reanalyses of the BIA.