Table 1Summary of the Manufacturer’s Economic Submission
View in own window
| Drug Product | Migalastat (Galafold) |
|---|
| Study Question | What are the anticipated costs and health consequences of the use of migalastat for the treatment of Fabry disease, compared with enzyme replacement therapy (ERT)? |
|---|
| Type of Economic Evaluation | Cost-utility analysis |
|---|
| Target Population | Patients with Fabry disease with mutations of their alpha-galactosidase A (GLA) gene determined to be amenable to treatment with migalastat |
|---|
| Treatment | Migalastat 123 mg (equivalent to 150 mg migalastat hydrochloride) orally once every other day at the same time of day |
|---|
| Outcomes | QALYs
LYs |
|---|
| Comparators | Blended ERTs: agalsidase alfa intravenously at a dose of 0.2 mg/kg over 40 minutes and agalsidase beta intravenously at a dose of 1 mg/kg over two hours |
|---|
| Perspective | Canadian Ministry of Health |
|---|
| Time Horizon | 50 years |
|---|
| Results for Base Case | Migalastat is dominant (i.e., less expensive and greater QALYs gained) compared with the blended ERT comparator |
|---|
| Key Limitations | CDR identified the following key limitations:
The clinical efficacy of migalastat is associated with some uncertainty. CDR Clinical Reviewers noted the results of the placebo-controlled trial (FACETS) indicated migalastat did not meet the primary (surrogate) end point, while the head-to-head study of migalastat and ERT (ATTRACT), which was used to support the assumption of non-inferiority of migalastat compared with ERT in the model, was associated with notable limitations, which resulted in uncertainty of the study findings. The generalizability of the findings to the Canadian setting is also uncertain. Inappropriate comparator. Use of a blended comparator in the base case is not appropriate. An analysis should have been presented for migalastat vs. each ERT individually. Calculation errors were identified with the disutility associated with dyspnea, the weighted cost of ERT and the annual cost of migalastat. These errors bias the comparison of cost in favour of migalastat. The submitted model overestimates patient survival. The predicted life expectancy of Fabry disease patients in the model is 81.54 years, which is much higher than that reported in the large international Fabry registry (66.9 years). The impact of this could not be tested by CDR. The disutility associated with ERT infusion is highly uncertain. The manufacturer assumed that ERT infusion was associated with a utility decrement of 0.048 from a UK study. The generalizability of this value to the Canadian population is uncertain. Additionally, the assumption of a significant difference in quality of life based on route of administration was not observed in the ATTRACT trial.
|
|---|
| CDR Estimate(s) |
CDR corrected errors that reduced the costs associated with ERT and increased the costs associated with migalastat. CDR considered migalastat compared with each ERT separately in the CDR reanalyses. Migalastat may be associated with a small incremental QALY gain due to a different adverse event profile compared with each ERT. If a difference in utility values due to the route of administration is considered acceptable, migalastat may be associated with a larger incremental QALY gain. Migalastat may be associated with a large incremental cost, or a large incremental cost saving, depending upon the comparator or comparator price used, patient weight, and discontinuation rate. In the CDR base case for migalastat vs. agalsidase alfa, the ICUR ranged from $200,487 per QALY to $55.9M per QALY depending on whether disutility for infusion was considered. In the CDR base case for migalastat vs. agalsidase beta, migalastat was the dominant strategy regardless of whether disutility was included or not. The results are highly sensitive to the price of ERT, disutility due to route of administration, patient weight, and discontinuation rates. Further, the results do not account for the uncertainty in comparative clinical effectiveness of migalastat.
|
|---|
CDR = CADTH Common Drug Review; ERT = enzyme replacement therapy; ICUR = incremental cost-utility ratio; LY = life-years; QALY = quality-adjusted life-year; vs. = versus.
View in own window
| Drug | Migalastat (Galafold) |
|---|
| Indication | Long-term treatment of adults with a confirmed diagnosis of Fabry disease [deficiency of alpha-galactosidase [alpha-Gal A]) and who have an alpha-Gal A mutation determined to be amenable by an in vitro assay. |
|---|
| Reimbursement Request | As per indication. |
|---|
| Dosage form | 123 mg oral capsules. |
|---|
| NOC date | September 5, 2017 |
|---|
| Manufacturer | Amicus Therapeutics, Inc. |
|---|
Background
Migalastat (Galafold) is indicated for the long-term treatment of adults with a confirmed diagnosis of Fabry disease (alpha-galactosidase A deficiency) who have an amenable mutation.1 The recommended dosage is 123 mg every other day at the same time of day. At the submitted price, migalastat at $1,700 per capsule is $310,250 annually per patient, which is similar to the publicly available prices of IV enzyme replacement therapy (ERT).2 The indication for migalastat differs with the indication for ERTs, given the requirement for an amenable mutation for migalastat.
The manufacturer submitted a cost-utility analysis comparing migalastat with a blended comparator of two ERTs (agalsidase alfa and agalsidase beta) in two patient populations: the base case, based on the mean age from the ATTRACT trial (start age in model = 49 years); and a scenario based on the product monograph (patients aged 18 years or older with Fabry disease and mutations of their alpha-galactosidase A gene determined to be amenable to treatment with migalastat; start age in model = 18 years). The model time horizon was 50 years with annual cycles, and undertaken from the perspective of the Canadian health care payer.2 The manufacturer assumed equivalence between migalastat and the blended ERTs based on a single randomized controlled trial (ATTRACT) that was stated to demonstrate the non-inferiority of migalastat and ERT on surrogate renal outcomes, i.e., measured glomerular filtration rate (GFR) and estimated GFR. The same transition probabilities for health states for each treatment group were therefore assumed and based on a Dutch cost-effectiveness study (Rombach et al.).3 The manufacturer assumed a treatment discontinuation rate of 1% for both migalastat and ERT. Patient weight was derived from unpublished data from the Canadian Fabry Disease Initiative. Adverse event rates were obtained from the ATTRACT trial. Health service utilization due to Fabry disease and associated complications were based on Rombach et al.3 Unit costs and utility values were obtained from published literature.
In its base case, the manufacturer reported that migalastat was dominant (more effective and less costly) compared with ERT (cost savings of $350,953; gain of 1.01 quality-adjusted life-years [QALYs]). Migalastat was also dominant in the scenario analysis for patients entering the model at age 18 years using all-male patient weights, and was associated with an incremental cost-utility ratio (ICUR) of $36,005 per QALY compared with ERT for patients entering the model at age 18 years using average patient weight from the ATTRACT trial.
Summary of Identified Limitations and Key Results
CADTH Common Drug Review (CDR) identified several limitations with the submitted analysis. First, the CDR Clinical Report indicates the clinical efficacy of migalastat is uncertain. In the placebo-controlled FACETS trial, migalastat did not meet its primary end point for the intention-to-treat population; and the comparative effectiveness of migalastat and ERT based on the ATTRACT trial was considered by the CDR clinical reviewers to be associated with uncertainty due to wide confidence intervals, concerns with imbalances in the trial populations, and use of surrogate outcomes as opposed to clinically meaningful outcomes (CDR Clinical Report). This uncertainty was not adequately captured in the model. Second, the use of a blended ERT comparator is not appropriate, particularly given the lack of information available regarding the market share of the comparator treatments. The assumption of disutility associated with ERT infusion is highly uncertain. The manufacturer assumed that ERT infusion was associated with an annual utility decrement of 0.048. This disutility value was based on a UK study that measured health utilities using a discrete choice experiment. It is unclear whether this utility estimate is generalizable to the Canadian population. Additionally, the use of a disutility for ERT infusion was not supported by the quality-of-life findings observed in the ATTRACT trial, which indicated negligible changes in health-related quality of life (as measured by the Short-Form 36-Item Health Survey [SF-36]). Furthermore, assumptions about how patients move within the model (e.g., transition probabilities between health states) do not reflect clinical evidence that suggests the risk of progression in Fabry disease increases with age; this results in an overestimation of life expectancy.4,5 Finally, the manufacturer did not test the impact of weight (in the probabilistic analysis), which influences ERT dosing as it is based on weight.
The model also included errors in the calculation of the annual migalastat cost, the weighted cost of ERT, and a disutility associated with dyspnea, which favour migalastat.
CDR attempted to address these issues, but was limited by the paucity of data, uncertainty associated with the available data, and appropriateness of the model structure. CDR conducted revised analyses that correct for errors in treatment costs and the disutility value of dyspnea, but was unable to assess the uncertainty associated with the comparative clinical effectiveness of migalastat and variation in transition probabilities.
CDR reanalyses, based on 10,000 simulations, included four different scenarios:
migalastat compared with agalsidase alfa with a disutility applied to ERT — ICUR = $200,487 per QALY
migalastat compared with agalsidase alfa without a disutility applied to ERT — ICUR = $55.9M per QALY
migalastat compared with agalsidase beta with a disutility applied to ERT — migalastat is dominant
migalastat compared with agalsidase beta without a disutility applied to ERT — migalastat is dominant.
As a stratified analysis of the ATTRACT trial based on the specific ERT was not provided, CDR was limited in the scope of the re-analysis for each comparator. Given the data limitations, the two ERTs were assumed to be equivalent, with the difference in cost of treatment the sole differentiator.
Additional scenario analyses were conducted to explore areas of uncertainty (treatment discontinuation, disutility associated with ERT infusion, and baseline patient weight). The results were highly sensitive to each of these parameters.
At the publicly available prices, a price reduction of 3.5% is required for migalastat to be less costly than agalsidase alfa. However, the clinical expert consulted by CDR indicated that the cost of ERT for different lysosomal diseases in Canada is currently under review. Any changes or differences in the total costs paid for ERT by the provinces to the costs used by the manufacturer will affect the cost-effectiveness results and subsequent price reduction analyses.
Conclusions
Based on the both the placebo-controlled trial (FACETS) and the comparative trial of migalastat and ERT (ATTRACT), the clinically meaningful effects of migalastat are associated with uncertainty (CDR Clinical Report). While there may be a preference for an oral treatment over an infusion, the QALY benefit associated with the oral treatment is likely to be overestimated in the model.
The cost of migalastat 123 mg every other day compared with ERT (dosed per product monograph recommendations) depends on the price of the comparator treatment and patient weight. Based on the submitted and publicly available prices only, in a patient weighing 75 kg, migalastat ($310,250) has a greater annual cost than agalsidase alfa ($299,821) and a lower annual cost than agalsidase beta ($312,186).
