TA906 · STA

Rimegepant for preventing migraine

Recommended with restrictionsCommittee DMay 2023

Rimegepant is recommended for preventing episodic migraine in adults who have at least 4 and fewer than 15 migraine attacks per month, if at least 3 preventative treatments have not worked. Treatment should be stopped after 12 weeks if the frequency of migraine attacks does not reduce by at least 50%. Where more than 1 treatment is suitable, the least expensive option should be chosen.

Source documents

Intervention

rimegepant (Vydura)
calcitonin gene-related peptide receptor antagonist · CGRP receptor antagonist · oral

Conditions

episodic migraineneurology_psychiatry · chronic
episodic migraineneurology_psychiatry · fourth-line preventative

Comparators

NameType Established Committee preferred
erenumabactive drugYes
fremanezumabactive drugYes
galcanezumabactive drugYes

Clinical trials

TrialDesignPhasePivotal
BHV3000-305RCT2/3Yes

Economic model

other (company)
Time horizon: 12 weeks decision tree phase plus Markov phase

ICER

Confidential (PAS-dependent) (rimegepant vs erenumab, fremanezumab and galcanezumab) · high uncertainty

Methodological decisions (8)

cost assumption

Healthcare resource use costs perspective: primary care vs secondary care. Company used primary care approach; ERG used secondary care approach

Company: Primary care approach reflecting potential for GP prescribing with one-off starting cost and 3-month follow-up cost plus neurologist referral cost for comparators

ERG: Secondary care approach. Not convinced by main rimegepant costs in primary care given committee's conclusions that specialist involvement needed. Clinical advice suggested neurologist follow-up more likely 6-monthly then yearly

Committee: Specialist costs in secondary care should be included alongside primary care costs. Rimegepant would most likely be started by a specialist due to its position in treatment pathway, possibly via shared care agreement for ongoing GP prescribing

ICER impact: increases

other

Company initially used BHV3000-305 trial outcome definition (50% reduction in moderate-to-severe MMDs in last 4 weeks); ERG preferred NMA definition (50% reduction in any severity MMDs as average over 12 weeks) for consistency across comparisons

Company: Response assessment at 12 weeks using trial definition from BHV3000-305 (moderate-to-severe MMDs, last 4 weeks)

ERG: Response should be defined as average over 12 weeks using any severity MMDs, consistent with NMA definition

Committee: Average over 12 weeks to ensure consistency across model inputs

ICER impact: uncertain_direction

population generalisability

Trial evidence excluded treatment-experienced patients but company proposed use after 3 failed treatments

Company: Evidence from comparator trials suggests rimegepant results may be conservative in refractory population

ERG: Evidence is uncertain and does not show substantial difference between refractory and non-refractory migraine

Committee: Clinical evidence from NMA not aligned with company positioning; took uncertainty into account in decision making

ICER impact: uncertain_direction

surrogate endpoint validity

Definition of response: proportion with ≥50% reduction in mean monthly migraine days compared with baseline. Company initially proposed endpoint at 12 weeks in last 4 weeks; ERG and committee preferred 12-week average across entire treatment period

Company: Response assessed at week 12 (last 4 weeks), based on clinical expert opinion that 85% of experts agreed on 12-week assessment

ERG: Response should be based on 12-week average over entire treatment period, for consistency with NMA definition and with trial data showing early improvements

Committee: 12-week average across entire treatment period for consistency across all model inputs

ICER impact: uncertain_direction

treatment effect duration

Timing of when NMA-derived comparative benefit should be applied in the economic model

Company: Option 1: Apply full 12-week benefit from baseline at week 4

ERG: Option 2: Use benefit observed before week 12 for responders, applied at week 4 using alternative regression allowing incremental improvements weeks 1-12

Committee: Option 2 was most appropriate method to apply, allowing for incremental improvements between weeks 1 to 12

ICER impact: increases

treatment effect duration

Timing of applying network meta-analysis (NMA) results in the economic model. Company initially applied full benefit from week 9-12; ERG and committee favoured earlier application

Company: Option 1: Apply full 12-week benefit from baseline at week 4. This was preferred by company in original base case

ERG: Option 2: Use benefit observed before week 12 for responders, applied at week 4. Allows for incremental improvements weeks 1-12 using alternative regression. ERG had concerns about MMD data for non-responders with option 1

Committee: Option 2, applying incremental benefits between weeks 1-12 rather than applying full benefit at week 4

ICER impact: uncertain_direction

utility source

Baseline EQ-5D utility values differed between treatment arms at baseline despite randomisation

Company: Used regression model with treatment arm and MMD as covariates; rimegepant baseline utility (0.6136) was higher than placebo (0.5976), p=0.1436

ERG: Baseline mapped EQ-5D scores should be included as covariate in regression to ensure baseline utilities are similar between arms

Committee: Baseline mapped EQ-5D values for each treatment arm should be the same; agreed with ERG's approach

ICER impact: increases

utility source

Baseline utility values: Company derived utilities by mapping MSQ-v2 from BHV3000-305 trial to EQ-5D using regression adjusted for treatment arm and MMD count

Company: Baseline utility values at 0.6136 for rimegepant (n=348) and 0.5976 for placebo (n=346), difference not statistically significant (p=0.1436). Include treatment arm in regression model

ERG: Baseline utilities favoured rimegepant which could bias downstream comparisons. Prefer including baseline mapped EQ-5D as covariate to ensure treatment arms have same baseline utility

Committee: Baseline mapped EQ-5D values for each treatment arm should be the same; agreed with ERG's approach of including baseline scores as covariate

ICER impact: decreases

Evidence gaps

no direct comparisonNo direct trial evidence comparing rimegepant with erenumab, fremanezumab or galcanezumab; indirect comparison through network meta-analysis used instead
short follow upClinical trial assessment period was 12 weeks

Commercial arrangement

none · confidential

Special considerations

Innovation acknowledged Equality issues raised

Cross-references

same drug — Rimegepant for acute treatment is being evaluated separately