Fremanezumab for preventing migraine

3.1 Migraine attacks usually last between 4 hours and 72 hours and involve throbbing head pain of moderate to severe intensity, which can be highly disabling. The patient experts explained that they are often accompanied by nausea, vomiting, sensitivity to light, sensitivity to sound or other sensory stimuli, numbness, confusion, loss of concentration and speech issues. Migraine can adversely affect quality of life, affecting people's ability to do their usual activities, including work. Some people with migraine have severe depression and suicidal thoughts. All of these can slow personal and professional development so that people feel they have unachieved potential. Chronic migraine is defined as 15 or more headache days a month with at least 8 of those having features of migraine. Episodic migraine is defined as less than 15 headache days a month. A clinical expert explained that the severity of the condition can vary over time. The committee concluded that migraine, particularly chronic migraine, is a debilitating condition that substantially affects both physical and psychological aspects of quality of life and employment.

3.2 The clinical experts explained that the aim of treatment is to reduce the frequency, severity or duration of migraine and improve quality of life. The committee was aware that in chronic migraine, a 30% reduction in migraine frequency is considered a clinically meaningful response to treatment. In episodic migraine, a 50% reduction is considered a clinically meaningful response. If clinical response is less than this, or the person is not able to have an adequate dosage for long enough or has adverse events, treatment is stopped and another oral preventive treatment is tried. The clinical experts explained that it is important for people to try a range of oral preventive treatments before considering more specialist treatment, such as botulinum toxin type A (for chronic migraine) or fremanezumab. A clinical expert noted that at least 5 different oral preventive treatments were available for migraine but that not all of these would be tried before offering fremanezumab. The clinical experts agreed that fremanezumab would usually be offered after 3 failed oral preventive treatments. This was because there was no clear evidence that using oral preventives after this was of benefit, and side effects may outweigh any benefits. The committee understood that some clinicians may choose to offer a fourth or fifth oral preventive before offering more specialist treatments. It concluded that an adequate trial of at least 3 oral preventive treatments represents usual NHS practice before more specialist treatment is considered. It further concluded that a clinically meaningful response was a 30% reduction (for chronic migraine) or a 50% reduction (for episodic migraine) in migraine frequency.

3.3 The company's submission focused on people with migraine for whom at least 3 previous preventive treatments had failed (defined as lack of a clinically meaningful response, intolerance to the treatment or the treatment was contraindicated or unsuitable). The company considered that fremanezumab would likely be used in NHS clinical practice at this point because of the unmet need for additional treatment options after 3 preventive treatments had failed. The company presented evidence for fremanezumab's clinical effectiveness compared with placebo for episodic migraine and compared with placebo and botulinum toxin type A for chronic migraine. The company considered that placebo was representative of best supportive care, because it comprised acute treatments that people would have for their migraine symptoms when preventive treatments had not worked. The committee recalled its discussion about using oral preventive treatments after 3 had failed (see section 3.2) and agreed that best supportive care was the most appropriate comparator in episodic migraine. It recognised that best supportive care would not reduce the frequency or severity of migraine and would increase the risk of medication overuse headache. It also recalled that botulinum toxin type A is recommended for people with chronic migraine whose condition has not responded to at least 3 prior preventive therapies. However, the committee was also aware that some people who are eligible for botulinum toxin type A are unable to have it or have to wait a long time for it. This is because few UK clinics are offering this treatment, and there are long waiting lists for it. The committee concluded that both botulinum toxin type A and best supportive care were relevant comparators in chronic migraine.

3.4 The company defined high-frequency episodic migraine as between 8 and 14 monthly headache days. The ERG noted that the company's high-frequency episodic migraine definition was not in line with other definitions in the literature (10 to 14 and 11 to 14 monthly headache days), highlighting that there was no consensus on the definition. The clinical experts explained that there is no internationally recognised classification of high-frequency episodic migraine and that it is not a clearly defined clinical subgroup. They also noted that the definition of high-frequency episodic migraine is arbitrary, and a person's quality of life is negatively affected irrespective of which type of migraine they have. The committee concluded that high-frequency episodic migraine is not a distinct subgroup and agreed not to consider it further.

3.5 The company's systematic literature review identified 3 double-blind randomised controlled trials evaluating fremanezumab:

3.6 The committee considered the generalisability of FOCUS to NHS clinical practice. FOCUS excluded people who had the most severe, unremitting headaches, clinically significant comorbidities or clinically significant psychiatric issues. Therefore, it agreed that people enrolled in FOCUS were on average healthier than people who may be eligible for fremanezumab in clinical practice. The committee also considered whether inadequate treatment response, as defined in FOCUS, reflected what would be considered treatment failure in clinical practice. FOCUS defined an inadequate treatment response as a lack of clinically meaningful improvement after at least 3 months of therapy, intolerance to the treatment or the treatment was contraindicated or unsuitable. The clinical experts explained that a contraindication would not necessarily represent a treatment failure. But the company clarified that only about 2% of recorded treatment failures were because of a contraindication. The committee also noted that some people may have had a clinically meaningful response to an oral preventive treatment before stopping because of adverse events. It also noted that valproic acid was not frequently used in the UK for migraine prevention, although about 1 in 3 people in FOCUS had previously had it. The committee concluded that FOCUS may not fully reflect those eligible for fremanezumab in clinical practice.

3.7 The committee understood that people with chronic migraine in the 225 mg monthly fremanezumab treatment group in both FOCUS and HALO CM had a 675 mg loading dose. It considered whether this loading dose could bias the clinical-effectiveness results for this group. The company noted that a loading dose was not included in fremanezumab's marketing authorisation because the 675 mg quarterly and 225 mg monthly dosages have equal efficacy. It also noted that having no loading dose simplified dosing, therefore benefiting patients and clinicians. The committee concluded that differences in dosing between the FOCUS and HALO CM trials and the marketing authorisation would not likely affect the generalisability of the results to clinical practice.

3.8 The company presented clinical-effectiveness results from FOCUS for the subgroup of people for whom 3 or 4 preventive migraine therapies failed to produce clinically meaningful improvement, were not tolerated, or were contraindicated or unsuitable. The baseline to week 12 subgroup results from FOCUS showed:

3.9 The duration of the blinded phase in the trials was 12 weeks for FOCUS, HALO EM and HALO CM. The company provided supporting data for fremanezumab's long-term effectiveness from the uncontrolled open-label HALO extension study. The committee recalled that the population in the HALO studies was less relevant than the population in FOCUS to the population of interest (see section 3.5), but acknowledged that no long-term evidence was available from FOCUS. People who had fremanezumab in HALO EM and HALO CM had the option to continue on a stable dose in the extension study, whereas those who had placebo could opt to be randomly assigned to either 675 mg fremanezumab quarterly or 225 mg monthly (with a 675 mg loading dose in HALO CM). The committee recognised that although the HALO extension study provided some longer-term clinical-effectiveness evidence for people having fremanezumab, comparative effectiveness could not be estimated because the extension study did not include a placebo group. The committee recognised that because not everyone in the trials continued to the extension phase, there was an additional risk of bias. This was because it considered that people not having benefit were more likely to drop out. The company said that the results suggested that treatment effectiveness was maintained long term with no evidence of waning. It noted similar results for people who previously had fremanezumab in HALO EM and HALO CM to those who had previously had placebo, and consistency in results between the 2 fremanezumab dosages (675 mg quarterly and 225 mg monthly). These results were considered academic in confidence by the company and cannot be reported here. The committee concluded that the long-term benefits of fremanezumab compared with best supportive care remained uncertain.

3.10 Clinical experts advised that botulinum toxin type A is ineffective for about 1 in 3 people with chronic migraine, based on real-world studies. The committee recalled that this group of people has high unmet need because of high disease burden and no further treatment options. At consultation, the company submitted additional evidence on the clinical effectiveness of fremanezumab compared with best supportive care for people with chronic migraine after 3 preventive treatments and botulinum toxin type A have failed. The evidence was based on a post-hoc subgroup analysis of FOCUS. The baseline to week 12 results showed improvements in key outcomes (as listed in section 3.8), which were in line with the results for the subgroup of people for whom 3 or 4 preventive migraine therapies failed (see section 3.8). They were considered academic in confidence by the company and cannot be reported here. The ERG highlighted the small population size in this analysis. It noted that the clinical-effectiveness evidence was weak and should be treated with caution. The company also submitted supporting analyses for 2 additional slightly larger populations:

3.11 There was no direct evidence comparing fremanezumab with botulinum toxin type A for chronic migraine. So the company did an indirect comparison using data from:

3.12 Health-related quality-of-life data was collected in FOCUS using the Migraine-Specific Quality of Life Questionnaire (MSQ) and the EQ‑5D‑5L. The committee was aware that in NICE's reference case and current position statement on the EQ‑5D‑5L, the EQ‑5D‑3L is preferred for base-case analyses. The company considered that the EQ‑5D‑5L was not sensitive to changes in quality of life with migraine because the questionnaire had to be completed on appointment days. This meant that it only captured quality-of-life data for people who were able to attend appointments. A person having a migraine on the day of their appointment would likely rearrange it and the effect of that migraine on quality of life would not be captured. The clinical experts explained that in clinical practice, they use the Headache Impact Test (HIT6) and Migraine Disability Assessment Test (MIDAS) to measure quality of life, so it was not known whether MSQ was the best available measure of quality of life. The company highlighted that the MSQ included a 4‑week recall period, which ensured the effect of migraine on quality of life was captured. The committee concluded that the rationale for using MSQ data was reasonable because the EQ‑5D‑5L was not sufficiently sensitive to changes in quality of life caused by migraine.

3.13 The company modelled the assessment period of 12 weeks (24 weeks for botulinum toxin type A) as a decision tree, and the post-assessment period as a Markov model. Episodic and chronic migraine were analysed separately, with each analysis using a dedicated set of input parameters. In the decision tree phase, people were grouped into those whose migraine:

3.14 The company's base-case model included a time horizon of 10 years. The company explained that it expected all meaningful differences in costs and quality-adjusted life years (QALYs) between treatments to be captured in this time horizon. It also noted that because there is no long-term natural history data, any long-term modelling beyond 10 years would be highly uncertain. The ERG highlighted that a time horizon of 10 years was a problem for predicting long-term safety and efficacy. However, it agreed with the company that extending the time horizon increased the uncertainty in extrapolating short-term evidence, and because of this it considered 10 years to be a reasonable time horizon. The committee understood that extending the time horizon could increase the uncertainty. But it noted that arbitrarily capping the time horizon could also increase uncertainty because long-term costs and benefits were not captured. It acknowledged that, although the average age of the subgroup from FOCUS was over 40 years, people much younger than this would have treatment in clinical practice. Therefore, it agreed this should be taken into account in the model time horizon. The committee concluded that it preferred a lifetime time horizon to ensure that all relevant costs and benefits associated with fremanezumab were captured. For the rapid review, the company used a lifetime time horizon in its updated model.

3.15 The company's model included a separate health state for people who stopped treatment. The discontinuation rate applied after each model cycle (4 weeks) was based on the number of people on fremanezumab who dropped out of the HALO extension study. The committee considered that the discontinuation rate (from all causes) was relatively high for what it understood to be a clinically effective and well-tolerated treatment. The ERG noted that the discontinuation rate in the HALO extension study was higher than that seen in the extension studies of another anti-calcitonin gene-related peptide (CGRP), erenumab (NICE has published guidance on erenumab for preventing migraine). The clinical experts noted that the additional injections given in the HALO trials to preserve the blinding of treatment allocation could explain why more people dropped out. The patient experts highlighted that most people would tolerate injections if the treatment was effective. The committee agreed that additional injections alone were unlikely to explain the higher-than-expected discontinuation rates. It also noted that the HALO extension study from which the discontinuation rate was calculated was an open-label study. This meant that treatment allocation was not blinded, so additional sham injections would not be necessary. It acknowledged that because treatment costs stop after discontinuation, an inflated discontinuation rate would affect the cost-effectiveness results. The committee concluded that the discontinuation rate was higher than expected and this could affect the cost-effectiveness results.

3.16 The ERG explained that assuming migraine frequency would revert to that of best supportive care after discontinuation from all causes was overly optimistic. This is because the migraine frequency of people having best supportive care was determined by the response to placebo in the clinical trials. It noted that this response was similar to that of people on fremanezumab. It also considered it unrealistic that a substantial treatment effect would be maintained indefinitely for people who are no longer having fremanezumab treatment. The clinical experts highlighted that there was no long-term evidence in people who have stopped treatment, but agreed that it seemed implausible that a substantial treatment benefit would be maintained. The committee agreed that this assumption was overly optimistic because an implausibly large benefit was maintained and costs were stopped. To account for this, the ERG did a scenario analysis. In this, people reverted to baseline migraine days after fremanezumab discontinuation (from all causes), and the treatment effect for people whose migraine responded to best supportive care diminished to baseline over 1 year. The committee agreed that this scenario was more in line with how the clinical experts expected treatment effectiveness could change after stopping treatment. The committee concluded that the company's post-discontinuation assumptions were overly optimistic. It agreed that it would consider the ERG's scenario in which people revert to baseline monthly migraine days after stopping fremanezumab, botulinum toxin type A or best supportive care.

3.17 The company's model included a negative stopping rule. So in the model, people whose migraine did not respond to treatment (a reduction in monthly migraine days from baseline of less than 50% for episodic migraine or less than 30% for chronic migraine) stopped treatment after assessment at 12 weeks (24 weeks for botulinum toxin type A). The committee concluded that it was appropriate to include a negative stopping rule at 12 weeks in the economic model if there was no response to treatment. It accepted the company's approach to modelling this. It agreed that any treatment benefit seen while on treatment (during the initial 12 weeks) would not be maintained after stopping the treatment.

3.18 The company's model applied a positive stopping rule by assuming 20% of people whose migraine responded to treatment would discontinue every 64 weeks (52‑week treatment period and 12‑week response assessment). After this period, treatment effect was maintained, but treatment costs were stopped indefinitely. The patient expert explained that, from their own experience, once fremanezumab was stopped, the benefit was maintained for only a short time before migraines returned to their pretreatment frequency and severity. The committee recalled that there was a lack of long-term effectiveness evidence for fremanezumab in the population of interest (see section 3.9). It recognised that there was no evidence but agreed it was unrealistic to assume that the treatment effect would be maintained indefinitely after stopping treatment. It also noted that any report of long-term treatment effectiveness could be affected by natural variation in the condition. The committee acknowledged that without long-term natural history data, this could not be fully understood. In response to consultation, the company revised its positive stopping rule. The proposed new stopping rule assumed that 15% of people whose migraine responded to treatment would stop every year. It also assumed that once treatment stopped, migraine frequency returned to pretreatment levels within 1 year. The committee was aware that clinical experts considered that successful treatment with fremanezumab would not be continued indefinitely (see section 3.2). The company did not provide any positive stopping criteria for chronic or episodic migraine. However, the committee noted clinical expert comments at consultation suggesting that people with chronic migraine would stop treatment when their migraines had reduced to 10 migraine days a month for at least 3 months. The committee was aware that no comments were received about positive stopping criteria in episodic migraine. Therefore, taking account of what it had heard from the clinical experts, the committee considered that there are no clear criteria for when people should stop treatment and understood that a positive stopping rule could be challenging to implement in clinical practice. It recognised that people may not be willing to stop treatment that is beneficial for them. It also recalled that no positive stopping criteria were used in FOCUS. Therefore, the committee concluded that it was not appropriate to apply the company's positive stopping rule in the model. But it acknowledged that treatment may not continue indefinitely after successful treatment and took this into account for decision making.

3.19 The utility values used in the model were generated from mapping MSQ results to the EQ‑5D‑3L using the Gillard et al. (2012) algorithm. The committee understood that the MSQ data was based on the full trial population, and not just on those for whom 3 or 4 treatments had failed. It also understood that the patient characteristics could not be included in the mapping algorithm because of data limitations. It agreed that this could limit the robustness of the mapped EQ‑5D‑3L utility values used in the economic model. It also noted concerns about the reliability of the utility values given the uncertainty of using data from the broader, full trial population. It noted that the mapping algorithm led to implausibly low utility values for certain patient populations, as confirmed by the clinical experts. The ERG explained that the company did not provide enough detail on how the utility values were extracted and mapped to assess their robustness. The committee concluded that the company's approach to calculating model utility values was reasonable but noted that the values were uncertain because of data limitations.

3.20 After mapping from MSQ to EQ‑5D‑3L, the company split the EQ‑5D utility values into 'on‑treatment' and 'off‑treatment' groups. Off‑treatment health state utility values were estimated using baseline (week 0) MSQ data. On‑treatment utility values were estimated from the week 4 and week 12 MSQ data. Off‑treatment utility values were applied for people on best supportive care and on‑treatment utility values were used for fremanezumab and botulinum toxin type A until people stopped treatment. The company highlighted that on‑treatment utility value benefits have been shown for people with migraine. It noted that applying treatment-specific utility values was consistent with the appraisal for botulinum toxin type A for chronic migraine. The ERG noted that the on‑ and off‑treatment utilities were not appropriately generated and applied in the model. It considered the company's approach was overly simplistic and did not account for possible improvements in quality of life related to being included in a clinical trial (placebo effect). It also explained that the on‑treatment utilities were not correctly applied in the model because of how the model was structured. The committee recalled that utility values were generated from MSQ data, which measured the impact of migraine on daily social and work-related activities, and emotional functioning. Therefore, it agreed that it was uncertain whether health-related quality-of-life benefits beyond those related to reducing monthly migraine days were not already adequately captured by the MSQ. It also noted that baseline (before treatment) fremanezumab utility values included a benefit over best supportive care, which it agreed was inconsistent with applying an on‑treatment utility value benefit. At the time of the original guidance (TA631; June 2020), the committee concluded that the company's additional on‑treatment utility value benefits should not be included in the economic model.

3.21 Since the publication of the original guidance for this topic (TA631), NICE has published guidance on galcanezumab for preventing migraine and erenumab for preventing migraine. Both are anti-CGRP drugs for migraine in adults, like fremanezumab. In these appraisals, the committee was satisfied that it had seen enough evidence to support using differential utilities for the treatment being appraised. That is, using different utilities for people on and off fremanezumab, as a result of taking into account improvements in utilities beyond a reduction in the monthly migraine days measure. For the rapid review of fremanezumab, the company presented evidence to support using differential utility values in its economic modelling. It focused on episodic migraine (specifically when 3 preventive treatments had failed) because fremanezumab was not recommended for episodic migraine in TA631. The company's clinical experts advised that there are improvements in utilities beyond a reduction in the monthly migraine days measure, which does not fully capture the duration and severity of migraines, and other factors that influence quality of life. The company reasoned that quality of life is affected by migraine between attacks, not just during them, so quality-of-life impairment could extend beyond monthly migraine days alone. Using differential utilities reflects further treatment benefits, for example improvements in photophobia and disability levels, less severe migraine attacks, and quicker recovery time. The company explained that a migraine-specific quality-of-life questionnaire was used, which found a significant improvement in the different domains for fremanezumab compared with placebo, in the full FOCUS population and the population for whom 3 or more treatments had failed. The committee noted that a reduction in the monthly migraine days measure would not capture the benefits to people with migraine of being able to plan their lives and remain in employment. It acknowledged that better day-to-day functioning overall would lead to increased quality of life and utility values, and that a similar rationale had been accepted by the committee in other appraisals.

3.22 The company's updated analyses for episodic migraine with 3 or more previous treatment failures used differential utilities, and the committee's preference for a normal distribution as in the NICE guidance on galcanezumab and erenumab. The company explained that in its regression analysis, fremanezumab treatment was a significant covariate in its preferred post-baseline model, and that its approach accounted for any placebo effect in the data. It did some face validity checks on the revised utility values generated by the new approach. The ERG believed the company's regression analysis seemed reasonable and consistent with previous migraine appraisals. However, the company did not provide a detailed statistical analysis plan. Because fremanezumab appeared to be a significant covariate, the ERG interpretation was that this may indicate fremanezumab has a utility benefit beyond reducing monthly migraine days alone. The ERG pointed out the similarities between the regression model used in the rapid review and those in the appraisals of galcanezumab and erenumab, which were used to justify the choice to use differential utilities for when people were on and off treatment. The ERG considered that the company's approach in its updated analysis addressed the limitations of previous regression models and used separate models for baseline and post-baseline quality-of-life data. The committee noted that the company should have provided a full statistical analysis plan alongside its updated analyses. It considered that fremanezumab may have a utility benefit beyond fewer monthly migraine days. So the committee concluded it was appropriate to take into account the company's updated analyses using differential utilities for its decision making in the rapid review.

3.23 The company based its resource use estimates on data from a European study of migraine burden by Vo et al. (2018). It noted a limitation of the study was that resource use estimates were based on monthly headache days, not migraine days, which it considered could underestimate the migraine cost burden. In the model, it assumed that resource use would be equivalent for both fremanezumab dosage schedules; monthly injections of 225 mg or 3 injections of 675 mg every quarter. The ERG noted that this could be a conservative assumption because quarterly administration is likely to be less resource intensive. The ERG also noted that resource use rates were not specific to the population of interest (that is, people who have had at least 3 failed preventive treatments) but based on the general migraine population. At consultation, the company submitted a revised base-case model, which included updated administration costs for botulinum toxin type A (£125 per administration as per the 2016/17 tariff; the corresponding cost for 2018/19 should be £116 but this has very limited impact on model results), as estimated by NICE and NHS England in their budget impact analysis. The committee concluded that, despite the limitations in the estimates of resource use, the costs included in the model were appropriate for decision making.

3.24 The company assumed that fremanezumab could be self-administered by subcutaneous injection. At the technical engagement stage, the clinical experts suggested that most people would be capable of self-administering fremanezumab. However, 1 expert noted that disabled people or people with a learning disability, older people and those who have a phobia of needles may need help. They also noted that additional services may be needed to train people how to administer treatment. The committee concluded that it was unlikely that everyone having fremanezumab would be capable of self-administering treatment. It agreed that applying administration costs for 10% of people having fremanezumab was reasonable, but acknowledged that this had little effect on the model results.

3.25 NICE's guide to the methods of technology appraisal notes that above a most plausible incremental cost-effectiveness ratio (ICER) of £20,000 per QALY gained, judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of certainty around the ICER. The committee will be more cautious about recommending a technology if it is less certain about the ICERs presented. The committee noted a high level of uncertainty, specifically:

3.26 The committee recalled that it had concluded that people with high-frequency episodic migraine were not a distinct group, so it should not consider the cost-effectiveness analysis for this population further (see section 3.4). The company's revised base-case ICER for fremanezumab compared with best supportive care for episodic migraine was below the range NICE normally considers an acceptable use of NHS resources. For the rapid review, the company's revised base case included the committee's preferred assumptions from the time of the original guidance (TA631; June 2020):

3.27 The committee was aware that many uncertainties associated with the episodic migraine evidence also applied to the chronic migraine evidence. However, it was aware that the eligible population and impact on NHS resources would be smaller for chronic migraine than for episodic migraine (see section 3.25). It also recalled that consultation comments from professional organisations and patient groups specifically highlighted the unmet need for the chronic migraine population. It therefore considered that an acceptable ICER would be within the range normally considered cost effective. The company's revised base-case fully incremental ICERs for fremanezumab compared with both best supportive care and botulinum toxin type A were within this range. The company's revised base case assumed the comparative effectiveness estimates from the network meta-analysis (see section 3.11) and the committee's preferred assumptions:

3.28 The committee recalled that for the whole chronic migraine population, including the population for whom botulinum toxin type A had failed, fremanezumab was likely to be a cost-effective use of NHS resources for preventing chronic migraine after 3 preventive treatments had failed (see section 3.27). The committee considered the company's additional evidence submitted at consultation for people with chronic migraine only after 3 preventive treatments and botulinum toxin type A have failed. The committee concluded that the most plausible ICERs for fremanezumab compared with best supportive care were within the range normally considered a cost-effective use of NHS resources, using both company and committee preferred assumptions.

3.29 The company and clinical and patient experts highlighted that migraine can be classed as a disability under the Equality Act (2010). Because migraine is most common in people of working age and affects more women than men, women may be disadvantaged in the workplace. In addition, there may be unequal access to specialist headache clinics in England. The committee considered these issues but concluded that no specific adjustments were needed to NICE's methods in this situation. No other equalities issues were raised after the original guidance for this topic was published (June 2020).

3.30 The company suggested that fremanezumab should be considered as an innovative treatment on the grounds that anti-CGRP therapies represent a step change in the management of migraine. The committee concluded that the modelling had adequately captured the benefits of fremanezumab. But it also acknowledged that fremanezumab administration may be considered more convenient and less unpleasant than administration of botulinum toxin type A.

3.31 The committee recognised the substantial burden that migraine has on quality of life and day-to-day functioning. It acknowledged that this could lead to psychosocial issues (see section 3.1). The committee recalled that the most relevant comparators for chronic migraine were botulinum toxin type A and best supportive care (see section 3.3). It considered that fremanezumab was a clinically effective treatment compared with placebo (see section 3.8). However, the committee considered that it was uncertain whether fremanezumab was more clinically effective than botulinum toxin type A and agreed that it was appropriate to also consider equal effectiveness (see section 3.11). It considered the revised base case provided by the company at consultation, which included a confidential simple discount patient access scheme for fremanezumab. The committee noted that the most plausible ICER for fremanezumab compared with best supportive care was within the range NICE usually considers a cost-effective use of NHS resources. Although there was uncertainty around the relative treatment effects of fremanezumab and botulinum toxin type A, the committee considered it likely that fremanezumab was a cost-effective use of NHS resources. Therefore, the committee recommended fremanezumab for use in the NHS for preventing chronic migraine in adults after 3 preventive treatments have failed. This includes the chronic migraine population for whom treatment with botulinum toxin type A has failed. Treatment with fremanezumab should be stopped if migraine frequency does not reduce by at least 30% after 12 weeks of treatment.

3.32 The committee recalled that the most relevant comparator for episodic migraine was best supportive care. It considered that the evidence showed that fremanezumab was clinically effective when compared with best supportive care. It also considered that the evidence in high-frequency episodic migraine was uncertain and did not consider it further because it is not a distinct subgroup. For the rapid review, the company presented new evidence supporting the use of differential utilities on and off treatment, specifically focusing on the population with episodic migraine after 3 preventive treatments have failed. It also submitted an updated commercial arrangement. The committee noted that the company should have provided a statistical analysis plan for the regression carried out. However, all the ICERs for scenarios using the updated model with differential utilities were less than £20,000 per QALY gained, and a scenario without differential utilities resulted in an ICER within the range usually considered an acceptable use of NHS resources. Therefore, the committee recommended fremanezumab for use in the NHS for preventing episodic migraines in adults after 3 preventive treatments have failed. Treatment with fremanezumab should be stopped if migraine frequency does not reduce by at least 50% after 12 weeks of treatment.