Teriflunomide for treating relapsing–remitting multiple sclerosis

Clinical effectiveness

4.4 The Committee discussed the clinical-effectiveness evidence from the trials. The Committee understood that the marketing authorisation included all people with relapsing–remitting multiple sclerosis, but that the trials included people who had 1 relapse in the last year or 2 in the last 2 years, and therefore were patients with more severe relapsing–remitting multiple sclerosis than the general population with relapsing–remitting multiple sclerosis; that is, they had active disease. It heard from the clinical specialists that the trial populations broadly represented patients who would be offered beta interferon or glatiramer acetate in the UK, in line with the Association of British Neurologists' guidelines. The Committee therefore concluded that the trial populations appropriately represented the decision problem because these would be the people who would receive treatment with disease-modifying treatment in the UK.

4.5 The Committee agreed that there was sufficient evidence to conclude that, compared with placebo, teriflunomide statistically significantly reduced annualised relapse rates in both the TEMSO and TOWER trials and the meta-analyses (see section 3.4), and that the proportion of people who experienced 3‑month sustained accumulation of disability (SAD) was reduced with teriflunomide compared with placebo and that this difference was statistically significant in the TEMSO trial and in the meta-analysis (see section 3.4). The Committee agreed, however, that there was no statistically significant difference between teriflunomide and placebo in 6‑month SAD in either of the placebo-controlled trials (see section 3.4). The Committee concluded that teriflunomide was clinically effective in reducing relapse rates compared with placebo, and that teriflunomide may have a beneficial impact on accumulation of disability compared with placebo.

4.6 The Committee discussed the appropriateness of reporting 3‑month SAD rather than 6‑month SAD. It heard from the clinical specialists that recovery from relapse may continue for up to 12 months, but on average, recovery from the disabling effects of a relapse will be seen within 3 or 4 months. The clinical specialists stated that 6‑month SAD is therefore a more robust outcome measure than 3‑month SAD for measuring disability progression. The Committee was also aware that sustained disability progression confirmed for 6 months was preferred by the European Medicines Agency in its draft guideline for the clinical investigation of medicinal products for the treatment of multiple sclerosis. The Committee was aware that most of the multiple sclerosis trials measured 3‑month SAD. It acknowledged that this outcome had been considered in previous multiple sclerosis appraisals and that it therefore should be considered in its decision-making. However, the Committee concluded that, although sustained disability progression confirmed for 6 months provides a more robust indication of the treatment effect when measuring disability progression, in light of the lack of 6‑month SAD data available, 3‑month SAD data should be considered.

4.7 The Committee discussed the mixed treatment comparison (MTC) presented by the manufacturer that compared teriflunomide with the beta interferons and glatiramer acetate. It noted that the manufacturer had presented a base-case MTC, which had excluded trials that recruited patients before the year 2000 ('post-2000'), and one that included all trials ('all years') (see section 3.6). The Committee acknowledged that trials carried out before the year 2000 were excluded because of changes in diagnostic criteria, which has resulted, in part, in changes in baseline relapse rates over time, but were concerned that important trials were excluded because of the cut-off date, including all trials comparing beta interferons with placebo (see section 3.29). The Committee agreed that an MTC should include all available evidence and that, in this case, adjusting the MTC for baseline relapse rates would account for any differences in relapse rates between trials. The Committee noted that the results from the adjusted 'all years' MTC, provided in response to consultation, were similar to those from the 'all years' MTC (see section 3.43). The Committee concluded that the adjusted MTC was the most appropriate and that, based on the MTC, there was no difference in effectiveness between teriflunomide and the beta interferons or glatiramer acetate (see section 3.43).

4.8 The Committee discussed the TENERE trial, which did not show any statistically significant differences in annualised relapse rate between teriflunomide and the active comparator Rebif‑44 (interferon beta-1a, 44 micrograms) (see section 3.5). The Committee heard from the manufacturer that the primary outcome of the trial was time to failure and that the trial was not a non-inferiority or equivalence trial. It had been powered to test the hypothesis that people will stay on an oral drug with a different side-effect profile longer than on an injectable drug. The Committee noted that the TENERE trial was designed to show the benefits of an oral drug and had not been designed to compare the effectiveness of teriflunomide with Rebif. Therefore, the Committee concluded that the effectiveness of teriflunomide compared with Rebif‑44 was still uncertain, and that the MTC results did not show a difference in effectiveness.

4.9 The Committee discussed the evidence for the effectiveness of teriflunomide in patients with rapidly evolving severe relapsing–remitting multiple sclerosis or highly active relapsing–remitting multiple sclerosis, and the comparisons with natalizumab and fingolimod, respectively. The Committee noted that the manufacturer had stated in its submission that these were not key comparators for teriflunomide. The Committee noted that the MTC results favoured natalizumab and fingolimod for the outcome annualised relapse rate, and that no difference was found between teriflunomide and the 2 drugs for 3‑month SAD. However, the Committee agreed that the MTC results, being based on the whole trial population, did not provide relevant information for the subgroups. The Committee concluded that the indirect comparisons carried out for the subgroups could not be considered reliable because of the small patient numbers and the inadequate methodology used (see sections 3.10 and 3.42). It therefore concluded that the evidence was insufficient to make any conclusions about the clinical effectiveness of teriflunomide compared with natalizumab and fingolimod in the respective subgroups. It noted that the clinical specialists confirmed that teriflunomide would not be routinely used in people with aggressive disease (see section 4.3). The Committee also noted that no comments to support the use of teriflunomide in these subgroups were received during consultation on the appraisal consultation document. The Committee concluded that the evidence presented by the manufacturer was insufficient to recommend teriflunomide for treating highly active relapsing–remitting multiple sclerosis or rapidly evolving severe relapsing–remitting multiple sclerosis.

4.10 The Committee was aware of the adverse effects associated with teriflunomide (see section 3.11). The Committee discussed the risk of teratogenicity with teriflunomide, and the long 'washout' period (2 years) needed for women to have stopped treatment before trying to conceive. It agreed that this was a particular concern because multiple sclerosis affects women of childbearing age. The Committee recognised, however, that none of the disease-modifying drugs are recommended in pregnancy, and therefore it was a concern for all the multiple sclerosis treatments (although the washout period with teriflunomide is longer). The Committee considered this a notable concern, but concluded that no additional monitoring for teriflunomide had been recommended than that already given for treatment with the disease-modifying therapies and therefore it would not need to be reflected in the modelling for teriflunomide.

Cost effectiveness

4.11 The Committee discussed the revised base case provided by the manufacturer, the ERG's critique of the manufacturer's additional analyses, and the comments from patient and professional groups, commentators, patient experts and clinical specialists received in response to the appraisal consultation document. The Committee understood the manufacturer's model to be structurally similar to models used in previous NICE technology appraisals. The Committee agreed that the manufacturer's revised base case reflected the Committee's preferred analyses. That is, it used the ERG's preferred scenario (see section 3.40), and the following changes:

4.12 The Committee discussed the external validation presented by the manufacturer in response to the appraisal consultation document. The Committee noted that the incremental cost-effectiveness ratios (ICERs) estimated by the manufacturer's revised model were substantially higher than those in NICE technology appraisal guidance 32 for most of the comparators, and that the manufacturer's model predicted slower disease progression and lower health-related quality of life than the UK risk sharing scheme (see section 3.52). The Committee noted the considerable uncertainty in the current analyses, and the analyses carried out for NICE technology appraisal guidance 32, and acknowledged that showing close convergence between the previous and present analyses was challenging. However, the Committee concluded that the manufacturer's model was sufficiently valid for decision-making in the current appraisal.

4.13 The Committee was aware from the original analyses that disease progression was the main driver of the ICERs; using the ERG's original preferred analyses, the ICER ranged from £6000 per QALY gained (deterministic, using the base-case MTC [post 2000]) to £107,000 per QALY gained (probabilistic, using the 'all years' MTC) for teriflunomide compared with glatiramer acetate. The Committee concluded that the most appropriate MTC to use was the 'all years' MTC adjusted for baseline relapse rate because it included all available studies and accounted for the differences in baseline relapse rates seen between studies. The Committee noted that this resulted in an ICER of £14,000 per QALY gained for teriflunomide compared with glatiramer acetate (as estimated by the ERG).

4.14 The Committee understood that withdrawal rates had a counterintuitive impact on the ICER, in that increasing the withdrawal rate reduced the ICER. The Committee noted from the ERG's exploratory analyses that, when the withdrawal rate from the 'all years' MTC rather than from the 'all years' MTC adjusted for baseline relapse was applied to the ERG's re-estimation of the manufacturers revised base-case analysis, the ICER for teriflunomide compared with glatiramer acetate increased from £14,000 to £23,000 per QALY gained (see section 3.50). In addition, when it was assumed that the withdrawal rates for glatiramer acetate and teriflunomide were equal, the estimated ICER was £33,000 per QALY gained. The Committee agreed that assuming withdrawal rates were equal was the most conservative approach, but that using the 'all years' MTC adjusted for baseline relapse rate to estimate withdrawal rates was appropriate.

4.15 The Committee discussed how the manufacturer's revised base case modelled the natural history of multiple sclerosis. It recognised that the placebo arms of the TEMSO and TOWER trials had been used to model the natural history of disease. The Committee noted that, in response to the consultation, the manufacturer had stated that using data from the placebo arms of the TOWER and TEMSO trials overestimated EDSS regression, especially at higher EDSS states, and that applying rates from the London Ontario data set for the higher EDSS states reduced the ICER. The Committee noted the inherent limitations associated with using the London Ontario data set to model the natural history of disease, namely, that it allowed only for movement to higher EDSS states, and that it reflected a cohort from the 1970s and 1980s. It agreed that it was appropriate for the model to allow movement to lower as well as to higher EDSS states, that is, to allow for the condition to both improve and get worse, which is in line with what is seen in clinical practice for the lower EDSS states. The Committee also heard from the clinical specialists that relapses are less likely to occur once patients are in higher EDSS states; and people are less likely to move to lower EDSS states. The Committee agreed that using the data from the placebo arms of the TOWER and TEMSO trials was appropriate for modelling disease progression.

4.16 The Committee heard from the clinical specialists that the health-related quality of life of people with multiple sclerosis was more closely related to EDSS state than to the clinical form of multiple sclerosis (that is, relapsing–remitting or secondary progressive). The clinical specialists stated that it is difficult to clearly identify when the condition becomes secondary progressive multiple sclerosis, and so it is also difficult to gauge the impact of progressing to secondary progressive multiple sclerosis on health-related quality of life. The Committee noted that, in the revised base case, the manufacturer had used clinical trial data to estimate utility values, where possible. For the higher EDSS states that could not be derived from clinical data, the manufacturer had extrapolated values using the differences in utility between EDSS states seen in Orme et al. (2007). The Committee agreed that this approach was appropriate.

4.17 The Committee discussed the disutility values incorporated in the model to reflect caregiver disutility, as in previous multiple sclerosis appraisals, and considered this to be appropriate. The Committee noted that the disutility values for adverse events included in the model were small (see section 3.15), and acknowledged the comment raised during consultation that a more extensive list of adverse events should be included in the model. The Committee understood that the disutility values did not have a large impact on the ICERs and therefore concluded that it did not need to consider the disutility values in the model further.

4.18 The Committee considered the duration of treatment benefit on disease progression, and noted that it remained constant over time in the manufacturer's original base case. That is, there was no option for the treatment benefit to decrease over time. It recognised that the long-term benefit was largely unknown but that it had a large impact on the outputs of the model. The Committee heard from the clinical specialists that they could not be confident that the treatment effect would not wane. The Committee acknowledged that, in the revised model, the manufacturer had included the assumption that the treatment effect decreased to 75% at 2 years and 50% at 5 years to explore a reduction of the long-term treatment effect. The Committee noted that applying treatment waning increased the ICER for teriflunomide compared with glatiramer acetate (from £10,000 to £13,000 per QALY gained, see section 3.45). The Committee agreed that it was important to include in the decision-making that the treatment effect could decrease over time but, given the uncertainty of how much the treatment effect would wane, the most plausible ICER was likely to lie between the estimates that included and excluded the modelled treatment waning effect.

4.19 The Committee considered the cost data used in the manufacturer's revised base case. It was concerned that the non-health costs contributed to a high proportion of the costs in the model, and it was unclear what proportion of these costs would include personal social services, and therefore would be appropriately included in the NICE reference case. The Committee acknowledged that the manufacturer had provided scenarios with and without non-health costs (with non-health costs: £13,000 per QALY gained compared with glatiramer acetate; without non-health costs: teriflunomide dominated glatiramer acetate). It recognised that the ERG's critique of the manufacturer's revised base case included a scenario using a cost midpoint, which increased the ICER slightly compared with including all non-health costs (deterministic ICER £3000 per QALY gained). The Committee understood that excluding all non-health costs was a conservative but arguably appropriate approach to adjusting the model's cost inputs to follow NICE's preferred perspective for analyses. The Committee concluded, however, that the most plausible ICER was likely to lie between the ICERs estimated with and without non-health costs, given the uncertainty about how much of the non-health costs from the cited sources were within the NICE reference case.

4.20 The Committee recognised that no specific sequence of disease-modifying treatments was standard practice in the NHS but that patients with active relapsing–remitting multiple sclerosis often receive more than 1 disease-modifying treatment over time. Therefore the Committee considered it important to explore how sensitive the ICERs were to the inclusion of more than 1 treatment. The Committee understood that teriflunomide may be added to the treatment pathway and, in this situation, the alternative treatment would be best supportive care. The Committee noted that the ERG's exploratory analysis comparing teriflunomide with best supportive care gave ICERs for teriflunomide of between £29,000 and £64,000 per QALY gained, depending on whether treatment waning or non-health costs were included. The Committee noted that these ICERs were lower than the ICERs for the disease-modifying treatments from the risk sharing scheme, when derived from the same model (the revised manufacturer's analysis). The Committee agreed that it was valuable to understand the impact on the ICER of including teriflunomide in a treatment sequence, or as an alternative to best supportive care (to account for adding to the treatment sequence), and agreed that for future technology appraisals in multiple sclerosis, a scenario in which a sequence is explored would be useful. However, the Committee concluded that the analysis of individual drugs (without a sequence) was the basis for decision-making in this appraisal because of:

4.21 The Committee discussed whether teriflunomide was innovative and noted the comments received during consultation on the appraisal consultation document. It recognised the limitations of the current treatments in terms of their side-effect profile and administration methods, and agreed that an oral treatment, with a different side-effect profile, would be beneficial for people with active relapsing–remitting multiple sclerosis. The Committee discussed whether there were health-related benefits associated with an oral treatment that were not captured in the modelling. It noted that the manufacturer's economic model assigned different disutilities for treatment-specific adverse events, including injection site reactions and that, therefore, some of the benefits of oral treatment had been captured, but possibly not all. The Committee concluded that teriflunomide was innovative and that additional health-related quality-of-life benefits related to the oral treatment may not have been captured fully.

4.22 The Committee considered the manufacturer's revised base-case results and the ERG's critique of these data, noting that teriflunomide dominated the beta interferons, and that the ICER for teriflunomide compared with glatiramer acetate was £14,000 per QALY gained. The Committee noted that excluding treatment waning reduced the ICER for teriflunomide compared with glatiramer acetate to £10,000 per QALY gained and that, when non-health costs were included, teriflunomide dominated the beta interferons and glatiramer acetate, irrespective of whether waning was included or excluded. The Committee acknowledged that when some non-health costs were included, as modelled by the ERG, the ICER for teriflunomide compared with glatiramer acetate was £3000 per QALY gained and that when withdrawal rates were assumed equal between teriflunomide and glatiramer acetate, as presented by the ERG, the ICER was £33,000 per QALY gained. The Committee concluded that teriflunomide dominated the beta interferons. For the comparison with glatiramer acetate, the Committee noted the varying ICERs from the different analyses and recognised that there were benefits not captured in the QALY, such as the oral administration of teriflunomide. The Committee concluded that, on balance, the most plausible ICER for teriflunomide compared with glatiramer acetate would be below £20,000 per QALY gained.

4.23 The Committee agreed there was insufficient evidence to appraise the cost effectiveness of teriflunomide compared with fingolimod for the highly active relapsing–remitting multiple sclerosis or with natalizumab for the rapidly evolving severe relapsing–remitting multiple sclerosis. The Committee therefore concluded that teriflunomide cannot be recommended for treating highly active relapsing–remitting multiple sclerosis or rapidly evolving severe relapsing–remitting multiple sclerosis. However, the Committee concluded that teriflunomide could be considered an effective use of NHS resources for treating relapsing–remitting multiple sclerosis in adults for whom beta interferons and glatiramer acetate would otherwise be considered as treatment options, that is, adults who have active relapsing–remitting multiple sclerosis, normally defined by 2 clinically significant relapses in the previous 2 years, and who do not have highly active relapsing–remitting multiple sclerosis or rapidly evolving severe relapsing–remitting multiple sclerosis, and only if the manufacturer provides teriflunomide with the discount agreed in the patient access scheme.

Summary of Appraisal Committee's key conclusions