Fingolimod (2) – Gilenya®

• Clinical trials
  • In its dossier, the pharmaceutical manufacturer refers to the directly comparative TRANSFORMS trial. The study design of this trial is multicentre, randomised, double-blind, parallel-group and actively controlled (double-dummy).
  • This trial included a total of 1,292 patients with relapsing-remitting multiple sclerosis who had experienced at least one relapse in the past year or two relapses in the preceding two years.
  • The study comprised three treatment arms: - fingolimod 0.5 mg (431 patients) - fingolimod 1.25 mg (426 patients) – INF-β 1a 30 µ intramuscularly (435 patients).

a) Patients with highly active relapsing-remitting multiple sclerosis with high disease activity who had not responded to a complete and appropriate course of treatment – normally lasting at least one year – course of at least one disease-modifying therapy (duration of prior treatment with at least one disease-modifying therapy (in this case: other than INF-β) ≥ 1 year)

• For patient population a), therefore, no additional benefit can be identified when considering the patient-relevant endpoints as a whole.
• mortality
  • No patients died in either treatment arm of the TRANSFORMS study.
  • Due to its size and duration, the study was not designed to detect differences in mortality.
  • Overall, the additional benefit of fingolimod is not proven for this endpoint.
• morbidity
  • For all endpoints listed under the category of morbidity, no statistically significant differences were observed between the two study arms in patient population a).
  • The benefit assessment did not reveal any statistically significant differences between the study arms with regard to the endpoints – relapses and disability progression.
  • No additional benefit of fingolimod is proven for these endpoints.
  • The endpoints – severity of disability, assessed using the MSFC-z score; fatigue, assessed using the mFIS; activities of daily living, assessed using PRIMUS Activities; and health status, assessed using the EQ-5D-VAS – are determined using questionnaires.
  • Due to the significant difference in the proportion of missing values between the two study arms (actual or potential non-response rate greater than 15 per cent), these results cannot be validly interpreted and are therefore not used for the benefit assessment.
  • Furthermore, regarding the endpoint of fatigue, there were uncertainties as to which questionnaire (modified FIS (mFIS), unidimensional FIS (U-FIS)) was used to assess this endpoint and how it was analysed.
  • The transfer of results from the non-validated version of the questionnaire (mFIS) to the validated final version (U-FIS), carried out by the pharmaceutical manufacturer, is methodologically inadequate and therefore cannot be interpreted.
  • This analysis of the fatigue endpoint is therefore not relevant for the benefit assessment.
  • An additional benefit of fingolimod is not proven for these endpoints.
• Health-related quality of life
  • For this endpoint too, assessed using the Primus Qol questionnaire, the results could not be validly interpreted due to the significant difference in missing values between the two study arms (actual or potential non-response rate greater than 15 per cent) and are therefore not used for the benefit assessment.
  • Additional benefit is not proven for this endpoint.
• Side effects
  • With regard to side effects, the pharmaceutical manufacturer did not provide a comprehensive analysis of all specific side effects in the dossier, but limited itself to those identified in accordance with the risk management plan.
  • The interpretability of the analyses of the specific side effects is limited by the minor size of the relevant patient population referred to in point a).
  • The benefit assessment decision lists all specific side effects occurring in at least 10% of patients in one of the two study arms.
  • Due to the small patient population, the p-values for an unconditional exact test were calculated by the IQWiG (CSZ method according to Martín Andrés A, Silva Mato A. Choosing the optimal unconditioned test for comparing two independent proportions. Computat Stat Data Anal 1994; 17(5): 555–574.).
  • Relative risks (RR) and associated 95% confidence intervals are only presented in the benefit assessment decision if the test result is statistically significant.
  • Side effects – nervous system disorders, investigations and back pain – occur with statistically significant greater frequency in the fingolimod treatment arm.
  • The clinical relevance of the finding for the SOC (system organ class) ‘Investigations’ is questionable, as the difference is primarily due to changes in laboratory values.
  • For the SOC ‘General disorders and administration site conditions’ and its subordinate PT (preferred term) ‘flu-like symptoms’, however, statistically significant results in favour of fingolimod are observed, subject to the aforementioned limitations on interpretability.
  • As the results on side effects are imprecise due to the small number of cases, no reliable conclusions regarding additional benefit can be drawn from them; consequently, it cannot be determined whether fingolimod causes greater or minor additional harm compared with IFN-β.

b) Patients with highly active relapsing-remitting multiple sclerosis with high disease activity who have not yet received adequate disease-modifying therapy (duration of prior treatment with at least one disease-modifying therapy (here: other than IFN-β) < 1 year)

• An additional benefit compared with the appropriate comparator therapy determined by the G-BA is therefore not proven.
• In the dossier, the pharmaceutical manufacturer presents analyses of all inadequately pre-treated patients who received INF-β as a continuation of their therapy.
• This patient cohort also includes patients who received INF-β as part of their prior treatment.
• This patient cohort has already been assessed by the G-BA in the previous benefit assessment decision and is not the subject of the current benefit assessment.
• The pharmaceutical manufacturer has not submitted any data for assessment regarding patients who were not pre-treated with INF-β but with another disease-modifying therapy (new indication).

Courtesy translation only, please refer to the German original.