Nusinersen (2) – Spinraza®

• Clinical trials
  • In the ENDEAR trial, 122 children with genetically confirmed 5q-associated SMA and two copies of the SMN2 gene, aged ≤ 7 months at the start of the trial and with an age at symptom onset of ≤ 6 months, were randomised in a 2:1 ratio and assigned either to treatment with nusinersen (N = 81) or a sham intervention (N = 41).
  • The randomised, double-blind, parallel-group CHERISH trial enrolled 126 patients with genetically confirmed 5q-associated SMA, aged between 2 and 12 years at the start of the study and with an age at symptom onset > 6 months, and were randomised in a 2:1 ratio to either treatment with nusinersen (N = 84) or the sham intervention (N = 42).
  • The NURTURE trial is recruiting 25 children with genetic evidence of 5q-associated SMA (of whom 15 had 2 copies of the SMN2 gene and 10 had 3 copies of the SMN2 gene), who had not yet developed any clinical symptoms of the disease at the time of study enrolment (presymptomatic patients), are being treated with nusinersen.

a) Patients with 5q-associated spinal muscular atrophy (5q-SMA) type 1

• Extent and probability of the additional benefit of nusinersen compared with best standard care (BSC): indication of a major additional benefit.
• mortality
  • The time from administration of the first dose of the study medication or randomisation until death was recorded over the entire study period. Thirteen (16%) patients died in the nusinersen group compared with 16 patients (39%) receiving BSC, indicating that the risk of death was statistically significantly reduced with nusinersen treatment.
  • The extent of this effect is assessed as a major improvement in overall survival.
  • For this endpoint, there is an effect modification by the characteristic of age at symptom onset (age at symptom onset ≤ 12 weeks / > 12 weeks): A statistically significant difference was observed for patients aged ≤ 12 weeks at symptom onset, but not for patients aged > 12 weeks at symptom onset.
• Morbidity – Long-term ventilation
  • For the endpoint of long-term ventilation, there is no statistically significant difference between the treatment groups. However, there are effect modifications by the characteristics of sex and duration of illness.
  • For patients with a disease duration of ≤ 12 weeks, there is a statistically significant difference in favour of treatment with nusinersen compared with best-supportive-care (BSC) treatment in the proportion of patients requiring long-term ventilation; the extent of this difference is classified as major.
• Morbidity – Achievement of motor milestones (Hammersmith Infant Neurological Examination – HINE, Subscale 2)
  • For the endpoint of achieving motor milestones, as measured by the HINE Subscale 2, there is a statistically significant difference in favour of treatment with nusinersen compared with best-supportive-care (BSC) treatment, whose extent is classified as major.
  • Taking into account the results of the long-term SHINE-ENDEAR study, it is evident that the improvement at the endpoint ‘achievement of motor milestones’ is sustained up to day 578 (approximately 1.5 years).
  • For this endpoint, there is an effect modification by the characteristic ‘duration of illness’ (≤ 12 weeks / > 12 weeks); a statistically significant difference in favour of nusinersen is observed only for patients with a duration of illness of ≤ 12 weeks.
• Morbidity – Motor function (Children’s Hospital of Philadelphia Infant Test of Neuromuscular Disease – CHOP INTEND)
  • For the endpoint of motor function, as measured by the CHOP INTEND, a statistically significant difference in favour of nusinersen compared with best-supportive care (BSC) was observed at all measurement time points throughout the study (day 64, day 183, day 302 and day 394), a statistically significant advantage in favour of nusinersen compared with best-supportive-care (BSC) treatment, whose extent is classified as non-quantifiable.
  • As the confidence interval for Hedges’ g lies entirely outside the irrelevance range [−0.2; 0.2] at all time points, the difference is interpreted as a relevant effect.
• Morbidity – Serious respiratory events
  • For the endpoint of serious respiratory events, there is no statistically significant difference between the treatment arms.
• Morbidity – Hospitalisations
  • For the endpoint of hospitalisations, there is no statistically significant difference between the treatment arms.
• Health-related quality of life
  • No data on quality of life were collected in the ENDEAR study.
• Side effects
  • An adverse event (AE) occurred in 96% of patients in the intervention group and in 98% of patients in the control group.
  • When recording SUEs and discontinuations due to AEs, a large number of events that are part of the symptoms of the underlying condition, or events that could be either a side effect or a symptom of the underlying condition (e.g. System Organ Class [SOC] respiratory, thoracic and mediastinal disorders) were also recorded. The results on SUEs and discontinuations due to AEs are therefore not interpretable.
• Overall review
  • For patients with 5q-SMA type 1, analyses from the head-to-head ENDEAR trial and the long-term SHINE-ENDEAR trial are available for comparing nusinersen with best standard care (BSC).
  • In summary, statistically significant benefits for nusinersen were observed in the mortality (overall survival) and morbidity endpoints ‘permanent ventilation’ and ‘achievement of motor milestones (HINE)’ as a statistical advantage for the patient population aged ≤ 12 weeks at symptom onset (mortality) and for the patient population with a disease duration of ≤ 12 weeks (morbidity endpoints).
  • For the endpoints of overall survival and ‘achievement of motor milestones’, the differences are also statistically significant for the overall patient population in the ENDEAR study.
  • Based on the results of the long-term SHINE-ENDEAR study, the improvement in the endpoint ‘achievement of motor milestones (HINE)’ is sustained up to day 578 (approximately 1.5 years).
  • For the endpoint ‘motor function (CHOP INTEND)’, a statistically significant advantage for nusinersen is also evident.
  • The results from the patient populations cannot be interpreted conclusively.
  • In summary, the effect modifications are taken into account for this benefit assessment; however, the additional benefit is derived for the overall population of the ENDEAR study (patients with SMA type 1).
  • Overall, there are exclusively positive effects for nusinersen compared with best standard care (BSC), with no corresponding disadvantages. However, it should be noted that the results regarding side effects can only be assessed to a very limited extent.
  • The positive effects of nusinersen compared with the appropriate comparator therapy in the category of mortality and in key morbidity endpoints are assessed as a significant improvement in treatment-related benefit that has not been achieved to date.
  • Based on these considerations, the information in the dossier and the results of the benefit assessment, the extent of the additional benefit of nusinersen compared with the appropriate comparator therapy (BSC) for the treatment of patients with SMA type 1 is classified as major.

b) Patients with 5q-SMA type 2

• Extent and probability of the additional benefit of nusinersen compared with BSC: a hint of a considerable additional benefit.
• mortality
  • Fatalities were recorded as part of the safety assessment. No patients died during the study.
• Morbidity – incidence of serious respiratory events
  • For this endpoint, serious AEs classified under the SOC ‘Respiratory, thoracic and mediastinal disorders’ were analysed post-hoc. No statistically significant difference was observed between the treatment groups.
• Morbidity – Achievement of motor milestones (Hammersmith Functional Motor Scale Expanded – HFMSE)
  • For the endpoint ‘Achievement of motor milestones (HFMSE)’, a statistically significant difference was observed in favour of treatment with nusinersen compared with best-supportive-care (BSC) treatment. The effect can be considered clinically relevant, as the confidence interval for Hedges’ g lies entirely outside the irrelevance range of -0.2 to 0.2, and its extent is classified as considerable.
  • For this endpoint, there is an effect modification by the characteristic of disease duration (< 25 months, ≥ 25 months to < 44 months, ≥ 44 months): There is a significant difference for patients with a duration of illness < 25 months, but not for patients with a duration of illness of ≥ 25 months to < 44 months and ≥ 44 months.
  • Taking into account the results of the long-term SHINE-CHERISH study, it is evident that the improvement in the HFMSE endpoint is sustained up to day 1410 (approximately 3.5 years).
• Morbidity – Upper limb motor function (Revised Upper Limb Module – RULM)
  • For the endpoint ‘Motor function of the upper limbs (RULM)’, there is a statistically significant difference in favour of treatment with nusinersen compared with best-supportive-care (BSC). The effect can be considered clinically relevant, as the confidence interval for Hedges’ g lies entirely outside the irrelevance range of –0.2 to 0.2, and its extent is classified as considerable.
  • For this endpoint, there is an effect modification by the characteristic of disease duration (< 25 months, ≥ 25 months to < 44 months, ≥ 44 months): There is a significant difference for patients with a duration of illness < 25 months, but not for patients with a duration of illness of ≥ 25 months to < 44 months and ≥ 44 months.
  • Taking into account the results of the SHINE-CHERISH long-term study, it is evident that the improvement in the RULM endpoint is sustained up to day 1410 (approximately 3.5 years).
• Morbidity – Disease-related hospitalisations
  • For the endpoint ‘disease-related hospitalisation’, there is a statistically significant difference in favour of treatment with nusinersen compared with best-supportive-care (BSC).
  • It should be noted that the observed effect in favour of nusinersen is largely attributable to the serious respiratory events included in this endpoint (11 events in 84 patients in the nusinersen arm, 14 events in 42 patients in the sham intervention arm).
• Health-related quality of life – Paediatric Quality of Life Inventory (PedsQL)
  • No statistically significant difference was observed between the treatment groups.
• Side effects
  • An adverse event (AE) occurred in 93% of patients in the intervention group and in 100% of patients in the control group.
  • Overall, however, based on the available data on the adverse event profile, no advantages or disadvantages can be inferred for nusinersen compared with best standard care (BSC).
• Overall view
  • For patients with 5q-SMA type 2, analyses comparing nusinersen with best standard care (BSC) are available from the head-to-head CHERISH trial and the long-term SHINE-CHERISH trial.
  • No deaths occurred in the CHERISH study. For the morbidity endpoints ‘achievement of motor milestones (HFSME)’ and ‘motor function of the upper limbs (RULM)’, Nusinersen showed statistically significant advantages for patients with a disease duration of < 25 months.
  • Based on the results of the long-term SHINE-CHERISH study, the improvement in both endpoints is sustained up to day 1410 (approximately 3.5 years).
  • For the endpoint ‘disease-related hospitalisation’, there is a statistically significant advantage of nusinersen compared with best-supportive care (BSC).
  • The results for the patient populations cannot be interpreted conclusively.
  • In summary, the effect modifications are taken into account for this benefit assessment; however, the additional benefit is derived for the overall population of the CHERISH study (patients with SMA type 2).
  • Overall, there are mainly positive effects for nusinersen compared with BSC. It should be noted, however, that the results regarding side effects can only be assessed to a very limited extent.
  • However, the effects of nusinersen compared with BSC are not called into question by the available data on the adverse reaction profile.
  • The positive effects of nusinersen on key morbidity endpoints are assessed as a significant improvement in treatment-related benefit that has not been achieved to date.
  • Based on these considerations, the information in the dossier and the results of the benefit assessment, the extent of the additional benefit of nusinersen compared with the appropriate comparator therapy, BSC, for the treatment of patients with SMA type 2 is classified as considerable.

c) Patients with 5q-SMA types 3 and 4

• Extent and probability of the additional benefit of nusinersen compared with BSC: The additional benefit is not proven.
• Following an assessment of the suitability of the registry data, these cannot be used due to ambiguities regarding the inclusion of data generated outside the German healthcare context, and regarding the information on the mean observation period, as well as due to a lack of sufficient comparability between patient populations and a high proportion of missing values for patient-relevant endpoints assessing motor function already at the start of the studies, the data cannot be taken into account for the benefit assessment.
• Similarly, the data from the comparison of individual arms across different studies are not suitable for the benefit assessment due to a selective patient population, insufficient consideration of relevant confounders and a failure to take into account the transferability of the data to the German healthcare context.
• Overall, it is therefore concluded that an additional benefit is not proven.

d1) Presymptomatic patients with 5q-SMA and 2 SMN2 gene copies

• Extent and probability of the additional benefit of nusinersen compared with best standard care (BSC): hint of a major additional benefit.
• Based on the ENDEAR study, this benefit assessment for patients with type 1 SMA provides an indication of a major additional benefit.
• It was therefore examined whether the additional benefit from the comparison of nusinersen versus BSC in patients with early-symptomatic treatment initiation (duration of illness ≤ 12 weeks) in the ENDEAR study could be extrapolated to presymptomatic patients.
• Across all benefit endpoints considered, there is a consistent pattern showing that starting treatment with nusinersen at the presymptomatic stage yields a better outcome than starting treatment at the early symptomatic stage.
• Despite the considerable uncertainties associated with evidence transfer, this can be used for the benefit assessment, taking into account the following specific circumstances: SMA is a rare neuromuscular disorder which, in its natural course, is associated with a high mortality rate.
• Approximately 80% of presymptomatic patients with 2 copies of the SMN2 gene develop a very severe Type 1 course of the disease, with a life expectancy of 1–2 years without treatment.

d2) Presymptomatic patients with 5q-SMA and 3 copies of the SMN2 gene

• Extent and probability of the additional benefit of nusinersen compared with best standard care (BSC): hint of a non-quantifiable additional benefit.
• Analogous to the evidence transfer for presymptomatic patients with 5q-SMA and 2 copies of the SMN2 gene, an assessment was made as to whether the additional benefit from the comparison of nusinersen versus BSC in patients with type 2 SMA from the CHERISH study could be transferred to presymptomatic patients.
• However, it is not reasonably possible to transfer evidence based on the available data, as the morbidity endpoints HFMSE and RULM, for which statistically significant advantages for nusinersen were demonstrated in the CHERISH study for patients with type 2 SMA, were not collected in the NURTURE study, or were only collected for patients who were over 2 years old at the time of data collection.
• Despite the highly limited evidence based on descriptive analyses and the associated very high levels of uncertainty, this evidence may, exceptionally, be used for the benefit assessment, taking into account the following specific circumstances: SMA is a rare neuromuscular disorder which, in its natural course, is associated with a high mortality rate.
• The evidence of an advantage from presymptomatic therapy in children with 3 SMN2gene copies are supported by the proven risk of motor neuron degeneration prior to the onset of the first clinical symptoms and the recognition that, once clinical symptoms have already appeared, treatment is only of limited success.

d3) Presymptomatic patients with 5q-SMA and more than 3 SMN2 gene copies

• Extent and probability of the additional benefit of nusinersen compared with best standard care (BSC): The additional benefit is not proven.
• No data are available for the assessment of the additional benefit in presymptomatic patients with 5q-SMA and more than 3 copies of the SMN2 gene.

Courtesy translation only, please refer to the German original.