CADTH Reimbursement Review

Nusinersen (Spinraza)

Sponsor: Biogen Canada Inc.

Therapeutic area: Spinal Muscular Atrophy

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

Spinal muscular atrophy (SMA) is a severe neuromuscular disease and is the leading genetic cause of infant death.^1,2 It is characterized by the degeneration of alpha motor neurons in the anterior horn of the spinal cord, leading to progressive muscle weakness.¹ The most common form of SMA, 5q SMA, makes up more than 95% of all cases and is an autosomal recessive disorder caused by homozygous deletion or deletion and mutation of the alleles of the survival motor neuron (SMN) 1 gene (SMN1).^3,4 A second survival motor neuron gene (SMN2) acts in a capacity similar to SMN1, but is usually not sufficient on its own to maintain motor neurons. The number of SMN2 genes usually determines the severity of SMA.^1-3,5 Genetic testing provides a definitive diagnosis of 5q SMA, and the first step is to test for an SMN1 gene deletion.¹

SMA is a rare disease, and estimates of its incidence and prevalence vary in different studies. Currently, the incidence of SMA in Canada is unknown, although it is estimated that SMA affects 1 in every 6,000 to 10,000 live births.^4,6-8 However, a recent review reported estimates of 700 to 2,140 active cases of SMA in Canada, with approximately 35 new cases per year.⁹

SMA presents in various ways, depending on age at onset. Adult-onset SMA presents as mild proximal muscle weakness, and it is more severe in the lower limbs than in the upper limbs.^1,2 SMA is divided into 4 clinical subtypes that vary in age at onset, highest motor milestone achieved, and prognosis. Of interest to this review are type II and type III SMA. In SMA type II, age at onset is 6 to 18 months, and patients have delayed motor milestones, respiratory issues, and the possibly of a shortened life expectancy. Patients with SMA type II achieve the milestone of sitting unsupported, but never walk independently.¹⁰ Patients with type II SMA make up about 20% to 30% of all SMA cases, and most patients with SMA type II have 3 copies of SMN2.¹¹ Onset of SMA type III occurs in patients from 18 months to 18 years of age. Type III SMA makes up about 10% to 20% of all SMA cases.⁴ These patients are able to walk independently at some point in their lives and typically have a normal life expectancy.¹⁰

In Canada, treatment options for 5q SMA consist of disease-modifying therapies (nusinersen [Spinraza], risdiplam [Evrysdi]), which stimulate the production of the SMN protein, and gene-replacement therapy (onasemnogene abeparvovec), which is a 1-time IV (IV) infusion that replaces missing or faulty SMN1 genes.

Nusinersen is an antisense oligonucleotide that increases the proportion of exon 7 inclusion in SMN2 messenger ribonucleic acid (RNA) transcripts by binding to a specific site in the SMN2 pre-messenger RNA, leading to the translation of the messenger RNA into functional full-length SMN protein.¹²

The recommended dose of nusinersen is 12 mg, administered by intrathecal injection via lumbar puncture at day 0, day 14, and day 8, and day 63, followed by maintenance doses every 4 months.¹²

Nusinersen was granted a Health Canada Notice of Compliance for the indication of 5q SMA on June 29, 2017. When nusinersen was initially reviewed by CADTH in 2017, it was recommended for reimbursement for patients with 5q SMA with 2 copies of the SMN2 gene and for those with a disease duration of less than 26 weeks and an onset of clinical signs and symptoms consistent with SMA from 1 week to 7 months of age.¹³ In 2019, nusinersen was reviewed as a resubmission, and a conditional positive recommendation was granted for patients with 5q SMA with 2 or 3 copies of the SMN2 gene, for patients with a disease duration of less than 6 months and symptom onset from 1 week to 7 months of age, and for patients 12 years and younger with symptom onset after 6 months of age who never achieved the ability to walk independently.¹⁴ Across Canada, most provinces and drug plans only reimburse nusinersen if initiated in patients 18 years or younger, with the exception of British Columbia, where nusinersen is reimbursed if initiated in patients 12 years or younger. It should be noted that patients are not expected to stop treatment with nusinersen when they reach 12 or 18 years of age.

The sponsor has submitted nusinersen for reassessment to expand reimbursement conditions to include adults with type II and type III SMA who are older than 18 years, regardless of ambulatory status. The objective of the current reassessment is to perform a systematic review of the beneficial and harmful effects of nusinersen in adults (≥ 18 years of age) with type II or type III 5q SMA.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient and clinician groups that responded to CADTH’s call for input and from clinical experts consulted by CADTH for the purpose of this review.

Patient Input

Patient input for the CADTH reassessment of nusinersen was received from 3 groups — Cure SMA Canada (CSMAC), Muscular Dystrophy Canada (MDC), and the Love for Lewiston Foundation — all of which are registered charities.

Patients who responded to surveys and participated in interviews conducted by CSMAC and MDC noted that as they approached adulthood, they experienced a decline in physical abilities, highlighting that they have lost the ability to, or are just barely able to walk as adults. Along with the loss of gross motor skills, patients noted a significant impact on activities of daily living due to a progressive loss of life skills and overall independence, including loss of the ability to dress themselves, feed themselves, swallow, turn over in bed, and transfer for the purpose of toileting. Additionally, patients reported a lack of energy and loss of strength in their voice, making communication difficult and affecting the ability to maintain employment, and an increase in hospitalizations and need for supportive equipment. The devastation of disease progression and loss of function in patients with full mental capacity has a severe negative impact on mental health and well-being. Patient groups noted that, coupled with the continued inability to access effective treatments, they experienced a significant increase in anxiety, depression, and self-harm, requiring additional mental health support. Last, with the loss of physical function, patients require alterations to their homes for accessibility, which has a considerable financial impact. Patients hope the treatment will stop disease progression and even reverse muscle atrophy, which they consider to be an improvement in terms of quality and quantity of life.

Patient and caregiver responders identified an unmet need for access to treatments in the adult population that offer stability and improved quality of life (QoL) through greater independence, improved strength (primarily in the arms and respiratory function), and halting of progression. With improvements in these facets, patients believe they can achieve greater independence and a better QoL. Patients also noted that some of the largest barriers to treatment and challenges with currently available treatment are unreasonable costs, the mode of delivery with intrathecal therapy, and potential harms.

Given the few options for adults with SMA, treatment is limited to nusinersen, risdiplam, alternative management of the disease, or no treatment. In the CSMAC survey, 41 (47%) patients provided information about their experience with SMA treatments. Of those, 32 (78%) were receiving nusinersen and 9 (22%) were receiving risdiplam. Many of the patients who participated in the CSMAC and MDC surveys and interviews were treatment-naive because of limited access to SMA treatments in Canada. Of the patients receiving nusinersen, 79% reported that they experienced improvements in respiratory function, endurance, upper limb and core strength, and voice, and 15% reported stabilization of their disease. The remaining 6% reported no stabilization or improvement. Patients were receiving treatment with nusinersen for 1 to 3.5 years. Negative experiences reported by patients receiving nusinersen included a wearing-off of treatment and a drop in function shortly before the next maintenance dose; both were rectified after subsequent treatment. Additional negative experience included temporary headaches, discomfort from intrathecal injections, and the travel and time off work required for treatment. Regardless, patients felt that the benefit of nusinersen, including gains in function, improved strength and energy, and disease stabilization, far outweighs the negative aspects of the treatment. Several patients from the MDC interviews revealed that they switched to risdiplam after the initiation of nusinersen because of limited access, financial constraints, and difficulties with intrathecal administration. Patients also reported seeking alternative ways to manage their SMA, such as physiotherapy, exercise, and traditional Chinese medicine.

Clinician Input

Input from Clinical Experts Consulted by CADTH

In adults with type II and III SMA, the clinical-expert panel identified an unmet need for treatments that can change the natural course of the disease, including the ability to reverse the weakness associated with motor neuron degeneration, as there are currently no disease-modifying treatments available for adults. Experts agree that the goals of treatment are dependent on the type of SMA, given the high degree of heterogeneity in the disease and disability, and that treatment should be individualized to specific manifestations of the disease.

Currently, the mainstay of disease-modifying treatment for treatment-naive adults with type II or III SMA is nonpharmacologic, and includes occupational therapy, physiotherapy, and speech-language pathology, which are aimed at supporting function, mobility, and independence, as well as supportive measures, such as ventilation, nutritional assistance, and assistive devices. If nusinersen is recommended for treatment-naive adults with type II or III SMA, the clinical-expert panel noted that it could become a first-line treatment. They also noted that risdiplam has recently been given a positive CADTH Canadian Drug Expert Committee (CDEC) recommendation in younger adults, so the treatment paradigm may shift in the future. As noted by 1 expert, nusinersen has received funding in Quebec for the treatment of adults with type II or type III SMA and is currently being used in this population. Regardless, the clinical experts suggested that for the treatment-naive adult population, there is no guidance on whether other medications should be tried before nusinersen.

The experts highlighted the lack of higher-level evidence (i.e., from randomized controlled trials [RCTs]) in this population to determine which adults with SMA are most likely to respond to treatment with nusinersen. The experts hypothesized that patients with higher functioning and who are ambulatory may demonstrate better responses because they have more nerves, leading to better function. As well, the clinical experts believe that patients without complex spines are more likely to have a better risk-benefit profile. However, most of the experts stressed that the earlier treatment is administered (i.e., in pre-symptomatic children), the greater the benefit observed, although there was some disagreement among panel members about whether to consider age a factor when assessing response. Conversely, the clinical-expert panel noted that patients least suitable for treatment with nusinersen are those with complex spines as a result of spinal fusion surgery, those who cannot tolerate lumbar puncture, and those who have previously been treated as infants or children (there is no evidence of benefit in these patients).

Clinically, numerous outcomes and measures are used to assess response to treatment. The clinical experts agreed that — given the variation in response to treatment and individualized treatment goals — several outcome measures can be used to assess the benefits of treatment, including motor-function and respiratory outcomes and outcomes such as bulbar function, strengthening of speech, and functional independence. The experts noted that in patients with type II or III SMA, disease progression occurs slowly, over the course of years; thus, the impact of treatment on these outcome measures is not likely to be seen over a short period of time.

The clinical-expert panel agreed that the main reasons for discontinuation would be progression or worsening of disease and any major complications or adverse events (AEs) related to therapy. One clinical expert noted that — based on experience with nusinersen in the adult population — the most common reason for discontinuation is a patient’s desire to stop because of lack of improvement or inability to tolerate the treatment. The panel agreed that all patients with SMA should receive care at a tertiary centre that has a variety of neuromuscular specialists, a multidisciplinary team, access to interventional radiology or neurosurgery, and the ability to admit patients who experience potential procedural or treatment-related complications.

Clinician-Group Input

CADTH received clinician-group input from the Neuromuscular Disease Network for Canada, a pan-Canadian network launched in 2020 to bring together clinical, scientific, technical, and patient expertise in neuromuscular disease, with the aim of improving the care, research, and treatment of neuromuscular diseases for all Canadians. Eight clinicians with experience treating SMA patients provided input.

The clinician group highlighted the 3 main disease-modifying treatments for SMA: nusinersen, risdiplam, and onasemnogene abeparvovec. The clinician group agreed that treatment goals for later-onset SMA would be to maintain current levels of motor function and strength, achieve disease stabilization (including the avoidance of ventilation), promote independence, and improve overall health-related quality of life (HRQoL). The clinician group highlighted the fact that risdiplam may be the only other treatment option for these patients. In this case, they noted that either nusinersen or risdiplam could be tried first. The clinician group explained that younger patients are most likely to derive benefit from nusinersen, and noted that it may be difficult to accurately identify which adults are most likely to derive benefit from nusinersen. The clinician group stated that a clinically meaningful response to treatment in adults is likely to consist of stabilization of motor and respiratory function, maintenance of independence, and a reduction in hospitalizations. Moreover, they noted that maintaining the ability to speak and avoiding the need for ventilatory support have profound impacts on patient QoL, autonomy, and the ability to maintain vocational and social roles. The clinician group emphasized that current provincial monitoring requirements are too frequent and there is significant variation between provinces. They agreed that patients should be assessed at treatment initiation, at 6 months, and annually thereafter to reduce patient burden and strain on health care resources, given the slow progression in functional decline. Last, the clinician group noted that nusinersen must be administered by or under the direction of health care providers experienced in performing lumbar puncture at designated treatment centres.

Drug-Program Input

The drug programs identified the following jurisdictional implementation issues: relevant comparators, considerations for initiation of therapy, considerations for continuation or renewal of therapy, considerations for discontinuation of therapy, generalizability, and system and economic issues. The clinical experts consulted by CADTH weighed evidence from the key studies submitted by the sponsor and clinical expertise to provide responses to drug-program implementation questions. Refer to Table 3 for more details.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

As part of the reassessment for nusinersen, the sponsor provided CADTH with 5 observational studies, 1 open-label extension study, and 1 critical review and meta-analysis. Four studies were included in the report^15-19 and the critical review and meta-analysis was summarized in the Other Relevant Evidence section.²⁰ The other 2 studies described in the report were considered outside the scope of this review and not discussed any further.^21,22

The study by Hagenacker et al. (2020)¹⁷ was a prospective, German, multi-centre, noncomparative observational study that evaluated the safety and effectiveness of nusinersen in 124 adults with 5q SMA. The study by Maggi et al. (2020)¹⁸ was an Italian, retrospective, noncomparative cohort study that evaluated the safety and effectiveness of nusinersen in 116 patients with type II or type III SMA. The EU registry study^15,16 was an observational, registry-based cohort study of combined data that evaluated the safety and effectiveness of nusinersen in 252 adults with 5q SMA from 3 prospective and retrospective European registries (SMArtCARE, CuidAME, and the International SMA Registry [ISMAR]) in 2 subcohorts: a before-and-after treatment comparison with nusinersen in 74 patients with type III SMA and 1 patient with type IV SMA; and a comparative dataset from 252 adults with type III SMA (235 who had been treated with nusinersen and 17 who had not). The study by Pera et al. (2021)¹⁹ was an observational, noncomparative, registry-based study from ISMAR in Italy of 144 ambulant and nonambulant type III SMA patients treated with nusinersen.

All studies included treatment-naive adults with SMA. Across studies, most patients had type III SMA (62% to 100%), with mainly 3 or 4 copies of SMN2. Type II SMA was infrequently reported, with only 45 and 13 type II patients in Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020),¹⁸ respectively. No type II patients were included in the EU registry study^15,16 or in the study by Pera et al. (2021).¹⁹ Across studies, 37% to 56% of patients were considered ambulatory. Baseline motor-function scores were high, with mean Hammersmith Functional Motor Scale Expanded (HFMSE) scores at baseline ranging from 20.74 to 30.75, mean Revised Upper Limb Module (RULM) scores at baseline ranging from 20.87 to 27.57, and mean 6-minute walk test (6MWT) distance at baseline ranging from 300.87 m to 323.03 m.^15-19

Effectiveness Results

HFMSE

In Hagenacker et al. (2020),¹⁷ the mean change from baseline in HFMSE scores was 1.73 points (95% confidence interval [CI], 1.05 to 2.41) at 6 months (n = 124), 2.58 points (95% CI, 1.76 to 3.39) at 10 months (n = 92), and 3.12 points (95% CI, 2.06 to 4.19) at 14 months (n = 57). The proportion of patients with an increase of 3 points in HFMSE score were 28%, 35%, and 40% at 6 months, 10 months, and 14 months, respectively. In Maggi et al. (2020),¹⁸ the mean change from baseline in HFMSE scores was 1.48 points (standard deviation [SD] = 2.28), 2.44 points (SD = 2.8), and 2.85 points (SD = 2.93) at 6 months (n = 103), 10 months (n = 75), and 14 months (n = 46), respectively, for all type III SMA patients. In all SMA patients at 6, 10, and 14 months, increases of 3 or more points in HFMSE score occurred in 28%, 38%, and 49% of patients, respectively. In the EU registry study,^15,16 the slope for the change in HFMSE score per week was –0.00006 points per week (95% CI, –0.00955 to 0.009428) before treatment with nusinersen, and was 0.2575 points per week (95% CI, 0.01038 to 0.04112) after treatment with nusinersen (n = 75). In the analysis comparing nusinersen-treated patients with untreated patients, the slope for the change in HFMSE score was 0.02907 points per week (95% CI, 0.01930 to –0.03884) in nusinersen-treated patients (n = 235), compared with −0.01129 points per week (95% CI, −0.03289 to 0.01031) in untreated patients (n = 17).^15,16 In the study by Pera et al. (2021),¹⁹ at 12 months, the mean change from baseline in HFMSE was 0.79 points (95% CI, –0.29 to 1.87) (n = 45), with the HFMSE results showing a decline in 11.1% of patients, stability in 53.3% of patients, and improvement in 35.6% of patients.

RULM

In Hagenacker et al. (2020),¹⁷ the mean change from baseline in RULM scores was 0.66 points (95% CI, 0.26 to 1.05) at 6 months (n = 120), 0.59 points (95% CI, 0.15 to 1.03) at 10 months (n = 90), and 1.09 points (95% CI, 0.62 to 1.55) at 14 months (n = 58). An increase of at least 2 points in RULM score was observed in 28 (23%) patients at 6 months, whereas 74 (64%) showed no meaningful change and 28 (23%) showed a decline. In Maggi et al. (2020),¹⁸ the mean change from baseline in RULM scores was 0.31 points (SD = 1.97), 0.61 points (SD = 2.08), and 0.86 points (SD = 2.18) at 6 months (n = 102), 10 months (n = 71), and 14 months (n = 44), respectively, for all type III SMA patients. Patients with type II SMA had a numerically greater change in mean RULM scores than those with type III with scores, at 0.8 points (SD = 1.95) at 6 months (n = 12), 1.67 points (SD = 1.8) at 10 months (n = 9), and 1.6 points (SD = 1.52) at 14 months (n = 5). A 2-point change in RULM score in all SMA patients at 6, 10, and 14 months was shown in 21%, 28%, and 35% of patients, respectively. In the EU registry study,^15,16 the slope for the change in RULM score was –0.00745 points per week (95% CI, –0.01401 to 0.0009) before treatment with nusinersen, and was 0.002569 points per week (95% CI, −0.00533 to 0.01047) after treatment with nusinersen (n = 75). In the analysis comparing nusinersen-treated with untreated patients, the slope for the change in RULM score was 0.01168 points per week (95% CI, 0.004957 to 0.01841) in nusinersen-treated patients (n = 235), compared with 0.003936 points per week (95% CI, −0.01030 to 0.01817) in untreated patients (n = 17). In the study by Pera et al. (2021), the mean change from baseline in RULM at 12 months was 0.07 points (95% CI, –0.48 to 0.63) (n = 55), with the RULM results showing a decline in 13.0% of patients, stability in 75.9% of patients, and improvement in 15.6% of patients.¹⁹

6MWT

In Hagenacker et al. (2020),¹⁷ the mean change from baseline in 6MWT was 22.12 m (95% CI, 8.7 to 35.6) at 6 months (n = 47), 31.14 m (95% CI, 15.2 to 47.1) at 10 months (n = 37), and 45.96 m (95% CI, 25.4 to 66.6) at 14 months (n = 25). In Maggi et al. (2020),¹⁸ change from baseline in 6MWT was only available for type III SMA walkers, who demonstrated a mean change in 6MWT of 14.66 m (SD = 27.57) at 6 months (n = 48), 26.45 m (SD = 34.6) at 10 months (n = 35), and 23.11 m (SD = 51.2) at 14 months (n = 24). The proportion of patients that achieved a minimum 30-metre improvement in 6MWT was 29% at 6 months, 46% at 10 months, and 42% at 14 months. In the EU registry study,^15,16 the slope for the change in 6MWT was –0.03399 m per week (95% CI, –0.4373 to 0.3694) after treatment with nusinersen (n = 75). In the analysis comparing nusinersen-treated with untreated patients, the slope for the change in 6MWT score was 0.2633 m per week (95% CI, 0.09922 to 0.42740) in nusinersen-treated patients (n = 235), compared with –0.7148 m per week (95% CI, –1.2789 to –0.1506) in untreated patients (n = 17). Mean change from baseline in 6MWT for adults in Pera et al. (2021)¹⁹ at 12 months was 0.52 m (95% CI, –19.85 to 20.89) (n = 17).

Other Effectiveness Outcomes

Respiratory outcomes of forced vital capacity (FVC)/forced expiratory volume in 1 second (FEV₁) were only evaluated in Maggi et al. (2020),¹⁸ with a mean change in percent-predicted FVC from baseline for all SMA type III patients at 14 months of 6.47% (SD = 9.22). Mean change in percent-predicted FVC at 14 months was not available for patients with type II SMA because of sample-size constraints. The mean change from baseline at 14 months in percent-predicted FEV₁ was 5.86% (SD = 9.22) for all type III SMA patients. The mean change in percent-predicted FEV₁ at 14 months was not available for type II SMA patients.

Other outcomes of interest to this review, including bulbar function, survival, hospitalization, HRQoL, anatomic-related outcomes, and caregiver burden, were not assessed in the included studies.

Harms Results

Harms were infrequently reported in the included studies, with all but 1 study (Pera et al. [2021]¹⁹) reporting the frequency of harms. When reported, the frequency of AEs in the included studies ranged from 30.0% to 41.4% and were considered mild to moderate by the investigators. The frequency of serious adverse events (SAEs) was low in all studies, when reported.

In Hagenacker et al. (2020),¹⁷ 2 patients withdrew from treatment at 10 months because of adverse drug reactions. In Maggi et al. (2020),¹⁸ nusinersen treatment was stopped in 2 (1.7%) type III SMA patients after 6 months because of a lack of subjective benefit and poor tolerability of repeated lumbar puncture. Withdrawals due to AEs (WDAEs) were not reported in the EU registry study^15,16 or in the study by Pera et al. (2021).¹⁹

The most frequently reported notable harms of interest to this review were lumbar puncture-related AEs; however, these were not reported in the EU registry study^15,16 or in the study by Pera et al. (2021).¹⁹ Post-procedural complications of headache (35% and 37.1%) and back pain (22% and 8.6%) were the most frequently reported AEs in the Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020)¹⁸ studies, respectively. The frequency of headache and back pain was not reported in the study by Pera et al. (2021),¹⁹ but they were noted as the most frequently occurring AEs. Frequency of other notable harms, including serious infections, renal toxicities, and coagulation abnormalities, was infrequently reported in the included studies.

No deaths were noted in any of the studies.

Critical Appraisal

No RCTs focusing on treatment-naive, adult, type II or III SMA patients were identified as part of the CADTH literature search, and all available and included studies were of observational design, focusing on real-world data, which have more limitations than RCTs.

The studies included in this reassessment are associated with lower internal validity because of the limitations in design, enrolled patient populations, and statistical analyses. The included studies shared a common limitation pertaining to the study design: they were noncomparative, so the results observed cannot be attributed to treatment with nusinersen. However, the EU registry study^15,16 included an untreated comparison group, albeit with a sample size of 17 patients. The studies included in this reassessment also suffer from a high level of selection bias, reporting bias, and information bias. In all studies, patients with SMA had to be able to complete at least 6 months of treatment with nusinersen, which selected for patients who were able to complete the induction dosing and who were able to tolerate and/or receive doses. Moreover, included patients were mostly SMA type III, with a seemingly higher functional status at baseline, according to ambulatory status and baseline motor scale scores. No or limited techniques were used to adjust for potential selection biases across studies. In all studies, important potential confounders and treatment-effect modifiers that were not identified or considered, which may have influenced the results, include training for the outcomes of interest, routine exercise and close observation, other routine care (such as physiotherapy and occupational therapy), as well as the placebo effect, and the extent to which uncontrolled confounders and treatment-effect modifiers influenced the results is unclear. In all studies, no protocol was identified, and it was not possible to determine whether sample sizes (ranging from 67 to 252 patients) were appropriate for the research question and objectives of each study. The EU registry study^15,16 conducted analyses on 2 groups: the first consisted of 235 nusinersen-treated patients and 17 untreated patients; and the second consisted of 75 patients assessed before and after treatment. Because of the limited population in the untreated group, the results observed could not be attributed to treatment with nusinersen. The proportion of patients lost to follow-up was infrequently reported in the included studies, although there was a notable proportion of patients with a lack of longer-term follow-up at 14 months, compared with earlier times of assessment. With the exception of the EU registry study,^15,16 no imputation of missing data was conducted, so missing data affected the validity of the results.

As previously mentioned, selection bias in the included populations was noted as a key limitation. Patients enrolled in the included studies consisted of mainly type III SMA (62.0% to 100.0%), with few type II patients (11.2% to 36.0%), which was noted by the clinical experts to be higher than what they see in clinical practice. The SMA patients included in the 4 studies were considered to be higher functioning, based on the high prevalence of type III disease, with most patients having 3 or 4 copies of SMN2, and the proportion of ambulatory patients (37.0% to 56.03%). Moreover, baseline motor-function scores were considered high, suggesting a population with less severe disease. As such, the included study populations were unrepresentative of the reimbursement request (lack of type II SMA patients), and the results may not be generalizable to adults with type II or III SMA in Canada. Given that up to half of all patients across studies were nonambulatory, the HFMSE and 6MWT may not be appropriate outcome measures in all patients, the clinical experts report, which further limits the evaluable population and the generalizability of the results. HRQoL and other patient-reported outcomes, which were outcomes important to patients, were not assessed in the included studies; therefore, the effect of nusinersen with respect to these outcomes remains unknown. The maximum follow-up time across studies was 14 months, which was considered insufficient to assess any clinically meaningful change in outcomes in adults with type II or III SMA because of the slowly progressing nature of the disease, as well as its natural history.

Indirect Comparisons

No indirect evidence was included in the sponsor’s submission to CADTH or identified in the literature search that matched the inclusion and exclusion criteria of this review.

Other Relevant Evidence

Other Sponsor Submitted Evidence

As part of the reassessment for nusinersen, the sponsor submitted a publicly available critical review and meta-analysis of patients with type II and III SMA. The objective of the meta-analysis by Coratti et al. (2021)²⁰ was to critically review literature that reported real-world data on motor function in type II and III SMA patients treated with nusinersen to establish possible patterns of efficacy by subdividing the results by SMA type, age (children versus adults), and type of assessment. Only results related to the adult population with type II or III SMA were of interest in this reassessment.

The meta-analysis was informed by a systematic review of existing literature. A total of 14,627 articles were identified. After study selection, 19 papers reporting data on efficacy in nusinersen-treated and untreated patients using structured assessments in type II and III SMA were selected. Pooled analyses were conducted at multiple levels. First, a rough evaluation on the overall benefit of treatment versus no treatment was run that included the largest available evidence, even if heterogeneous. Next, the effect size was estimated using random-effect models, and heterogeneity among studies was quantified by the I² coefficient. Then, meta-regression analysis was undertaken to identify possible sources of heterogeneity among studies. Motor-function outcomes included HFMSE, RULM, 6MWT, Medical Research Council (MRC) scale for muscle strength, and Children’s Hospital of Philadelphia–Adult Test of Neuromuscular Disorders (CHOP-ATEND).²⁰ Meta-regression was not conducted for the MRC and CHOP-ATEND outcomes and are not summarized. Meta-regression analyses were employed with a random-effects model using aggregate-level data. Only studies with complete data available (sample size, mean, SD, or 95% CI) were included in the meta-analysis.²⁰

In the meta-regression analysis, pooled mean changes from baseline in HFMSE score, RULM score, and 6MWT in the adult population were 1.87 points (95% CI, 1.05 to 2.68), 0.64 points (95% CI, 0.27 to 1.01), and 20.28 m (95% CI 1.17 to 39.40), respectively.²⁰

The meta-analysis was based on an adequately conducted and reproducible systematic literature search. It was unclear if the inclusion and exclusion criteria for population, outcomes, and study design were pre-specified. A quality assessment of the included studies was conducted using the Risk of Bias Assessment Tool for Nonrandomized Studies). No interpretation on the quality of studies was conducted; however, as all studies were observational, most studies were considered to suffer from a high level of bias in the selection of participants. All publicly available studies summarized in the review conducted by CADTH (Hagenacker et al. [2020],¹⁷ Maggi et al. [2020],¹⁸ and Pera et al. [2021]¹⁹) were included in the submitted meta-analysis. Outcomes included in the meta-analysis were appropriate and relevant to the Canadian context, with HFMSE, RULM, and 6MWT most commonly included in studies, although there were differences in reporting and time of assessment. Most of the included studies had a follow-up time ranging from 10 to 14 months, which was considered by the clinical experts consulted by CADTH to be insufficient to observe clinically meaningful changes in the motor function of adults. There was considerable heterogeneity in the studies, given the inclusion of both ambulant and nonambulant type II and type III SMA patients, as well as the inclusion of both adults and children. Additionally, the included studies had small sample sizes, limiting the conclusions that can be drawn from individual studies. Overall, there was a moderate to considerable level of heterogeneity in the included studies across outcomes, with I² values ranging from 43% to 71%. Pooled estimates of mean change for motor-function outcomes favoured nusinersen treatment in the adult population; however, the pooled estimates generally displayed wide 95% CIs, particularly for the 6MWT, and in many cases crossed the zero meridian, indicating a high level of variability in these cohorts and substantial imprecision in estimates of treatment effect. Additionally, the change from baseline in motor-function outcomes was minor, and in discussion with the clinical experts, there is uncertainty about what constitutes a clinically meaningful change in the adult population for these outcome measures. Moreover, given the nature of the included studies and the limitations defined for the studies included in both the systematic review conducted by CADTH and the meta-analysis, the observed effects cannot be attributed to nusinersen.

Evidence Identified From the Literature

Eight non-comparative, observational studies were identified in the literature search that met all inclusion criteria of the systematic review, with the exception of study design, as they consisted of descriptive observational studies. As with the studies provided by the sponsor, the effectiveness of nusinersen in these studies is highly uncertain due to the noncomparative study design, selection bias, and relatively small sample sizes of adults with type II and III SMA.

Conclusions

Four noncomparative, observational studies were included in the reassessment for nusinersen in adults with type II or III SMA. The observational nature and lack of a well-defined concurrent comparator in the included studies significantly limits the ability to establish causal relationships in treatment effect with nusinersen.

In all of the studies, selection bias in the included populations and relatively small sample sizes were noted as key limitations. All studies included mostly type III SMA adults with higher physical functioning at baseline because of their SMA type, number of SMN2 copies, ambulatory status, and higher baseline scores for motor-function outcomes. Input from clinical experts noted the populations were not reflective of the reimbursement request, particularly due to the lack of type II SMA patients, or their clinical practice.

Although there was a generally consistent positive effect of nusinersen on motor-function outcomes, the magnitude of the treatment effect with nusinersen was variable and often not clinically meaningful and, given the limitations of these studies in study design, statistical analysis, duration, and the heterogeneous natural history of adults with SMA, results in all studies were considered highly uncertain and may not be generalizable to the Canadian population. Harms associated with nusinersen were generally mild to moderate in severity, and AEs related to the lumbar puncture procedure were the most frequently reported. However, the reporting of AEs was inconsistent and infrequent, and based on the study design and associated biases, may be under-reported.

Although the amount of real-world data for nusinersen is relatively high, the overall quality of studies remained a concern. Most of the identified evidence could not provide conclusive evidence demonstrating the effectiveness of nusinersen in adults with type II or III SMA. Overall, it remains unclear if nusinersen resulted in clinically meaningful improvements or disease stabilization, which were considered important outcomes to patients. Additionally, since HRQoL was not assessed in any of the studies, the effect of nusinersen on this important outcome in adults is unknown.

Introduction

Disease Background

SMA is a severe neuromuscular disease and is the leading genetic cause of infant death.^1,2 It is characterized by the degeneration of alpha motor neurons in the anterior horn of the spinal cord, leading to progressive muscle weakness.¹ The most common form of SMA, 5q SMA, makes up more than 95% of all cases and is an autosomal recessive disorder caused by homozygous deletion or deletion and mutation of the alleles of the survival motor neuron (SMN) 1 gene (SMN1).^3,4 Whereas deletion or mutation of the SMN1 gene results in SMN protein deficiency, a second nearby gene, the SMN2 gene, produces a small amount of functional SMN protein. The number of available SMN2 gene copies and the extent of the expression of these genes modulates the severity of the disease.^1-3,5

SMA is a rare disease, and estimates of its incidence and prevalence vary between studies. Currently, the incidence of SMA in Canada is unknown, although it is estimated that SMA occurs in 1 of every 6,000 to 10,000 live births.^4,6-8 One study that examined the Cure SMA database (a voluntary registry that is 1 of the largest patient-reported repositories in the world) reported the birth prevalence in the US at about 1 in 20,000 live births.²³ A recent review reported estimates of 700 to 2,140 active cases of SMA in Canada, with approximately 35 new cases per year.⁹

SMA presents in various ways, depending on age of onset. Infants present with severe hypotonia and feeding difficulties, whereas later onset in young children may present as difficulty with stairs and frequent falls.¹⁰ Adult-onset SMA presents as mild proximal muscle weakness.² Genetic testing gives a definitive diagnosis for 5q SMA. The first step is to test for SMN1 gene deletion.¹ In 2020, the Government of Canada and Muscular Dystrophy Canada, in collaboration with Novartis Canada, developed a National Newborn Screening in SMA program to help improve early diagnosis and timely access to treatment for infants born with SMA.²⁴

Deficiency in SMN results in defects in multiple components of the motor system, including the motor neurons.² Electrophysiological studies and clinical findings in patients with SMA show that patients typically experience a sharp decline in motor function, with motor-unit loss soon after symptom onset, followed by a long plateau period of relative stability in motor function.^2,25 Clinical-expert input indicated to CADTH that motor-function decline is irreversible, aside from possible gains in strength and gross motor abilities in infants still undergoing normal muscle hypertrophy in the first 2 years of life. Muscle weakness tends to be symmetric, proximal rather than distal, and more severe in the lower limbs than in the upper limbs.¹

SMA is divided into 4 clinical subtypes that vary in age of onset, highest motor milestone achieved, and prognosis. Although the subtypes provide a convenient means of classifying patients, it should be noted that patients exist along a continuum of disease severity, with overlap of symptoms between subtypes. This spectrum is represented in Figure 1.

Figure 1: Continuous Spectrum of SMA Phenotype

MM = minimal manifestations; SMA = spinal muscular atrophy; SMN = survival motor neuron.

Source: Talbot K, Tizzano EF. The clinical landscape for SMA in a new therapeutic era. Gene Ther. 2017; 24:529 to 533. Licensed under: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Type I SMA presents by the age of 6 months and is the most common genetic cause of infant mortality. These patients never achieve the motor milestone of sitting unsupported, and generally do not survive past 2 years of age, owing to respiratory failure.^1-3 Almost all patients with SMA type I have 2 or 3 copies of SMN2, giving rise to a broad range of phenotypes.¹¹

In SMA type II, age of onset is 6 to 18 months and patients have delayed motor milestones, respiratory issues, and the possibly of a shortened life expectancy. Patients with type II SMA achieve the milestone of sitting unsupported, but never walk independently. Symptoms generally appear 6 to 18 months after birth, and most patients survive past the age of 25,^5,10 with life expectancy improved by aggressive supportive care.¹⁰ Patients with type II SMA represent about 20% to 30% of SMA cases, and most patients with SMA type II have 3 copies of SMN2.¹¹ In addition to the inability to walk independently, common symptoms are fine tremors of the upper extremities, tongue fasciculation, difficulty swallowing, joint contractures, and scoliosis.¹ Scoliosis and weak intercostal muscles can cause restrictive lung disease.³ There is a range in severity, and weaker patients require noninvasive ventilation.¹ Difficulty swallowing is less common than in patients with type I SMA, and difficulty with feeding comes from masticatory muscle weakness. In a study that examined 1,966 patients in the Cure SMA database (with data available from 2010 to 2016), median survival for those with type II SMA was 59.9 years.²³

Patients with SMA type III have an age of onset from 18 months to 18 years and experience muscle weakness. Type III SMA makes up about 10% to 20% of SMA cases.⁴ These patients are able to walk independently at some point in their lives and typically have a normal life expectancy.¹⁰ Most patients with type III SMA have 3 or 4 copies of SMN2.¹¹ Patients with SMA type III have little or no respiratory weakness.³ Ambulatory patients may exhibit abnormal gait characteristics due to proximal weakness,¹⁰ and patients who lose the ability to walk often develop scoliosis.¹

SMA type IV is a very rare form of adult-onset SMA with various degrees of muscle weakness. These patients retain the ability to walk, have a normal life expectancy, and do not suffer from respiratory or nutritional issues.¹ Common to all types of SMA is a progressive decline in muscle function.

Standards of Therapy

There is no cure for SMA, and treatment options are dependent on SMA type and extent of symptoms. Treatment options for 5q SMA available in Canada consist of disease-modifying therapies (nusinersen, risdiplam), which stimulate the production of the SMN protein, and gene-replacement therapy (onasemnogene abeparvovec), which is a 1-time IV (IV) infusion that replaces missing or faulty SMN1 genes.

In addition to treatment with disease-modifying therapies, current standards of practice involve clinical monitoring and surveillance, anticipatory management of symptoms, and attempts to improve overall QoL. SMA patients are monitored for growth, gastrointestinal function, nutrition, respiratory complications, and orthopedic complications (i.e., scoliosis and/or contractures). These standards of practice include respiratory management, secretion mobilization in patients with weak cough, management of swallowing difficulties, and various multidisciplinary strategies to manage gross motor functions and spinal deformities (e.g., physiotherapeutic, orthopedic, surgical).

Drug

Nusinersen is an antisense oligonucleotide that increases the proportion of exon 7 inclusion in SMN2 messenger RNA transcripts made, through binding to a specific site in the SMN2 pre-messenger RNA. This leads to the translation of the messenger RNA into functional full-length SMN protein.¹²

Nusinersen 2.4 mg/mL solution is administered via intrathecal injection by lumbar puncture. The recommended dose of nusinersen is 12 mg in a 5 mL solution, administered as nusinersen sodium.

It is administered in a regimen of 4 loading doses, with the first 3 administered at 14-day intervals (day 0, day 14, and day 28), with the fourth loading dose approximately 30 days after the third loading dose (day 63). After the fourth loading dose, a maintenance dose should be administered once every 4 months.¹²

Nusinersen was granted a Health Canada Notice of Compliance for the indication of 5q SMA on June 29, 2017. Nusinersen was initially reviewed by CADTH in 2017, and was recommended for reimbursement for patients with 5q SMA with 2 copies of the SMN2 gene and for those with a disease duration of less than 26 weeks and an onset of clinical signs and symptoms consistent with SMA from 1 week to 7 months of age.¹³ In 2019, nusinersen was reviewed as a resubmission, and a conditional positive recommendation was granted for patients with 5q SMA with 2 or 3 copies of the SMN2 gene, for patients with a disease duration of less than 6 months and symptom onset from week 1 to 7 months of age, and for patients 12 years or younger with symptom onset after 6 months of age who never achieved the ability to walk independently.¹⁴ Across Canada, most provinces and drug plans only reimburse nusinersen if initiated in patients 18 years or younger, with the exception of British Columbia, where reimbursement for nusinersen is only provided when initiated in patients 12 years or younger. It should be noted that patients are not expected to stop treatment with nusinersen upon reaching 12 or 18 years of age.

As part of the reassessment for nusinersen, the sponsor is requesting that the reimbursement criteria for nusinersen be expanded to include adults (> 18 years) with type II or type III SMA, regardless of ambulatory status.

Recently, the European Medicines Agency recognized real-world clinical findings that support the use of nusinersen for stabilization or improvement in motor function for some adults with type II and III SMA.²⁶

The characteristics of treatments for 5q SMA available in Canada are summarized in Table 2.

Stakeholder Perspectives

Patient-Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

Patient input for the CADTH reassessment of nusinersen was received from 3 groups — CSMAC, MDC, and the Love for Lewiston Foundation — all of which are registered charities. The information provided by CSMAC was collected with an online survey conducted in November 2020 that consisted of open-ended questions, rating scales, and forced-choice options, and semi-structured interviews conducted in December 2021 with 88 adults from Ontario (35%), Quebec (24%), British Columbia (19%), Alberta (12%), Saskatchewan (4%), Manitoba (4%), New Brunswick (1%), and Nova Scotia (1%); in addition, 1 respondent resided outside of Canada. The information provided by MDC was gathered through a health care experience survey conducted by MDC neuromuscular service support staff who completed semi-structured virtual interviews (e.g., phone or Zoom sessions) from December 2021 to January 4, 2022 with 60 adults with SMA; 20 participants resided in Quebec and the rest resided in Alberta, British Columbia, Manitoba, New Brunswick, Nova Scotia, Ontario, Prince Edward Island, and Saskatchewan. The information provided by the Love for Lewiston Foundation, which works to identify SMA individuals and families that require funding due lack of support and adequate resources, highlights personal experiences and opinions of families affected by SMA.

Respondents to the CSMAC and MDC surveys and interviews noted that as they approached adulthood, they experienced a decline in physical abilities, highlighted by a complete or partial loss in the ability to walk. Along with the loss of gross motor skills, patients noted a significant impact on activities of daily living due to a progressive loss of life skills and overall independence, including a loss of the ability to dress themselves, feed themselves, swallow, turn over in bed, and transfer for the purpose of toileting. Additionally, patients reported a lack of energy and a loss the strength in their voice, making communication difficult and affecting their ability to maintain employment, and an increase in hospitalizations and the need for supportive equipment. The devastation of disease progression and loss of function in a person with full mental capacity has a severe negative impact on mental health and well-being. Patient groups noted that, coupled with the continued inability to access effective treatments, these patients experience a significant increase in anxiety, depression, and self-harm, requiring additional mental health support. Last, with the loss of physical function, patients must alter their homes for accessibility, which has a considerable financial impact. Patients hope the treatment will stop disease progression and even reverse muscle atrophy, which they consider is an improvement in terms of quality and quantity of life.

Patient and caregiver responders identified an unmet need for treatments in the adult population that offer stability and improved QoL through greater independence, improved strength (primarily in the arms and respiratory function), and a halting of progression. Patients believe, with improvements in these facets, they can achieve greater independence and a better QoL. Patients also noted that some of the largest barriers to treatment and challenges with currently available treatment are the unreasonable costs, the mode of delivery with intrathecal therapy, and the potential harms of treatment.

Given the few treatment options available for adults with SMA, experience with treatment was limited to nusinersen, risdiplam, alternative management of the disease, or no treatment. In the CSMAC survey, 41 (47%) patients provided information about their experience with SMA treatments; of those, 32 (78%) were receiving nusinersen and 9 (22%) were receiving risdiplam. Many of the patients from the CSMAC and MDC surveys and interviews were treatment-naive as a result of limited access to SMA treatments in Canada. Of the patients receiving nusinersen, 79% reported that they experienced improved respiratory function, endurance, upper limb and core strengths, and voice, with 15% reporting stabilization of their disease. The remaining 6% reported no stabilization or improvement. Patients were receiving treatment with nusinersen for 1 to 3.5 years. Negative experiences reported by patients receiving nusinersen included a wearing-off of treatment and a drop in function shortly before the next maintenance dose, which was subsequently rectified after treatment. Additional negative experience included temporary headaches, discomfort from intrathecal injections, as well as the travel and time off work required to receive treatment. Regardless, patients felt that the benefit of nusinersen, including gains in function, improved strength and energy, and disease stabilization, far outweighs the negative aspects of this treatment. Several patients from the MDC interviews revealed that they switched to risdiplam after the initiation of nusinersen because of limited access, financial constraints, and difficulties with the intrathecal administration. Patients also reported seeking alternative ways to manage their SMA, such as physiotherapy, exercise, and traditional Chinese medicine.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). In addition, as part of the nusinersen review, a panel of 4 clinical experts from across Canada was convened to characterize unmet therapeutic needs, assist in identifying and communicating situations where there are gaps in the evidence that could be addressed with the collection of additional data, promote the early identification of potential implementation challenges, gain further insight into the clinical management of patients living with a condition, and explore the potential place in therapy of the drug (e.g., potential reimbursement conditions). A summary of this panel discussion follows.

Unmet Needs

All the clinical experts agree that the main limitation and unmet need in adult type II and type III SMA patients is that there are no disease-modifying treatments available that can change the natural course of the disease, such as the ability to reverse the weakness associated with motor neuron degeneration and, as such, cannot address key outcomes for these patients. Currently, only nonpharmacologic treatments are mostly available in Canada, and are aimed at supporting function, mobility, and independence. Supportive measures, including ventilation and nutritional assistance, can help prolong survival, and assistive devices can help manage weakness; however, the effect is limited and there is a need for treatments that go beyond the benefits provided by the current standard of care.

The clinical experts noted the high degree of heterogeneity in the disease and disability. They agreed that the goals of treatment are dependent on the type of SMA and should, therefore, be individualized to the specific manifestations of the disease. Patients with type II SMA are by definition nonambulatory, with more significant and rapid progression than patients with type III SMA, and a shortened lifespan. Hence, prolonging life would be an important outcome metric for patients with type II SMA. Type III patients generally have a normal lifespan, although motor disability remains. The loss of motor function in the lower limbs translates to the loss of ambulation, resulting in a transition from walking to requiring assistive devices (such as a cane, walker, and subsequently a wheelchair) and, in more affected patients, loss of the use of the upper limbs may occur; both have a major impact on QoL. The clinical experts noted that there is no 1 outcome measure that fits all patients, and ambulation should not be considered the gold standard of an unmet need for type II and III patients. For some severely disabled, nonambulatory patients, being able to use the upper limbs can mean being able to use noninvasive ventilation (i.e., BiPAP), being able to use a computer or cell phone and communicate with the outside world, or even being employed. Additional parameters of importance include the ability to speak and swallow independently.

Place in Therapy

Currently, the treatment of adults with type II or III SMA is similar to that of patients with other neuromuscular disorders and consists of a multidisciplinary approach, including the provision of assistive devices, as needed (such as canes, walkers, and wheelchairs), physiotherapy to prevent contractures, and speech-language pathology for patients with dysphagia.

Given that there are currently no other disease-modifying drugs reimbursed by public drug plans for adults in Canada with type II or III SMA, the approach to the treatment of adults varied among the panel members. If nusinersen is recommended in this population, it could be considered a first-line treatment. Clinical experts noted that risdiplam has recently been given a positive CDEC recommendation for younger adults, so the treatment paradigm may shift in the future. One clinical expert noted that nusinersen is reimbursed for the treatment of adults with type II or type III SMA and is currently being used in this population in Quebec based on observational studies from Germany and Italy.

The clinical experts noted that nusinersen would not be used in patients who cannot tolerate the lumbar puncture due to pain or discomfort. As well, logistical issues, including care, travel, and numerous testing requirements may prevent some patients from being able to receive nusinersen. Moreover, it was noted that some patients may have already undergone spinal fusion surgery, impeding access to the intrathecal sac and making treatment with nusinersen increasingly complicated, although interventional radiology has assisted in making treatment more accessible to patients with complex spines.

The clinical experts suggested that for the adult, treatment-naive population, there is no guidance on whether other medications should be tried before nusinersen.

Patient Population

The diagnosis of 5q SMA requires molecular genetic profiling to identify biallelic variants in SMN1 or increases in SMN2 copy number. The experts highlighted the fact that early diagnostic tests often do not provide accurate results on SMN2 copy number, so repeat genetic results to confirm the SMN2 copy number should be considered.

The experts highlighted the lack of higher-level evidence (i.e., from RCTs) in this population to determine which adults with SMA are most likely to respond to treatment with nusinersen. The experts hypothesized that patients with higher functioning and who are ambulatory may demonstrate better responses because they have more nerves, leading to better function. As well, the clinical experts believed that patients without complex spines are more likely to have a better risk-benefit profile. However, most of the experts stressed that the earlier treatment is administered (i.e., in pre-symptomatic children), the greater the benefit observed, as shown in the CHERISH and SUNFISH studies for nusinersen and risdiplam, although there was some disagreement among the panel members as to whether age should be considered when determining response.

Conversely, the patients least suitable for treatment with nusinersen are those with complex spines due to spinal fusion surgery, those who cannot tolerate lumbar puncture, and those who have previously been treated as infants or children, as there is no evidence of benefit in these patients.

Assessing Response to Treatment

Clinically, there are numerous outcomes and measures to determine response to treatment. The experts noted that there is heterogeneity in the way patients are treated and assessed in Canada. Smaller centres may only be equipped to conduct standard neurologic examinations and may not have the personnel or equipment necessary to conduct the battery of validated functional tests conducted in tertiary neuromuscular centres.

The clinical experts agreed that — given the variation in response to treatment and individualized treatment goals — several outcome measures should be used to determine the benefits of treatment. The most common outcomes used in adults with SMA include the HFMSE, RULM, 6MWT, MRC strength score, and respiratory function (FEV₁, FVC, maximal inspiratory pressure, maximal expiratory pressure); all have been validated and have concordance in clinical practice and trials. They noted that other outcomes, such as bulbar function, strengthening of speech, or functional independence, may be important but are not routinely assessed.

The experts noted that in patients with type II and III SMA, disease progression occurs slowly, over the course of years, so the impact of treatment on these outcome measures is not likely to be seen over a short period of time. As such, they agreed that patients should be seen annually, and that more frequent visits would place an undue burden on the patient and not provide a meaningful assessment of change that could be used to inform clinical decisions. It was noted, however, that at the initiation of treatment, patients may be seen every 6 months.

Discontinuing Treatment

The clinical-expert panel agreed that the main reasons for discontinuation would be progression or worsening of disease, as well as any major complications or AEs related to therapy. The experts explained that trying to determine whether patients have achieved stabilization is difficult, given that SMA progresses so slowly that it can be unclear whether progression is due to a lack of therapeutic efficacy or the natural history of the disease. One clinical expert noted that, based on experience, the most common reason for discontinuation of nusinersen was a patient’s desire to stop, owing to a lack of improvement or inability to tolerate the treatment due to the time investment and overall burden associated with painful injections, risk of bleeding or infection, or nerve damage or meningitis.

Prescribing Conditions

The panel agreed that all patients with SMA should be referred to a tertiary centre with a variety of neuromuscular specialists, a multidisciplinary team, and access to interventional radiology or neurosurgery, as most patients will have complex spines. The clinical experts also noted that these centres would have to be able to admit patients who experience potential procedural or treatment-related complications. Additionally, respiratory therapists, physiotherapists, occupational therapists, and nutritionists are essential members of the care team and can help with nonpharmacological therapies and assess appropriate outcome measures.

One panel member contended that once patients have passed the diagnosis and initial-treatment phases, it may be reasonable to provide some services (i.e., physiotherapy and occupational therapy) at less specialized centres during the more chronic phase of the disease; however, patients would still be required to attend a specialized treatment centre to receive injections.

Clinician-Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups.

CADTH received clinician-group input from the Neuromuscular Disease Network for Canada, a pan-Canadian network launched in 2020 to bring together clinical, scientific, technical, and patient expertise in neuromuscular disease with the aim of improving the care, research, and treatment of neuromuscular diseases for all Canadians. Eight clinicians with experience treating SMA patients provided input to this submission.

The clinician group highlighted the 3 main disease-modifying treatments for SMA: nusinersen, risdiplam, and onasemnogene abeparvovec. The clinician group agreed that treatment goals for later-onset SMA would be to maintain current levels of motor function and strength, achieve disease stabilization (including the avoidance of ventilation), promote independence, and improve overall HRQoL. The clinician group highlighted the fact that risdiplam may be the only other treatment option for these patients, and that either nusinersen or risdiplam could be tried first. The clinician group explained that younger patients are most likely to derive benefit from nusinersen, and noted that it may be difficult to accurately identify which adults are most likely to derive benefit from nusinersen. The clinician group stated that a clinically meaningful response to treatment in adults is likely to consist of stabilization of motor and respiratory function, maintenance of independence, and a reduction in hospitalizations. Moreover, they noted that maintaining the ability to speak and avoiding ventilatory support have profound impacts on patient QoL, autonomy, and the ability to maintain vocational and social roles. The clinician group emphasized that current provincial monitoring requirements are too frequent and that there is significant variation among provinces. They agreed that patients should be assessed at treatment initiation, at 6 months, and then annually thereafter to reduce patient burden and strain on health care resources, given the slowly progressive functional decline that takes place over years. Last, the clinician group noted that nusinersen must be administered by or under the direction of health care providers experienced in performing lumbar puncture at designated treatment centres.

The clinician group provided anecdotal reports of 12 patients from Quebec with type II or III SMA. The CADTH clinical team has reviewed these reports, but they are not included in this report as they do not meet the review protocol criteria.

Drug-Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 3.

Table 3: Summary of Drug-Plan Input and Clinical-Expert Response

6MWT = 6-minute walk test; AEs = adverse events; CDEC = CADTH Canadian Drug Expert Committee; FVC = forced vital capacity; HFMSE = Hammersmith Functional Motor Scale Expanded; RCT = randomized controlled trial; RULM = Revised Upper Limb Module; SMA = spinal muscular atrophy.

Clinical Evidence

The clinical evidence included in the review of nusinersen is presented in 3 sections. The first section, the systematic review, includes key studies provided in the sponsor’s submission to CADTH, as well as studies that were selected according to an a priori protocol. The second section includes indirect evidence selected from the literature that met the selection criteria specified in the review. No indirect evidence was submitted by the sponsor. The third section includes any additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of nusinersen 2.4 mg/mL solution for intrathecal injection for the treatment of adults (> 18 years) with type II and type III 5q SMA.

Methods

Studies selected for inclusion in the systematic review included the key studies provided in the sponsor’s submission to CADTH and those meeting the selection criteria presented in Table 4. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Table 4: Inclusion Criteria for the Systematic Review

AEs = adverse events; FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; HRQoL = health-related quality of life; RCTs = randomized controlled trials; SAEs = serious adverse events; SMA = spinal muscular atrophy; WDAEs = withdrawal due to adverse events.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.²⁸

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) via Ovid, and Embase (1974‒) via Ovid. All Ovid searches were run simultaneously as a multi-file search. Duplicates were removed using Ovid deduplication for multi-file searches, followed by manual deduplication in Endnote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was nusinersen. The following clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, the WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on January 6, 2022. Regular alerts updated the search until the meeting of the CDEC on April 27, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.²⁹ Included in this search were websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the pre-determined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

A focused literature search for network meta-analyses dealing with SMA was run in MEDLINE All (1946–) on January 6, 2022. No limits were applied to the search.

Findings From the Literature

A total of 6 studies were identified for inclusion in the systematic review (Figure 2). The included studies are summarized in Table 5. A list of excluded studies is presented in Appendix 2.

Figure 2: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

No RCTs focusing on treatment-naive adults with type II or III SMA were identified as part of the CADTH literature search, and all available and included studies were of observational design and focused on real-world data; such studies have more limitations than RCTs.

As part of the reassessment for nusinersen, the sponsor provided CADTH with 7 studies: 1 prospective noncomparative cohort study (Hagenacker et al. [2020]¹⁷), 1 retrospective noncomparative cohort study (Maggi et al. [2020]¹⁸), 1 real-world cohort comprised of prospectively and retrospectively collected data from of 3 European registries, 1 retrospective chart review (Konersman et al. [2021]²²), 1 retrospective registry-based review (Pera et al. [2021]¹⁹), 1 open-label extension of trials investigating nusinersen (SHINE trial), and 1 critical review and meta-analysis of patients with type II or III SMA (Coratti et al. [2020]²⁰). No additional studies meeting the inclusion criteria for the systematic review were identified by the CADTH review team.

Hagenacker et al. (2020)

The study by Hagenacker et al. (2020)¹⁷ was a prospective, multi-centre, noncomparative observational study of 124 patients with 5q SMA treated with nusinersen from 10 German neurologic centres that aimed to investigate the safety and effectiveness of nusinersen in adults with 5q SMA over a 6-, 10-, and 14-month period. There was no funding source reported for this study.

Maggi et al. (2020)

Maggi et al. (2020)¹⁸ was a retrospective, noncomparative cohort study of 18 Italian secondary or tertiary care centres for SMA aimed at investigating the safety and effectiveness of nusinersen on motor function in a cohort of 116 adults with type II or type III SMA. No funding source from public, commercial, or not-for-profit sectors was reported for this study.

EU Registry Study

This observational, registry-based cohort study was conducted using combined data from 252 adults from 3 European registries (SMArtCARE from German-speaking countries; CuidAME from Spain; and ISMAR from Italy, UK, and US) with the aim of assessing the safety and effectiveness of nusinersen in adults with 5q SMA and at least 6 months of follow-up, and informing data analyses of the impact of nusinersen on motor function.^15,16 Two subcohorts were used for the statistical analyses: a before-and-after nusinersen treatment group of 75 patients with type III or IV SMA; and a comparative dataset from 252 adults with type III SMA (235 who had been treated with nusinersen and 17 who had not). Further details on the analysis populations are provided in the summary of the Statistical Analysis section.

Pera et al. (2021)

The study by Pera et al. (2021)¹⁹ was a noncomparative, longitudinal registry-based study of ISMAR in Italy, the UK, and the US, and aimed at reporting treatment outcomes in a cohort of 144 ambulant and nonambulant type III SMA patients treated with nusinersen.

Multiple foundations and organizations were acknowledged in the funding information, with support from the sponsor to the International SMA Consortium registry.

Although the enrolled population for this study included children and adolescents, details for the adult population were provided, so results for this group were presented and summarized in this report.

Konersman et al. (2021)

The study by Konersman et al. (2021)²² was a noncomparative retrospective chart review of 35 older SMA patients (aged 5 to 58 years) in the US conducted from April 2017 to June 2019. The objective of this study was to determine the effectiveness of nusinersen at improving motor function, to report on methods of administration, and to detail the AEs seen in adolescents and adults with SMA.

No stratified or subgroup analyses were conducted for the population of interest defined in Table 4. Therefore, this study will not be further summarized because the relevant population for the reassessment could not be evaluated.

SHINE

The SHINE trial (Study CS11)²¹ was identified as a key study by the sponsor. It is an ongoing open-label extension study of SMA patients who previously participated in investigational studies to evaluate the long-term safety, tolerability, and efficacy of nusinersen. All participants received 12 mg (5 mL) of nusinersen. The study consisted of a screening phase, a treatment period that included a loading dose and maintenance treatment, a post-treatment follow-up period, and an end-of-study evaluation for all patients who completed an index study of nusinersen. A modified maintenance dosing regimen was used in Study CS11; the index studies had different maintenance schedules. After the loading-dose period, nusinersen is administered intrathecally once every 4 months to all participants. The planned duration of participation in the study is 5 years after administration of the modified maintenance dosing regimen. As of the data cut-off, 190 patients were included in the ongoing SHINE trial as part of the later-onset analysis group; data from 28 of these adults informed the efficacy analysis.

Although a subgroup of later-onset adults with type II or III SMA was used in a post hoc analysis, these patients had previously been enrolled in index studies for nusinersen (studies CS2 and CS12) and, therefore, received treatment with nusinersen when they were younger than 18 years. Results provided in this analysis are confounded by the age at which treatment was initiated and cannot be used to assess treatment benefit in the treatment-naive adult population. Although identified as a key study by the sponsor, the SHINE later-onset analysis will not be further summarized or appraised because it is not relevant to the reassessment of nusinersen.

Coratti et al. (2021)

Coratti et al. (2021)²⁰ was a critical review and meta-analysis of patients with type II and III SMA treated with nusinersen that was included in the list of key studies submitted by the sponsor. This study is further summarized in the Other Relevant Evidence section. This study was also identified in the literature search; however, the study design did not meet the eligibility criteria of the systematic review conducted by CADTH because the population combined children (< 18 years) and adults (≥ 18 years), even though a subgroup analysis for adults was presented.

It is worth noting that all primary studies of the population of interest included in the meta-analysis that were publicly available were also identified in the CADTH systematic review (Hagenacker et al. [2020],¹⁷ Maggi et al. [2020],¹⁸ Konersman et al. [2021],²² and Pera et al. [2021]¹⁹).

Populations

Inclusion and Exclusion Criteria

Limited information on inclusion and exclusion criteria were provided for each study. In all studies submitted by the sponsor, a common inclusion criterion was a confirmed diagnosis of SMA, with genetic documentation of 5q SMA homozygous deletions of exons 7, 8, or both, or with compound heterozygous mutations.^15-19

In the studies by Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020),¹⁸ patients were required to be adults (≥ 18 years), with nusinersen treatment initiation after they reached the age of 18 years. However, in the EU registry study^15,16 and Pera et al. (2021)¹⁹ studies, patients younger than 18 years were also included.

Additionally, when reported, patients were required to have type II or type III SMA (Maggi et al. [2020]¹⁸ and Pera et al. [2021]¹⁹). In Pera et al. (2021),¹⁹ type III SMA was further subdivided into type IIIA or IIIB, depending on age at symptom onset (before or after 3 years). In the Hagenacker et al. (2020)¹⁷ and EU registry study,^15,16 cases of type I and type IV SMA were included. In Hagenacker et al. (2020),¹⁷ 1 patient with type I SMA and 1 patient with type IV SMA were enrolled, but they were only included in the 10-month analysis. In the EU registry study^15,16 submitted by the sponsor, patients of all SMA types were included; however, the analysis only included patients with type III or IV SMA, and only the adult population with type III SMA was of interest for this review, so only results for this population will be summarized. In most cases, to be eligible for inclusion in the analyses, patients had to have undergone a minimum of 6 months of treatment.

Baseline Characteristics

Hagenacker et al. (2020)

Baseline characteristics for patients in the Hagenacker et al. (2020)¹⁷ study were provided separately for patients with 6, 10, and 14 months follow-up; these are summarized in Table 6. A total of 139 patients completed the 6-month assessment. The primary end point analysis included 124 (89%) patients with a treatment period of at least 6 months, 92 (66%) patients with a treatment period of 10 months, and 57 (41%) patients with a treatment period of 14 months. Patient with 6 months of follow-up ranged in age from 16 to 65 years, with a mean age of 36 years (SD = 12). Most patients were SMA type III (77 [62%]) or type II (45 [36%]). In the 6-month analysis, 2 patients (2%) were SMA type I, whereas in the 10-month analysis, 1 patient (1%) each was included in the type I and type IV SMA groups. Most patients had 3 (48 [39%]) or 4 (41 [33%]) SMN2 copies, but SMN2 status was unknown in 24 (19%) patients. Only 46 (37%), 35 (38%), and 23 (40%) patients were ambulant at baseline in the 6-, 10-, and 14-month analyses, respectively, and 20% to 25% had prior spondylodesis. According to the authors, most patients had low HFMSE scores at baseline, defined as a score of less than 35 points (69% at 6 months, 64% at 10 months, and 61% at 14 months).

Table 6: Summary of Baseline Characteristics (Hagenacker et al. [2020])

6MWT = 6-minute walk test; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Hagenacker et al. (2020).¹⁷

Maggi et al. (2020)

A summary of baseline characteristics of patients in the study by Maggi et al. (2020)¹⁸ are summarized in Table 7. Of the 116 patients enrolled, 103 (88.8%) had type III SMA — equally distributed between type III sitters (n = 51) and type III walkers (n = 52) — and 13 (11.2%) had type II SMA. Enrolled patients had a median age at onset of 3 years (range = 0 to 17) and a median age at the beginning of treatment (T0) of 34 years (range = 18 to 72). Median age at treatment initiation was 40 years for sitters and 33 years for walkers. The majority of patients were male (58.62%), and most patients had 3 (n = 36 [31%]) or 4 (n = 54 [46.6%]) copies of SMN2, but some had 2 copies (n = 5 [4.3%]) and some copy numbers were unknown (n = 21 [18.1%]).

Ventilatory support at treatment initiation was required by 21 (18.1%) patients; of these, 10 (76.9%) had type II SMA, 8 (15.7%) were type III sitters, and 3 (5.8%) were type III walkers. Prior surgery for scoliosis was reported in 16 (13.8%) patients overall, and in 61.5% of type II patients, 13.7% of type III sitters, and 1.9% of type III walkers.¹⁸

Table 7: Summary of Baseline Characteristics (Maggi et al. [2020])

6MWT = 6-minute walk test; FEV₁ = forced expired volume in 1 second; FVC = forced vital capacity; HFMSE = Hammersmith Functional Motor Scale Expanded; max = maximum; min = minimum; N/A = not available; RULM = Revised Upper Limb Module; SMA = spinal muscular atrophy; T0 = treatment initiation.

Note: All patients carried homozygous SMN1 exon 7 deletions, except 3: 1 with a nonsense and 2 with a missense mutation on the other allele.

^aHFMSE was available for all patients; remaining assessments were not available for all patients.

^bSMN2 copy number was not available in 21 patients.

^cA further patient stopped ventilatory support before T0 due to poor tolerance.

^dTwo patients used ventilatory support due to obstructive sleep apnea and a further patient refused ventilatory support although indicated.

^ePatient able to walk for few steps with cane.

Source: Maggi et al. (2020).¹⁸

EU Registry Study

Baseline characteristics for the cohort of patients from the EU registry study^15,16 are summarized by registry and the pooled population in Table 8. Baseline characteristics for all patients in each registry for all ages and for SMA types III and IV were provided; however, as previously mentioned, only characteristics for adult type III populations were summarized.

All adult type III SMA patients in the SMArtCARE registry were treated with nusinersen (n = 151). Adult type III patients who were treated with nusinersen and who were untreated were included in the ISMAR (n = 53, n = 5, respectively) and CuidAME (n = 24, n = 9, respectively) registries.^15,16

The mean age of type III adults treated with nusinersen across registries was 35 to 39 years (range = 18 to 71 years). In untreated type III SMA patients, the mean age was 36 years in the CuidAME registry and 46 years in the ISMAR registry. The age at diagnosis of SMA was not reported. In general, most patients were male across registries, particularly those treated with nusinersen (60% to 63%); in the untreated population, the majority were female (60% in ISMAR, 56% in CuidAME). In type III adults, most had 4 copies of SMN2 (38% to 67% in treated and untreated patients across registries). Across registries, most patients were not fully ambulatory (54% of nusinersen-treated patients and 29% of untreated patients), most did not have scoliosis (75% to 100% of treated and untreated patients), and few required noninvasive ventilation (6% to 7%). No information on feeding support was provided in the EU registry study.^15,16

Pera et al. (2021)

Baseline characteristics for the study by Pera et al. (2021)¹⁹ are summarized in Table 9. Only 67 patients (47%) in the Pera et al. (2021)¹⁹ cohort were adults (≥ 18 years). Baseline characteristics for the adult population were not reported in the article.

Treatment Exposure

As the studies were observational in nature, none had an intervention. Instead, the exposure of interest in these studies was nusinersen. Given the design of all the studies submitted by the sponsor, the only treatment in all studies was nusinersen 12 mg administered by intrathecal injection, although the specific dose was not mentioned in the study by Pera et al. (2021).¹⁹ The dosing regimen for nusinersen consisted of induction injections on days 1, 14, 28, and 63, with maintenance injections every 4 months thereafter, in accordance with the product label, and was only reported in the studies by Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020).¹⁸ Dosing regimen was not described in the EU registry study^15,16 or in Pera et al. (2021).¹⁹

In Hagenacker et al. (2020),¹⁷ intrathecal injections were administered by a trained neurologist or neuroradiologist via conventional, fluoroscopy-guided or CT (CT)-guided lumbar puncture.¹⁷ In Maggi et al. (2020),¹⁸ intrathecal injections were primarily performed with standard lumbar access in 85 (82.5%) patients, CT-guided procedures in 4 (3.9%) patients, and X-ray-guided procedures in 14 (13.6%) patients with type III SMA. A total of 7 (8.2%) patients with type III SMA shifted from manual to imaging-guided techniques during treatment. Only 1(7.7%) patient with type II SMA (N = 13) was managed without imaging guidance, whereas 7 (53.8%) received intrathecal injection via CT-guided procedures and 5 (38.5%) received intrathecal injection with X-ray-guided approaches.¹⁸

In the EU registry study,^15,16 a comparison between nusinersen-treated and untreated adults was made in 1 subcohort; however, no information on the intrathecal administration or procedures were provided. No information on nusinersen dosing or administration was provided in the study by Pera et al. (2021).¹⁹

Limited information on concomitant medications and cointerventions was available.

Outcomes

A list of effectiveness end points identified in the CADTH review protocol that were assessed in the studies included in this review is provided in Table 10. These end points are further summarized in the following sections, when information was available. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

In both Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020),¹⁸ available outcomes were assessed at months 6, 10, and 14. Outcomes in Pera et al. (2021)¹⁹ were evaluated at 12 months. All patients included in the EU registry study^15,16 had to have a treatment duration of at least 6 months; however, the specific time of assessment of effectiveness outcomes were not reported.

Table 9: Summary of Baseline Characteristics (Pera et al. [2021])

6MWT = 6-minute walk test; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Pera et al. (2021).¹⁹ This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 International Licence. Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8351459/

Table 10: Outcomes Evaluated in the Included Studies

6MWT = 6-minute walk test; AEs = adverse events; FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; HFMSE = Hammersmith Functional Motor Scale Expanded; HRQoL = health-related quality of life; RULM = Revised Upper Limb Module; SAEs = serious adverse events.

Sources: Hagenacker et al. (2020),¹⁷ Maggi et al. (2020),¹⁸ Sponsor submission,^15,16 Pera et al. (2021).¹⁹

Motor-Function Outcomes

Motor function was the most commonly reported outcome in the included studies, with the HFMSE, RULM, and 6MWT the most frequently used measures to evaluate motor function in type II and III SMA adults.

The HFMSE consists of 33 itemized motor functions that assess activities of daily living. Each item is scored on a scale from 0 to 2, with higher scores indicating better motor function, up to a maximum of 66 points. According to Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020),¹⁸ a score change of at least 3 points is considered a clinically meaningful improvement and indicated a response to treatment. Additional information on the measurement properties and minimally important difference in HFMSE results are described in Table 26 of Appendix 4.

The RULM consists of 20 items with a maximum of 37 points, with higher scores indicating better arm function. Score changes of at least 2 points were considered to be clinically meaningful and indicated a response to treatment, according to the Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020).¹⁸ Additional information on the measurement properties and minimally important difference in RULM results are described in Table 26 of Appendix 4.

The 6MWT measures the maximum distance a patient can walk in 6 minutes. In Maggi et al. (2020),¹⁸ responders were defined as patients who improved from baseline by at least 30 m. Additional information on the measurement properties and minimally important difference in 6MWT results are described in Table 26 of Appendix 4.

Additional motor-function outcomes assessed included timed-function tests (timed run or walk 10 m, timed rise from floor, timed rise from chair, and timed climb 4 standard steps) in Maggi et al. (2020),¹⁸ with all results expressed as velocities; however, no information on these methods of assessment were provided.

Respiratory Outcomes

Respiratory outcomes were not frequently assessed in the included studies, with only Maggi et al. (2020)¹⁸ evaluating pulmonary function with the percent-predicted FEV₁ and FVC. No description of these outcomes was provided.

Other Outcomes of Interest

The included studies did not provide information on any additional outcomes of interest defined in Table 4, including bulbar function, feeding requirements, survival, hospitalization, HRQoL, anatomic-related outcomes, or caregiver burden.

Safety Outcomes

Safety was not assessed using defined measures in any of the studies. In Hagenacker et al. (2020),¹⁷ AEs were reported as adverse drug reactions according to MedDRA (version 21.1). In Maggi et al. (2020),¹⁸ safety evaluations included vital signs, clinical and laboratory findings, and patient-reported AEs, all categorized by severity and relationship to nusinersen.¹⁸ The EU registry strudy^15,16 reported the frequency of AEs, specifically in the adult population; however, AEs were not recorded in a standardized manner and the MedDRA classification was not used. AEs were not recorded in the study by Pera et al. (2021).¹⁹

Statistical Analysis

Hagenacker et al. (2020)

Statistical analyses were based on mean comparisons from baseline to months 6, 10, and 14 for primary and secondary end points, with the corresponding 95% CIs, and with the Wilcoxon signed-rank test. A difference of 0.31 or greater in a treatment-effect size can be detected with a power of 80% and with a 2-sided alpha of 0.05. This estimation was considered suitable by the authors for the primary end point and for the descriptive analysis of the secondary end points. A mixed model was used to estimate the treatment effect on the HFMSE score. The model used sex, time, age, spondylodesis, and HFMSE baseline score as fixed effects and the patient as a random effect. Outliers were not removed because there were no indications of incorrect measurements. No imputation of missing data was done for the 6-month, 10-month, or 14-month analyses. No alpha adjustment was done for potentially inflated type I error in the secondary end point analyses.¹⁷

A protocol for this study could not be located; however, in the article by Hagenacker et al. (2020),¹⁷ it was stated that preplanned subgroup analyses included analyses by HFMSE baseline score (≥ 35 versus < 35) or previous spondylodesis (yes versus no). Post hoc subgroup analyses were based on ambulant versus nonambulant status, and on SMA type (II versus III). Subgroup analyses were done by Mann–Whitney U test.

Sensitivity analyses were conducted by replacing missing baseline values for the primary end point using the existing values at later time points, such that these values were included in the analysis with a change from baseline of 0, and thus replaced conservatively. An additional sensitivity analysis also considered replacing the missing baseline values with values 5 points higher than at a later time points.¹⁷ The results and methods of these sensitivity analyses were not reported.¹⁷

Maggi et al. (2020)

No formal power calculation was presented for this study. The magnitude of change in outcomes from baseline at months 6, 10, and 14 was assessed in SMA type II and type III groups, including 2 type III subgroups of sitters and walkers. Walkers were defined as patients able to take at least a few steps independently or with aids (e.g., cane) but without the assistance of others. Responders to treatment with nusinersen were defined as patients who improved from baseline by at least 3 points on the HFMSE, 2 points on the RULM total score, or 30 m on the 6MWT. Responders in at least 1 of the 3 outcomes were defined as overall responders to treatment with nusinersen.¹⁸

No protocol for this study could be located. Change from baseline in study outcomes were summarized as mean (SD) or median (range), as appropriate. For comparisons of quantitative and ordinal variables between times of assessment, the Wilcoxon-Mann–Whitney test or Student’s t-test was used. For correlations between quantitative and/or ordinal variables, Spearman’s method was used. Distributions of categorical variables were compared with the Chi-square test. The effect of age, sex, and SMN2 copy number on treatment response was explored with logistic regression. Statistical significance was set at P < 0.05. No adjustments for multiplicity were done.¹⁸

No pre-specified sensitivity analyses were reported.

EU Registry Study

No protocol was provided for this study. No formal power calculation was reported for this study, and sample size was based on enrolment in the registries.

Two sets of data analyses were performed for the registry-based cohorts: a piecewise linear mixed model of pre- and post-treatment initiation comparisons; and an analysis of treated versus untreated patients based on a mixed-effect model. The first set of analyses explored the effects of treatment in the group of 75 adults with type III or IV SMA with data available before and after treatment initiation, and used piecewise linear mixed modelling of pre- and post-treatment outcomes to assess changes in HFMSE and RULM. For this analysis, the slope before the start of treatment was compared with the slope after treatment initiation in the treated cohort of patients (time spline). A piecewise linear mixed model was developed to consider the impact of treatment on functional ability scores, which estimated slopes of change over time separately in each treatment group, permitting assessment of whether the trajectory of the outcome over time differed between treated and untreated patients. Results were expressed as estimated change in points/week (95% CI). From the registries, 75 of 235 (31.91%) treated type III or type IV adults with at least 1 visit before the start of treatment and with at least 6 months of follow-up after treatment initiation were retained in the piecewise mixed models. These included:

• 38 patients from the Italian registry

• 26 patients from the German registry

• 11 patients from the Spanish registry.

The second set of analyses explored the effects of treatment by comparing outcomes among 235 patients treated with nusinersen and 17 untreated patients, using linear mixed-effects modelling to assess changes in disease progression, measured by changes in HFMSE and RULM scores over time. Results were presented for the overall type III population and the adult population (type III and IV SMA). Subgroup analyses were presented for type III fully ambulatory and type III not fully ambulatory patients. No pre-specified sensitivity analyses were reported.

The imputation of missing values for HFMSE and RULM were based on the scale total scores. Any missing baseline values were imputed using median within stratu, considering nonmissing baseline records. For post-baseline visits, flanked by nonmissing visits, missing values were imputed using linear interpolation. Only actual visits with nonmissing data were imputed for each patient. For the HFMSE, if more than 6 item scores were missing for an individual, then the total score imputed was as if all the 33 items were missing. For the RULM, if 3 or more items were missing for an individual, then the total score imputed was as if all 19 items were missing.¹⁵ At the final assessment, if the date was available and at least 1 item was nonmissing, the last observation carried forward (LOCF) method was used to impute missing HFMSE, RULM, and 6MWT scores. The multiple imputation method was also used, in which baseline missing data were handled with a chained equation with 10 imputations and stratified by registry.

The dependency in the data due to repeated measures was accounted for by a random intercept per individual, and an autoregressive covariance R matrix was used as correlation structure. The default estimation method (restricted maximum likelihood) was used for the covariance parameters. The Kenward-Roger method was used to compute degrees of freedom for the tests of fixed effects. The structure of the models was kept uniform with regard to the fixed- and random-effects structure. In consultation with 2 clinical experts, the EU registry study^15,16 investigators included the following potential confounders in the regression model: age at onset of symptoms (onset at ≥ 3 years versus onset at < 3 years), sex (male versus female), number of SMA gene copies, motor function (not fully ambulatory or fully ambulatory), disease duration, registry (Germany, Italy, or Spain), age at baseline, and baseline score value.

Pera et al. (2021)

No protocol for this study could be located. Two analyses were conducted. First, all patients with a follow-up at 1 year of treatment were analyzed to evaluate 12-month changes in functional measures. Comparisons of measurements from baseline to 12 months were performed using estimates and 95% CI of before-and-after differences and the Wilcoxon signed-rank test. Twelve-month changes were also analyzed, subdividing the cohort by functional status (sitters versus walkers), SMA type III subtype (IIIA or IIIB), and age (children younger than 18 years versus adults). Details on 12-month trajectories grouped changes as stable (± 2 points), improved (more than 2 points), or declined (more than –2 points).¹⁹ For this review, only results for the overall adult population and for subgroup data for patients older than 20 years are presented.

A second analysis was added as no imputation was performed on missing data, and a number of patients failed the 12-month assessment because of restricted access to hospitals during the COVID-19 pandemic or other reasons. The second analysis consisted of a mixed model to estimate changes in the whole type III population and exclude possible selection bias. The model was set up with measurements at baseline (age, sex, time, disease duration, SMN2 copy number, disease onset, and SMA function) as fixed effects and the patient as a random effect. To make inferences about mean slopes by age at onset, the model was expanded to include appropriate main-effect and interaction terms in the model.¹⁹

Sensitivity analyses were performed for HFMSE and RULM after exclusion of the 3 nonsitter patients. An additional sensitivity analysis was conducted after the exclusion of 27 patients with baseline RULM scores equal to 37.¹⁹ No further information on the sensitivity analyses were provided.

Results

Patient Disposition

Hagenacker et al. (2020)

Of the 173 patients assessed for eligibility, 139 (80%) completed the 6-month assessment, 105 (61%) completed the 10-month assessment, and 61 (35%) completed the 14-month assessment at the time of analysis. Fifteen patients were excluded from the 6-month time of assessment because of missing baseline or 6-month time-of-assessment values, resulting in 124 patients available for the 6-month analysis. Thirty-four patients were subsequently excluded before 10 months — 30 who had not yet reached the 10-month assessment and 4 who withdrew (2 because of AEs and 2 who chose to) — and 13 were missing baseline values, resulting in a total of 92 patients in the 10-month assessment. A further 48 patients were excluded from the 14-month assessment — 44 and 4 did not have 14-month or baseline values, respectively, resulting in a total of 57 patients in the 14-month analysis.¹⁷

Maggi et al. (2020)

Of the 149 patients screened, 33 were excluded, resulting in a total of 116 included patients. Patients were excluded due to age (n = 5), disease onset after 18 years of age (n = 4), inability to complete the treatment loading phase because of side effects (n = 2), clinical trial enrolment (n = 1), no 6-month follow-up at the time of data collection (n = 21).¹⁸

EU Registry Study

A quality assessment of the individual registries was conducted; however, no information was provided on the quality of the databases in these registries. Data from the SMArtCARE, ISMAR, and CuidAME registries were combined to create a cohort of SMA patients from multiple countries. The SMArtCARE registry is an indication-specific registry that was initiated before the approval of nusinersen in Europe, but it did not start enrolment until the launch of nusinersen. The SMArtCARE registry aimed to evaluate all people with 5q SMA, regardless of their current treatment. Data collection took place as part of a patient’s regular, clinically recommended routine visits, which depended on the treatment they were receiving. Standardized results were collected during routine visits at regular intervals of 4 (nusinersen treatment) or 6 months (maximum time frame recommended by guidelines). The ISMAR registry is an ongoing collaboration between 3 large national networks in 16 locations in Italy, the UK, and the US to gain understanding of the disease and response to treatments. The data for the registry were collected as part of regular, clinically recommended routine visits, depending on the treatment the patient was receiving. The CuidAME registry collected data from 6 clinics relevant to the care of people diagnosed with 5q SMA, and included all patients with SMA, regardless of the treatment they were receiving.

The SMArtCARE registry included 240 type III SMA patients, the ISMAR registry included 114 type III SMA patients, and the CuidAME registry included 55 type III SMA patients. Overall, in the cross-registry analysis, 252 adults with type III SMA (235 treated with nusinersen and 17 untreated) were included in the analysis. Full details of the patient flow in the registries can be found in Figure 3.

Figure 3: Patient Flow in the EU Registry Study

FU = follow-up; RCT = randomized controlled trial; SMA = spinal muscular atrophy.

Note: a) Patient flow from the SMArtCARE registry in Germany; b) Patient flow from the ISMAR registry in Italy, the UK, and the US; c) Patient flow from the CuidAME registry in Spain.

Source: Sponsor submission.^15,16

Pera et al. (2021)

A summary of patient disposition was not provided in the study by Pera et al. (2021).¹⁹ A total of 144 SMA type III patients were enrolled in the study based on data collected from ISMAR.

Exposure to Study Treatments

Reporting of drug exposure varied in the included studies, given the study designs. However, patients all received nusinersen 12 mg according to standard protocols for at least 6 months. The follow-up duration in the Hagenacker et al. (2020)¹⁷ and Maggi et al. (2020)¹⁸ studies was up to 14 months, with assessments conducted at 6, 10, and 14 months.

In the EU registry study^15,16 submitted by the sponsor, patients were required to have at least 6 months of treatment with nusinersen. In the SMArtCARE registry, the median length of follow-up in adults with type III SMA (n = 151) was 329 days (10.82 months; range = 168 to 930 days), with a mean of 5.01 (SD = 1.53) visits. In the ISMAR registry, the median length of follow-up in adults with type III SMA was 489 days (16.08 months; range = 161 to 757 days) in treated patients (n = 53), and 567 days (18.64 months; range = 520 to 700 days) in untreated patients (n = 5). Treated patients had a mean of 9.40 (SD = 5.88) visits, whereas untreated patients had a mean of 8.20 (SD = 8.29) visits during the entire follow-up period. In the CuidAME registry, the median follow-up in treated (n = 24) and untreated (n = 9) adults with type II SMA was 412 days (13.55 months; range = 167 to 706 days) and 413 days (13.58 months; range = 189 to 1,462 days), respectively. The mean number of visits was 4.96 (SD = 1.60) in treated patients and 4.22 (SD = 1.86) in untreated patients.

In Pera et al. (2021),¹⁹ patients treated with nusinersen were required to have a minimum of 12 months of follow-up, and the mean duration of follow-up was 1.83 years (SD = 0.61).

Effectiveness

Only effectiveness outcomes and analyses of subgroups identified in the review protocol are reported here. Refer to Appendix 3 for detailed effectiveness data.

Motor-Function Outcomes

HFMSE

Hagenacker et al. (2020): Results for HFMSE in the Hagenacker et al. (2020)¹⁷ study are summarized in Table 11. The mean HFMSE scores at all time points were greater than at baseline, with mean differences of 1.73 points (95% CI, 1.05 to 2.41) at 6 months (n = 124), 2.58 points (95% CI, 1.76 to 3.39) at 10 months (n = 92), and 3.12 points (95% CI, 2.06 to 4.19) at 14 months (n = 57). Compared with baseline, 35 (28%) of 124 patients at 6 months, 33 (35%) of 92 patients at 10 months, and 23 (40%) of 57 patients at 14 months had a greater than 3-point increase in HFMSE score, which the authors determined to be clinically meaningful.

Subgroup analyses of mean change in HFMSE score from baseline were consistent with the primary analysis at all time points in patients in all subgroups. However, changes were numerically larger in patients with type III SMA, ambulatory patients, and those without prior spondylodesis (Refer to Table 12).

In the subgroup analyses, more type III patients than type II patients had a greater than 3-point increase in HFMSE scores at all time points (23 [30%] versus 1 [2%] at 6 months, 19 [32%] versus 2 [7%] at 10 months, and 15 [41%] versus 1 [5%] at 14 months).

Results for sensitivity analyses, in which missing baseline values were replaced with existing values at later time points and missing baseline values were replaced with values 5 points higher than the values at later time points, remained consistent with the primary analysis (not shown).

Maggi et al. (2020): Results for HFMSE from the study by Maggi et al. (2020)¹⁸ are summarized in Table 13. In SMA type III patients, the mean change from baseline in HFMSE score at 6 months (n = 103), 10 months (n = 75), and 14 months (n = 46) was 1.48 points (SD = 2.28), 2.44 points (SD = 2.8), and 2.85 points (SD = 2.93), respectively. In patients with type II SMA (n = 13), the mean change from baseline in HFMSE score reached a high of 1.2 points (SD = 2.7) at 14 months; however, the median HFMSE change was zero at all assessment time points.

Table 12: Subgroup Analysis of Changes in HFMSE and RULM Scores vs. Baseline (Hagenacker et al. [2020])

CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module; SD = standard deviation.; vs. = versus.

Note: P values were not adjusted for type I error.

*Data are mean difference (SD; 95% CI) or mean difference (SD).

^†For 6-month, 10-month, or 14-month values vs. baseline.

Source: Hagenacker et al. (2020).¹⁷

Clinically meaningful improvements in HFMSE (i.e., a 3-point change), as concluded by the authors, at all time points are summarized in Figure 4. The proportion of patients who had a 3-point or greater change from baseline in HFMSE score at 6 to 14 months ranged from 28% to 49% for all SMA patients, 8% to 20% for patients with type II SMA, 31% to 52% for patients with type III SMA, 27% to 58% for type III SMA sitters, and 35% to 48% for type III SMA walkers.

Figure 4: Responder Analyses for Change From Baseline in HFMSE, RULM, and 6MWT Scores in Adults With type II or type III SMA (Maggi et al. [2020])

6MWT = 6-minute walk test; HFMSE = Hammersmith Functional Motor Scale Expanded; NA = not applicable; RULM = Revised Upper Limb Module; SMA = spinal muscular atrophy; T6 = 6 months of treatment; T10 = 10 months of treatment; T14 = 14 months of treatment.

Note: Responders were defined as having a change in HFMSE score from baseline of at least 3 points, a change in RULM score from baseline of at least 2 points, and a change in 6MWT distance from baseline of at least 30 m. “Overall” response was defined as a clinically meaningful response in at least 1 measure.

Source: Maggi et al. (2020).¹⁸

EU Registry Study: Results for the change in the HFMSE total score for the pre- and post-treatment nusinersen analysis are summarized in Table 14. The model results showed that before the start of nusinersen treatment, the HFMSE score decreased by an average of 0.00006 points per week. After treatment with nusinersen, the slope for the change in score per week was 0.2575 (95% CI, 0.01038 to 0.04112) (n = 75).^15,16

Results for the change in the HFMSE total score between nusinersen-treated (n = 235) and untreated (n = 17) patients are summarized in Table 15. The slope for the change in nusinersen-treated patients was 0.02907 per week (95% CI, 0.01930 to 0.03884), compared with −0.01129 per week (95% CI, −0.03289 to 0.01031) in untreated patients.^15,16 The slopes were similar regardless of the imputation method for missing data (LOCF or multiple imputation).

Pera et al. (2021): Results for the change in HFMSE in adults in Pera et al. (2021)¹⁹ are summarized in Table 16. In the primary analysis, the results for the mean change in HFMSE score from baseline for the adult population at 12-months (n = 45) was 0.79 points (95% CI, –0.29 to 1.87). Supplemental results based on descriptive statistics for all adults older than 20 years with type IIIA and type IIIB SMA varied, with mean changes from baseline at 12 months of 0.26 points (SD = 2.66) for ambulant patients (n = 19), and 1.58 points (SD = 3.91) for nonambulant patients (n = 26).

Table 14: Pre- and Post-Treatment With Nusinersen (n = 75; EU Registry Study)

6MWT = 6-minute walk test; CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module.

^aPiecewise linear mixed models adjusted for age at onset of symptoms (onset ≥ 3 years vs. onset < 3 years), sex (male vs. female), number of SMA gene copies, motor SMA function (not fully ambulatory or fully ambulatory), disease duration, registry (Germany, Italy, or Spain), age at baseline, and baseline score value.

Source: Sponsor submission.^15,16

Table 15: Nusinersen-Treated vs. Untreated (Best Supportive Care Alone) Patients (EU Registry Study)

6MWT = 6-minute walk test; CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module.

^aStandard Linear Mixed Model adjusted for age at onset of symptoms (Onset ≥ 3 years vs. Onset < 3 years), sex (Male vs. Female), number of SMA gene copies, motor SMA function (non-fully ambulatory; fully ambulatory), disease duration, registry (Italy; Germany; Spain), age at baseline, and baseline score value

Source: Sponsor submission.^15,16

Table 16: Changes in HFMSE (12 Months vs. Baseline) in the Overall Adult and the Greater Than 20 Year SMA type III Populations (Pera et al. [2021])

CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded; N/A = not applicable; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Pera et al. (2021).¹⁹ This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 International Licence. Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8351459/

Stability of HFMSE scores at 12 months in adults older than 20 years are summarized in Table 19. In all adults (> 20 years), the HFMSE results declined in 11.1% of patients, remained stable in 53.3%, and improved in 35.6%, based on the definitions used to categorize stability. These results were similar in walkers and sitters.

In an additional comparison with an external untreated cohort, patients older than 20 years treated with nusinersen (n = 45) had a mean 12-month change in HFMSE of 1.02 points (SD = 3.47), compared with –1.65 points (SD = 3.42) in the untreated external cohort (n = 49).

RULM

Hagenacker et al. (2020): Results for RULM from the Hagenacker et al. (2020)¹⁷ study are summarized in Table 11. The mean RULM scores at all time points were similar to those at baseline, with mean differences of 0.66 points (95% CI, 0.26 to 1.05) at 6 months (n = 120), 0.59 points (95% CI, 0.15 to 1.03) at 10 months (n = 90), and 1.09 points (95% CI, 0.62 to 1.55) at 14 months (n = 58). At the 6-month follow-up, a greater than 2-point increase in RULM score, which was considered clinically meaningful by the authors, was observed in 28 (23%) patients, whereas 74 (64%) patients showed no meaningful change and 18 (15%) and 10 (8%) showed a decline of 1 point or more, or a decline of 2 points or more, respectively.

Subgroup analysis for RULM scores in type II and type III patients is presented in Table 12.

Maggi et al. (2020): Results for RULM scores in Maggi et al. (2020)¹⁸ are summarized in Table 13. In SMA type III patients, the mean change from baseline in RULM score at 6 months (n = 102), 10 months (n = 71), and 14 months (n = 44) was 0.31 points (SD = 1.97), 0.61 points (SD = 2.08), and 0.86 points (SD = 2.18), respectively. Patients with type II SMA had changes in mean RULM scores of 0.8 points (SD = 1.95) at 6 months (n = 12), 1.67 points (SD = 1.8) at 10 months (n = 9), and 1.6 points (SD = 1.52) at 14 months (n = 5).

Clinically meaningful improvements in HFMSE, according to the authors, at all time points are summarized in Figure 4. The proportion of patients with a 2-point change in RULM score at 6 to 14 months ranged from 21% to 35% for all SMA patients, 25% to 60% for patients with type II SMA, 21% to 32% for patients with type III SMA, 29% to 53% for type III SMA sitters, and 11% to 16% for type III SMA walkers.

EU Registry Study: Results for the change in RULM score from the pre- and post-treatment analysis in the EU registry study^15,16 are summarized in Table 14. The model results showed that before the start of nusinersen treatment, the RULM score decreased by an average of 0.00745 points per week (95% CI, –0.01401 to 0.0009). After treatment with nusinersen, the slope for the change in score per week was 0.002569 points per week (95% CI, −0.00533 to 0.01047). The same trend was seen when patients with a ceiling effect were excluded from the analysis (not shown).

Results comparing the change in RULM between nusinersen-treated (n = 235) and untreated (n = 17) patients are summarized in Table 15. The slope in nusinersen-treated patients was 0.01168 points per week (95% CI, 0.004957 to 0.01841), compared with 0.003936 points per week (95% CI, −0.01030 to 0.01817) in untreated patients.

Pera et al. (2021): Results for the change in RULM scores in adults in Pera et al. (2021)¹⁹ are summarized in Table 17. In the primary analysis, the mean change from baseline in RULM score in adults (n = 55) at 12-months was 0.07 points (95% CI, –0.48 to 0.63). In the subgroup of nonambulant patients older than 20 years (n = 33), the mean change from baseline in RULM score at 12 months was 0.48 points (SD = 2.12). In the subgroup of ambulant type III SMA patients 20 years and older (n = 21), the mean change from baseline at 12 months in RULM score was –0.62 points (SD = 2.11).

Stability of RULM scores at 12 months are summarized in Table 19. In all adults (> 20 years), RULM scores declined in 13.0% of patients, remained stable in 75.9%, and improved in 15.6%. These results were generally similar in ambulant and nonambulant patients.

Table 17: Changes in RULM (12 Months vs. Baseline) in the Overall Adult and the Greater Than 20 Year SMA type III Populations (Pera et al. [2021])

CI = confidence interval; N/A = not applicable; RULM = Revised Upper Limb Module; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Pera et al. (2021).¹⁹ This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 International Licence. Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8351459/

Table 18: Changes in 6MWT (12 Months vs. Baseline) in the Overall Adult and the Greater Than 20 Year SMA type III Populations (Pera et al. [2021])

6MWT = 6-minute walk test; CI = confidence interval; N/A = not applicable; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Pera et al. (2021).¹⁹ This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 International Licence. Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8351459/

Table 19: 12-Month Trajectories Grouped as Stable ( + 2 Points), Improved (>  + 2), or Declined (< −2) HFMSE and RULM Scores in Patients Greater Than 20 Years (Pera et al. [2021])

CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded; RULM = Revised Upper Limb Module; SD = standard deviation; SMA = spinal muscular atrophy.

Source: Pera et al. (2021).¹⁹ This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 International Licence. Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8351459/

6MWT

Hagenacker et al. (2020): Results for the change in 6MWT in Hagenacker et al. (2020)¹⁷ are summarized in Table 11. The mean changes from baseline in 6MWT were 22.12 m (95% CI, 8.7 to 35.6) at 6 months (n = 47), 31.14 m (95% CI, 15.2 to 47.1) at 10 months (n = 37), and 45.96 m (95% CI, 25.4 to 66.6) at 14 months (n = 25).

Maggi et al. (2020): Results for the change in 6MWT in Maggi et al. (2020)¹⁸ are summarized in Table 13. Results for SMA type II patients, type III sitters, and all-type III SMA patients were not available, Walkers with type III SMA had a mean change in 6MWT of 14.66 m (SD = 27.57) at 6 months (n = 48), 26.45 m (SD = 34.6) at 10 months (n = 35), and 23.11 m (SD = 51.2) at 14 months (n = 24). The proportion of patients achieving a minimum improvement in 6MWT of 30 m was 29% at 6 months, 46% at 10 months, and 42% at 14 months (Figure 4).

Results for the rise from chair timed function test showed an increase in velocity at 6 months (0.02 second^–1), 10 months (0.04 second^–1), and 14 months (0.06 second^–1). Ten-metre run or walk speed increased only at 6 months ( + 0.07 m/s).

EU Registry Study: Results for the change in 6MWT from the pre- and post-treatment analysis in the EU registry study^15,16 are summarized in Table 14. The model results showed that before the start of nusinersen treatment, the 6MWT distance decreased by 0.2393 m per week (95% CI, –0.349 to –0.1297). After treatment with nusinersen, the slope for the change in 6MWT was –0.03399 m per week (95% CI, −0.4373 to 0.3694).

Results comparing the change in 6MWT between nusinersen-treated and untreated patients are summarized in Table 15. The slope in nusinersen-treated patients (n = 235) was 0.2633 m per week (95% CI, 0.09922 to 0.4272), compared with –0.7148 m per week (95% CI, −1.2789 to –0.1506) in untreated patients (n = 17).

Pera et al. (2021): Changes in 6MWT for adults in Pera et al. (2021)¹⁹ are summarized in Table 18. The mean 6MWT distance from baseline to 12 months for adults (n = 17) was 0.52 m (95% CI, –19.85 to 20.89).

Respiratory Outcomes

FVC/FEV₁

Hagenacker et al. (2020): Changes in lung function were not evaluated in the Hagenacker et al. (2020)¹⁷ study.

Maggi et al. (2020): Results for respiratory outcomes in the study by Maggi et al. (2020)¹⁸ are summarized in Table 13. The mean change in percent-predicted FVC in all type III SMA patients was 0.53% (SD = 7.77) at 6 months (n = 35), 4.77% (SD = 11.79) at 10 months (n = 17), and 6.47% (SD = 9.22) at 14 months (n = 15). In SMA type II patients, the mean change in FVC was –0.25% (SD = 2.06) at 6 months, 0.75% (SD = 2.5) at 10 months, and was not available at 14 months due to small sample sizes. Mean change in percent-predicted FVC showed an increase over time for type III sitters and walkers.

In all SMA patients, the percent-predicted FEV₁ change from baseline was 0.01% (SD:8.45) at 6 months (n = 32), 6.93% (SD = 18.37) at 10 months (n = 18), and 5.86% (SD = 9.22) at 14 months (n = 14). Change from baseline in the percent-predicted FEV₁ at 6 months in subgroups of type II SMA patients, type III SMA patients, type III sitters, and type III walkers showed varied results at the 3 evaluation time points.

EU Registry Study: Changes in lung function were not evaluated in the EU registry study.^15,16

Pera et al. (2021): Changes in lung function were not evaluated in the Pera et al. (2021)¹⁹ study.

Bulbar Function

Outcomes related to bulbar function or feeding support were not included in the studies.

Survival

Outcomes related to survival were not included in the studies.

Hospitalization

Outcomes related to hospitalization were not included in the studies.

HRQoL

Outcomes related to HRQoL were not included in the studies.

Anatomic-Related Outcomes

Anatomic-related outcomes were not included in the studies.

Caregiver Burden

Outcomes related to caregiver burden were not included in the studies.

Harms

Only harms identified in the review protocol are reported in the following sections.

AEs and SAEs

Hagenacker et al. (2020)

A total of 82 (47%) patients who received at least 1 injection had AEs due to drug reactions or procedure-related complications (Table 20). Throughout the 14-month period, the most frequently reported AEs were headache (35%), back pain (22%), and nausea (11%).¹⁷

No SAEs were reported.

Table 20: Adverse Drug Reactions and Procedure-Related Complications Recorded in Participants (n = 173), Classified According to MedDRA Version 21.1 (Hagenacker et al. [2020])

Source: Hagenacker et al. (2020).¹⁷

Maggi et al. (2020)

Overall, AEs were reported in 48 (41.4%) patients, with the most frequently reported AE being post-procedural headache (n = 43 [37.1%]). Headache was orthostatic, mild to moderate in intensity, and spontaneously resolved in a few days, except in 5 patients who required hospitalization. Two patients with type III SMA reported transient worsening of existing hand tremor. The investigators generally considered AEs mild or moderate and to be related to nusinersen administration procedures.¹⁸

EU Registry Study

AEs and SAEs at 6 and 11 months in the EU registry study^15,16 are summarized in Table 21. In patients who received nusinersen, 78 (34.67%) experienced at least 1 AE at 6 months, compared with 86 (38.22%) at 11 months. At least 1 SAE occurred in 8 (3.79%) and 10 (4.74%) patients at 6 and 11 months, respectively.

Table 21: Safety in the Adult Population (EU Registry Study)

AE = adverse event; CI = confidence interval; IRR = incidence rate ratio; n = number of patients with event; N = number of patients assessed; SAE = serious adverse event.

^aDifference in time ≤ 10%.

^bModel did not converge.

Source: Sponsor submission.¹⁵

Pera et al. (2021)

The most frequent AEs were all related to the procedure, and included headache, nausea, and back pain,¹⁹ although frequency was not reported. No SAEs were reported at the time of data collection.

WDAEs

WDAEs were minimal in the included studies. In Hagenacker et al. (2020),¹⁷ 2 patients withdrew from treatment at 10 months because of adverse drug reactions. In Maggi et al. (2020),¹⁸ nusinersen treatment was stopped in 2 (1.7%) type III SMA patients after 6 months due to lack of subjective benefit and poor tolerability of repeated lumbar puncture. WDAEs were not reported in the EU registry study^15,16 or the study by Pera et al. (2021).¹⁹

Mortality

Survival and mortality were not reported in any of the included studies.

Notable Harms

Serious Infections

Serious infections were not reported in any of the included studies; however, as noted in Table 20, 1 (1%) patient experienced an infection at injection 2, and upper airway infections were experienced by 1 (1%) patient at injection 1 and by 2 (1%) patients each at injections 4 and 6 in the Hagenacker et al. (2020)¹⁷ study, although the severity was unknown.

Lumbar-Puncture-Related AEs

Hagenacker et al. (2020): After lumbar puncture, headache (61 [35%]) and back pain (38 [22%]) were the most frequently reported AEs in the Hagenacker et al. (2020)¹⁷ study at 14 months, occurring in up to 1-fifth of patients.

Maggi et al. (2020): As mentioned previously, 2 (1.7%) patients with type III SMA discontinued treatment due to poor tolerability of repeated lumbar puncture.¹⁸ Additionally, post-procedural headache was reported in 43 (37.1%) patients. Headaches were generally characterized as mild to moderate in intensity and resolved spontaneously; however, hospitalization was required in 5 patients. Headache was associated with manual procedures in 34 cases and with imaging-guided techniques in 9 cases, with no difference in headache probability between the 2 approaches. Lumbar pain was reported by 10 (8.6%) patients, 7 of whom underwent imaging-guided lumbar puncture.

The frequency of lumbar-puncture-related AEs was not reported in the EU registry study^15,16 or in the study by Pera et al. (2021).¹⁹

Coagulation Abnormalities

No coagulation abnormalities were reported in the included studies.

Renal Toxicity

In the study by Maggi et al. (2020),¹⁸ 1 case of renal colic requiring hospitalization was reported in a patient with type II SMA after the 10-month assessment. No other included studies reported events of renal toxicity.

Critical Appraisal

No RCTs focusing on treatment-naive adults with type II or III SMA were identified as part of the CADTH literature search, and all available and included studies were of observational design, focusing on real-world data, which have more limitations than RCTs.

The studies included in this reassessment consisted of 1 prospective (Hagenacker et al. [2020]¹⁷) and 3 retrospective, observational, noncomparative studies, comprising cohorts of patients who had been treated with nusinersen at individual centres or who were enrolled in a registry. Overall, the included studies suffer from high levels of selection bias, reporting bias, and information bias. There is a high potential for selection bias because of the observational design of the studies and the way patients were chosen, and a risk of reporting and information bias because of the methods used for data collection. All studies were also noncomparative in nature, negating the ability to draw a statistical association between the reported results and nusinersen because of small sample sizes (although this is reasonable for a rare disease), the slow and minimal progression of SMA in adults, and the restriction to mostly high-functioning type III SMA patients. Additional details on the internal and external validity of the included studies follow.

Internal Validity

Hagenacker et al. (2020)

The study by Hagenacker et al. (2020)¹⁷ was a prospective, multi-centre, uncontrolled, noncomparative, observational cohort study. Given the noncomparative study design, there was no control group for comparison, so the results observed cannot be attributed to treatment with nusinersen. Moreover, there was a lack of blinding of both patients and outcome assessors. There is a risk of bias because of the subjective nature of the outcome measurements, as patients and providers were aware of the treatment.

This was a prospective cohort study, and selection criteria limited patients to those with confirmed 5q SMA and prior nusinersen treatment for at least 6 months. This selected for patients who were able to complete the induction dosing and who were able to tolerate and/or receive doses. No other selection criteria were defined, and the authors noted that as nusinersen is approved for the treatment of a severe, chronic, progressive disease without limitations on age or disease classification, further controls were not warranted. Evidence of selection biases was noted in baseline characteristics — the study included higher-functioning, ambulatory, type III SMA patients — and is discussed in detail in the External Validity section. As such, the risk of selection bias is likely, and no adjustment techniques were applied to correct for potential selection biases.

Missing data also impacted the validity of the results. A total of 173 patients were screened for enrolment. The proportion of patients lost to follow-up was not reported, although of the 139 patients enrolled, only 57 (41.0%) were available at 14-month follow-up. Thus, there was a high level of variation in the number patients available at different follow-up times, resulting in a lack of data at later time points. No imputation of missing data was done for the 6-month, 10-month, or 14-month analyses, nor for the amount of missing data during the observation period for each outcome reported. Assessment of attrition was not conducted, and it is unclear whether attrition was random (i.e., whether nusinersen was discontinued due to lack of efficacy or AEs); therefore, there is an increased risk that the treatment effect of nusinersen is overestimated.

Methods for statistical analysis of pre- and post-treatment comparisons for the primary outcome of HFMSE using nonparametric tests at selected time points, and the conducted subgroups analyses, were reportedly pre-specified, although no protocol was identified. A list of potential confounders was not described; however, a mixed model that included sex, time, age, spondylodesis, and baseline HFMSE score as a fixed effect and the patient as random effect was only used to estimate effect on the primary outcome. No rationale for these variables was provided. Other potential confounders and treatment-effect modifiers that were not identified or considered, even though they may influence outcomes, include training for the outcomes of interest, routine exercise and observation, other routine care (such as physiotherapy and occupational therapy), and the placebo effect. The extent to which uncontrolled confounders and treatment-effect modifiers influenced the results is unclear.

Maggi et al. (2020)

The study by Maggi et al. (2021)¹⁸ was a retrospective, multi-centre, noncomparative, observational cohort study. The aim and outcome measures of this study were clearly described, although no protocol was identified. There was no comparator group in this study, so the results observed cannot be attributed to treatment with nusinersen. Because patients and outcome assessors were not blinded to treatment, there is a risk of bias in outcome assessment.

As with the Hagenacker et al. (2020)¹⁷ study, a key limitation with the Maggi et al. (2020)¹⁸ study was the potential for selection bias. A similar population of eligible patients was included (i.e., adults with confirmed 5q SMA), and reasons for exclusion were noted; however, given that it was a retrospective study, patients were likely selected after treatment with nusinersen and, therefore, the study is at high risk of selection bias. It was not clear whether or how the potential for selection bias was assessed, and it was not reported whether adjustment techniques were applied to correct for potential selection biases. There was potential for information bias in this study because of the retrospective design and the lack of reporting on data quality and accuracy. No consideration of potential information bias was noted in this study and, therefore, it is unclear what effect this may have had on the results. No information was provided on the methods used to account for missing data, nor were amounts of missing data reported. The authors noted that missing data were mostly limited to timed and pulmonary function tests.

A total of 116 patients were enrolled in the study; however, just 17% to 65% of patients (depending on the analysis group and outcome measure) completed outcome assessments at 14-month follow-up. There were also notably small sample sizes for certain subgroups, particularly patients with type II SMA (n = 13), making it very difficult to interpret results for these patients.

Both parametric and nonparametric tests were conducted on the population. No consideration was given to potential confounders in this study. Logistic regression was used to identify the effects of predictor variables (age, sex, SMN2 copy number) on treatment response; however, no further information on this process (e.g., selection of variables in the model) or the way results of the regression were used was provided. Potential confounders and treatment-effect modifiers that were not identified or considered, even though they may influence outcomes, include training for the outcomes of interest, routine exercise and observation, other routine care (such as physiotherapy and occupational therapy), and the placebo effect. The extent to which uncontrolled confounders and treatment-effect modifiers influenced the results is unclear.

EU Registry Study

Data from the EU registry study^15,16 submitted by the sponsor was based on data from 3 registries — SMArtCARE, ISMAR, and CuidAME — which collected data from multiple centres in Germany, Italy, Spain, the UK, and the US. Data in these registries were collected both prospectively and retrospectively. Although registries can provide large collections of data on real-world patients and settings, there are concerns with the quality of data from registries due to limitations in the availability of key elements (patient characteristics, exposures, outcomes), as well as the reliability of data accuracy, completeness, provenance, or traceability, and no information on accuracy was reported.

A key issue is the way patients are enrolled in the registry, which can increase the potential for selection bias. It was unclear if these registries represented patients with different self-care practices, socioeconomic backgrounds, or levels of supportive care than patients in the entire population of interest for the reassessment (type II and III SMA). According to the sponsor, the inclusion and exclusion criteria for patients in each registry were identical, including a genetically confirmed diagnosis of 5q SMA. Patient were enrolled in each registry through SMA clinics; however, information on the enrolment process was not provided. It was unclear if patients received nusinersen before or after enrolment in the registries. Additional selection biases were noted in the baseline characteristics — the study included ambulatory, high-functioning, type III SMA patients — and are discussed in detail in the External Validity section.

The sponsor conducted a quality assessment of each individual registry using the IQWiG tool, an abbreviated version of the Reporting of Studies Conducted Using Observational Routinely-Collected Data (RECORD) tool, and the National Institutes of Health Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies, although no accuracy or agreement between the registry sample and reabstracted records was provided. The results of this quality assessment were not appraised. No overarching protocol was submitted for this study; however, the sponsors provided individual publications for the development of each of the registries. The ISMAR registry did not share a single protocol at all centres, so there may be differences in collection methods or procedures across institutions that are unaccounted for.

A statistical analysis plan for the EU registry study^15,16 was submitted in response to a request from CADTH. The sponsor pooled patients from the 3 registries into 2 analysis groups (one that compared 235 treated patients with 17 untreated patients, and 1 pre- and post-treatment analysis that consisted of 75 patients); however, no information on the pooling methods or justification for the pooling of patients from registries was provided. Regardless of pooling, differences in treated (N = 235) and untreated (N = 17) populations — the German SMArtCARE registry had no untreated patients, the Italian ISMAR registry had 5 untreated type III adults, and the Spanish CuidAME registry had 9 untreated type III adults — highlighting the uncertainty around how patients were enrolled in the registries and potential bias in the selection of patients. Due to the limited number of patients in the untreated groups, the results observed cannot be attributed to nusinersen.

Two separate analyses were presented for the EU registry study.^15,16 In an unpublished report, the sponsor provided baseline data and motor-function outcome results for all type III SMA patients and for adults with type III SMA, with subgroups of fully ambulatory and not fully ambulatory patients. First, a standard mixed-effect model used LOCF and multiple-imputation methods to account for missing data, with interaction terms between treatment and time in the model. However, no information on the modelling approach was provided. No information was presented on the amount of missing data. The LOCF method of imputing missing data may have increased the chance of bias because it cannot be confirmed, based on the available information, whether missing data occurred at random. Also, the lack of details reported about the amount of missing data and the time contribution of each patient make it difficult to determine the effects of LOCF imputation. The sponsor also conducted a multiple imputation model on the missing, pooled baseline characteristics, as well as on the outcomes of interest; however, no information on the amount of missing data was provided and, therefore, it is unclear which missing data were considered. Overall, results using the LOCF and multiple-imputation methods were similar, suggesting that there were minimal differences related to missing data; however, the CADTH review team noted that despite the lack of details on how the multiple imputation was conducted, this method is preferred. The sponsor noted that these analyses were adjusted for any potential predefined confounders, although it is unclear to what extent uncontrolled confounders and treatment effect-modifiers influenced the results.

In a separate, draft manuscript, the second analysis consisted of a piecewise linear mixed-effect model, which was used to evaluate pre-and post-treatment comparisons. The model included relevant confounders, including age, sex, number of SMN2 copies, ambulatory status, disease duration, baseline scores, and registry. Other potential confounders and treatment-effect modifiers that were not identified or considered, even though they may influence outcomes, include training for the outcomes of interest, routine exercise and observation, other routine care (such as physiotherapy and occupational therapy), and the placebo effect. The population used for the model included 75 SMA adults — 74 with type III SMA and 1 with type IV — with at least 6 months of follow-up, although the actual assessment times for these patients was unknown. Missing values at baseline (HFMSE, RULM, and 6MWT) were imputed using linear interpolation based on pre-treatment measures.

In the unpublished draft manuscript, results for the standard mixed-effect model and piecewise linear mixed model were presented as the difference between slopes in points per week for each measure for the populations before and after treatment (n = 75), as well as in 235 treated (228 with type III and 7 with type IV SMA) and 17 untreated patients (14 with type III and 3 with type IV SMA), using a time-spline methodology. Using this method, the difference between the slope before treatment start and the slope after treatment start (time spline) was assessed. In both the standard linear-effects model and the piecewise linear mixed model, although weeks were used as the time variable, it was unclear what the duration of follow-up was. The CADTH review team, as well as the clinical experts consulted by CADTH, considered weeks as the unit of measure to be inappropriate for type II and III SMA. The sponsor attempted to extrapolate the results over 1 year, but CADTH considered the results of this extrapolation to be highly uncertain and inappropriate, given the already short duration of follow-up.

Pera et al. (2021)

Pera et al. (2021)¹⁹ was a noncomparative, registry-based analysis of SMA patients collected from ISMAR data. No published protocol was specified or identified, and no sample size calculation was provided. Minimal information was provided on eligibility criteria for these patients, and data were based on what was available in the registry. No information on the quality or accuracy of the registry was provided. This study enrolled patients outside of the reassessment reimbursement request; patients younger than 18 years were included. Only 67 (47%) patients were in the population of interest for this review. The authors attempted to reduce any selection bias by conducting a mixed-model assessment to estimate changes in the entire type III population, using baseline measurements (age, sex, time, disease duration, SMN2 copy number, disease onset, and SMA function) as fixed effects and the patient as a random effect. Other potential confounders and treatment-effect modifiers that were not identified or considered include training for the outcomes of interest, routine exercise and observation, other routine care (such as physiotherapy and occupational therapy), and the placebo effect. The extent to which uncontrolled confounders and treatment-effect modifiers influenced the results is unclear.

Relevant subgroups based on SMA type (type IIIA and IIIB), age (children versus adult), and functional status (sitters versus walkers) appeared to be pre-specified; however, no information on the analysis of these subgroups was provided. No imputation of missing data was performed, and the amount of missing data was not reported, so the impact of missing data on the results could not be determined. Given the noncomparative nature of the study, the results observed cannot be attributed to nusinersen.

No baseline characteristics were reported solely for the adult population; thus, specific selection biases in this population could not be determined.

External Validity

Hagenacker et al. (2020)

Hagenacker et al. (2020)¹⁷ enrolled 139 adults with genetically confirmed 5q SMA; mostly type II or type III SMA. The proportion of patients with type III SMA was 62% to 65%, most had 3 SMN2 gene copies (36% to 39%), and 37% to 40% were ambulant. These values were higher than what the clinical experts see in clinical practice. As well, there was a lower-than-expected proportion of patients with prior spinal fusion surgery (20% to 25%). As such, the baseline characteristics of the included population suggested a selective population of patients that limits the generalizability of the results. The clinical experts consulted by CADTH reported that these proportions did not reflect the type II and III SMA adult population they see in practice; instead, they would expect more type II patients who are nonambulatory (approximately 75%) and more patients (approximately 65% to 70%) who have undergone spinal fusion surgery. Overall, the CADTH review team and clinical experts considered the patients included in the study by Hagenacker et al. (2020)¹⁷ to be of lower disease severity, reducing the generalizability of this study.

There was potential survivorship bias in this study, as evidenced by the percent of ambulant patients with type III SMA and the higher baseline HFMSE, RULM, and 6MWT scores at baseline in the 14-month group than in the 6-month group. The follow-up duration of this study was only 14 months. In discussion with the clinical experts, it was determined that the follow-up duration of 14 months was inadequate to capture the true effects of treatment in type II and III SMA patients, although the direction of bias is unclear.

In this study, all nusinersen injections were conducted by trained neurologists or neuroradiologists with conventional, fluoroscopy-guided or CT-guided lumbar puncture. It was unclear how many patients required guided intrathecal injections, which may have implications on clinical practice in Canada, given the spinal complexity of these patients.

The outcomes chosen for this study were based on validated measures of motor function and are routinely used in studies of SMA and in Canadian clinical practice. All patients were evaluated according to the recommendations of the SMArtCARE real-world data collection initiative. As such, all evaluators were properly trained, minimizing potential variation in the way motor-function outcomes were measured. However, the HFMSE scale was the primary outcome of this study and, by definition, is to be used in patients who can both sit and walk. Considering that only 37% to 40% of the population was ambulatory, neither HFMSE nor 6MWT may be an appropriate measure of improvement in the patients studied. HRQoL and other patient-reported outcomes, which were outcomes important to patients, were not assessed in the study by Hagenacker et al. (2020).¹⁷

As noted by the authors, the natural disease course of SMA might be influenced by the type and level of supportive care provided to the patient, and all included patients had access to supportive care before and during nusinersen treatment, although it was unclear what supportive care consisted of. The clinical experts emphasized the importance of patients with SMA receiving care from a multi-dimensional care centre that employs physiotherapists and other health providers. However, it is not clear what proportion of patients in Canada have access to and receive this level of care, so the results may not be generalizable to these patients.

Maggi et al. (2020)

As with the Hagenacker et al. (2020)¹⁷ study, although the enrolled population consisted of adults with type II and III SMA, they were a selective cohort of patients. The study by Maggi et al. (2020)¹⁸ enrolled 116 adults with type II and III SMA, but only 13 (11.2%) had type II SMA. As noted by the experts, type II patients represent the majority of adults they see in practice; thus, the population in this study is less generalizable to type II patients in Canada because of the limited inclusion in the study. Approximately half of all type III patients included were sitters; the other half were walkers. Most patients had 4 copies of SMN2 (46.6%), which was noted by the clinical experts to be unreflective of the eligible population. Compared with type III walkers, more patients who were considered sitters required ventilatory support (15.7% versus 5.8%) and had undergone surgery for scoliosis (13.7% versus 1.9%). The clinical experts consulted by CADTH indicated that most patients would likely be not fully ambulatory, and most would have undergone prior spinal surgery (including spinal fusion); thus, this population may not be reflective of Canadian patients. Additionally, baseline motor-function scale scores were considered to be higher than expected, indicating better baseline function in these patients, particularly SMA type III walkers, suggesting a bias in the patients selected for inclusion in this study. Most ventilatory support at baseline occurred in type II (76.9%) and type III sitter (15.7%) patients. However, it was unclear whether this referred to the proportion of patients that required night-time BiPap, or if some other method of ventilation was required. As in the Hagenacker et al. (2020)¹⁷ study, the CADTH review team and clinical experts considered the included population to have less severe disease than most adults with type II and III SMA.

Nusinersen was administered by trained professionals, mostly with standard lumbar access, which, according to the clinical experts consulted by CADTH, reflects the type III nature of patients; however, they noted that this may not be reflective of the general SMA population, because most of these patients would have complex spines, requiring interventional radiology to administer doses.

The outcomes selected for this study were appropriate and were in line with standard practice; however, there was significant attrition in the proportion of patients completing the outcome measures between baseline and 14 months, resulting in uncertainty and limited generalizability of the results. As previously noted, HFMSE and 6MWT may not be appropriate outcome measures in all patients, considering that half of all patients were nonambulatory. HRQoL and patient-reported outcomes, which were outcomes important to patients, were not assessed in the study by Maggi et al. (2020)¹⁸ and, therefore, the effect of nusinersen on these outcomes remains unknown.

This study was a retrospective cohort study that was limited to 14 months of follow-up. As in the Hagenacker et al. (2020)¹⁷ study, this duration of follow-up was considered to be too short to observe the true effects of treatment in type II and III SMA patients due to the slowly progressing nature of the disease.

EU Registry Study

As previously mentioned, the EU registry study^15,16 enrolled patients from 3 registries from multiple centres in Germany, Italy, Spain, the UK, and the US. Overall, there were differences in the proportions of patients from each registry, as 157 (67%), 54 (23%), and 24 (10%) treated patients, and 0, 6 (35%), and 11 (65%) untreated patients came from the German SMArtCARE, Italian ISMAR, and Spanish CuidAME registries, respectively. As such, there may be differences in disease management for treated and untreated patients across regions, which may affect the generalizability of results, given the larger contributions of patients from specific countries and registries. Although all patients were likely similar with respect to indicated disease, based on registry enrolment, there was considerable heterogeneity across registries in baseline characteristics with regard to age, age at symptom onset, ambulatory status, and proportion of patients with scoliosis. Nearly all patients included in the study were adults, had type III SMA, and had 4 copies of SMN2 (45%), which was noted by the clinical experts consulted by CADTH to be unreflective of practice. Despite the fact that included populations were generally evenly split between ambulatory (48%) and nonambulatory (52%) type III patients, many of the patients included in the registry analysis did not have scoliosis (76%) or prior spinal fusion surgery. Additionally, only 15 (6.2%) patients required noninvasive ventilatory support, although it was unclear what this consisted of. Overall, there was considerable selection bias for higher-functioning patients, based on baseline scale scores that were considered high, indicating greater motor function and less severe disease. The CADTH review team and clinical experts consulted by CADTH considered there to be a high degree of selection bias in the patients included in the EU registry study,^15,16 which may affect generalizability to Canadian adults with type II or III SMA.

Most patients in the analysis of treated and untreated adults had type III SMA (n = 228); however, there were also 10 type IV patients, which was not reflective of the reimbursement request. Comparative analysis of treated and untreated patients only included 17 patients in the untreated group, 3 of whom had type IV SMA. The characteristics and small sample size of the untreated group make it very difficult to generalize the results or draw conclusions from the analysis.

A total of 74 adults with type III SMA and 1 patient with type IV SMA were included in the pre- and post-treatment analysis. Although the few type IV SMA patients included are unlikely to affect the results of either analysis, the small sample size limits the conclusions that can be drawn. Furthermore, there were no patients with type II SMA included in the study, so any results cannot be generalized to the type II population.

There was no information provided on the method of administration of nusinersen, the dosing regimen in the registries, or background care, so the CADTH review team was unable to determine whether how and where patients were treated had any potential impact on outcomes.

It was unclear what the overall time of assessment was for the EU registry study.^15,16 Regardless, the CADTH review team, as well as the clinical experts consulted by CADTH, considered weeks as the unit of measure to be inappropriate for type II and III SMA. The sponsor attempted to extrapolate the results over 1 year; however, CADTH considered the results of this extrapolation to be highly uncertain and inappropriate, given the short duration of follow-up relative to the natural history of SMA and the unverified assumptions required to make this extrapolation.

Outcomes included in the EU registry study^15,16 were appropriate and matched other published studies; however, as previously mentioned, the conclusions that can be drawn from the HFMSE and 6MWT measures are limited, as nearly half of the patients were nonambulatory. HRQoL and patient-reported outcomes, which were outcomes important to patients, were not assessed in the EU registry study, so the effect of nusinersen with respect to these outcomes remains unknown.

Pera et al. (2021)

The study by Pera et al. (2021)¹⁹ was based on the data collected from the ISMAR registry, so there may be some overlap in patients included in this study, the EU registry study,^15,16 and the Maggi et al. (2020)¹⁸ study. It is unclear what effect this may have on the interpretation of the results. The Pera et al. (2021)¹⁹ study included 144 patients with type III SMA, but also included children and adults. Overall, only 67 patients included in this study met the reimbursement request criteria of adults with type II or III SMA. No baseline characteristics of the target population were included, so CADTH is unable to assess the potential for selection bias or external validity in these patients.

No information on the dose, regimen, or setting of nusinersen administration were provided, so it was unclear whether the recommended dosing administration procedure was followed. Outcome measures were consistent with the other studies, including HFMSE, RULM, and 6MWT, and were administered by trained clinical evaluators. In the overall population (including adults and children), nearly half of all patients were sitters; thus, as previously mentioned, the HFMSE and 6MWT may not be appropriate. As in the other included studies, baseline scores for HFMSE, RULM, and 6MWT were considered high for the adult type III population, which may not be reflective of the Canadian population, and did not include type II SMA patients. HRQoL and patient-reported outcomes, which were outcomes important to patients, were not assessed in the study by Pera et al. (2021),¹⁹ so the effect of nusinersen with respect to these outcomes remains unknown.

The time of assessment of this study was 12 months, which was considered too short to assess changes in adults with type II or III SMA and generally too short to notice any meaningful change.

Indirect Evidence

No indirect evidence was included in the sponsor’s submission to CADTH or identified in the literature search conducted by CADTH that matched the review protocol criteria.

Other Relevant Evidence

Other Sponsor-Submitted Evidence

As part of the reassessment for nusinersen, the sponsor submitted a publicly available critical review and meta-analysis of patients with type II and III SMA. The objective of this section is to summarize and critically appraise methods and findings of the submitted meta-analysis, which evaluated efficacy patterns of nusinersen on motor function in the existing literature in real-world datasets of patients with type II and III SMA.²⁰

Description of Study

The sponsor submitted a publicly available critical review and meta-analysis, informed by a systematic review that evaluated motor-function outcomes in patients with type II or III SMA. A total of 14,627 articles were preliminarily selected based on title. A total of 30 full-text articles were included in the analysis.²⁰

Methods

Objectives

The objective of the meta-analysis by Coratti et al. (2021)²⁰ was to critically review the literature reporting real-world data on motor function in type II and III patients treated with nusinersen to establish possible patterns of efficacy by subdividing results by SMA type, age (children versus adults), and type of assessment. Only results related to the adult population with type II or III SMA were of interest to this reassessment.

Study Selection Methods

The meta-analysis was informed by a systematic review of existing literature conducted up to January 2021. PRISMA guidelines were applied to the critical review, including research on online databases for peer-reviewed journal (PUBMED, MEDLINE, Web of Science, CINAHL, PsycINFO, and Embase) and manual research on reference lists of included articles. Eligible publications reporting data on nusinersen-treated patients were grouped into categories by SMA type (type II, type III, or a combination of the 2 types) or age (children versus adult). A similar approach was used to classify data from untreated patients. When needed, recalculation of the mean and SD for age or motor outcome was performed from papers reporting full data.²⁰

Screening and data collection were conducted by 5 authors. Full-text review was conducted by that group and by the senior author to determine the full eligibility of articles. Data extracted included outcome measures, target population (size, type, age category, age range at treatment, mean age at treatment), magnitude of changes at 10, 12, or 14 months from baseline (mean, SD or 95% CI). Studies reporting data on type I SMA only were excluded. All electronic searches were limited to the English language.²⁰ No additional information on the PICOS framework or on inclusion criteria for the review and meta-analysis were provided.

Analysis Methods

Data looking at reported changes in individual motor-function outcome measures in the different treated groups, subdivided by age category (adults, children), motor function (ambulant, nonambulant), and SMA type were analyzed. Pooled analyses were conducted at multiple levels: first, a rough comparison of the overall benefit of treatment versus no treatment was run, which included the largest amount of available evidence, even if heterogeneous. The effect size was estimated using random-effect models, and heterogeneity among studies was quantified with the I² coefficient. Subgroup analyses were conducted to verify and estimate the influence of different categories (age, SMA type, and motor function) on the pooled results of the treated population. Meta-regression analysis was undertaken to identify possible sources of heterogeneity among studies. Meta-regression analyses were employed, with a random-effects model using aggregate-level data. Only studies with complete data available (sample size, mean, SD or 95% CI) were included in the meta-analysis.²⁰

Results

Summary of Included Studies

The PRISMA flow diagram for study selection in the Coratti et al. (2021)²⁰ meta-analysis is summarized in Figure 5. A total of 14,627 articles were selected based on title, 788 of which were related to nusinersen treatment. After the review of the abstracts, 9,221 were excluded. After the review of 55 full-text papers, 30 publications were selected and analyzed. Fifteen studies reported on adults treated with nusinersen and 10 reported on natural history in adults.

After the exclusion of reviews, commentaries, and individual case reports, 19 papers reporting data on efficacy that used structured assessments in type II and III SMA were selected. In 4 of the 19 papers, type I SMA patients were also included, with 1 of the 4 papers describing data on type I SMA separately. According to the authors, none of the data for outcomes of interest (HFMSE, RULM, and 6MWT) included type I patients. Data from clinical trials were not included in the review. The 19 papers were reviewed using the Risk of Bias Assessment Tool for Nonrandomized Studies. Twelve papers reported data in untreated patients.²⁰

A summary of specific demographic or baseline patient characteristics from the included trials was not provided, so assessment of heterogeneity related to patients, outcomes, or study design was not possible.

Figure 5: Search and Selection Process (PRISMA Framework)

6MW = 6-minute walk; HFMSE = Hammersmith Functional Motor Scale Expanded-2; MFM = motor-function measurement; MRC = Medical Research Council; RULM = Revised Upper Limb Module; SMA = spinal muscular atrophy.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Results of Meta-Analysis

Motor-function outcomes for the adult subgroup in the meta-analysis include HFMSE, RULM, 6MWT, MRC Scale for Muscle Strength, and CHOP-ATEND.²⁰ Meta-regression was not conducted for the MRC and CHOP-ATEND outcomes and are not summarized.

HFMSE: Of the 13 publications that reported HFMSE results for patients treated with nusinersen, 9 focused on adults, 2 focused on adult and pediatric populations, and 2 focused only on the pediatric population. Five publications reported HFMSE results in untreated patients. Results for change in HFMSE score in the individual publications are summarized in Figure 6. Individual data on HFMSE were available in 3 papers; therefore, to subdivide the populations according to SMA type or age group, and to eliminate the problem of missing data, the authors recalculated mean change over time.²⁰

Figure 6: HFMSE Results by Reporting Author in the Adult Population

Dashed line = 95% confidence interval; Plain line = standard deviation; Square = SMA type II; Circle = SMA type III, Diamond = ambulant SMA type III; Triangle = nonambulant SMA type III; Square + circle + triangle = mixed phenotypes; Bold font = median value; italic = mean value; Light blue = approximately 10 months from initiation of drug; Red = approximately 12 months from initiation of drug; Grey shade = SMA type II; White shade = SMA type III; Striped shade = mixed phenotypes.

* Mean/median values of the baseline population, non excluding drop-outs at 10, 14, or 24 months of follow-up.

** Mean/median values of the baseline population of both SMA type II and III combined.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Results of the meta-regression analysis for the mean change in HFMSE in the adult population are summarized in Figure 7. Heterogeneity in the included studies for adults was moderate (I² = 68%). The pooled mean change in HFMSE score in adults was 1.87 points (95% CI, 1.05 to 2.68).²⁰

RULM: Thirteen papers reported RULM results for patients treated with nusinersen: 9 in adults only, 2 in children only, and 2 in adults and children. Five publications reported RULM results in untreated patients. Individual data on RULM were available in just 1 paper; therefore, to eliminate the problem of missing data, the authors have recalculated mean change over time.²⁰ Results for change in RULM score in the individual publications are summarized in Figure 8.

Results of the meta-regression analysis for the mean change in RULM in the adult population are summarized in Figure 9. Heterogeneity in the included studies was moderate (I² = 43%). The pooled mean change in RULM score in adults was 0.64 points (95% CI, 0.27 to 1.01).²⁰

6MWT: Eight publications were identified that reported data on patients treated with nusinersen, and 1 publication reported data from untreated patients. Of the 8 publications on nusinersen, 6 focused on adults, 1 focused on children only, and 1 focused on both adults and children.²⁰ Results for change in 6MWT score in the individual publications are summarized in Figure 10.

Results of the meta-regression analysis for the mean change in 6MWT in the adult, ambulatory population are summarized in Figure 11. Heterogeneity in the included studies was considerable (I² = 71%). The pooled mean change in 6MWT score in adults was 20.28 m (95% CI, 1.17 to 39.40).²⁰

Figure 7: Meta-Regression Analysis of Change in HFMSE Score From Baseline

CI = confidence interval; HFMSE = Hammersmith Functional Motor Scale Expanded.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Figure 8: RULM Results by Reporting Author in the Adult Population

Dashed line = 95% confidence interval; Plain line = standard deviation; Square = SMA type II; Circle = SMA type III, Diamond = ambulant SMA type III; Triangle = nonambulant SMA type III; Square + circle + triangle = mix phenotypes; Bold font = median value; Italic = mean value; Light blue = approximately 10 months from initiation of drug; Red = approximately 12 months from initiation of drug; Grey shade = SMA type II; White shade = SMA type III; Striped shade = mixed phenotypes.

* Mean/median values of the baseline population, non excluding drop-outs at 10, 14, or 24 months of follow-up.

** Mean/median values of the baseline population of both SMA type II and III combined.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Figure 9: Meta-Regression Analysis of Change in RULM Score from Baseline

CI = confidence interval; RULM = Revised Upper Limb Module.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Figure 10: 6MWT Results by Reporting Author in the Adult Population

Dashed line = 95% confidence interval; Plain line = standard deviation; Bold font = median value; Italic = mean value; Light blue = approximately 10 months from initiation of drug; Red = approximately 12 months from initiation of drug.

* Mean/median values of the baseline population, non excluding drop-outs at 10, 14, or 24 months of follow-up.

** Mean/median values of the baseline population of both SMA type II and III combined.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Figure 11: Meta-Regression Analysis of Change in 6MWT Score From Baseline

6MWT = 6-minute walk test; CI = confidence interval.

Source: Coratti et al. (2021).²⁰ This work is licensed under the Attribution 4.0 International Licence (https://creativecommons.org/licenses/by/4.0/). Full text available here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8515709/

Critical Appraisal

The meta-analysis was based on an adequately conducted and reproducible systematic literature search that included planned searches of multiple databases and hand searches of references. Information on screening and data collection was provided, but it was unclear if studies were selected independently or in duplicate. Moreover, it was unclear if the inclusion and exclusion criteria related to population, outcomes, and study design were pre-specified. A quality assessment of the included studies was conducted using the Risk of Bias Assessment Tool for Nonrandomized Studies. No interpretation on the quality of studies was conducted; however, as all studies were observational, most were noted to suffer from a high level of bias in the selection of participants. The publicly available studies summarized previously (Hagenacker et al. [2020],¹⁷ Maggi et al. [2020],¹⁸ and Pera et al. [2021]¹⁹) were included in the submitted meta-analysis.

Outcomes included in the meta-analysis were appropriate and relevant to the Canadian context, with HFMSE, RULM, and 6MWT (in ambulant patients) most commonly included in studies, although there were differences in reporting and time of assessment. The authors noted that details on respiratory function or safety concerns were not systematically addressed in all the studies, so the impact of nusinersen on these outcomes cannot be determined. Most of the included studies had follow-up times ranging from 10 to 14 months. As previously mentioned, in discussion with the clinical experts consulted by CADTH, this short follow-up duration is insufficient to observe clinically meaningful changes in motor-function outcomes, so there is uncertainty in the magnitude of benefit.

Key features of the study design, inclusion or exclusion criteria, and baseline characteristics in the included studies were not provided, so a complete assessment of potential sources of clinical and methodological heterogeneity was not possible. The authors noted that there was considerable heterogeneity in the studies, given the inclusion of both ambulant and nonambulant type II and type III SMA patients, as well as the inclusion of both adults and children. Additionally, the included studies had small sample sizes, limiting the conclusions that can be drawn from individual studies. To explore the sources of heterogeneity, meta-regression was used, generally adjusted by age, SMA type, and treatment. Overall, there was a considerable level of heterogeneity in the overall cohorts, including both adult and pediatric populations, with I² ranging from 68% to 90%. In the adult subgroup, there was a moderate to considerable level of heterogeneity in the included studies across outcomes, with I² values ranging from 43% to 71%. No further information on sources of clinical, methodological, or statistical heterogeneity were explored.

Results and effect sizes for each outcome were presented for each study, with effect size estimated with random-effect models, which was considered appropriate because of the high level of between-study heterogeneity. Pooled estimates of mean change for motor-function outcomes — 1.87 points (95% CI, 1.05 to 2.68) for HFMSE, 0.64 points (95% CI, 0.27 to 1.01) for RULM, and 20.28 m (95% CI, 1.17 to 39.40) for 6MWT — favoured nusinersen treatment. Although there is a general positive association between nusinersen treatment and motor-function outcomes, it is unclear to what extent these changes are clinically meaningful. There was variation in what the included studies considered a clinically meaningful change, with some defining clinically meaningful improvements as a 3-point change in HFMSE score, a 2- to 3-point change in RULM score, and a 30-metre change in 6MWT, so the results of the meta-analysis may not demonstrate clinically meaningful improvements in motor-function outcomes. As previously mentioned, in discussion with the clinical experts, there is uncertainty about what constitutes a clinically meaningful change in the adult population for these outcome measures. Moreover, as previously discussed, given the nature of the included studies and the limitations defined for the studies included in both the systematic review conducted by CADTH and the meta-analysis, the observed effects cannot be attributed to nusinersen; therefore, the results of the meta-analysis are uncertain. Last, the estimates generally displayed wide 95% CIs, particularly for the 6MWT, and in many cases crossed the zero meridian, indicating a high level of variation in these cohorts and substantial imprecision in estimates of treatment effect. In addition, the cause of the wide CIs was unclear, but it is believed to be due to unaccounted-for heterogeneity and low sample sizes, particularly for the subgroups of interest that were evaluated (i.e., adults only). No sensitivity analyses were conducted. The authors noted the high number of missing study details on baseline functional status, and scores and other variables did not allow for detailed statistical analysis or meta-analysis, which the authors noted would have helped them to better understand the possible treatment effect of a number of variables, such as age, SMN2 copies, and functional ability at baseline. Although suggesting that the benefit of nusinersen is favourable in motor function, the results of the meta-analysis are unclear and provide imprecise estimation of the true effect of nusinersen on these outcomes in adults with type II or III SMA.

Evidence Identified From the Literature

Eight studies were identified in the literature search that met all inclusion criteria of the systematic review except study design, as they were all noncomparative observational studies. These 8 studies are briefly summarized in the following sections.

Brakemeier et al. (2021)

The study by Brakemeier et al. (2021)³⁰ was a prospective, noncomparative, German, single-centre, observational study that included 22 patients with molecularly confirmed 5q SMA who had been on therapy for at least 6 months. Included patients had type II or III SMA, documented bulbar dysfunction before treatment initiation, and no percutaneous enteroscopic gastrostoma. Patients were treated with nusinersen 12 mg on days 1, 14, 28 and 63, and then consecutively at 4-month intervals, in accordance with the product monograph. The study aimed to assess bulbar function in adults with type II and III SMA at 6 and 14 months. Patients ranged in age from 20 to 72 years, with a mean age of 38.5 years (SD = 14.2). Most patients were male (13 [59%]), had type II SMA (12 [54%]), had 3 copies of the SMN2 gene (14 [64%]), and had no presence of spondylodesis (12 [54%]). Symptom duration ranged from 14 to 52 years, with a mean duration of 34.3 years (SD = 11.9).

The primary outcomes were pre- and post-treatment comparisons of the bulbar subscore of the Amyotrophic Lateral Sclerosis Functional Rating Scale Revised (ALSFRS-R) and the Sydney Swallow Questionnaire (SSQ) scores at baseline, after 6 months of therapy, and after 14 months of therapy. The article stated that pre- and post-treatment comparisons in outcomes were made using the Wilcoxon signed-rank test with an alpha of 0.05 or less. There were 16, 20, and 18 patients measured using ALSFRS-R at each time point, respectively. The mean bulbar subscores from the ALSFRS-R were 10.50 (SD = 1.21) at baseline, 10.50 (SD = 1.32) at 6 months, and 10.94 (SD = 0.97) at 14 months. No significant change in bulbar function measured with the bulbar subscore of the ALSFRS-R was found between baseline and 6 months (z = –0.302; P = 0.763), or between baseline and 14 months (z = –1.406; P = 0.160). A total of 18, 15, and 13 patients completed the SSQ at baseline, 6 months, and 14 months, respectively. The mean SSQ scores were 23.59 (SD = 18.85) at baseline, 24.12 (SD = 19.32) at 6 months, and 19.95 (SD = 17.86) at 14 months. No significant change was found in the SSQ score between baseline and 6 months (z = –0.210; P = 0.834), or between baseline and 14 months (z = –0.392; P = 0.695). No significant changes in bulbar function were observed in subgroup analyses in which pre- and post-treatment comparisons of the 2 bulbar scores for type II and III SMA patients were calculated separately. Secondary outcomes included pre- and post-treatment comparisons of HFMSE and RULM scores. Improvement in HFMSE was observed from baseline to 6 months (z = –2.236; P = 0.025), but the change from baseline to 14 months was not significant (z = –1.248; P = 0.212). Similarly, no significant pre- and post-treatment differences were observed for the change in RULM score between baseline and 6 months for 22 patients (z = –1.932; P = 0.053), or between baseline and 14 months for 18 patients (z = –0.256; P = 0.798).³⁰

The study used an observational, noncomparative design. A clear hypothesis to be tested was not stated, so a direct cause-and-effect relationship could not be established for nusinersen. It was unclear if the study had a pre-specified protocol, and there was no mention of how patients were identified or selected to participate in the study, which prevented assessment of the population enrolled for selection bias and external validity. The study did not provide clear information regarding patients who discontinued treatment or withdrew from the study. The number of patients completing outcome measures at various time points was provided, but reasons for drop out or noncompletion were not reported. The approach for handling missing data was not mentioned, so the impact of attrition bias and missing data should be considered when interpreting the study results. The study included patients with type II and III SMA who had been on nusinersen for at least 6 months, and 19 patients had been on therapy for 14 months at the time of data analysis; however, there was no reporting of the time patients had contributed before and after the data analysis, or whether the patients’ time had been considered into the analysis. Moreover, the study did not provide details of the therapy patients received; for example, there was no mention of whether there were other treatments or whether such treatments were administered before nusinersen, at the time nusinersen was started, or during nusinersen treatment. Thus, it is unclear whether the patients were truly naive to disease-modifying therapy, what supportive treatments they received, or whether the time patients’ contributed was considered in data analysis. Last, the funding source for this study was Biogen.

Duong et al. (2021)

Duong et al. (2021)³¹ was a prospective, noncomparative, multi-centre, observational study that included 42 patients with genetic confirmation and clinical symptoms of SMA in the US. Included patients were 17 years or older at the start of nusinersen treatment. Patients were treated intrathecally with doses of 12 mg of nusinersen, based on the recommended schedule (baseline, day 14, day 28, day 58, and every 4 months thereafter). The study aimed to evaluate changes in clinical outcome measures and to obtain safety information related to nusinersen treatment. Patient ages ranged from 17.7 to 66.1 years, with a mean age of 33.7 years (SD = 12.6). Most patients were male (24 [57.1%]), had type III SMA (24 [57.2%]), had 3 copies of SMN2 (26 [72.2%]), were nonambulant (17 [40.5%] nonsitter and 14 [33.3%] sitter), and had no presence of spinal fusion (26 [61.9%]). Patients had a symptom duration that ranged from 12.1 to 60.6 years, with a mean duration of 30.9 years (SD = 12.3).

Motor function was measured with HFMSE, RULM, CHOP-ATEND, Spinal Muscular Atrophy Functional Rating Scale (SMAFRS), 6MWT, and timed up and go scores. Baseline assessments occurred in the 5 months after the first dose, and post-treatment assessments were done up to 24 months after nusinersen initiation. The study used linear mixed-effects models to estimate the mean annual rate of change (slope) for each outcome.

The slopes of all motor-function measures showed change in a positive direction. However, greater changes were observed in CHOP-ATEND (slope = 3.59 points per year [95% CI, 0.67 to 6.51]; n = 24) and SMAFRS (slope = 1.44 points per year [95% CI, 0.04 to 2.83]; n = 31] than in HFMSE (slope = 0.86 points per year [95% CI, −0.52 to 2.24]; n = 31), RULM (slope = 0.11 points per year [95% CI, −0.45 to 0.67]; n = 39), 6MWT (slope = 3.29 m per year [95% CI, −28.04 to 34.62]; n = 10), and timed up and go (log slope = −0.10 [95% CI, −0.21 to 0.01]; n = 8). Ventilatory function was measured using percent-predicted FVC, maximal expiratory pressure, and maximal inspiratory pressure. The slopes for the changes from baseline were 0.75% per year (95% CI, −1.87 to 3.38) for FVC, 6.38 cm H₂O per year (95% CI, 2.52 to 10.25) for maximal expiratory pressure, and −5.50 cm H₂O per year (95% CI, −11.47 to 0.47) for maximal inspiratory pressure. In the subgroup analysis, greater changes were observed in patients who were SMA type II, nonsitters, and had 3 copies of SMN2 than in other subgroups in general. For example, greater scores changes were observed in CHOP-ATEND in patients who were SMA type II (slope = 3.75 points per year [95% CI, −0.16 to 7.66]), nonsitters (slope = 6.44 points per year [95% CI, 2.25 to 10.62]), and had 3 copies of SMN2 (slope = 3.62 points per year [95% CI, −0.10 to 7.35]) than in patients who were SMA type III (slope = 3.26 points per year [95% CI, −1.34 to 7.86]), had 4 copies of SMN2 (slope = 3.05 points per year [95% CI, −5.96 to 12.05]), and sitters (slope = −0.29 points per year [95% CI, −5.07 to 4.50]).

The most frequently reported AE was post-lumbar-puncture headache (n = 6 [14.3%]). Other less frequent AEs included injection-site pain, nausea and/or vomiting, lightheadedness, and anxiety. No cases of hydrocephalus, bleeding and/or bruising, or renal compromise were reported.³¹

The primary limitation of the study was the noncomparative observational design. The study recruited participants through the Pediatric Neuromuscular Clinical Research network; however, there was no mention of inclusion or exclusion criteria for study participants. In addition, the study included participants with good prognostic factors; for example, the majority of patients were nonsitters (40.5%) and ambulant (26.2%), with 3 copies of SMN2 (72.2%) and no spinal fusion (61.9%). Therefore, the study was at high risk for selection bias. The study lacked a comparator group, making it difficult to determine whether the observed effects in the study population can be attributed to nusinersen. The study did not specify any hypotheses and did not report any sample size or power calculations; thus, it is not clear whether the sample size of 42 was sufficient to detect changes in the targeted population, or in subgroup analyses of patients with different SMA types, SMN2 copy numbers, ambulatory status, or spinal fusion status. The baseline clinical and disease characteristics (SMN2 copy number, proportion of ambulatory patients, and presence of spinal fusion) were suggestive of an over-representation of patients with higher-functioning disease and may compromise the representativeness of the study sample to the general population of adults with SMA. Thus, the primary results and results of subgroup analyses must be considered with caution and may not be generalizable. The lack of blinding in this study may lead to expectation bias in favour of nusinersen, as both patients and investigators were aware of the assigned treatment. The SMA Foundation and Cure SMA were the funding sources for this study.

Elsheikh et al. (2021)

Elsheikh et al. (2021)³² was a prospective, open-label, noncomparative, observational study conducted in the US that assessed the safety and treatment effect of nusinersen in 19 nonambulatory adults with confirmed 5q SMA. All patients were reportedly given intrathecal nusinersen in accordance with product-monograph dosing. At baseline, mean age of the patients was 39.7 years (SD = 13.9; range = 21.3 to 64.8). Only 7 patients (36.8%) were male. Most patients (16 [84.2%]) had 3 copies of SMN2, and 1 (5.3%) and 2 patients (10.4%) had 2 and 4 copies, respectively. Nearly half (9 [47%]) of the included patients had type II SMA, and 10 (53%) had SMA type III. The mean age at SMA onset was 27 years (SD = 34). The mean age at loss of ambulation in SMA type III patients was 25.8 years (SD = 18.3). A total of 10 (52.6%) patients had spinal fusion, and 12 (63.2%) were on ventilatory support at baseline.

The primary outcome of the study was the mean change from baseline in FVC at 2 months, 6 months, 10 months, and 14 months, which were –0.02 L (95% CI, –0.11 to 0.08), –0.02 L (95% CI, –0.11 to 0.07), –0.02 L (95% CI, –0.11 to 0.07), and 0.02 L (95% CI, –0.09 to 0.12), respectively. Key secondary outcomes were measured at the same time points and included mean change from baseline in HFMSE (0.77 points [95% CI, –0.29 to 1.83]; 0.74 points [95% CI, –0.3 to 1.78]; 0.32 points [95% CI, –0.73 to 1.36]; and 0.11 points [95% CI, –1.11 to 1.32], respectively), mean change from baseline in RULM (1.31 points [95% CI, 0.24 to 2.39]; 0.89 points [95% CI, –0.16 to 1.95]; 0.95 points [95% CI, –0.1 to 2], and 0.27 points [95% CI, –0.96 to 1.5], respectively).

AEs reported include transient headache, nausea, dizziness, back and neck pain, urinary tract infection, and upper respiratory infection. There was no report of clinically significant vital-sign abnormalities. High baseline protein/creatinine ratio was seen in 4 patients, but no significant change related to treatment was noted.³²

A key limitation of the study was the noncomparative observational design. In the absence of a comparator group, it is difficult to determine whether the observed effects in the study population can be attributed to nusinersen. Furthermore, given that no blinding was in place, patients and outcome assessors were aware of the interventions and outcomes of the study. Expectations of treatment may interfere with performance or assessment of outcomes. There were also concerns with selection bias in this study. It was unclear if the sampling method was determined a priori, or how patients were chosen. There was about an equal proportion of SMA type II and type III patients enrolled in the study, which appeared to be inconsistent with the observation that SMA type II is the predominant subtype in adults in Canada. Type III patients generally have better baseline function and are more responsive to interventions than type II patients. Thus, the higher proportion of type III SMA patients may result in bias toward nusinersen. The study also had a significant loss to follow-up. Of the 19 patients enrolled, only 12 patients were studied at all time points for most outcomes. Reasons for loss to follow-up were not mentioned. In addition, when assessing the HFMSE score, the study assigned a HFMSE score of 0 to 13 patients who were not scorable due to phenotypic severity. This imputation method may lead to bias in favour of the intervention, as these patients had no change in score throughout the study and were assumed to be stable.

Mix et al. (2021)

The study by Mix et al. (2021)³³ was a prospective, noncomparative, observational study of 26 patients with genetically confirmed 5q SMA in Germany. Included patients were a mix of adults (n = 14) and children (n = 12) treated with nusinersen. In addition, the investigators included a comparator group of 22 untreated individuals who were neurologically healthy. The patients and the comparator group were mean-matched for sex, age, and education. The study aimed to assess physical function and psychological well-being, including QoL and depressiveness in patients with SMA at baseline, day 60, and day 180 (6 months) of nusinersen treatment. For the 14 adults, ages ranged from 21.9 to 60.8 years, with a median age of 44.2 years. Most patients were male (10 [71%]), had type III SMA (9 [64%]), and were nonambulant (6 [43%] nonsitters and 4 [29%] sitters). Age at disease onset ranged from 0 to 17 years, with a median age of 1 year.

Well-being and physical-function measurements were compared between patients and the comparator group at baseline and 6 months. Psychological well-being was measured with the Anamnestic Comparative Self-Assessment (ACSA) for global QoL and the 36-item Short Form Health Survey (SF-36) for HRQoL. Depressive symptoms were assessed with the 12-item ALS Depression Inventory (ADI-12). For adults, no significant differences in ACSA were observed in the comparator group, and there were no data reported on the baseline comparisons of ADI-12 and SF-36 between adults and comparator group. Four adults (28.6%) showed at least mild depressive symptoms on a scale from 25.0% (least depressive) to 100% (most depressive) at baseline. At 6 months, no difference was observed in ACSA and ADI-12 between adults and the comparator group. Concerning SF-36 scores in adults, changes were observed in 3 of 9 dimensions, including role-limitations due to physical health (P = 0.034), general health (P = 0.021), and health transition (P = 0.002) over 6 months. Physical function was evaluated with the ALSFRS-R for all patients and with the HFMSE for patients with type II or III SMA. For motor-function assessments, no data were reported on the baseline and 6-month comparison of ALSFRS-R and HFMSE between adults and matched patients in the comparator group.³³

No inclusion and exclusion criteria were specified in the study by Mix et al. (2021),³³ and 50% of the participants had type III SMA, 50% were nonsitters, and 23% were ambulant (walkers), which increases the potential for selection bias. Moreover, there was no mention of the demographic characteristics of the comparator group, or how they were identified and selected to participate in the study. The study did not report any sample size or power calculations that tied to a hypothesis; thus, it is not clear whether the sample size was sufficient to detect changes in the targeted population and in subgroup analyses of patients with different ages. As such, primary and subgroup analyses must be considered with caution. In addition, patients reported their psychological well-being status in a retrospective manner (from the previous 2 to 4 weeks); this approach increases the risk of recall bias. The study had a relative short follow-up period (6 months), which is insufficient to observe the longer-term benefits and harms of nusinersen treatment in patients with a chronic and heterogeneous disease. As such, there is limited confidence in the ability to extrapolate these results in the longer term. Furthermore, there was limited reporting of within-group and between-group changes in outcome measures. For example, measurements of ACSA, SF-36, ADI-12, HFMSE, and ALSFRS-R scores were taken at baseline, day 60, and day 180, but the study only reported Chi-square and P values, with brief summaries of the trend in change from baseline to day 180; outcome data from day-60 assessments were not reported. The study was funded by various founding agencies, such as a clinician scientist fellowship sponsored by the Charcot Foundation for ALS research, the Bundesministerium für Bildung und Forschung, and the Deutsche Forschungsgemeinschaft.

Kizina et al. (2020)

Kizina et al. (2020)³⁴ was a prospective, observational, cohort study that assessed fatigue in 28 adults with genetically confirmed 5q type II and III SMA in Germany. All patients had 3 to 5 copies of the SMN2 gene and a report of disease progression over the 12 months before treatment with nusinersen was started. Patients with psychosocial stress, a flu-like infection during the previous week, or depression were excluded, as were those using sedative medication. Intrathecal nusinersen was administered to all patients in accordance with the product monograph. Fatigue severity scale (FSS) and motor-function scores (6WMT and HFMSE) were measured at baseline, and at 6 and 10 months. The correlation between the change in FSS and the change in motor-function scores were also assessed. At baseline, the mean age of patients was 37 years (SD = 12). A total of 18 (64%) patients were male, 10 (36%) had type II SMA, and 18 (64%) had type III SMA. Most patients (18 [64%]) had 3 or 4 copies (9 [32%]) of the SMN2 gene. Most patients were nonambulatory (18 [64%]), but 10 (36%) were ambulatory.

The mean difference in FSS from baseline was –0.69 points (SD = 1.10) at 6 months and –0.70 points (SD = 1.56) at 12 months. In the subgroup analyses, the mean difference in FSS from baseline at 6 months and at 12 months was –0.93 points (SD = 1.15) and –0.89 points (SD = 1.86), respectively, in ambulatory patients, and –0.56 points (SD = 1.08) and –0.59 points (SD = 1.40), respectively, in nonambulatory patients. The correlation coefficient between the difference in FSS from baseline and the difference in HFMSE from baseline was –0.19567 (P = 0.3183) at 6 months and –0.18663 (P = 0.3513) at 10 months. The correlation coefficient between the differences in FSS and the differences in 6WMT from baseline was –0.68294 (P = 0.0295) at 6 months and –0.52727 (P = 0.1173) at 10 months.³⁴

A key limitation of the study was the lack of a comparator. The study did not control for known prognostic factors or effect modifiers of SMA in the analyses, so it is impossible to establish a causal link between the treatment and outcomes. It is also noteworthy that the distribution of SMA subtypes (type II vs III) in the study was not in line with the distribution in Canada, according to feedback from the clinical experts consulted by CADTH. The majority of patients had SMA type III. It appeared that the study population did not reflect a random sampling of the SMA population, suggesting a risk of selection bias. Another limitation was that FSS has not been validated in the SMA population. The scale is subject to a high risk of reporting bias, given that FSS is a patient-reported outcome,³⁵ and fatigue is highly subjective. Because no blinding was in place, a patient’s experience of fatigue can be influenced by a perception of the treatment. In addition, given that the scale used a recall period of 1 week and was administered only 3 times during the study — at baseline, 6 months, and 10 months — it is unclear if the data are reflective of patients’ treatment experiences throughout the study.

Moshe-Lilie et al. (2020)

Moshe-Lilie et al. (2020)³⁶ was a retrospective chart review that examined the change in motor function in 22 adults with genetically confirmed type II and III SMA between 2017 and 2019 at a medical centre in the US. Ten patients received treatment with nusinersen and 12 were untreated. The sum of MRC scores of upper and lower limb strength was recorded at baseline and at 4, 6, 12, and 24 months. The patients had a median age of 26 years (range = 20 to 71). Most patients were female (15 [68%]), nonambulatory (20 [91%]), had 3 copies of the SMN2 gene (13 [59%]), and symptom onset in childhood (14 [64%]). Most patients had type III SMA (13 [59%]) and the remainder had type II (9 [41%]). Scoliosis was present in 17 (77%) patients, and ventilator support was required by 10 (45%) patients. It was unclear if the patients in the nusinersen group had received any prior nusinersen treatment.

The change from baseline in mean % MRC was the primary outcome of interest. In the treated group, the mean % MRC change from baseline was 2.5% at 12 months and 3.9% at 24 months. In the untreated group, the mean % MRC change was not reported but, according to the authors, % MRC was stable in most untreated patients. A decline of 2.5% to 3.8% in % MRC was noted in 3 nusinersen-treated patients. Of the 3 treated patients for whom HFMSE scores were available, 2 remained stable and 1 had a 12-point improvement at 12 months. In terms of safety, lumbar-puncture headache was observed in 5 patients, 2 of whom required a blood patch. One patient was hospitalized with bacterial meningitis and required long-term antibiotics. One patient died of respiratory failure in the setting of pneumonia shortly after treatment initiation. Three patients discontinued treatment for reasons that included recurrent pneumonia, lack of improvement, and proteinuria.³⁶

The primary limitation of the study was the risk of selection bias. All charts reviewed were sourced from 1 neuromuscular group. Having a single source may introduce selection bias, as the patients under the care of this team may share common characteristics, (such as treatment history, disease severity, and level of supportive care) that can bias estimation of the treatment effect. It is unlikely that the selected patients were representative of the SMA patient population, considering that the majority (59%) had SMA type III, which does not align with the comment from the clinical experts that SMA type II is the predominant SMA subtype seen in clinical practice. Furthermore, there was no mention whether any matching or adjustments for plausible prognostic factors were made between treated and untreated patients. It is impossible to assess whether there were imbalances in the baseline characteristics between the 2 groups, as only pooled data were presented. Any imbalances in prognostic variables will likely result in a biased conclusion. Another limitation was that there was very limited reporting of outcome data. Other than a trajectory plot to illustrate the % MRC of each patient over time, most outcome data were incomplete and were summarized descriptively. For example, the difference in mean % MRC from baseline, which was the key outcome, was reported in the treated arm but not in the untreated arm. Minimal information on measurements of HFMSE and RULM scores was provided. The qualitative descriptions provided limited value for objective interpretation of the outcomes. The risk of reporting bias was high, given the limited availability of data and the presence of selective reporting.

Yeo et al. (2020)

The study by Yeo et al. (2020)³⁷ was a prospective, noncomparative, observational study that included 6 adults with molecularly confirmed 5q type III SMA from a single centre in the US. Patients with limb contractures, a history of spinal fusion, spine surgery or significant scoliosis that affected function, or respiratory insufficiency that required more than nocturnal BiPAP or continuous positive airway pressure were excluded from the study. Patients were treated with nusinersen, with loading doses of 12 mg at baseline, 2, 4, and 8 weeks, followed by maintenance doses of 12 mg at 4-month intervals (120 days). The study aimed to determine the impact of nusinersen on motor outcomes in a targeted cohort of adults with type III SMA for at least 12 months. Primary outcomes were HFMSE and RULM scores. Secondary outcomes were the PedsQL Multidimensional Fatigue Scale, the SMAFRS, and the 6-minute and 10-metre walk tests. Outcome measures and evaluations were conducted every 3 to 4 months, before each maintenance dose. The mean follow-up duration was 17 months, with a range of 14 to 21 months. Patient ages ranged from 24.9 to 56.5 years, with a median age of 29.9 years. Most patients were male (5 [83%]), had 3 copies of SMN2 (3 [50%]), and were ambulant (4 [67%]). Age at symptom onset ranged from 1 to 14 years, with a median of 8 years. Other patient characteristics included a median HFMSE score of 35 (range = 21 to 53), and a median RULM score of 31.5 (range = 22 to 37).

Mean changes in HFMSE and RULM scores over 14 months were 2 points (range = 1 to 5) and 1.8 points (range = 0 to 3), respectively. Over the course of the study (at various points until 21 months), individual changes greater than 2 points were observed in the HFMSE score of 3 (50%) patients and in the RULM score of 2 (33%) patients. Of the 4 ambulatory patients able to complete the 6MWT and 10-metre walk test, no statistically or clinically meaningful changes were observed in either test from baseline to the final study visit. The study reported a decline in the SMAFRS total score and no changes in the PedsQL. A total of 12 AEs were reported: 4 patients developed post-lumbar-puncture headache, 2 developed vertigo, 4 experienced fall-related injuries (including 1 SAE that required hospitalization), and 2 developed recurrent pressure sores (including 1 SAE that required hospitalization).³⁷

The main limitations of the study are the noncomparative, observational design and the lack of reporting on any sample size or power calculations tied to a hypothesis; thus, it is not clear whether the sample size was sufficient to detect changes in the targeted population, which may limit the conclusions and generalizations that can be drawn from this study. All patients included in the study had SMA type III, and patients with worse prognostic factors (such as limb contractures, spinal fusion or scoliosis, and respiratory insufficiency) were excluded, which resulted in a higher-functioning population and compromised the representativeness of the study sample to the general adult SMA population. The study did not deploy any method of blinding, so both the patients and investigators were aware of the nusinersen treatment. The expectation of treatment-induced improvements may have influenced the behaviour of patients or outcome assessors, leading to expectation bias in favour of nusinersen. The study reported limited data on outcome measures using line graphs and brief description, and the investigators didn’t provide statistical analysis of outcomes. For example, HFMSE, RULM, PedsQL, SMAFRS, 6MWT, and 10-metre walk test were assessed at various time points until 21 months, but the study only reported the mean change and range for HFMSE and RULM from baseline to 14 months and brief summaries of the trend of change for PedsQL, SMAFRS, 6MWT, and 10-metre walk test from baseline to final visits; there was no mention of data on outcome measures at other time points. Although the study described the statistical analysis in the analysis plan section, it provided no statistical test results. Moreover, limited data were presented for outcomes of interest at various time points. The selective reporting of data may lead to distortion, and prevent readers from interpreting the study results objectively. Therefore, the study reported incomplete data and was subjected to a high risk of reporting bias.

Walter et al. (2019)

Walter et al. (2019)³⁸ was a prospective, single-centre, observational study that evaluated the safety and treatment effects of nusinersen in 19 adults with 5q SMA type III in Germany. Patients were treated with intrathecal nusinersen regimen, per the product monograph, for up to 300 days. Patients were monitored within the SMArtCARE registry. Motor-function scores, pulmonary function, and laboratory biomarkers were assessed at baseline and days 63, 180, and 300. The mean age at baseline was 35.1 years (SD = 11.72; range = 18 to 59). The mean disease duration was 24.3 years (SD = 12.1). The mean age at symptom onset was 10.8 years (SD = 9.2). Most patients had 4 copies of SMN2 (79%), and most patients were ambulatory (12 [63%]); only 7 (37%) patients were wheelchair-dependent and only 3 (15.7%) patients had scoliosis. The study did not indicate whether the patients had received any prior nusinersen treatment.

The mean RULM score improved from 32.32 (SD = 7.39) at baseline to 33.06 (SD = 7.33) on day 300. The mean HFMSE score was 35.16 (SD = 21.14) at baseline and 39.50 (SD = 20.58) on day 300. The mean 6WMT increased from 369.50 m (SD = 126.62) at baseline to 377.75 m (SD = 156.60) on day 300. The mean percent-predicted FVC at baseline (94.54 [SD = 15.45]) was similar to the level on day 300 (99.54% [SD = 12.42]). Among the 11 patients who reported AEs, there were 4 cases of post-lumbar-puncture headache, 1 case of fatigue, and 7 cases of lower back pain. Laboratory tests were normal and there was no evidence of nephrotoxicity.³⁸

As in the other studies, the noncomparative design was a key limitation. All enrolled patients were treated with nusinersen; there was no comparison arm. No definitive conclusion can be drawn regarding the treatment effect of nusinersen, as the outcomes were confounded by factors such as physical exercise and learning effect, which are known effect modifiers of SMA. It was also unclear if random sampling of patients was involved; the sampling method was not described. It is, therefore, unclear whether the patients selected were reflective of the type III SMA population. Although the study captured a group of patients with large variations in disease duration (range = 6 to 52 years), the distribution was not even across the spectrum. Of the 19 patients, 14 (73.7%) had less than 30 years of SMA history. Having a high proportion of patients with shorter disease history can result in biased results, as these patients generally have more functional reserves, which allow for the detection of any potential treatment responses. Moreover, the baseline motor-function scores in this study were high, suggesting that patients enrolled had less severe disease. Another limitation was that the duration of follow-up was only up to 300 days, which was inadequate to capture meaningful changes in functional outcomes because SMA patients generally have very slow disease progression.

Discussion

Summary of Available Evidence

As part of the reassessment for nusinersen, 7 studies were submitted to CADTH, of which 4 were summarized and appraised; Hagenacker et al. (2020)¹⁷ was a prospective, multi-centre, noncomparative observational study; Maggi et al. (2020)¹⁸ was a retrospective, noncomparative cohort study; Pera et al. (2021)¹⁹ was a noncomparative, registry-based study of the ISMAR; and the EU registry study^15,16 submitted by the sponsor was a noncomparative, registry-based cohort study from 3 European registries.

There was a high degree of selection bias in the included patients. Overall, the studies enrolled mostly adults with type III SMA (62% to 100%), with very few type II patients. Moreover, across studies, the majority of patients had 4 or more copies of SMN2, with a high degree of variation in the proportion of ambulatory and nonambulatory patients. The primary outcome in all studies was change in HFMSE over time; other measures included RULM and 6MWT. Baseline functional measures for the HFMSE, RULM, and 6MWT in all studies were considered high, indicating a high level of motor functionality in included patients. The median follow-up for each study was not frequently reported; however, most patients were followed for a minimum of 6 months and up to 14 months, which was considered immature for SMA. In addition, the noncomparative nature of these studies prevents the determination of any association between nusinersen and the reported results.

In addition to the key studies reviewed, CADTH identified 8 studies that included the relevant population; however, they did not meet the study design criteria for inclusion in the systematic review.

Interpretation of Results

Effectiveness

Nusinersen has previously been reviewed by CADTH in 2017 and 2019, and the CDEC recommended reimbursement of the drug with conditions for patients with 5q SMA with 2 or 3 copies of the SMN2 gene and have had a disease duration of less than 6 months and symptom onset from 1 week to 7 months of age, and for patients who are 12 years or younger with symptom onset after 6 months of age who never achieved the ability to walk independently on the basis of multiple comparative and noncomparative RCTs (ENDEAR, CHERISH, and NURTURE).^13,14 The clinical experts and patients identified an important unmet need for treatments in adults with SMA naive to disease-modifying therapy that will reverse damage, halt or stabilize muscle weakness related to disease progression, and maintain independence and QoL by improving strength.

Overall, the numerous limitations in the reviewed studies, identified in the critical appraisal, prevent generalization of the results to adults who have type II or III SMA and who have never been treated with nusinersen. The key limitations were the relatively small sample sizes (ranging from 67 to 252 patients), selection bias leading to study populations unrepresentative of the target population or the reimbursement request (e.g., a lack of type II SMA patients), and the generally noncomparative study designs that did not allow for the establishment of cause and effect or associations between nusinersen and the observed changes in outcomes measures. As well, the notable rates of drop-outs for key outcomes between baseline and 14 months and the relatively short duration of follow-up for a chronic, slowly progressing disease were important limitations that hindered interpretation of the results.

Most of the experts agreed that there is limited evidence in this population and highlighted the lack of RCTs, noting that all available evidence comes from observational studies, precluding the use of nusinersen in adults. Despite the small sample sizes, the clinical experts consulted by CADTH considered it reasonable to conduct a comparative, placebo-controlled RCT in adults with type II and III SMA. However, several of the clinical-expert panel members discussed the lack of a clear biologic mechanism through which SMN2-modifying therapy such as nusinersen could benefit adults with SMA in whom the production of the SMN protein had already decreased significantly, making the reversal of muscle function impossible.

Most of the included studies did not include a comparator, other than evaluating pre- and post-treatment outcomes. An analysis of nusinersen-treated and untreated patients was conducted in the EU registry study^15,16; however, only 17 SMA patients (14 type III, and 3 type IV) were included in the untreated group, which was considered insufficient to determine whether nusinersen demonstrated an effect in these patients.

All of the main studies provided by the sponsor and summarized in this reassessment shared a common outcome of change from baseline in motor function, measured with the HFMSE, RULM, or 6MWT. These measures are well known and validated in the SMA population and are regularly used in Canadian clinical practice. Although the HFMSE and 6MWT are appropriate measures to determine motor function in SMA, they are generally limited to ambulatory patients who can stand and sit. Given that in all the included studies there was heterogeneity in the proportion of ambulant and nonambulant patients, these measures are subject to even smaller sample sizes, affecting the interpretation of the results across studies. In addition to outcomes that evaluated motor function, clinically important outcomes identified by patients and clinicians, including the maintenance of independence and QoL, were not evaluated in any of the included studies; therefore, the effect of nusinersen on these outcomes remains uncertain. Other important outcomes, including vocalization, swallowing issues, ventilation, or nutritional requirements, were not available in the included studies, which many patients and clinicians cited as having a significant impact on HRQoL and daily life. Only Brakemeier et al. (2021)³⁰ (summarized in the Other Relevant Evidence section) evaluated the effect of nusinersen on bulbar function.

Most patients enrolled in the studies had type III SMA with 3 or 4 copies of the SMN2 gene, and appeared to have higher functioning on baseline scale scores and ambulatory status. As hypothesized in the clinical-expert input, this population may be more likely to experience a benefit due to an elevated number of motor units and nerves; however, it remains unclear if clinical significance was reached, and the magnitude of effect is still uncertain. Combined with the identified limitations in the data, there is a high degree of uncertainty about the benefit of nusinersen in the requested population.

Across studies, the baseline scores for motor-function outcomes in nusinersen-treated patients — considered to be high by the clinical experts consulted by CADTH, based on the patients seen in their practice — ranged from 20.74 in Hagenacker et al. (2020)¹⁷ to 30.75 in Pera et al. (2021)¹⁹ for the HFMSE, from 20.87 in Hagenacker et al. (2020)¹⁷ to 29.0 in Pera et al. (2021)¹⁹ for the RULM, and from 300.87 m in the EU registry study^15,16 to 323.03 m in Pera et al. (2021),¹⁹ suggesting that the enrolled patients had less severe disease and greater physical function. Changes from baseline in these motor-function outcomes varied across studies, with a mean change from baseline in HFMSE score ranging from 0.79 points (95% CI, –0.29 to 1.87) at 12 months in Pera et al. (2021)¹⁹ to 3.12 points (95% CI, 2·06 to 4·19) at 14 months in Hagenacker et al. (2020).¹⁷ In the EU registry study,^15,16 the change in HFMSE was 0.02907 points per week (95% CI, 0.01930 to 0.03884) in nusinersen-treated patients, compared with –0.01129 (95% CI, −0.03289 to 0.01031) in untreated patients, and 0.2575 points per week (95% CI, 0.01038 to 0.04112) after the initiation of nusinersen; however, before initiation of nusinersen, the HFMSE was declining at –0.00006 points per week (95% CI, –0.0096 to 0.0094). As such, there is uncertainty as to whether patients were truly declining before nusinersen, and whether the slope improved in a meaningful way after nusinersen treatment. For the RULM, mean change from baseline ranged from 0.07 in Pera et al. (2021)¹⁹ at 12 months to 1.09 points in Hagenacker et al. (2020)¹⁷ at 14 months. In the EU registry study,^15,16 the change in RULM was 0.01168 points per week (95% CI, 0.004957 to 0.01841) in nusinersen-treated patients, with a post-treatment change of 0.002569 points per week (95% CI, −0.00533 to 0.01047). The post-treatment results were not statistically significant, and it is unclear whether this change represents a clinically meaningful improvement in the change in RULM score after treatment. Moreover, the comparison group consisted of only 17 untreated patients, making it difficult to interpret these comparisons. For the 6MWT, the mean change from baseline ranged from 0.52 m in Pera et al. (2021)¹⁹ at 12 months to 46.0 m in Hagenacker et al. (2020)¹⁷ at 14 months. In the EU registry study,^15,16 the change in 6MWT was 0.2633 points per week (95% CI, 0.09922 to 0.42740) in nusinersen-treated patients, with a post-treatment change of –0.03399 points per week (95% CI, –0.4373 to 0.3694). Despite extrapolation of the EU registry study^15,16 results to a 1-year timespan, it could not be concluded that these results were due to treatment with nusinersen, and they were not considered clinically meaningful by the clinical experts consulted by CADTH. In all studies, the mean changes in motor-function scores over the short study durations were considered small, and according to the clinical-expert panel, are likely to be noise due to the natural history of the disease. Coupled with the wide CIs, the motor-function results are considered imprecise, and do not represent meaningful change that can be attributed to nusinersen.

Despite the limitations related to study design, sample sizes, and selection bias, the totality of evidence identified and submitted consistently suggests a positive association in motor-function outcomes in adults treated with nusinersen throughout the studies. However, it is uncertain whether any of the observed changes can be attributed to nusinersen or whether they can be considered clinically meaningful. As to the natural history of the disease, the clinical experts consulted by CADTH noted that individual variations in disease progression are observed among adults, with phases of clinical worsening but also periods of stabilization, during which patients may retain vital motor functions; for some patients, this translates into preservation of functional independence. The reported interpretations in each study that nusinersen led to improvement or stabilization cannot be made without a clear hypothesis statement or a design or analysis to test this outcome specifically; thus, any changes seen in patients across studies must consider the individual variance and heterogeneity, and must be interpreted with the utmost caution. Furthermore, the maximum time of assessment across studies was 14 months, which was considered immature to determine the effectiveness of nusinersen or capture clinical changes in adults with type II or III SMA because of individual disease heterogeneity, natural history, and periods of disease stabilization.

The clinical experts consulted by CADTH noted that in SMA or any neuromuscular disease, there are many additional factors to consider when interpreting scale scores. Potential confounders and treatment-effect modifiers that are likely to influence outcomes and that are routinely seen in clinical practice, according to the clinical experts consulted by CADTH, were not considered, including training for the outcomes of interest, routine exercise and observation, and other routine care (such as physiotherapy and occupational therapy). It has been well documented in other chronic diseases that there are notable learning and encouragement effects associated with the motor-function measures, particularly the 6MWT,^39-42 although no training effect has been described for 6MWT in natural-history studies,³⁸ and the clinical experts stressed that routine exercise is known to have a profound effect on the maintenance of neuronal connectivity and muscle. These factors are expected to heavily influence the results of the included motor-function outcome measures, limiting the generalizability of results to Canadian patients. It should also be noted that in Canada, many SMA patients are potentially under a system of renewal for nusinersen, whereby reimbursement is contingent on the demonstration of improvements in motor-function scales; thus, patients are motivated to show improvements, which may further influence the outcome results. Another main concern of the clinical experts consulted by CADTH was evidence for the use of nusinersen in treatment-naive adults with type II or III SMA; the studies did not consider pathobiology or the natural history of the disease. The experts noted that nusinersen works best in young children with SMA, and emphasized the lack of data in older patients with more advanced disease. The experts highlighted the well-established pathobiology of SMA, particularly the demonstrated post-natal progression of motor denervation, reflected by compound muscle action potential and motor unit number estimation values in all SMA types in early ages by Swoboda et al. (2005).²⁵ The experts also noted that these results were compounded by the subgroup analysis for the CHERISH study, which demonstrated no statistically significant functional benefit with nusinersen in HFMSE scores in patients older than 4 years.⁴³ As such, the clinical experts consulted by CADTH indicated that pathobiology, natural history, trial data, and the often difficult administration of nusinersen make it unclear whether it (or any other currently available medications) would have a clinically meaningful effect in adults with type II or III SMA.

Clinically meaningful improvements in the motor-function outcomes have been established; however, there is some uncertainty about what constitutes a clinically meaningful improvement in adults with type II or III SMA. The experts also noted that there is a discrepancy in the reflection of the actual limitations of these measures on daily life; improvements in the scales may be shown, but might not translate into changes in functional ability. As noted in Appendix 4, an increase in HFMSE score greater than 2 points is unlikely in untreated patients with type II or III SMA.⁴⁴ However, patients and caregivers consider a 1-point increase meaningful,⁴⁵ and Pera et al. (2017)¹⁹ considered a 3-point change to be clinically meaningful.⁴⁶ With RULM, 2 points was considered to be a clinically meaningful improvement, and with 6MWT, an improvement of 23 to 45 m is considered to be clinically meaningful in other chronic conditions. The clinical experts consulted by CADTH noted that there is some heterogeneity in what can be considered a meaningful change in motor function, and explained that these minor changes in HFMSE or RULM would only result in minimal improvements, which may not cross important disease-severity thresholds that affect patient function or QoL.

A single meta-analysis of published evidence submitted by the sponsor evaluated the effect of nusinersen treatment on key motor-function outcomes. Although there was a general positive association between nusinersen and HFMSE, RULM, and 6MWT scores, the results were imprecise because of wide 95% CIs, and there was notable heterogeneity in the studies included in the meta-analysis with regard to study design (all observational), the included populations (both adults and children with type II and III SMA of various ambulatory ability), and small sample sizes (which limited the ability to draw any conclusions about the effectiveness of nusinersen). Moreover, it was uncertain if the pooled point estimates for the adult subgroup were considered clinically meaningful.

Eight studies were identified in the literature search that met all inclusion criteria of the systematic review except study design; all were noncomparative observational studies. As with the studies provided by the sponsor, the effectiveness of nusinersen in these studies is highly uncertain because of the noncomparative study design, selection bias, and relatively small sample sizes of adults with type II and III SMA.

Harms

The safety of nusinersen has been examined in the children and adolescents with SMA, as noted in the previous submissions to CADTH.^13,14 However, it was important for this review to consider the harms of nusinersen associated with the intrathecal administration method, given that patients with type II and III SMA generally have complex spines (some had already undergone spinal-fusion surgery).^47,48 It was not expected that harms would vary considerably by SMA type, but patients with type II SMA are more likely to have complex spines due to the inability to walk.⁴⁸ Because of the smaller sample sizes in the studies, separate analyses of AEs by SMA type were not conducted in any study.

Harms results varied and were infrequently reported in the available evidence. Across studies, the overall frequency of AEs was low, occurring in up to 47% of patients across studies, and the reported AEs were mostly related to procedural complications (most frequently headache and back pain), which are relatively common with this procedure. It should be noted that AE reporting is not standardized in registries, so these results should be interpreted with caution and may be biased. SAEs were rarely reported. As noted by the clinical experts, any AEs that occur with nusinersen will likely occur immediately after administration.

The frequency of attrition in the included studies likely affected the reporting of AEs and complications. As previously noted, survivorship bias may be present in the study populations, in which patients who are more likely to tolerate treatment remain on treatment for the duration of the study. Overall, only 4 patients across all included studies withdrew due to AEs; however, given the study designs, it is unclear whether more patients withdrew due to AEs but were not reported. In terms of safety, the duration of follow-up of 14 months for most studies is likely sufficient to observe nusinersen-related AEs; however, the effect of long-term exposure in adults with type II or III SMA is unknown.

Notable harms of interest for this reassessment were guided, in part, by the product monograph for nusinersen, and included serious infections, lumbar-puncture-related AEs, coagulation abnormalities, and renal toxicity, which are generally considered to be important in this patient population receiving disease-modifying therapy. Minimal results were provided for the notable harms of interest to this review, or for harms in general, which may be limited due to the design and aims of the included studies, as well as availability of information.

Overall, the clinical experts consulted by CADTH indicated that the potential for harms in adults related to the method of administration and the frequent need for interventional-radiology-guided administration in practice is a key part of their decision to use nusinersen in adults. The experts noted that there are limited intraspinous spaces available for lumbar puncture, which reduces the area for the various nusinersen injections. Additionally, the complexity of the spine can be a barrier for the intrathecal administration of nusinersen. Procedural complications in patients with complex spines were not reported in any of the included studies, and the proportion of patients requiring guided administration was only reported in 1 study (Maggi et al. [2020]¹⁸). The experts emphasized that the administration of nusinersen in patients with complex spines increases the difficulty and risk of the procedure. Administration of nusinersen using interventional-radiology-guided techniques allows for patients with complex spines to receive treatment, but the experts expressed concern about the frequency of exposure to ionizing radiation in patients who requiring X-ray-, CT-, or fluoroscopy-guided injections. Although the cumulative radiation is known to be low in the intrathecal, interventional-radiology-guided administration of nusinersen, clinical experts expressed concern that the effect of long-term exposure due to repeated administration over a significant period of time is unknown, and they acknowledged a need for more safety data in this area.

Conclusions

Four noncomparative, observational studies were included in the reassessment of nusinersen for adults with type II or III SMA. The observational nature and lack of a well-defined concurrent comparator in the included studies significantly limits the ability to establish causal relationships between treatment effects and nusinersen.

In all of the studies, selection bias in the included populations and relatively small sample sizes were noted as key limitations. All studies included mostly type III SMA adults with higher physical functioning at baseline based on SMA type, a higher number of SMN2 copies, better ambulatory status, and higher baseline scores for motor-function outcomes. Input from clinical experts noted that the populations were not reflective of the reimbursement request, particularly due to the lack of type II SMA patients, or to their clinical practice.

Although there was generally a consistent positive effect of nusinersen on motor-function outcomes, the magnitude of the treatment effect with nusinersen varied and was often not clinically meaningful. Given the limitations in study design, statistical analysis, duration, and the heterogeneous natural history of adults with SMA, results in all studies were considered highly uncertain and may not be generalizable to the Canadian population. Harms associated with nusinersen were generally mild to moderate in severity and were related to the administration procedure, with lumbar-puncture-related AEs the most frequently reported. However, the reporting of AEs was inconsistent and infrequent and, because of study designs and associated biases, may be under-reported.

Although the amount of real-world data for nusinersen is relatively high, the overall quality of studies remained a concern. Most of the identified evidence could not provide conclusive evidence demonstrating the effectiveness of nusinersen in adults with type II or III SMA. Overall, it remains unclear if nusinersen resulted in clinically meaningful improvements or disease stabilization, which were considered important outcomes by patients. Additionally, since HRQoL was not assessed in any studies, the effect of nusinersen on this important outcome in adults is unknown.

References

1.D’Amico A, Mercuri E, Tiziano FD, Bertini E. Spinal muscular atrophy. Orphanet J Rare Dis. 2011;6:71. PubMed

2.Tisdale S, Pellizzoni L. Disease mechanisms and therapeutic approaches in spinal muscular atrophy. J Neurosci. 2015;35(23):8691-8700. PubMed

3.Kolb SJ, Kissel JT. Spinal muscular atrophy. Neurol Clin. 2015;33(4):831-846. PubMed

4.Verhaart IEC, Robertson A, Wilson IJ, et al. Prevalence, incidence and carrier frequency of 5q-linked spinal muscular atrophy - a literature review. Orphanet J Rare Dis. 2017;12(1):124. PubMed

5.Farrar MA, Park SB, Vucic S, et al. Emerging therapies and challenges in spinal muscular atrophy. Ann Neurol. 2017;81(3):355-368. PubMed

6.Prior TW, Snyder PJ, Rink BD, et al. Newborn and carrier screening for spinal muscular atrophy. Am J Med Genet A. 2010;152a(7):1608-1616. PubMed

7.Lunn MR, Wang CH. Spinal muscular atrophy. Lancet. 2008;371(9630):2120-2133. PubMed

8.Ogino S, Wilson RB. Genetic testing and risk assessment for spinal muscular atrophy (SMA). Hum Genet. 2002;111(6):477-500. PubMed

9.McMillan HJ, Gerber B, Cowling T, et al. Burden of spinal muscular atrophy (SMA) on patients and caregivers in Canada. J Neuromuscul Dis. 2021;8(4):553-568. PubMed

10.Arnold WD, Kassar D, Kissel JT. Spinal muscular atrophy: diagnosis and management in a new therapeutic era. Muscle Nerve. 2015;51(2):157-167. PubMed

11.Wadman RI, Stam M, Gijzen M, et al. Association of motor milestones, SMN2 copy and outcome in spinal muscular atrophy types 0-4. J Neurol Neurosurg Psychiatry. 2017;88(4):365-367. PubMed

12.Spinraza (nusinersen): solution for intrathecal injection 2.4 mg/mL nusinersen as nusinersen sodium [product monograph]. Mississauga (ON): Biogen Canada Inc.; 2020 Apr 20.

13.CADTH Drug Reimbursement Expert Review Committee final recommendation: nusinersen (Spinraza — Biogen Canada Inc.). Ottawa (ON): CADTH; 2017: https://www.cadth.ca/sites/default/files/cdr/complete/SR0525_Spinraza_complete_Dec_22_17.pdf. Accessed 2022 Apr 4.

14.CADTH Drug Reimbursement Expert Review Committee final recommendation: nusinersen (Spinraza — Biogen Canada Inc.). Ottawa (ON): CADTH; 2019: https://www.cadth.ca/sites/default/files/cdr/complete/SR0576-Spinraza-Resubmission-Mar-1-19.pdf Accessed 2022 Apr 4.

15.Drug Reimbursement Review sponsor submission: Spinraza (nusinersen injection), solution for intrathecal injection 2.4 mg/mL nusinersen as nusinersen sodium [internal sponsor’s package]. Toronto (ON): Biogen Canada Inc.; 2021 Dec 1.

16.Bovis F, Sormani MP. Registry analyses spinal muscular atrophy. In: Drug Reimbursement Review sponsor submission: Spinraza (nusinersen injection), solution for intrathecal injection 2.4 mg/mL nusinersen as nusinersen sodium [internal sponsor’s package]. Toronto (ON): Biogen Canada Inc.; 2021 Dec 1.

17.Hagenacker T, Wurster CD, Gunther R, et al. Nusinersen in adults with 5q spinal muscular atrophy: a non-interventional, multicentre, observational cohort study. Lancet Neurol. 2020;19(4):317-325. PubMed

18.Maggi L, Bello L, Bonanno S, et al. Nusinersen safety and effects on motor function in adult spinal muscular atrophy type 2 and 3. J Neurol Neurosurg Psychiatry. 2020;91(11):1166-1174. PubMed

19.Pera MC, Coratti G, Bovis F, et al. Nusinersen in pediatric and adult patients with type III spinal muscular atrophy. Ann Clin Translat Neurol. 2021;8(8):1622-1634. PubMed

20.Coratti G, Cutrona C, Pera MC, et al. Motor function in type 2 and 3 SMA patients treated with Nusinersen: a critical review and meta-analysis. Orphanet J Rare Dis. 2021;16(1):430. PubMed

21.Clinical Study Report: ISIS 396443-CS11. An open-label extension study for patients with spinal muscular atrophy who previously participated in investigational studies of ISIS 396443 [internal sponsor’s report]. Cambridge (MA): Biogen MA Inc.; 2021 Jan 4.

22.Konersman CG, Ewing E, Yaszay B, Naheedy J, Murphy S, Skalsky A. Nusinersen treatment of older children and adults with spinal muscular atrophy. Neuromuscul Disord. 2021;31(3):183-193. PubMed

23.Belter L, Jarecki J, Reyna SP, et al. The Cure SMA membership surveys: highlights of key demographic and clinical characteristics of individuals with spinal muscular atrophy. J Neuromuscul Dis. 2021;8(1):109-123. PubMed

24.Muscular Dystrophy Canada. MDC Statement. 2020; https://muscle.ca/2020/12/04/mdc-statement/. Accessed 2022 Apr 4.

25.Swoboda KJ, Prior TW, Scott CB, et al. Natural history of denervation in SMA: relation to age, SMN2 copy number, and function. Ann Neurol. 2005;57(5):704-712. PubMed

26.Prodcut information: Spinraza (nusinersen): solution for intrathecal injection 2.4 mg/mL nusinersen as nusinersen sodium. (European public assessment report). London (GB): European Medicines Agency; 2022: https://www.ema.europa.eu/en/documents/product-information/spinraza-epar-product-information_en.pdf. Accessed 2022 Apr 4.

27.EVRYSDI: risdiplam powder for oral solution 60 mg/bottle (0.75 mg/mL after reconstitution), oral or enteral [product monograph]. Mississauga (ON): Hoffmann-La Roche Limited; 2021 Apr 14.

28.McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40-46. PubMed

29.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2022 Apr 8.

30.Brakemeier S, Stolte B, Thimm A, et al. Assessment of bulbar function in adult patients with 5q-SMA type 2 and 3 under treatment with nusinersen. Brain sci. 2021;11(9):20. PubMed

31.Duong T, Wolford C, McDermott MP, et al. Nusinersen treatment in adults with spinal muscular atrophy. Neurol. 2021;11(3):e317-e327. PubMed

32.Elsheikh B, Severyn S, Zhao S, et al. Safety, tolerability, and effect of nusinersen in non-ambulatory adults with spinal muscular atrophy. Front Neurol. 2021;12:650532. PubMed

33.Mix L, Winter B, Wurster CD, et al. Quality of life in SMA patients under treatment with nusinersen. Front Neurol. 2021;12:626787. PubMed

34.Kizina K, Stolte B, Totzeck A, et al. Fatigue in adults with spinal muscular atrophy under treatment with nusinersen. Sci Rep. 2020;10(1):11069. PubMed

35.Mattsson M, Moller B, Lundberg I, Gard G, Bostrom C. Reliability and validity of the Fatigue Severity Scale in Swedish for patients with systemic lupus erythematosus. Scand J Rheumatol. 2008;37(4):269-277. PubMed

36.Moshe-Lilie O, Visser A, Chahin N, Ragole T, Dimitrova D, Karam C. Nusinersen in adult patients with spinal muscular atrophy: Observations from a single center. Neurology. 2020;95(4):e413-e416. PubMed

37.Yeo CJJ, Simeone SD, Townsend EL, Zhang RZ, Swoboda KJ. Prospective cohort study of nusinersen treatment in adults with spinal muscular atrophy. J Neuromuscul Dis. 2020;7(3):257-268. PubMed

38.Walter MC, Wenninger S, Thiele S, et al. Safety and treatment effects of nusinersen in longstanding adult 5q-SMA Type 3 - a prospective observational study. J Neuromuscul Dis. 2019;6(4):453-465. PubMed

39.Chan WLS, Pin TW. Practice effect and cueing of 2-minute walk test, 6-minute walk test and 10-metre walk test in frail older adults with and without dementia - Recommendations to walk tests protocols. Exp Gerontol. 2019;124:110648. PubMed

40.Morales Mestre N, Audag N, Caty G, Reychler G. Learning and encouragement effects on six-minute walking test in children. J Pediatr. 2018;198:98-103. PubMed

41.Spencer L, Zafiropoulos B, Denniss W, Fowler D, Alison J, Celermajer D. Is there a learning effect when the 6-minute walk test is repeated in people with suspected pulmonary hypertension? Chron Respir Dis. 2018;15(4):339-346. PubMed

42.Wu G, Sanderson B, Bittner V. The 6-minute walk test: how important is the learning effect? Am Heart J. 2003;146(1):129-133. PubMed

43.Mercuri E, Darras BT, Chiriboga CA, et al. Nusinersen versus sham control in later-onset spinal muscular atrophy. N Engl J Med. 2018;378(7):625-635. PubMed

44.Mercuri E, Finkel R, Montes J, et al. Patterns of disease progression in type 2 and 3 SMA: Implications for clinical trials. Neuromuscul Disord. 2016;26(2):126-131. PubMed

45.McGraw S, Qian Y, Henne J, Jarecki J, Hobby K, Yeh WS. A qualitative study of perceptions of meaningful change in spinal muscular atrophy. BMC Neurol. 2017;17(1):68. PubMed

46.Pera MC, Coratti G, Forcina N, et al. Content validity and clinical meaningfulness of the HFMSE in spinal muscular atrophy. BMC Neurol. 2017;17(1):39. PubMed

47.Fujak A, Raab W, Schuh A, Richter S, Forst R, Forst J. Natural course of scoliosis in proximal spinal muscular atrophy type II and IIIa: descriptive clinical study with retrospective data collection of 126 patients. BMC Musculoskelet Disord. 2013;14:283-283. PubMed

48.Wijngaarde CA, Brink RC, de Kort FAS, et al. Natural course of scoliosis and lifetime risk of scoliosis surgery in spinal muscular atrophy. Neurology. 2019;93(2):e149-e158. PubMed

49.O’Hagen JM, Glanzman AM, McDermott MP, et al. An expanded version of the Hammersmith Functional Motor Scale for SMA II and III patients. Neuromuscul Disord. 2007;17(9-10):693-697. PubMed

50.Pera MC, Coratti G, Forcina N, et al. Content validity and clinical meaningfulness of the HFMSE in spinal muscular atrophy. BMC Neurol. 2017;17(1):39, 2017.

51.Glanzman AM, O’Hagen JM, McDermott MP, et al. Validation of the expanded Hammersmith Functional Motor Scale in spinal muscular atrophy type II and III. J Child Neurol. 2011;26(12):1499-1507. PubMed