Evidence in focus: Nusinersen use in spinal muscular atrophy

Level of evidence

Level of evidence statements are provided in table 2. The first study was a Class IV, phase 1, open-label study of nusinersen administration to medically stable patients aged 2–14 years with types 2 and 3 SMA (NCT01494701; NCT01780246).⁶ Four ascending single-dose levels (1, 3, 6, and 9 mg) were examined in 6–10 participant cohorts. The primary outcomes were safety and tolerability. The exploratory clinical assessment was the Hammersmith Functional Motor Scale Expanded (HFMSE), a 33-item scale that grades SMA-specific motor function (e.g., related to sitting, crawling, squatting, climbing stairs) on a 3-point (0–2) scale.⁷ In the 1-, 3-, and 6-mg groups, there were no significant HFMSE score changes from baseline to day 29, day 85, or 9–14 months (mean changes in the 1-mg group: +1.0 points at day 29, −1.7 points at 9–14 months; the 3-mg group: +1.0 points at day 29, +0.5 points at 9–14 months; and the 6-mg group: +0.7 points at day 85, +2.5 at 9–14 months). The 9-mg group demonstrated improved HFMSE scores from baseline at day 85 (mean change +3.1 points, p = 0.016). After 9–14 months, the mean increase was 5.8 points (32.8%, p = 0.008).⁶

A phase 2, open-label study investigated nusinersen doses of 6 mg or 12 mg dose equivalents in infants with SMA symptom onset between 3 weeks and 6 months of age, a SMN1 homozygous gene deletion or mutation, and either 2 or 3 SMN2 copies (NCT01839656) (table 2).⁸ This study was rated Class III based on the use of a well-defined historical control cohort. Outcomes included event-free survival and change from baseline according to results from 2 motor function assessments: the motor milestones portion of the Hammersmith Infant Neurological Exam–Part 2 (HINE-2) and the Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP-INTEND) motor function test. The HINE-2 motor milestones portion rates 8 items (from head control to walking) on a 5-point (0–4) scale in children younger than 2 years.⁹ CHOP-INTEND includes 16 items relating to motor skills (capturing neck, trunk, and proximal and distal limb strength/function) rated on a 5-point (0–4) scale.¹⁰ Participants receiving 6 mg equivalent loading doses on days 1, 15, and 85 were switched to 12-mg doses starting on day 253. The interim efficacy analysis included all participants who completed the loading doses and day 92 assessment; follow-up duration ranged from 2 to 32 months. Change in HINE-2 from baseline to last visit was significant for the 12-mg group and the 6- and 12-mg cohorts combined (p < 0.001, p = 0.0002, respectively; n = 19). CHOP-INTEND scores increased from baseline to last visit, with a mean increase of 11.5 points in the 6- and 12-mg cohorts combined (p = 0.008, n = 18) and 15.2 points in the 12-mg group (p = 0.0013). This contrasts with the natural history controls, who had a mean decline of 1.27 points annually (95% confidence interval [CI] 0.21–2.33). Study participants with 2 copies of SMN2 (n = 17) had a higher probability of ventilation-free survival compared with natural history controls who had 2 copies of SMN2 (n = 23) (log-rank test, p = 0.0014). Only 7 of the 20 total participants reached the endpoint of death or permanent ventilation prior to the interim analysis.⁸ Final study results are not available.

A Class I, phase 3, randomized, double-blind, sham-controlled study investigated the use of nusinersen in symptomatic patients with infantile-onset SMA (NCT02193074, ENDEAR).¹¹ The study was terminated early after a positive interim analysis. Eighty-one infants were randomized to nusinersen treatment (12-mg dose equivalent) and 41 to the sham control. The primary endpoints were a motor milestone response (defined as improvement in at least 1 category or improvement in more categories than worsening on the HINE-2) and event-free survival (time to death or use of permanent assisted ventilation). In both the interim and final analyses (performed using last visit data in patients enrolled for at least 6 months), more infants in the nusinersen group had a motor milestone response (interim: 41% vs 0%, p < 0.001; final: 51% vs 0%, significance not tested). Event-free survival was higher in the nusinersen group than the control group by the final analysis cutoff date (61% vs 32%, p = 0.005; hazard ratio 0.53, 95% CI 0.32–0.89). Several secondary endpoints also demonstrated treatment benefits.¹¹

A Class I, phase 3, randomized, double-blind, sham-controlled study investigated the use of nusinersen in children with SMA who had symptom onset after 6 months of age (NCT02292537, CHERISH).¹² Eighty-four patients were randomized to nusinersen (12 mg) and 42 patients received the sham control. The study was terminated after a positive interim analysis. The primary endpoint was the least-squares mean change in the total HFMSE score from baseline to month 15. In both the interim and final analyses, change in HFMSE scores at 15 months were significantly greater in the nusinersen group vs the control group (interim analysis: 4.0 vs −1.9, mean difference 5.9, 95% CI 3.7–8.1; final analysis: 3.9 vs −1.0, mean difference 4.9, 95% CI 3.1–6.7). In addition, more children receiving nusinersen than the sham procedure had a 3-point or greater increase in their HFMSE score at 15 months (57% vs 26%, p < 0.001).¹²

A Class IV, open-label, phase 1/2 study in children aged 2–15 years with later-onset SMA assigned 8–9 patients to each of 4 cohorts receiving 3-, 6-, 9-, or 12-mg doses of nusinersen (NCT01703988) (table 2). Safety and tolerability results are available only on clinicaltrials.gov.

Additional studies are underway (table 3). As demonstrated above, the regimen for loading and maintenance nusinersen dosing has evolved from varied dose ranges in phase 1 and 2 studies to focusing on 12-mg equivalent dosing in phase 3 studies.

Safety

Most information regarding possible adverse effects (AEs) is derived from clinical trial data presented to the FDA (<200 patients).¹³ No serious AEs were classified as definitely treatment-related. Several AEs were more common in nusinersen-treated patients, including proteinuria, decreased platelet counts, atelectasis, lower and upper respiratory tract infections, and constipation. The FDA recommended that clinicians regularly monitor patients receiving nusinersen for coagulopathy, thrombocytopenia, and proteinuria. The incidence of clinically significant thrombocytopenia, platelet dysfunction, or nephrotoxicity across trials with similarly modified antisense oligonucleotides was very small,^11,14,15 but one patient treated with an antisense oligonucleotide (inotersen) in a study of individuals with familial amyloid polyneuropathy died due to thrombocytopenia-related intracranial hemorrhage.¹⁶ Transient AEs related to intrathecal administration, including headache, vomiting, back pain, and post-LP syndrome, were reported in 40%–50% of patients. One rare but serious AE, communicating hydrocephalus, has been observed outside of clinical trials in five treated patients as of July 2018.¹⁷

Populations for use

The phase 3 nusinersen trials (NCT02193074/ENDEAR,¹¹ NCT02292537/CHERISH¹²) enrolled children with SMA subtypes I and II who were in the early and middle symptomatic phases without significant scoliosis or contractures that would limit the potential for measured functional improvement on selected outcome scales (table 2). In these populations, nusinersen appears to be beneficial and to pose little risk, resulting in proposals to incorporate nusinersen into standards of care.²² Less is known about how nusinersen affects a wider range of patients, such as those with milder SMA forms or those who are already severely disabled with muscle atrophy and fixed contractures. It is believed that nusinersen may be most effective when started as soon as possible after diagnosis (genetic or clinical), but NURTURE results and future research are needed. Preliminary NURTURE findings suggest that nusinersen use in presymptomatic infants with 2 or 3 copies of SMN2 may allow for the greatest improvements on the HINE-2 motor milestones,²³ but results are open-label and preliminary. All nusinersen trials are of relatively short duration; long-term implications of treatment (safety, efficacy, and long-term prognosis in the context of treatment) remain to be established. Older patients with later-onset type III and type IV SMA may prove responsive to nusinersen as they carry more SMN2 copies, but the cost–benefit ratio is different in these populations. Monitoring through nusinersen expanded access programs will inform some of these ongoing questions (e.g., results of real-world use, open-label efficacy in populations such as adults, and long-term effects/risks) but will reflect Class IV data.

Patients with more advanced disability, severe muscle atrophy, joint contractures, and skeletal deformities likely have less potential for functional improvement with nusinersen. These patients may still desire treatment if there is reasonable hope for stabilization and maintenance of an ability that is crucial to independence and quality of life, such as controlling the joystick of a power wheelchair or making communicative gestures. Such patients also face the greatest practical challenges to LP because of severe neuromuscular scoliosis and complex hardware from prior spinal fusion and instrumentation. Joint contractures make LP positioning more challenging and put patients at greater risk of injury. Advanced respiratory insufficiency or ventilator dependence may make interventionists unwilling to perform procedures with less than full anesthesia team support and general anesthesia use. In some children, use of spinal ultrasonography, fluoroscopy, conical CT guidance, and a transforaminal approach increases rates of successful intrathecal access.^24,25 In rare cases, laminectomies and intraspinal catheter and subcutaneous reservoir placements are performed by orthopedic surgeons to facilitate treatment. In patients in whom lumbar access is impossible, nusinersen delivery has been accomplished via fluoroscopic-guided lateral cervical puncture at the C1–2 interspace. These alternative administration routes have not been evaluated for drug delivery effects and are associated with additional procedural risks.

Patient preferences

While many patients and families are pursuing nusinersen treatment outside the clinical trial populations, some patients and families decline nusinersen even when insurance coverage is likely. Reasons to defer treatment include lack of sufficient information with which to weigh treatment risks and benefits, caution about using a medication with a novel mechanism of action, lack of long-term safety data, risks related to repeated LPs, desire to participate in other clinical trials, and cumulative radiation and anesthesia exposure risks. Families also may be discouraged by financial and logistical burdens directly or indirectly related to treatment, such as taking time off from work or having to travel to distant administration sites. Weight placed on these considerations varies according to phenotypic severity and individual circumstances and values.

Infrastructure for use

Nusinersen administration requires a collaborative approach among neurologists, pharmacists, anesthesiologists, interventional radiologists, therapists, patients, and families. Prescribing neurologists must be comfortable with patient selection and counseling. In addition to discussing potential benefits and risks, physicians should explain that nusinersen is not a cure. Successful treatment will improve disease trajectory, but patients and families will still face life with a chronic disease. Repeated (e.g., every 6 months) time-consuming evaluations of motor function by physical therapists familiar with the scales used in SMA are required to monitor therapy response and maintain insurance approval for continued therapy. These evaluations require expertise and equipment that are not universally available and often require visits to specialized multidisciplinary clinics. Individuals receiving nusinersen also require laboratory access for AE monitoring.

Coverage considerations

In the United States, nusinersen's list price is $125,000 per injection, equaling approximately $750,000 for drug administration during the first year and $375,000 annually in subsequent years, not including physician, therapist, or facility fees associated with drug administration²⁶ or indirect costs for patients and families (e.g., travel and lost wages). The high per-treatment and lifetime costs mean that US patients face significant hurdles in getting treatment approved by insurers, although processes are becoming more streamlined. Some patients, particularly those with SMA forms not included in the phase 3 trials, face persistent difficulties if not denials from insurers. After initial approval, some insurers ask physicians to submit the results of therapy assessments/motor function tests used in the phase 3 trials to justify ongoing treatment.^22,27 Insurers may also require evidence that patients' conditions have not deteriorated to a point of requiring permanent ventilator support.²⁷ Even when coverage is approved, patients and families face billing challenges related to in-network and out-of-network costs, as nusinersen administration and related assessments are often only available at select centers.

Several US insurers have declined requests to cover nusinersen in patients with later-onset type III and IV SMA, classifying the treatment as experimental. There are concerns that high treatment costs are not justified by the clinical benefit in this population. Variations in how physicians and insurers are weighing the uncertainties about the quality of life and value to society provided by nusinersen use in different patient groups has led to discrepancies in treatment availability in different clinics and states.²⁸ Some patients choose to travel out of state or to change insurance providers to obtain treatment.

In December 2017, the Canadian Agency for Drugs and Technologies in Health (CADTH) recommended nusinersen reimbursement for the treatment of 5q SMA in infants symptomatic before 7 months of age, based on ENDEAR evidence.^11,29 Similar to US insurers, treatment discontinuation is recommended if there is no demonstrated improvement or maintenance of function as assessed by the HINE-2 or if permanent invasive ventilation is required. The coverage decision is conditional on SMA specialist care, a substantial reduction in treatment price, and collection of “real-world evidence” on nusinersen use.²⁹

Cost-effectiveness

No US nusinersen cost-effectiveness analyses are published. A technical brief for the Institute for Clinical and Economic Review Orphan Drug Assessment & Pricing Summit noted that high nusinersen costs “are likely to result in cost-effectiveness estimates that exceed commonly cited thresholds in the US (i.e., $50,000–$150,000 per quality-adjusted life-year [QALY] gained), even if substantial cost offsets from reduced supportive care needs are realized.”²⁶ A CADTH reanalysis of the manufacturer cost–utility model found that nusinersen was unlikely to be cost-effective in Canada at the submitted price, with costs per QALY of $9.2 million for SMA type I, $24.4 million for SMA type II, and $7.4 million per QALY for SMA type III (with different levels of certainty). CADTH noted that even in the scenario of a 95% price reduction, incremental cost–utility ratios would still exceed $400,000 per QALY.²⁹

Using traditional cost-effectiveness thresholds for treatments of rare conditions is controversial, with challenges including small populations on which to base judgments, novel therapeutic mechanisms of action, clinical trial feasibility, quality of life assessments in diseases affecting infants and young children, incorporation of the effect of quality of life factors on families and caregivers, assessment of both quality and extension of life, and the distribution of high costs across a small population.²⁶ At societal and clinician levels, there remain unanswered questions about the balance of improving the lives of individual patients vs the societal responsibility to control health care costs and maximize benefits for all.

Changing approach

The discovery and approval of nusinersen is fundamentally altering the landscape of SMA diagnosis and treatment. With emerging treatments that prevent disease manifestations and improve function, systems for early diagnosis are needed to maximize potential benefits. The Cure SMA advocacy group lobbied the US Congress to include SMA in the Recommended Uniformed Screening Panel for newborn screening, with several states initiating pilot programs in 2018. If NURTURE shows nusinersen efficacy in preventing or lessening clinical manifestations in presymptomatic infants, early diagnosis and treatment will be further emphasized. Caution is needed in interpreting genetic testing results in presymptomatic individuals, however. SMN2 copy number is correlated with SMA subtype but is not completely predictive,³⁰ partly related to phenotypic modifiers.³¹

Nusinersen will redefine the natural history of SMA. No evidence is currently available to counsel patients and families about what to expect for long-term prognosis and function after nusinersen treatment, a critical gap that will be addressed in part by ongoing open-label follow-up studies. Historical treatment protocols developed for guiding optimal SMA care will need revision as approaches and natural history are redefined.

Other novel SMA treatments also are under investigation. Targets for therapeutic intervention include increasing SMN protein production either through alternative SMN2 splicing (nusinersen, RG7916, branaplam) or gene replacement of SMN1 (AVXS-101), preventing motor neuron death by maintaining neuronal mitochondrial integrity (olesoxime), or increasing muscle endurance through a fast skeletal muscle troponin activator that amplifies muscle response (CK-2127107). These investigations need to explore the therapeutic effect beyond the motor neuron alone, as research suggests that low SMN levels affect muscle, afferent sensory nerves, brain, heart, blood vessels, liver, pancreas, bone, lungs, and intestines, with implications for treatment timing and therapeutic targets.^2,32