A multidisciplinary approach to dosing nusinersen for spinal muscular atrophy

Abstract

Background

In December 2016, nusinersen gained FDA approval as the first pharmacologic treatment for spinal muscular atrophy (SMA), a disorder of motor neurons and the leading genetic cause of infant mortality. Nusinersen's intrathecal delivery requirement, strict dosage protocol, and accelerated FDA approval presented a challenge to health care centers hoping to implement treatment of patients with SMA. Scheduling logistics, combined with the specific ventilatory, anesthetic, and spinal access needs of this patient population, requires extensive coordination of care. This complexity, in addition to the high cost of treatment, may lead to overburdening of an institution's dosing resources, causing delays in treatment initiation and limiting patients' access to therapy and may result in barriers to coverage.

Methods

We initiated a comprehensive stepwise protocol to maximize patient inclusion, as well as safety and efficiency outcome measures. This retrospective cohort study reviews the dosing process.

Results

As a result of immense collaborative efforts involving care coordination of patients and families, in addition to health providers in the divisions of neurology, anesthesiology, pulmonology, orthopedics, interventional radiology, physical therapy, and neurosurgery, we have successfully dosed 62 SMA patients. Throughout this process, we have improved anesthetic techniques, as well as minimized procedural complications and missed scheduled doses.

Conclusion

We present here recommendations for safe and effective implementation of nusinersen utilizing a multidisciplinary approach, based on our 1 and a half year experience at a tertiary care children's hospital.

Spinal muscular atrophy (SMA) is a neurodegenerative disorder caused by biallelic mutations of the SMN1 gene.¹ There are 3 main subsets in childhood.^1,2 Type I presents prior to 6 months of life and is notable for inability to sit and early need for airway support.^1–3 Type II has onset prior to 18 months of age and is associated with inability to stand.^1–3 Type III presents after 18 months, with eventual loss of ambulation.^1–3 Disease severity correlates with copy number of the homologous SMN2 gene, which produces some SMN proteins.^4,5 Nusinersen acts by altering splicing of SMN2 pre-mRNA to produce a fully functional SMN protein.^6–8 Clinical studies of nusinersen led to statistically significant changes in motor function in patients with SMA, leading to FDA approval in December 2016.^9–13

Nusinersen's intrathecal administration, scheduling requirements, and cost challenge health care centers' attempts to implement treatment.^14,15 Dosing regimen involves 4 doses in the first 2 months (days 1, 15, 29, and 59), followed by maintenance dosing every 4 months.¹⁶

Scheduling requires coordination among multiple subspecialty services. Pediatric patients often require sedation, and for those with baseline respiratory support needs, anesthesiology is optimally present.¹⁵ Patients with scoliosis may require interventional radiology (IR) or surgery for access. Each dose of nusinersen costs $125,000, with projected annual costs of $750,000 for the first year and $375,000 for each successive year for the medication alone.¹⁶ Insurance authorization must be obtained in advance because of policy limitations, despite broad FDA approval. Insurance carriers may require proof of response to therapy to ensure ongoing coverage.

Methods

Prior to initiation of care following nusinersen's approval, a comprehensive electronic medical record review was conducted to identify patients with a confirmed diagnosis of SMA at our center. Initially, 152 patients were identified and contacted, of which 73 planned to dose at Children's Hospital of Philadelphia (CHOP). A subset of patients identified in chart review did not plan to dose at CHOP for the following reasons: patient deceased, patient and/or family was not interested in the medicine, or patient and/or family chose to be dosed elsewhere.

A collaborative interdepartmental team composed of health care providers from neurology, orthopedics, pulmonary, palliative care, ethics, and teams familiar with triaging based on limited resources such as transplant and oncology, as well as patient advocacy and a parent of a child with SMA, devised a comprehensive stepwise protocol to maximize patient inclusion and safety. Patient-directed characteristics taken into account and developed from natural history data in SMA cohorts included risk of developing life-threatening bulbar symptoms either imminently or over a period of months. This included patients with type 1 SMA, both with and without tube feeding or ventilatory support but with the likelihood of continuing to worsen rapidly over months, as well as more severe type 2 SMA patients. Consideration was also given to patients deemed likely to have substantial clinical deterioration over a period of months, including new diagnosis type 1, young patients with type 2 SMA requiring nocturnal ventilatory support, and patients with type 3 SMA about to become nonambulatory. Other factors accounted for in the triage system included level of care needed for the dosing procedure, siblings of different functional status in the same family, and whether care was established at CHOP prior to nusinersen's approval or if a new diagnosis or referral of care occurred in CHOP's catchment region.

The majority of patients were scheduled in a multidisciplinary nusinersen screening clinic that included evaluations by a neurologist, pulmonologist, anesthesiologist, and physical therapist. This clinic was specifically added to help expedite the evaluations, which would have otherwise been limited by scheduling constraints in our regular neuromuscular clinic. Child life was also present, to initiate discussions regarding the injection process as well as blood draws. Patients with type 1 SMA who had minimal motor function at baseline were screened by their primary neuromuscular provider as the risk/benefit conversation in this population was more nuanced given available research data at the time. Clinics occurred weekly for approximately 4 months. Screening clinics allowed for patient-specific determinations of level of care. This included who would administer nusinersen (IR or neurology), as well as the level of anesthesia support required for a safe sedation based on a patient's respiratory baseline (potential sites included operating room [OR] or sedation suite). The involvement of multiple specialties prevented the need for multiple separate trips for patients and allowed for direct conversations on patient management among the multiple disciplines.

The screening pulmonary assessment served 2 primary purposes: to assess the patient's respiratory status and to ensure that the patient had adequate respiratory support to safely tolerate procedural sedation or anesthesia. This plan included airway clearance and when appropriate, noninvasive ventilation. For the patients who had a successful plan for ventilatory support (either nasal or tracheal), the plan was confirmed and made clear to the anesthesia team for procedural sedation/anesthesia and to respiratory therapy for postprocedural support. For patients not on ventilatory support, if the patient was assessed as likely to need ventilatory support within the subsequent months, he or she was referred for a sleep study to formally assess for nocturnal hypoventilation and the need for ventilatory support. If the patient was assessed as not likely to need ventilatory support, he or she was advanced to the procedure without ventilatory support. Patients with tracheostomy or more than nocturnal noninvasive ventilation were given their home level of support during anesthesia; for those with only nocturnal ventilation, a respiratory plan was decided on a case-by-case basis with anesthesia and pulmonary.

While it was felt that the likelihood of needing airway clearance for removal of lower airway secretions was low, a plan was assessed for each patient to use a mechanical in-exsufflator (Cough Assist, Hill-Rom) to help with lung re-recruitment as part of the postprocedural recovery process. If the patient was naïve to the Cough Assist, a trial was arranged beforehand to assess the proper effective and well-tolerated treatment settings.

During the screening visit, in addition to initial evaluations, patients and their families gave consent for the lumbar puncture procedure, enrolled in the Biogen Start program, and had baseline laboratories drawn prior to initiating insurance approval. Preprocedure laboratories drawn included platelets, coagulation studies, and urine protein per pharmaceutical dosing instructions. Six patients, whose clinical data are not otherwise included here, were seen in screening clinic and were not subsequently treated during the study time for the following reasons: insurance requiring treatment in-state, patient had a fused spine and did not want laminectomy or transforaminal dosing, patient moved, or patient wanted a dosing technique that was not currently being offered at our institution. Given the recurrent nature of the injections, the team was careful to initiate treatments in a manner that would not overwhelm the system once a large number of patients had begun therapy. Discrete event computer simulation (DES) was run to optimize treatment start dates, minimize delayed doses because of limited resources like weekend coverage and holidays, and keep patient volume manageable on any given day. Estimated capacities in each of the dosing contexts were modeled in a 4-step process: (1) decomposition of the system into component elements (resources, locations, path networks, etc.), (2) development of flowcharts that describe the flow of simulated patients through the system, (3) coding of the model system into a DES computation engine (in this case, Med Model 2014; Promodel Corp, Allentown, PA), and finally (4) validation of the simulation. A greedy algorithm modeling from day 0 to 1,095 days (3 years) determined the dosing day for each patient that maximized available resources and updated this calculation with each new patient once their dosing location was determined and approval was in place.

Possible locations for nusinersen administration included the OR for patients at the highest risk, IR for patients with anticipated difficulty of the lumbar puncture procedure because of scoliosis or prior spinal surgery, and procedural rooms within the sedation suite with the following options: planned need for sedation, planned no need for sedation, or option for sedation if needed. For follow-up injections, laboratories were checked with intravenous catheter (IV) placement prior to procedure. Platelet level was confirmed prior to injection. Injections were delivered in accordance with packaging insert instructions. Anesthetics varied from anesthesia standby to general anesthesia. For patients not requiring sedation or prior to sedation (IV insertion, etc.), Child Life services were available. All patients, regardless of anesthetic need, were required to lay flat for 1 hour following injection to minimize headache side effect.

Standard protocol approvals, registrations, and patient consents

Approval from the CHOP Institutional Review Board was obtained to conduct this retrospective study.

Data availability

De-identified participant data will be shared at the request of other investigators.

Results

Sixty-two patients with genetically confirmed diagnosis of SMA were treated with nusinersen at our children's hospital from January 2017 to June 2018. There was no significant difference in sex among those treated. Type 3 SMA patients made up 40% of those treated, followed by type 2 SMA patients (34%) and then type 1 SMA patients (26%). The youngest patient treated was 2 months of age, and 94% of patients treated were 18 years of age or younger at time of first dose. Nearly all patients (97%) had received all 4 loading doses by the cutoff date for this retrospective chart review of June 1, 2018. Two patients received all dosing treatments while inpatient in the pediatric intensive care unit, and 2 patients received loading doses at an outside institution and then transferred care. These 4 patients were included in the demographics data but were not included in statistics related to the patient dosing experience (table 1).

Table 1.

Patient demographics

Screening clinic visits were completed over the course of 5 months, with priority given in accordance with the above recommendations. In evaluating wait time to first dose, most patients (62%) were able to be treated within 2 months from the date of their nusinersen screening clinic visit. One-quarter of the patients had a wait of greater than 3 months from screening visit to first dose, in most cases because of delays in insurance approval. Some also had more prolonged waits so that the appropriate sleep study could be done prior to anesthesia. Wait time for laminectomy and evaluations for spinal fusion patients also led to delays in treatment initiation. Time required for insurance approval and subsequent time to first dose varied among the SMA types, with type I SMA patients having the shortest wait (median 2.5 weeks) and type 2 SMA patients having the longest wait (median 12.5 weeks). There were no clinically important dosing delays, defined as greater than 1 week between the first and second or second and third doses, or greater than 2 weeks between subsequent doses. Because of capacity constraints, and because we do not administer doses on Saturday or Sunday, it was not possible to give all doses on the exact day recommended by the manufacturer. Some doses were adjusted slightly because of colds or other respiratory exacerbations.

Location and anesthesia requirement varied among patients and changed over time (Figure).

Figure. Flowchart for dosing location and anesthesia requirement.

IR = interventional radiology; OR = operating room; PACU = post-anesthesia care unit.

Two patients were changed to IR because of difficulty performing lumbar puncture. About 20% of patients received their initial dose in the OR. All of these patients were able to receive subsequent treatments outside of the OR, with most able to transition to the sedation suite (table 2). Thirteen total patients received an inhalational anesthetic with sevoflurane for at least one of their procedures. The vast majority of these patients (75%) were managed with a laryngeal mask airway, with the remainder through a tracheostomy. Most patients received an IV anesthetic consisting of midazolam, propofol, or dexmedetomidine with a natural airway. Thirteen patients tolerated all procedures with only local anesthesia and midazolam, and 10 patients who initially needed anesthesia needed only midazolam and anesthesia standby for subsequent injections. In total, of the 49 patients initially requiring some level of sedation, nearly 40% were able to reduce the level of sedation needed on subsequent doses, and 27% required no sedation on subsequent doses (table 3).

Table 2.

Patient dosing experience

Table 3.

Factors related to change in anesthesia requirement

After approximately 6 months of experience with dosing nusinersen across a wide range of patient ages and anesthesia needs, we were able to reduce recovery time from 4 to 2 hours following initial dose, and from 2 to 1 hour after subsequent doses. No doses were delayed because of thrombocytopenia. One patient had a short delay because of abnormal coagulation profile and was evaluated prior to injection by hematology. There were no urinary protein abnormalities beyond baseline. Complications included pain at injection site (n = 10), headache (n = 9), and vomiting (n = 2). No patients received a blood patch or were readmitted for post-lumbar puncture headache during the study time frame. However, several patients were managed at home with fluids, oral caffeine, and over-the-counter analgesics. Laminectomy was performed on 3 patients: this was planned in 2 cases, and because of failed lumbar puncture attempt in the third, 1 patient required a transforaminal approach.

Discussion

Reviewing our first 62 nusinersen cases has provided key insights into the patient enrollment and dosing process. There were a number of challenges to the enrollment process, beginning with limitations in the number of patients able to be seen each week in screening clinic. The enrollment process was further complicated by the need for IR vs neurology administration and the need for a backup system in place to prevent missed doses. We used computer simulation to devise a dosing schedule for a group of patients prior to initiation of dosing to address these issues. Perhaps the greatest unforeseen challenge, however, was the need for insurance prior authorization and lack of consensus for patient eligibility from insurance companies, despite broad FDA approval in the product label. The multiple insurances and varying ways of addressing the new medication staggered the patient enrollment more so than initial visits in screening clinic. We made every effort to maximize the efficiency of access to treatment, while prioritizing safety of patients.

Once enrolled, patients tolerated anesthesia better than anticipated. There was considerable anxiety surrounding the first dose, but we were able to decrease sedation in patients as young as age 6 with the use of Child Life services. SMA patients with restrictive lung disease are particularly vulnerable to postanesthesia respiratory complications. However, all of our patients tolerated sedation and/or anesthesia without incident and were able to be discharged home after a brief period of observation. Transition from the OR to the sedation suite allowed for more flexible scheduling and decreased costs. Lastly, for patients with spinal fusion requiring laminectomy with anesthesia, follow-up injections were straightforward and did not require additional anesthesia. Transforaminal dosing was also possible in patients with complex spinal anatomy. We did not use an Ommaya reservoir or any other delivery device for any patients.

The preprocedural screening clinic visit was critical to evaluate individual patient needs and to ensure patient safety by crafting an appropriate plan for ventilatory support and airway clearance to optimize the recovery process and minimize the length of stay. This allowed for appropriate use of hospital resources. It also served as an opportunity to reach out to “lost patients” and reconnect. Its multidisciplinary setting allowed for direct conversations between the appropriate groups on patient care. The model system was continually tweaked during the first year of treatment. In general, our policies were more conservative initially, but once patients were tolerating the procedure, observation times were decreased and anesthesia was provided more frequently in the sedation suite, rather than the OR. As additional medications are developed with intense administration requirements in vulnerable populations, similar explorations of the appropriate level of care are likely to occur. Insurance led to our most substantial delays in treatment. The presence of an insurance policy could have been used to help put those that could get approval first into clinic; however, this was untenable as few companies had policies when treatment started because of the very recent nature of the drug approval. Helping to separate those that would need IR from those that could have neurology administration may have also led to less wait time, as they are 2 different resources. However, scheduling of the procedure was rarely a limiting factor, as insurance issues meant that approvals came through slowly.

An organized approach and team planning framework were key to the seamless execution of treatment plans and attaining optimal nusinersen dosing work-flow. Stratification of anesthesia resources of general anesthesia vs monitored anesthesia care vs standby vs nonsedated is particularly relevant to future emerging therapies in complex pediatric populations.

By employing DES, we were able to minimize dosing delays while simultaneously addressing the needs of both our patient population and our institution's capabilities. Such systems are particularly relevant given the recent addition of SMA to the Recommended Uniform Screening Panel, in response not only to currently available nusinersen, but anticipation of possible gene therapies in the not so distant future. While a situation identical to nusinersen may not arise, there are an increasing number of treatments for rare disorders that carry with them increasing cost as well as difficulty of administration, often in specifically vulnerable populations.

This is a large, diverse cohort of predominantly pediatric SMA patients treated with nusinersen. Our experience demonstrates how a comprehensive perioperative program utilizing a multidisciplinary approach can optimize the screening process and dosing experience for nusinersen therapy and allocate resources both ethically and efficiently. Although patients with SMA are particularly vulnerable to postanesthesia respiratory issues, patients tolerated their brief anesthetic well and many were able to reduce the level of sedation needed on subsequent treatments.

Appendix. Authors

All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Study funding

This study was funded in part by a grant from CureSMA.

Disclosure

C.D. Zingariello reports no disclosures. J. Brandsema received an educational grant from Biogen to support a CHOP neuromuscular fellowship in 2019–20; is a Speaker on Demand for Biogen; and serves as a site PI or subinvestigator for several industry-sponsored protocols and serves on the Advisory Boards for Biogen, AveXis, PTC Therapeutics, Marathon, Sarepta, Alexion, and Cytokinetics. E. Drum, A.A. Henderson, and S. Dubow report no disclosures. A.M. Glanzman has a licensed patent and receives royalties for the CHOP INTEND scale; is a consultant for Biogen and Synios; and performs clinical trial training and review for Biogen, Roche, Avexis, and Synios. O. Mayer serves on the Advisory Board and Speakers' Bureau for Biogen. S.W. Yum serves on the advisory board for PTC Therapeutics. E.A. Kichula serves on the advisory board for PTC Therapeutics and Biogen and is a subinvestigator for a number of clinical trials. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.