Ultrasound-guided Nusinersen Administration for Spinal Muscular Atrophy Patients With Severe Scoliosis: a Retrospective Single-arm Cohort Study

Background This observational study describes our experience delivering nusinersen through lumbar puncture with real-time ultrasound guidance in spinal muscular atrophy (SMA) patients with severe scoliosis. Intrathecal nusinersen via real-time ultrasound-guided lumbar puncture was given to 3 patients who had severe thoracic and lumbar scoliosis: a 34-year-old female with type 3 SMA, a 28-year-old male with type 2 SMA, and a 14-year-old girl with type 3 SMA. Lumbar puncture was performed without sedation under ultrasound guidance using a 22G echogenic needle in the interlaminar aspect of the L4-L5 or L5-S1 interspace and a full dose of nusinersen (12 mg/5 mL) was injected after visualizing free cerebrospinal uid (CSF) ow. Patients completed their 4 loading doses and 1 maintenance dose of nusinersen. All 15 procedures were successful and well tolerated.

clinically effective in infants with type 1 SMA and children or adults with later-onset SMA (types 2 and 3) for improving motor function [10,11]. This drug cannot pass the blood-brain barrier and, therefore, has to be repeatedly administered intrathecally. For most infants or young children with SMA, intrathecal delivery via a landmark-guided lumbar puncture technique has been shown to be safe, feasible and well tolerated [11]. However, in SMA patients with progressive scoliosis, standard lumbar puncture may be challenging [12][13][14][15]. Previous studies showed that computed tomography (CT)-or uoroscopy-guided techniques could facilitate intrathecal injection of nusinersen in SMA patients with aberrant spinal anatomies [13][14][15]. However, potential cancer risks existing from associated ionizing radiation of computed CT or uoroscopy scans -in particular for children, who are more radiosensitive than adultsshould not be ignored [16]. The real-time ultrasound-guided technique, with the advantage of avoiding radiation exposure, has been reported to be successfully used by anesthesiologists for central neuraxial block in patients with moderate to severe lumbar scoliosis [17,18].
In this report, we describe our clinical experience of using the real-time ultrasound-guided technique to deliver nusinersen repeatedly in 3 SMA type 2 and 3 patients with severe scoliosis, through 15 lumbar punctures.

Study participants
The institutional review board approved this retrospective study with a waiver for informed consent. From October 30, 2019 to July 23, 2020, patients with SMA and severe scoliosis seen by the neurology physician at our institution were referred to the anesthesiology department for intrathecal nusinersen administration because of the expected di culty in obtaining intrathecal access (Fig. 1).

Pre-puncture preparation
The purpose, methodology, and potential complications of the ultrasound-guided puncture procedure, and the bene ts and side effects of nusinersen, were discussed with the patients and their parents prior to initiation of therapy and each procedure. Informed consent was obtained. Spinal 3D-CT reconstructions were obtained to ascertain intrathecal access at various lumbar levels and to facilitate pre-procedure planning ( Fig. 1A-C).

Ultrasound-guided lumbar puncture
An additional movie le shows the whole process of puncture in more detail (see Additional le 1). All the punctures were performed in the operating room at the anesthesiology department. Upon arrival, standard monitoring and peripheral venous access were established for the patients. As it is important for patients to be able to communicate any concerns during the lumbar puncture procedure, sedation (intravenous midazolam) was used only when necessary to reduce the patient's severe anxiety prior to some of the procedures. Each patient was placed in the lateral decubitus position with the knees exed to improve lumbar space access. The patient's head and body were supported by tted padding. A pre-procedural ultrasound scan of the lumbar spine based on a scan protocol for scoliosis, rst described by Karmakar [19], was performed using a low-frequency (8-3MHz), curvilinear transducer (Sonosite X-Porte; Sonosite Inc., Bothell, WA, USA). Brie y, the scan followed 3 steps. In the rst step, the positions of the spinous processes (SPs) of the lumbar spine on the patient's back were palpated and marked using a skin marking pen (Fig. 1C). Next, a line connecting the SPs was drawn, referred to as the "SP line". The last step was to perform a paramedian sagittal scan by placing the transducer 1 to 2 cm lateral and parallel to the "SP line" with the transducer tilted slightly medially during the scan, so that the ultrasound beam was insonated in a paramedian sagittal oblique plane; this approach ensures that the ultrasound signal enters the spinal canal through the window of the interlaminar space (Figs. 2A and C). According to ultrasonography, the interlaminar spaces were visible with an acceptable acoustic window and returned good hyperechoic dura signals ( Fig. 2A and Additional le 1), then the position of the transducer was marked using a skin marking pen. After sterile preparation of the posterior back with iodophor and the ultrasound transducer draped in a sterile plastic sheath, 2 mL of 1% lidocaine was administered as a local anesthetic. A 10-cm, 22G, echogenic needle with a short-bevel tip (SonoLong Nanoline; Pajunk Inc., Geisingen, Germany) and stylet was advanced in-plane in a caudal-to-cephalad direction (Fig. 2B) toward the interlaminar space. The needle was traced on the ultrasound screen in real time (Fig. 2C). If dural puncture was not achieved after 3 needle passes (de ned as forward advancement of the needle, i.e., needle redirection without exiting the skin, including the rst pass), the operator had options to change to an alternative interlaminar space level. The exact localization of the needle tip (penetrating the dura) was con rmed on ultrasound image and by free ow of clear cerebrospinal uid (CSF) after needle stylet removal (Fig. 2C, D, and Additional le 1). After extraction of 5 mL CSF, 12 mg (12 mg/5 mL) of nusinersen was administrated through the needle. Subsequently, the needle was removed, and patients were placed in a comfortable position for 1 hour of observation.

Data collection
Patient characteristics including age, sex, genotype, subtype of SMA, personal history, clinical signs and symptoms, and results of CSF examination were obtained from the electronic medical records. Details of punctures (including number of interspace level adjustments and insertion attempts per procedure) were recorded. Adverse events including, but not limited to, pain or discomfort leading to need for sedation, puncture of visceral organs, post-lumbar puncture syndrome, respiratory disturbance, bleeding, and spinal ischemia were also obtained.

Results
Real-time ultrasound-guided intrathecal injections of nusinersen were performed in 3 SMA patients with severe scoliosis in the anesthesiology department at the Peking Union Medical College Hospital between October 2019 and July 2020. Clinical characteristics of the patients are described in Table 1. Two patients were diagnosed with type 3 SMA and 1 with type 2 SMA. All 3 patients completed the 4 loading doses and 1 maintenance dose, which accounts for a total of 15 consecutive punctures. No patients required intravenous sedation or general anesthesia. All punctures used the paramedian sagittal in-plane approach with a success rate of 100%, assuring correct delivery of the drug in all patients. Among all the 15 procedures, 14 (91.67%) were performed successfully at the interspace level of L4-L5, 1 at the level of L5-S1. In detail, for the rst puncture of Patient 1, we initially chose to inject at the level of interspace L2-L3 based on the 3D-CT image, but failed after 2 attempts because of the steep angle of the needle path, which was the only path available due to the curvature of the vertebral column. Following that, we changed to the level of L5-S1 based on the ultrasound image and succeeded after 2 attempts. The second puncture of Patient 1 and the rst puncture of Patient 2 failed at the rst attempts at level L5-S1, due to inability to withdraw CSF for uncertain reasons, even though the needle tips were seen underneath the hypoechoic dorsal dura on the ultrasound image. Finally, both punctures were successful at the second attempts at level L4-L5. For all the following procedures, level L4-L5 was selected for the rst attempt and all succeeded.
All procedures were technically successful and well tolerated. Minor complications included Patient 3 reporting positional headache and nausea after the second puncture, which was resolved by bed rest. No severe complications were recorded. CSF tests and laboratory tests including platelet counts, coagulation pro le, and renal and liver function tests remained unremarkable in all patients.

Discussion
In this study, we present the results of real-time ultrasound-guided lumbar intrathecal administration of nusinersen in SMA patients with severe scoliosis. The success rate was 100% with no major complications.

Technique
Lumbar puncture is a routine practice performed by anesthesiologists. The application of ultrasound in assisting lumbar puncture can be traced back to the 1970s [20], and was rst used in patients with abnormal spinal anatomy in 1999 [21]. With improvements in ultrasonic imaging quality and greater understanding of the sonoanatomy of the spine among investigators, techniques of real-time ultrasoundguided spine puncture were investigated [22][23][24][25] and reported to be successfully used in patients with challenging spinal anatomy in 2010 [26]. In the current study, we chose in all procedures to use the realtime ultrasound-guided in-plane approach, advancing the needle in the plane of the ultrasound beam, through the longitudinal paramedian sagittal oblique interlaminar view, which was rst described by Karmakar [23].
During lumbar spinal sonography, the paramedian sagittal oblique interlaminar approach possesses the advantage of providing a better sonographic window into the vertebral canal from the interlaminar space compared with the interspinous space (median) approach [27,28], and is more successful in patients with a di cult anatomy [29,30]. Performers of the technique should be familiar with the sonoanatomy of the interlaminar, the intrathecal and the surrounding bony structures. Sonographically, the interlaminar space is the gap between the adjoining lamina and is the "acoustic window" through which the neuraxial structures are visualized within the spinal canal ( Fig. 2A). The ligamentum avum appears as a hyperechoic band across adjacent lamina ( Fig. 2A). The posterior dura is the next hyperechoic structure anterior to the ligamentum avum ( Fig. 2A), and the epidural space between the ligamentum avum and the posterior dura is hypoechoic. The ligamentum avum and the posterior dura may also be seen as a single linear hyperechoic structure, which is referred to as the "posterior complex". The intrathecal space with the CSF is the anechoic space anterior to the posterior dura. The deeper linear hyperechoic structure is the "anterior complex" composed of the anterior dura, the posterior longitudinal ligament and the posterior surface of the vertebral body, which are of similar echogenicity and closely opposed to each other. To accurately obtain this paramedian sagittal oblique interlaminar window in SMA patients with complex deformity of the spine, we recommend to strictly follow the currently used 3-step scan protocol [19].
The key to the real-time ultrasound-guided in-plane technique used in this study is in obtaining and maintaining alignment of the ultrasound beam with the needle and the acoustic window into the interlaminar space during the whole procedure, which is crucial to guarantee the success of lumbar puncture [25]. Advanced skills in hand-eye coordination and needle-transducer manipulation will de nitely be required to successfully perform this technique [25]. In the current study, there was no e ux of CSF in 2 punctures, though the hypoechoic "needle tips" were seen in the intrathecal space on the ultrasound image. We speculate that imprecise alignment of the needle along the ultrasound beam might have been responsible. This means that the hypoechoic dots we saw on the ultrasound image may actually represent the transverse section of the needle body, and, instead, the real needle tip had been advanced too deep into the anterior epidural space. There were another 2 failed attempts (in Patients 1 and 2) in which the intrathecal structures were clear in the ultrasound image, but we were unable to insert the needle through the interlaminar space. This may have been due to the steep needle-puncturing angle.
A steeper needle-puncturing angle will inevitably attenuate the imaging quality of the needle and make it di cult to manipulate the needle based on the ultrasound image. Therefore, when determining the optimal puncture level during the pre-procedure scan, both the imaging quality of the given interlaminar space and the predicted needle-puncturing angle should be considered.

Safety
Here, we successfully and safely administered nusinersen intrathecally in all 15 procedures. No severe adverse events were noted, and only 1 patient reported positional headache and nausea after the second injection, which was resolved by bed rest. To improve the visualization of the needle under ultrasound we chose to use a 22G echogenic needle. The incidence of headache overall in this study was 8.3%, which appears to be consistent with the post-lumbar puncture headache rate of ~ 10% in children from previous literature [31]. In terms of safety, a prominent advantage of the current technique is its inherent radiationfree property when compared with traditionally used uoroscopy-or CT-guided technique. As therapy with nusinersen is long term, the cumulative radiation exposure from repeated puncture may be of concern, because of its deleterious effects for human health in inducing cancers (e.g. colon cancer, lung cancer and leukemia) [13,16]. Additionally, by using the real-time ultrasound-guided technique, advancement of the needle could be visualized during the whole procedure, which is particularly helpful in avoiding unnecessary vascular or visceral injuries.

Limitations
The current study has several limitations. First, this is a retrospective evaluation of a method that has been used in only a few patients. However, the objective of this study was to evaluate the feasibility of performing real-time ultrasound guidance for SMA patients with severe scoliosis, which we have successfully demonstrated. Second, our experience is mainly based on patients with severe scoliosis but not previous spinal surgery. Spinal surgery for rod placement or fusion is a common treatment for SMA patients to control their progressive scoliosis [32], which may present greater challenges for lumbar puncture since the interlaminar approach may not be feasible in these complex conditions. Ultrasound has seldom been used in this population of patients, except for one report describing ultrasound-guided cervical puncture for di cult lumbar nusinersen delivery in SMA patients with previous spinal instrumentation [33]. However,successful use of ultrasound assistance or real-time guidance to facilitate a di cult interlaminar lumbar puncture for other purposes in patients with spinal fusion instrumentation [21,26,34] has been reported previously. Similarly, the transforaminal lumbar injection technique, which has been used under CT guidance for intrathecal access in patients with spinal fusion where a interlaminar approach is often not possible, has also been reported to be successfully performed with ultrasound guidance in other circumstances [35,36]. Therefore, the value of ultrasound guidance in SMA patients with spinal fusion is worthy of attention and exploration.

Conclusion
The real-time ultrasound-guided technique can be successfully used to assist lumbar nusinersen administration in SMA patients with severe scoliosis. The advantage of the current technique is that it is free of exposure to radiation, provides real-time images, and improves patient comfort during the whole procedure. A sound knowledge of the sonoanatomy of the spinal deformity and skilled manipulation techniques are needed for practitioners to apply this technique safely and effectively.

Declarations
Ethics approval and consent to participate This study protocol was approved by the Institutional Review Boards of the Peking Union Medical College Hospital (No. S-K 1378). And all the subjects or their legal representatives gave written informed consent.

Consent for publication
Consent for publication was obtained from all involved persons (or from their legal representatives), whose images were used in the present manuscript.

Availability of data and materials
All study data, including raw and analyzed data, and materials will be available from the corresponding author on reasonable request.