*Case Report* **Improvement in Fine Manual Dexterity in Children with Spinal Muscular Atrophy Type 2 after Nusinersen Injection: A Case Series**

**Minsu Gu and Hyun-Ho Kong \***

Department of Rehabilitation Medicine, Chungbuk National University Hospital, Cheongju 28644, Korea; msgrehab@gmail.com

**\*** Correspondence: jimlight@hanmail.net

**Abstract:** Although nusinersen has been demonstrated to improve motor function in patients with spinal muscular atrophy (SMA), no studies have investigated its effect on fine manual dexterity. The present study aimed to investigate the ability of nusinersen to improve fine manual dexterity in patients with SMA type 2. A total of five patients with SMA type 2 were included. The Hammersmith Functional Motor Scale (expanded version) (HFMSE) and Purdue Pegboard (PP) tests were used to evaluate gross motor function and fine manual dexterity, respectively, until 18 months after nusinersen administration. HFMSE scores improved by 3–10 points (+13–53%) in all patients following nusinersen administration. PP scores also improved in all patients, from 4 to 9 points (+80–225%) in the preferred hand and from 3 to 7 points (+60–500%) in the non-preferred hand. These results suggest that nusinersen treatment improved both gross motor function and fine manual dexterity in children with SMA type 2. Addition of the PP test may aid in evaluating the fine manual dexterity essential for activities of daily living in these patients.

**Keywords:** spinal muscular atrophy (SMA); nusinersen; fine manual dexterity

#### **1. Introduction**

Spinal muscular atrophy (SMA) is an autosomal recessive disorder that affects the motor neurons in the anterior horn of the spinal cord, resulting in muscle atrophy and loss of muscle strength [1]. SMA is caused by insufficient production of SMN protein due to deletion or mutation of the survival motor neuron 1 (SMN 1) gene [2,3].

Recently, nusinersen targeting the SMN gene has been used as a treatment for SMA. Previous research has consistently demonstrated that nusinersen treatment improves both gross motor function as measured using the Hammersmith Functional Motor Scale (expanded version) (HFMSE) and upper extremity motor function as measured using the Revised Upper Limb Module (RULM) in patients with later-onset SMA [4].

Although the RULM also includes some items related to hand dexterity, such as picking up coins and tearing a piece of paper, it is difficult to assess quantitative changes in dexterity because the tool utilizes a three-point scale (0, 1, 2). Previous studies only addressed changes in the total RULM score following administration of nusinersen, without performing subgroup analysis for specific items [4,5]. Therefore, while such studies were able to confirm improvements in general upper limb motor function following nusinersen administration, they were unable to confirm whether patients exhibited improvements in fine manual dexterity. The present study is the first to investigate and demonstrate the effect of nusinersen on fine manual dexterity in patients with SMA type 2.

#### **2. Materials and Methods**

#### *2.1. Patients*

A total of five patients with 5q SMA, confirmed based on SMN1 genetic documentation, were included in this study. All patients had a clinical classification of SMA type 2 and

**Citation:** Gu, M.; Kong, H.-H. Improvement in Fine Manual Dexterity in Children with Spinal Muscular Atrophy Type 2 after Nusinersen Injection: A Case Series. *Children* **2021**, *8*, 1039. https:// doi.org/10.3390/children8111039

Academic Editor: Rudolf Korinthenberg

Received: 9 October 2021 Accepted: 10 November 2021 Published: 11 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

received neither permanent ventilator support nor enteral feeding. Between May 2019 and December 2019, the patients were referred to the Department of Rehabilitation Medicine to evaluate functional changes before and after nusinersen administration.

Nusinersen was administered intrathecally at a dose of 12 mg on days 0 (1st), 14 (2nd), 28 (3rd), and 63 (4th) according to the protocol, following which, it was administered once every 4 months for maintenance. Functional evaluations were performed before starting nusinersen treatment, between the 3rd and 4th doses, and before administration during the maintenance period. Patients underwent follow-up for a total of 18 months.

This study was approved by the Institutional Review Board of Chungbuk National University (CBNUH 2021-08-011), who waived the requirement for informed consent due to the retrospective nature of the study.

#### *2.2. Functional Assessments*

The 33-item HFMSE was developed to evaluate gross motor function related to daily living in patients with SMA type 2 or 3. Each item is scored from 0 (no response) to 2 (full response), with total scores ranging from 0 to 66 [6,7].

The Purdue Pegboard (PP) test is a standardized assessment of fine manual dexterity that is mainly used to evaluate functional abnormalities in patients with neurological impairment or developmental delay. There are normative data for most age groups, as well as reference data for preschool children over the age of 2 years, 6 months [8].

The PP test was used to evaluate fine manual dexterity in each hand. The PP test assesses the patient's ability to pick up pegs one at a time from a cup on top of the pegboard and insert them into the holes as quickly as possible. The test was first performed using the preferred hand followed by the non-preferred hand, and the number of pegs inserted within 30 s was measured [8].

The assessments were performed by one trained clinical evaluator, and training was conducted to establish reliability before data collection began.

#### **3. Results**

#### *3.1. Baseline Characteristics*

Baseline characteristics for the five patients included in the study are summarized in Table 1. All five patients were female and had SMA type 2. Genetic sequencing analysis indicated that the SMN2 gene copy number was 3 in all 5 patients. The age at the onset of SMA symptoms ranged from 12 to 14 months, while the age at SMA diagnosis ranged from 2 to 24 months. The age at initiation of nusinersen treatment ranged from 12 to 14 months (Table 1).


**Table 1.** Baseline characteristics of the study patients.

SMA—spinal muscular atrophy; SMN—survival motor neuron.

#### *3.2. Efficacy Results*

#### 3.2.1. Hammersmith Functional Motor Scale (Expanded Version)

Baseline HFMSE scores before nusinersen administration ranged from 10 to 40 points. At 18 months after nusinersen administration, HFMSE scores had improved in all patients, ranging from a minimum of +3 points to a maximum of +10 points (+13–53%). Although

**∆**

**Δ**

Δ Δ

**Δ**

there were variations among patients, gross motor functions related to trunk control such as lying, rolling, sitting, crawling, and kneeling [9] tended to improve (Table 2 and Figure 1). **∆**

**∆**


**∆**

**Table 2.** Change in HFMSE scores from baseline after nusinersen injection.

HFMSE—Hammersmith Functional Motor Scale (expanded version); ∆—amount of change. ∆

3 40 +3 0 +1 +1 +1 46 +6 (+15%) 4 13 0 +3 0 0 0 16 +3 (+23%) 5 30 +3 0 +1 0 0 34 +4 (+13%)

**Figure 1.** Change in HFMSE score from baseline to 18 months after nusinersen injection. HFMSE—Hammersmith Functional Motor Scale (expanded version).

3.2.2. Purdue Pegboard Test

In the preferred hand, PP scores before and 18 months after the initiation of nusinersen administration ranged from 2 to 7 and from 6 to 14, respectively. The PP score of the preferred hand improved in all patients, ranging from +4 to +9 (+80–225%) when compared with the baseline score (Table 3 and Figure 2a).

In the non-preferred hand, PP scores before and 18 months after the initiation of nusinersen administration ranged from 1 to 6 and from 6 to 13, respectively. The PP score of the non-preferred hand also improved in all patients, ranging from +3 to +7 (+60–500%) when compared with the baseline score (Table 3 and Figure 2b).

Shown in Figure 3, most subjects—except for patient 3—had lower PP scores than normative data of the same age and sex before nusinersen administration; however, the PP scores in both hands of patients 2, 3, and 5 after nusinersen administration (at 18 months) improved to the normal range.

Δ Δ

Δ Δ

Δ Δ

∆

**Figure 2.** Change in Purdue Pegboard score from baseline to 18 months. (**a**) Preferred hand; (**b**) non-preferred hand.

**Figure 3.** Comparison of Purdue Pegboard score changes in patients versus normative data. (**a**) Preferred hand; (**b**) non-preferred hand.


**Table 3.** Change in Purdue Pegboard scores from baseline.

HFMSE—Hammersmith Functional Motor Scale (expanded version); PP—Purdue Pegboard; ∆—amount of change.

#### **4. Discussion**

In this study, we investigated the effects of nusinersen in five patients with SMA type 2. Gross motor function as measured using the HFMSE improved in all patients after 18 months of treatment, and the functional improvements were mainly related to trunk control. Moreover, fine manual dexterity, as evaluated using the PP test, significantly improved in all patients following nusinersen treatment, and there was no difference between the preferred and non-preferred hands.

In the present study, we observed little or no improvements in standing, walking, running, or jumping ability as measured using the HFMSE; however, scores for items related to trunk control, such as lying, rolling, sitting, crawling, and kneeling improved following treatment [9]. In a previous study by Rosenblum et al., development of trunk control was identified as a prerequisite for upper extremity function and manual dexterity in healthy children [10]. Wang et al. also noted that development of trunk control in preterm infants improved fine motor skills [11]. This is because trunk stability plays an important role in upper limb motor function [12]. In accordance with previous findings, the improvements in PP score observed in this study are presumed to include the effect of improved trunk control, as measured using the HFMSE.

In addition, a recent study by Bram et al. showed significant improvements in hand grip strength and hand motor function in adult patients with SMA type 3 and 4 treated with nusinersen [5]. Similarly, administration of nusinersen may have improved fine motor function of the hand itself in our patients with SMA type 2.

In comparing the fine manual dexterity of SMA type 2 patients treated with nusinersen versus normal children, most SMA patients before receiving nusinersen had lower PP test scores than normal children of the same age and sex. However, the scores improved to the normal range observed from healthy children in most patients treated for 18 months (Figure 3) [8,13]. In a previous study, it was reported that when nusinersen was administered to infants during the pre-symptomatic stage, most patients achieved the motor milestone within the window for healthy children [14]. Similar to the results of the previous study, it is postulated that nusinersen administration rapidly improved fine manual dexterity in SMA patients; thus, they were able to reduce the gap with the normal fine motor milestone.

Previous studies have demonstrated that patients with SMA treated with nusinersen exhibit improvements not only in HFMSE scores but also in upper arm skill, as assessed using the RULM [4]. The RULM is designed to evaluate upper limb functions closely related to activities of daily living [15]. Although the RULM has the advantage of evaluating the overall function of the upper extremities, it has limitations in quantitatively measuring fine manual dexterity. Considering that most patients with SMA type 2 cannot stand alone or walk with assistance, even if they receive nusinersen treatment, fine manual dexterity is important for activities of daily living in these patients. Simply adding the PP test to the battery of existing evaluation tools may aid in providing a more detailed assessment of fine manual dexterity in patients with SMA.

Since this study was conducted only on patients with SMA type 2, the effect of nusinersen on fine manual dexterity in patients with other subtypes of SMA could not be identified. In a previous study, comparing the effects of nusinersen on SMA types 2 and 3, the change in the HFMSE score was greater in SMA type 2 than in SMA type 3 (+10.8 points versus +1.8 points) while the change in the upper limb module (ULM) score was also greater in SMA type 2 [16]. Considering the results of this previous study, it is presumed that the benefits of nusinersen on fine manual dexterity will also be greater in type 2 SMA than in other later-onset SMA types; however, this remains to be confirmed through further studies.

The present study had some limitations. Although nusinersen improved fine manual dexterity in a small number of subjects with SMA, this effect may not be statistically significant in large-scale studies. The statistical significance of the results of this study will be confirmed through a large multicenter study. Additionally, only the PP test was used to evaluate fine manual dexterity. Use of other tools to evaluate fine manual dexterity may have yielded more robust findings. Despite these limitations, this study is meaningful in that it is the first to report improvements in fine manual dexterity after nusinersen administration in patients with SMA type 2. Nonetheless, further studies including larger numbers of patients are required to verify our findings.

#### **5. Conclusions**

Changes in PP scores from baseline to 18 months confirmed that nusinersen treatment improved fine manual dexterity in patients with SMA type 2. Simply adding the PP test to the existing battery of evaluation tools may help to provide a more thorough assessment of the fine manual dexterity essential for daily living activities in patients with SMA type 2.

**Author Contributions:** Conceptualization, M.G. and H.-H.K.; methodology, H.-H.K.; software, M.G.; formal analysis, M.G. and H.-H.K.; investigation, H.-H.K.; resources, H.-H.K.; data curation, M.G.; writing—original draft preparation, M.G.; writing—review and editing, H.-H.K.; visualization, M.G.; supervision, H.-H.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** This study was approved by the Institutional Review Board of the Chungbuk National University (IRB number: CBNUH 2021-08-011, date of approval: 18 August 2021).

**Informed Consent Statement:** Patient consent was waived due to the type of study (retrospective, based on existing data).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Meaghann S. Weaver 1,2 , Alice Yuroff 3 , Sarah Sund 3 , Scott Hetzel <sup>3</sup> and Matthew A. Halanski 1, \***

<sup>1</sup> Children's Hospital and Medical Center, Omaha, NE 68114, USA; meweaver@childrensomaha.org


**\*** Correspondence: mhalanski@childrensomaha.org; Tel.: +402-955-4160; Fax: +402-955-6330

**Abstract:** The purpose of this study was to explore early changes in patient and family caregiver report of quality of life and family impact during the transitional period of nusinersen use. Communication; family relationships; physical, emotional, social, and cognitive functioning; and daily activities were measured using Pediatric Quality of Life modules (Family Impact Modules and both Patient and Proxy Neuromuscular-Specific Reports) pre- and post-nusinersen exposure. A total of 35 patients with SMA (15 Type 1, 14 Type 2, and 6 Type 3) were grouped according to nusinersen exposure. When analyzed as a whole cross-sectional clinical population, no significant differences were found between the initial and final surveys. Nusinersen therapy was associated with improved communication and emotional functioning in subsets of the population, particularly for patients on maintenance therapy for longer duration. Several unexpected potentially negative findings including increases in family resources and trends towards increases in worry warrant further consideration. Further research is warranted to explore the impact of novel pharmaceuticals on quality of life for children with SMA longitudinally to optimize clinical and psychosocial outcomes.

**Keywords:** spinal muscular atrophy; quality of life; child neurology; patient-reported outcomes; neuromuscular

#### **1. Introduction**

Spinal muscular atrophy (SMA) is an autosomal-recessive, progressive neuromuscular disease associated with extensive morbidity related to muscular atrophy and proximal muscle weakness with risk for early mortality. In the past, children with SMA Type I seldom survived beyond the first few years of life even with mechanical respiratory support [1]. With the recent introduction of novel pharmaceutical interventions such as nusinersen [1,2] and gene therapies [3], children with SMA now have potentially increased lifespans and improved quality of life (QOL). A modified 2'-O-methoxyethyl antisense oligonucleotide by the name of nusinersen was approved by the Food and Drug Agency in December of 2016 with subsequent evidence of high efficacy and safety [4,5]. Quality of life outcomes associated with nusinersen use have been less studied.

A paucity of data exists on how children with SMA depict quality of life from their own report or how family caregivers of children with SMA perceive the diagnosis impacts the child and family before, during, and after the early phases of introducing nusinersen. As survival may be prolonged through medical advancements, learning about the child's QOL remains a compassionate, competent clinical care priority [6,7]. This knowledge can help clinicians partner with the child and family for symptom or support interventions intended to further support lived experiences. QOL is defined as "an individual's perception of his/her position in life in the context of culture and value systems in which he/she lives and in relation to wellness, goals, expectations, standards, and concerns" [8]. By investing in the subjective perspective of pediatric patients and their family caregivers before and after introduction of a therapy such as nusinersen, clinical teams are then positioned to

**Citation:** Weaver, M.S.; Yuroff, A.; Sund, S.; Hetzel, S.; Halanski, M.A. Quality of Life Outcomes According to Differential Nusinersen Exposure in Pediatric Spinal Muscular Atrophy. *Children* **2021**, *8*, 604. https:// doi.org/10.3390/children8070604

Academic Editor: Rudolf Korinthenberg

Received: 15 June 2021 Accepted: 13 July 2021 Published: 17 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

better appreciate how therapies may trend with enhanced or burdened overall perceptions of health or wellness.

In 2016, prior to the widespread use of nusinersen at our institution, we began a study to evaluate the sensitivity of the Caregiver Priorities and Child Health Index of Life with Disabilities (CPCHILD™) questionnaire and PedsQL™ 3.0 Neuromuscular Module NMM (PedsQL) outcome measures to detect uniqueness between the between patient and proxy measurements and between SMA types [9]. We extended collection of the PedsQL outcome measures through 2019 to detect early changes in PEdsQL measurements in our clinical population during the transitional period of nusinersen use. This study highlights the early differences in PedsQL according to nusinersen exposure.

#### **2. Materials and Methods**

#### *2.1. Participants and Setting*

The University of Wisconsin-Madison Minimal Risk Institutional Review Board approved the study methodology and the ethics of implementation of this health sciences study in October 2015. Children and their family caregiver proxy were enrolled from November 2016 to September 2019. Eligibility criteria included patients with a diagnosis of SMA currently younger than age 18 receiving care at the outpatient neuromuscular clinic.

A letter was sent to eligible children/families providing details about the voluntary research study opportunity. The letter was mailed one to two weeks prior to the eligible subject's scheduled outpatient visit. The study coordinator then offered to meet with the patient and family caregiver through an informed consent process at the clinic visit. iPads linked wirelessly to the RedCAP© study database served as the survey response collection modality. The initial study included cross-over assessment of both the PedsQL™ 3.0 Neuromuscular Module (NMM) (for child-report and proxy-report) coupled with the PedsQL™ Family Impact Module (FIM) or CPCHILD™ questionnaire [9]. We continued collection of outcome measures until 2019 using the PedsQL measures as these appeared the most sensitive in our previous study in this population [9].

The electronic medical record was reviewed in a retrospective nature to determine the nusinersen status at the time of each initial questionnaire to produce four cohorts: (1) No Intent of Treatment, which included patients who did not start nor proceeded with any nusinesen treatment at time of final outcome measure; (2) Intent of Treatment, which included patients not on any treatment at time of initial questionnaire but began therapy after the initial assessment; (3) Loading Phase, which included patients within the first two months of treatment (received four intrathecal infusions) at the first quality of life (QOL) assessment; and (4) Maintenance Phase, which included patients on maintenance schedule (infusion every 4 months) at the time of initial QOL assessment. Cohorts 2–4 were receiving maintenance dosing at the time of their final survey (Figure 1). Pairwise differences between initial and final scores were analyzed within each cohort.

#### *2.2. Methods*

The PedsQL 3.0 Neuromuscular Module (NMM) includes 25 items covering core dimensions: (1) About My Neuromuscular Disease (17 items with emphasis on physical functioning), (2) Communication (3 items), and (3) About Our Family Resources (5 items). Child self-report and family proxy-reports are summarized for the past month. The PedsQL™ NMM maintains Cronbach's coefficient alpha scores >0.77 for each scale dimension in SMA cohorts [10–12].

**Figure 1.** Nusinersen exposure at initial and final study timepoints. Legend—Transitions between nusinersen exposure during study initiation and follow-up timepoint including Cohort 1 which remained nusinersen naïve throughout; Cohort 2 which transitioned from no nusinersen to maintenance dosing; Cohort 3 which transitioned from loading phase to maintenance dosing; and Cohort 4 which remained on maintenance dosing with longest steady exposure to nusinersen.

The 36 item PedsQL Family Impact Module (FIM) measures parental perceptions of parental self-reported physical functioning (6 items), emotional functioning (5 items), social functioning (4 items), cognitive functioning (5 items), communication (3 items), and worry (5 items). The parent is reporting on his/her own well-being rather than the child's well-being over the past month. The PedsQL FIM explores the impact of the child's SMA diagnosis and neuromuscular health on family daily activities (3 items) and family relationships (5 items). In validation studies, Cronbach's coefficient alpha scores were >0.82 for PedsQL FIM scales [13].

#### *2.3. Statistical Analysis*

Survey total scores and sub-scores were calculated. Results were summarized using mean (SD) at each timepoint and the difference of means between the initial and final survey reported. Separate analyses were then performed based on SMA type or initial nusinersen status for those receiving nusinersen. Subjects in Cohorts 2 and 3 (no nusinersen at baseline and those in the loading phase) were also pooled together to assess any changes occurring once maintenance dosing is achieved. This combined cohort was then compared with those at on maintenance dosing at the time of initial survey. Comparisons of survey scores across two-level factors utilized t-tests, while comparisons of survey scores across three-level factors utilized ANOVA models. Due to the magnitude of testing, *p*-values were Benjamini–Hochberg corrected to control for false discovery rate [14]. Significant ANOVA *p*-values resulted in post hoc pairwise t-tests with Holm-adjusted *p*-values [15]. All tests had an adjusted alpha level of 0.05 and were conducted using R for Statistical Computing Version 3.5 [16].

#### **3. Results**

#### *3.1. Participants*

A total of 35 patients with SMA: 15 Type 1, 14 Type 2, and 6 Type 3, with a respective average age at initial survey (2.7+/−2.1), (11.2+/−5.9), (10.2+/−2.7) years and an average 1.8 (+/−0.5) years between surveys were analyzed. Five patients were in Cohort 1, the non-treatment control cohort containing two patients with Type 1, one patient with Type 2, and two patients with Type 3. Cohort 2 had *n =* 8 that had not started nusinersen at the time of the first survey, Cohort 3 had *n =* 11 that were in the loading phase of nusinsersen at the time of the first survey, and Cohort 4 had *n =* 11 subjects already at the maintenance phase of treatment at the time of initial PedsQL. All patients in Cohorts 2–4 were on maintenance dosing at the time of the final QOL survey.

#### *3.2. Collective Cohort*

When analyzed as a whole cross-sectional clinical population (pooling Cohorts 2–4), no significant differences were found between the initial and final surveys for family impact (Table 1), child self-report (Table 2), or proxy family caregiver report (Table 3).


**Table 1.** PedsQL—Family Impact Module.

\* Reported mean (SD); *p*-value from paired *t*-tests.

**Table 2.** Child Self-Repot PedsQL.


\* Reported as mean (SD); *p*-values are from paired *t*-tests.

**Table 3.** Proxy -Report (Family Caregiver) PedsQL.


\* Reported as mean (SD); *p*-values are from paired t-tests.

After sub-analyzing the data by SMA type and cohort (nusinersen status at the initial survey), several significant differences and trends were identified (Tables 4 and 5). In the Family Impact Module, improvements in emotional functioning were observed for children (*n* = 8) that progressed from no treatment to maintenance therapy (56.2+/−7.5→65.4+/−15.3, *p* = 0.014).


**Table 4.** Significant Difference in Quality of Life Scales by Cohort.

SMA 1, 2, and 3 reflect diagnostic subtype. colors are useful to reveal statistical significance reached vs. close.


**Table 5.** Significant Difference in Quality of Life Scales by Timeframe.

Abbreviations—0 = No Nusinersen, L = Loading Phase, and M = Maintenance Dosing.

Patients on the maintenance dosing at the time of the initial questionnaire and therefore on maintenance therapy for longer duration (increased time exposure to nusinersen) demonstrated significant improvements in communication (43.3+/−19.6→54.2+/−25, *p* = 0.041). Per the Parental-PedsQL, improvements in communication trended towards improvement in patients initiating therapy and reaching maintenance dosing (Cohort 2) (53.1 +/−34.5→67.7+/−32.6, *p* = 0.089) and became significant when pooled with Cohort 3 (45.2+/−34.5→53.9+/−33.7, *p* = 0.04) demonstrating an improvement in communication scores when maintenance dosing was achieved and sustained.

Improvements in daily activities (39.1+/−28.1→52.6+/−32.5, *p* = 0.065) and PedsQL Family Impact Total Score (59.2+/−21.2→64.2+/−21.8, *p* = 0.081) domains trended towards significance in the population with SMA Type 2.

Patients on the maintenance dosing at the time of the initial questionnaire (again, with increased time exposure to nusinersen) demonstrated significant improvements in PedsQL Family Impact Total Score (45.7+/−13→52.6+/−26.1; *p* = 0.027).

While the majority of findings demonstrated improvements for the patents and families undergoing nusinersen treatment, several unaccepted adverse findings became apparent. First, the Family Resources Domain in the Child-PedsQL (N = 4) was significantly higher at follow up in the SMA Type 1 cohort (36.7+/−15.3→71.2+/−31.7, *p* = 0.003) and appeared to most affect those progressing from the loading to the maintenance phase (52.5+/−26→71.7+/−15.4; *p =* 0.015) (N = 7) and trended similarly in the Parental-PedsQL for patients with SMA Type 3 (58.8+/−12.5→68.8+/−15.5, *p* = 0.066).

Increases in the Worry domain also trended towards significance in our Family Impact Module for the SMA Type 1 cohort (45+/−12.7→50.8+/−12.6; *p =* 0.054) and surprisingly for those on the maintenance dosing for the entirety of this study (Cohort 4) 44.5+/−16.1→57.9+/−23.2, *p =* 0.056.

#### **4. Discussion**

In the PedsQL scale, a change of 5 in the Standard Error of the Mean (SEM) has been pre-determined to represent a minimally clinically important difference [17,18]. Thus, while noted change did not reach statistical significant difference when the at-large group was analyzed, MCID was reached for communication and family resources according to child self-report.

The major domains impacted during nusinersen treatment between the no treatment cohort and all SMA types were communication and emotional functioning. The major domain impacted during nusinersen treatment between SMA Type 1 and other SMA types was worry. Of interest, improvements in daily activities and Family Impact total score were significant only in patients with SMA Type 2. This may be due to the number of patients with Type 2 included in this study or may be due to other factors since there was no difference observed in baseline and follow up. Prior analyses by SMA type have revealed benefits in axial, proximal, and distal motor function, particularly for those with more severe forms of the disorder [3,4].

In this cross-sectional clinical study, utilizing patient reported outcome measures validated in the SMA population, nusinersen therapy was found to improve communication and emotional functioning in subsets of the population. However, several unexpected potentially negative findings including increases in family resources and trends towards increases in worry (particularly in those on the medication for the longest period of time) warrant further consideration.

#### *4.1. Improvement in Communication and Emotional Functioning*

This study revealed the benefits of nusinersen on psychosocial function beyond physiologic metrics, recognizing the importance of family communication for starting the medication and in goal setting for sustaining the medication. Nusinersen has been shown to prolong survival in infants with SMA [19–22] and improve motor function [23,24] and yet the impact of treatment options on family-based communication quality and satisfaction has been under-explored. The PedsQL FIM specifically asks about the experience of the family in communicating with the child's doctors and nurses about how they feel in addition to questions about communicating with friends and other extended family members. In a qualitative study of 19 parents engaged in decision making for their children with SMA, the most important factor for parental decision making was "honest communication with physicians" [25]. For parents in Germany whose children received nusinersen via an expanded access program, "good communication and trusting relationships with medical and non-medical staff at the hospital helped caregivers cope with the uncertainties associated with the treatment" [26]. Fifty-one parents of Swedish children with SMA emphasized the desire for health care professionals to not only possess knowledge but to provide knowledge [27], seemingly as a means to foster family communication and concordance in family decision making.

A population-based study among 34 Danish parents of children with severe SMA revealed the prioritized importance parents place on provider communication that specifies what SMA entails, the treatment options, and prognosis [28]. Among 95 parents of children with severe SMA in Denmark and Sweden, bereaved parents were significantly more satisfied with care than non-bereaved parents (81% vs. 29%), with noted emphasis on communication as part of care coordination [29].

While medical outcomes matter, families also highly regard and uphold the process of communication as formative in their family experience. Introducing nusinersen as a treatment option necessarily results in engagement about current and anticipated research findings, potential benefits and harms, and experiences of other families. This treatment-dialogue has potential to improve knowledge and empower communication within families.

#### *4.2. Increase in Use of Family Resources*

This study revealed that use of family resources was perceived as significantly increased for children with Type 1 SMA receiving nusinersen according to child self-report. Parents in this study did not document parallel perception of increased use of family

resources according. This speaks to the pediatric patient's awareness of the investment of family time and finances to nusinersen as a biomedical intervention. The ways in which children with complex care needs may internally compare their resource requirements as compared to healthy peers or siblings, and how this translates into a child's sense of self (whether the child views herself as worthy or as burdensome) or perception of stress (whether the child carries undue fear about fiscal wellness for others in the family) have been under-explored and even under-recognized by health systems. Parents in this study may have normalized resource utilization out of deep regard for their child's access to the intervention and inability to place a resource measure on the infinite value of their child's life.

Parents of children starting nusinersen describe striving for longer duration of life and improved quality of life [30] in the setting of invasive treatment and complex care with frequent hospital-based procedures. Parents of children with SMA starting nusinersen have reported worries about the high cost and maintaining adequate insurance coverage; potential side effects, risk factors, and adverse events; and treatment time [31]. The indirect care coverage costs and foregone parental employment add to the direct medical costs along with the hidden cost of mental health strain [32]. In a study of 64 parents of children with SMA, family finances were depicted as an under-recognized and yet realistic family concern [33]. In a study of parents of children w/ SMA Types 2 and 3 in Australia, parents described: "significant financial and caregiving burdens, adjusted career choices and limitations on career progression and a complex landscape of access to funding, equipment, support and resources" [32].

From a health system perspective, the average annual cost of SMA1 "ranged from \$75,047 to \$196,429 per year" [34]. The "incremental cost-effectiveness ratio (ICER) of nusinersen compared to standard of care in SMA1 ranged from \$210,095 to \$1,150,455 per quality-adjusted life years (QALY) gained." [34] In a health resource comparison study, patients in the SMA Type I group (*n =* 349) and SMA Type 1 nusineran group (*n =* 45) "experienced an average of 59.4 and 56.6 days with medical visits per-patient-per-year (PPPY), respectively, including 14.1 and 4.6 inpatient days." [35] Regardless of pharmaceutical or hospital-use economic impact, families of children receiving motor, speech, and survival benefit from nusinersen speak of the miraculous impact of the medication, which exceeds a describable cost value for those children and families.

#### *4.3. Increase in Worry*

An important finding from this study was how worry started at the lowest in the maintenance cohort, but worry notably increased longitudinally. Prior studies have shown worry peak at time of decision making about starting a new medication with unknown outcomes and concern for side effects. While nusinersen has been shown to prolong survival in infants with SMA [19–22] and improve motor function [23,24], the parents involved in this study engaged in treatment decision making prior to the more recent accumulation of outcomes-based data and thus were venturing into the unknown.

Guilt regarding genetic diagnoses and uncertainties introduced by new therapies compound the underlying unpredictable trajectory of SMA [36,37], resulting in realistic worry at medication start. Parents of children starting nusinersen report worrying about "making difficult treatment choices" as well as "reactions, side effects, and worsening quality of life" [33,38]. A qualitative study of German parents of children with SMA Type 1 depicted "significant uncertainty and stress among caregivers prior to the actual treatment. Further, concerns persisted that nusinersen could not be approved or that the child could be excluded due to an insufficient treatment response" [26]. While medical teams may consider nusinersen generally well tolerated and efficacious, parents depict worry about their child not responding to nusinersen, requiring treatment interruptions, and experiencing complications [39].

As data show that earlier initiation of treatment is associated with more efficacy on functionality (such as ambulation) [40], family caregivers recognize time-sensitive decision

making which may compound the sense of worry or urgency at initiation. Secondary spine and thorax deformities are frequent in children with SMA [41], adding worry for many families about not only the frequency of sedation but also the lumbar puncture itself [42]. Parents weigh the hoped-for benefits of nusinersen with concern about the child's discomfort. Even for parents of children with SMA who did not experience nusinersenrelated adverse events, realities of disease-specific adverse events such as cough, respiratory infections, and weakness continue to cause concern [4]. This study revealed that worry did not dissolve or mitigate with time, but instead seemed to increase longitudinally. This pattern of sustained worry as captured in this study hints at ongoing concern that patients and families have about whether the medication will continue working, whether there will be a delayed side effect, and the extent to which benefit may be sustained.

#### *4.4. Study Strengths and Limitations*

Strengths of this study include access to not only proxy-report but also pediatric patient-reported outcomes, now recognized as the gold standard for drug impact reporting [43]. Additional study strength includes use of quality of life metrics validated for this population and obtainment of surveys at more than one timepoint. Study limitations include single-site enrollment. This study did not control for whether children had missed any doses of nusinersen or adverse event/side effect profile of medication administration.

#### **5. Conclusions**

Nusinersen has offered a form of medical hope to children with SMA and their family caregivers with measurable impact on motor function and ambulation, despite the cost and challenges with administration. As the science advances to now include gene therapy, an interim goal would be additional treatment options with less burden on patients for SMA such as oral administration or one-time infusions. The lived experience of children with SMA receiving nusinersen warrants attentiveness towards ways to continually improve their quality of life. This includes consideration of ways to support family emotion and economic burden as well as foster family-centric communication. Future studies would ideally explore the impact of nusinersen and novel pharmaceutical interventions on functional abilities chronologically and longitudinally with correlated quality of life and family impact metrics.

**Author Contributions:** Conceptualization, A.Y., S.S. and M.A.H.; methodology, A.Y., S.S. and M.A.H.; software, S.H.; formal analysis, S.H.; investigation, M.S.W., A.Y., S.S., S.H. and M.A.H.; data curation, A.Y., S.S., S.H. and M.A.H. writing—original draft preparation, M.S.W. and M.A.H.; writing—review and editing, A.Y. and M.A.H.; supervision, M.A.H.; project administration, S.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors would like to acknowledge support from the UW-ICTR PCORI Pilot Program.

**Institutional Review Board Statement:** This study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the local Institutional Review Board. The committee name is University of Wisconsin—Madison Minimal Risk IRB (Health Sciences). The ID for this study is 2015-1039. Initial Approval: 10 December 2015.

**Informed Consent Statement:** Informed consent and/or assent (age-dependent) was obtained from all subjects involved in this study.

**Data Availability Statement:** Data can be made available upon reasonable request to senior author.

**Acknowledgments:** Gratitude to the patients and families involved in this study. The authors would like to acknowledge support from the UW-ICTR PCORI Pilot Program.

**Conflicts of Interest:** The authors declare no conflict of interest. Weaver contributed to this paper in a private capacity. No official support or endorsement by the U.S. Department of Veterans Affairs is intended, nor should be inferred.

#### **References**


## *Article* **Sagittal Plane Deformities in Children with SMA2 following Posterior Spinal Instrumentation**

**Matthew A. Halanski 1 , Rewais Hanna 2 , James Bernatz 2 , Max Twedt 1 , Sarah Sund 2 , Karen Patterson 2 , Kenneth J. Noonan 2 , Meredith Schultz 3 , Mary K. Schroth 4 , Mark Sharafinski <sup>2</sup> and Brian P. Hasley 1, \***


**Abstract:** This is a retrospective radiographic review to assess post-operative sagittal plane deformities in patients with Spinal Muscular Atrophy type 2 that had been treated with posterior spinal instrumentation. Thirty-two patients with a history of either spinal fusion (N = 20) or growing rods (N = 12) were identified with an average of 7.6 (2.1–16.6) years post-operative follow-up. Forty percent (13/32) of the patients were identified as having obvious "tucked chin" (N = 4), "tipped trunk" (N = 9), or both (N = 3). Sacral incidence was the only parameter that was statistically significant change between pre-operative or immediate post-operative measurements (66.9 ◦ vs. 55.2 ◦ *p* = 0.03). However, at final follow-up, the post-operative thoracic kyphosis had decreased over time in those that developed a subsequent sagittal deformity (24.2 ◦ ) whereas it increased in those that did not (44.7 ◦ , *p* = 0.008). This decrease in thoracic kyphosis throughout the instrumented levels, resulted in a greater lordotic imbalance (30.4 ◦ vs. 5.6 ◦ , *p* = 0.001) throughout the instrumented levels in the group that developed the subsequent cervical or pelvic sagittal deformities. In conclusion, sagittal plane deformities commonly develop outside the instrumented levels in children with SMA type 2 following posterior spinal instrumentation and may be the result of lordotic imbalance that occurs through continued anterior growth following posterior instrumentation.

**Keywords:** spinal muscular atrophy; posterior spinal fusion; kyphosis; sagittal plane deformity

#### **1. Introduction**

Spinal Muscular Atrophy (SMA) is the most common fatal genetic disease affecting the pediatric population (1 in 6–10,000 live births). Classically, before the widespread use of disease modifying agents, children with this disease experienced progressive weakness and early mortality. SMA is classified into three types based on the onset of disease: type 1 has symptoms starting before 6 months of age, type 2 has onset between 6–18 months of age, and type 3 has onset after 18 months of age [1]. Children with type 1 never sit and without intervention have a life expectancy <2 years, type 2 sit but do not walk and survive into the second decade, and type 3 ambulate and have a life expectancy into adulthood [1]. Respiratory failure is the most frequent cause of death in children with SMA type 1 or 2 [2].

Our institutional experience in treating severely affected children with SMA (types 1 and 2) [3–9] with spinal deformities [10,11] has led to clinical observations that a sub-set of children with SMA type 2 (upright wheelchair sitters) developed very characteristic sagittal plane deformities following spinal instrumentation that resulted in either: (1) a "tipped trunk" deformity, in which the entire (fused) and unsupported trunk leans forward

**Citation:** Halanski, M.A.; Hanna, R.; Bernatz, J.; Twedt, M.; Sund, S.; Patterson, K.; Noonan, K.J.; Schultz, M.; Schroth, M.K.; Sharafinski, M.; et al. Sagittal Plane Deformities in Children with SMA2 following Posterior Spinal Instrumentation. *Children* **2021**, *8*, 703. https:// doi.org/10.3390/children8080703

Academic Editor: Rudolf Korinthenberg

Received: 1 July 2021 Accepted: 2 August 2021 Published: 16 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

causing the abdomen to rest on the anterior thighs in the sitting position, resulting in a very prominent buttock posteriorly that complicates seating support of the lumbar and thoracic spine (Figure 1a), or a (2) a "tucked chin" deformity in which the angle of the jaw appears retracted (Figure 1b). The primary purposes of this study were to (1) screen lateral radiographs to determine the prevalence of these deformities in our postinstrumentation SMA type 2 population (2) use radiographic measurements to objectively characterize the deformities and (3) to set out to determine whether the sagittal deformities were present before, immediately after, or if they developed slowly over time following posterior spinal instrumentation. Additionally, we attempt to identify factors associated with deformity development. Our hypotheses were that these deformities developed slowly over time following spinal instrumentation and that a loss of thoracic kyphosis at the time of instrumentation would contribute to the deformity development.

**Figure 1.** Clinical photos of "trunk tip" (**a**) and "tucked chin" (**b**) deformities. With the "trunk tip" deformity, note the space between the back of the chair and the posterior chest wall.

#### **2. Materials and Methods**

A radiographic review of SMA Type 2 patients that had undergone posterior spinal instrumentation (instrumented fusion or growing rod insertion) was performed. As we did not have standardized lateral clinical photographs of every child treated at our institution at each clinical encounter, we used the most recent lateral scoliosis radiograph as a proxy to their physical clinical examination or photographs to identify those patients who had the characteristic sagittal deformities that we have observed in either the cervical spine or trunk. The overall sagittal alignments were graded as 0 (normal), 1 (borderline), 2 (obvious) then classified as cervical ("tucked chin"), pelvic ("tipped trunk"), or both (Figure 2). Patients were then grouped into two cohorts those with or without obvious deformities (grade 0 or 1 vs. grade 2). These radiographs were reviewed and scored by a fellowship trained pediatric orthopedic spinal deformity surgeon (MAH).

**Figure 2.** Examples of scoring the latest available radiographs in this study. (**Left**) Normal (0,0) scoring of no tucked chin or trunk tip. (**Middle**) Transitional scores of 1 for slight visual tucked chin (top) and prominent buttock (bottom). (**Right**) Obvious tucked chin (top) and tipped trunk (bottom) scoring a 2 in our system. Only those scoring a 2 (denoted in yellow) were included in our Deformity Cohort, these were compared against those scoring 0 or 1 (white).

Prevalence of the deformities was then determined from this data. Demographic comparisons including sex, age at surgery, and post-operative length of follow-up between cohorts was then performed. Being a tertiary referral center for these patients, some of the patients had their procedure performed at outside institutions. As such, exact operative dates were not available for every child, however the year of surgery was able to be deduced from available records. In such instances (4/32) the operative date was assigned to be December thirtieth of the operative year, to assure we were not over-estimating followup. To objectively characterize the deformities, "tucked chin" deformities were assessed by cervical sagittal Cobb angles and apex of deformity, while "tipped trunk" deformities were assessed by sagittal balance (C7 plumb line distance to anterior S1 endplate), Sacral Inclination (SI), and Seated Sacral Femoral Angle (SSFA) (a new measurement defined by a line tangential to the posterior sacrum and a line parallel with the anterior femoral shaft) (Figure 3). These values were then compared between those identified with and without the deformities in our screening. The same radiographic measurements described above were then performed on available pre-operative and post-operative radiographs, to assess whether these deformities were present pre-operatively, appeared in post-operative period as a result of surgery, or developed throughout the follow-up period. Finally, pelvic obliquity, coronal deformity, thoracic and lumbar sagittal Cobb angles, instrumentation levels, and hip status (reduced, subluxated, dislocated) were assessed between those with and without the obvious sagittal deformities to identify factors associated with the development of these deformities. All radiographic measurements were made using digital radiographs and measurement tools available through our clinical picture archiving and communicating system (PACS) (McKesson, San Francisco, CA, USA).

**Figure 3.** Examples of radiographic sagittal parameters used to objectiv **Figure 3.** Examples of radiographic sagittal parameters used to objectively characterize deformities. These include Sacral Incidence (SI, Yellow); Cervical Kyphosis (CK, Yellow); Sagittal Balance (SB, Black); Seated Sacral Femoral Angle (SSFA, White).

Statistical analysis to compare the radiographic variables between those with deformity and those without was performed using an unpaired Student *T*-Tests. All categorical variables were assessed using Fisher's Exact test. Significance for all statistical comparisons were defined as *p* < 0.05.

#### **3. Results**

Prevalence of Obvious Deformities in our Population: Thirty-two patients with SMA type II with a history of either spinal fusion (N = 20) or growing rods (N = 12), performed between 1993 and 2015, were identified with an average of 7.6 (2.1–16.6) years post-operative follow-up. Latest lateral radiographs for each of the patients were used to grade the deformities. Obvious "tucked chin" (cervical kyphosis (N = 4)), "tipped trunk" (N = 9), or both (N = 3) deformities resulted in a total of 13/32 (40%) of the patients were identified as having a deformity, the breakdown of the scoring of the 32 can be found in Table 1. Those with deformities had significantly longer follow-up (10 (3.2–16.6) years versus 5.9 (2.1–14.7) years; *p* < 0.01) than those that did not (Table 2). No significant differences were found in the presence of the deformities between those with spinal fusion versus those with growing rods (*p* = 0.76).

**Table 1.** Results of screening most recent lateral radiographs for evidence of clinically recognized deformities (N = 32).



**Table 2.** Breakdown of cohorts studied and available radiographs.

Radiographic Characterization of Deformities: Of the 32 patients, 26 had upright follow-up radiographs available for review. As these deformities fall outside the region of instrumentation, not all measurements could be made on every film as some films were focused only on the instrumented levels (Table 3). Cervical kyphosis was greater at final follow-up (53◦ (37–61◦ ) vs. −24◦ (−70–11◦ ), *p* < 0.001) in those identified with a tucked chin, with the kyphotic apices in these four being located at C1–2, C2–3,C4,C5; much more proximal than classic proximal junctional kyphosis. Those identified with an obvious tipped trunk demonstrated a more positive sagittal balance (63 mm (0–165 mm) vs. 16.4 mm (−48–59 mm), *p* = 0.04) and an increased anterior tilt of the entire pelvis (demonstrated by the increased sacral inclination (SI) 74.1◦ (60–94◦ ) vs. 46.8◦ (32–66◦ ), *p* < 0.0001) and the decreased seated sacral femoral angle (SSFA) (2◦ (−13.9–8.2◦ ) vs. 35◦ (8–58◦ ), *p* < 0.0001) (Table 4).

Temporal Appearance of Deformities: The lack of standard adequate pre-operative radiographs limits interpretation of the pre-operative status of the deformities, however, no significant differences were found in mean cervical kyphosis, sagittal balance, SSFA between cohorts. Immediate post-operative radiographs also failed to demonstrate a difference between cohorts in these measurements. Post-operative SI was the only measurement found to be significantly different greater 69◦ (52–88◦ ) vs. 55◦ (28–77◦ ) *p* = 0.03, in those with an ultimate tucked chin or tipped trunk deformity (Table 4).

Variables Contributing to the Deformities: No differences in any of the pre-operative Cobb angles or immediate post-operative coronal or lumbosacral Cobb angles were identified over the instrumented segments between groups (Table 5).


**Table 3.** Demographic differences between those found with and without obvious deformities on screening.

**Table 4.** Sagittal parameters between those identified with and without obvious deformities on screening. Sagittal measurements were only performed on upright radiographs. These data indicate that subjective screening identified objectively measured differences. N = (number with upright lateral that allowed for each measurement/total number with upright radiographs).


**Table 5.** Temporal analysis of sagittal parameters between those that ultimately developed a deformity or did not. Lack of adequate upright radiographs limits interpretation of the data, however, from the available data no significant differences were found.


The lack of adequate standardized radiographs severely limits the interpretation of the preoperative data. Interestingly, the final thoracic kyphosis (throughout the instrumented levels) was significantly less in those that developed a subsequent sagittal deformity 24◦ (−12–46◦ ) than in those that did not (45◦ (9–87◦ ) *p* = 0.008; while lumbar lordosis was the same (Table 6). This resulted in a significantly greater overall lordotic imbalance (Defined as a sum of thoracic kyphosis-lumbar lordosis) throughout the instrumented levels in the spines that developed subsequent deformity compared to those that did not (−30◦ (−70◦–(−)0.3◦ ) vs. −6 ◦ (−33.9◦–52.8◦ ), *p* = 0.001). Visual inspection of residual plots produced from the linear mixed effects analysis (performed to determine statistical differences in sagittal measurements over time) failed to reveal any obvious deviations from homoscedasticity or normality and indicated a significant effect of thoracic kyphosis on the presence of deformity. Descriptive analysis did not reveal any obvious differences in the levels of instrumentation (Table 7) or in hip status (Table 8).


**Table 6.** Analysis of spinal parameters at each time point, comparing those with and without spinal deformities.

ˆ N = number with radiographs; \* Includes supine films; \*\**p*-value = No deformity versus deformity.


**Table 7.** Comparison of instrumentation between those with and without deformities.

\* Proximal implants revised. ˆ Distal implants removed and later revision.

**Table 8.** Comparison of hip status between those with and without deformities.


#### **4. Discussion**

Scoliosis is common in all types of spinal muscular atrophy (up to 92% of patients with type 1 and 2 and 50% Type 3) and spinal deformities occur earlier with increased disease severity (Type 1 < 2 years of age, Type 2: 1–7 years of age, Type 3: 4–14 years of age) [12,13]. Posterior instrumented fusions [14–17] or distraction-based growing systems [10,11,18] have been recommended for progressive scoliotic curves in the 50–60 degree range [19]. Due to the relatively rare nature of the disease, most previous studies have grouped sub-types of SMA patients [13,16,20] or included other neuromuscular diagnoses in their analyses and reports [21–24]. These studies have focused on determining if such procedures were safe and effective [14,23] and how they affected pulmonary status [11,20,25,26], patient function and satisfaction [24,27]. While the effects of early fusion on coronal curve progression have been reported [17], no studies to date have described the effects that spinal stabilization has on the sagittal alignment above or below the instrumented levels in children with SMA.

In this study, we describe obvious deformities which occur in the sagittal plane of children with SMA type 2, above and below previously instrumented segments. For years, we have noticed these clinical deformities in our SMA population, however for the most part they have only caused seating issues, especially in those with a tipped trunk as the prominence of the buttock makes spine support difficult (Figure 1). Prior to this work, we had presumed that the children with the tipped trunk were either instrumented in excessive lordosis or perhaps there had been a gradual increase in lumbar lordosis with either continued growth or subtle loss of pelvic fixation given the known low bone density in these children [28–32]. However, after one child in our cohort required surgical treatment for severe cervical kyphosis with neurologic symptoms, we set out to critically assess how many others had such sagittal plane deformities: looking above and below the instrumented levels. In doing so, we demonstrated that these deformities are relatively common and that they are not associated with changes in lumbar lordosis (Table 6 and Figure 4).

**Figure 4.** Examples of the cervical (**A**,**B**) and trunk deformities (**A**,**C**) developing over time.

Similar changes in sagittal alignment have been described cephalad [33–39] and/or caudal [40–42] to posterior instrumentation in children with adolescent idiopathic scoliosis. However, we are not aware of any other reports describing these changes we have observed in the SMA population. Interestingly, the opposite cervical deformity (hyperextension) has been reported following posterior spinal fusion in children with Duchenne Muscular Dystrophy [43].

From our data, it appears, that children with SMA type 2 are sensitive to hypokyphosis (or excessive overall relative lordosis (subtracting the lumbar lordosis from the thoracic kyphosis) following spinal instrumentation. Interestingly, little difference in kyphosis was found between cohorts immediately after surgery, but rather, that kyphosis lessened over time in those with a deformity yet increased in those without a deformity. As the cohort that developed these deformities had significantly longer follow-up, time will tell if more of these deformities develop in the remainder of these patients. As thoracic kyphosis lessened over time in the deformity group, these findings suggest that there may have been subtle anterior growth or crankshafting following the initial posterior spinal instrumentation (Figure 4). Why the average kyphosis increased over time in those without a deformity and why certain children developed cervical deformities and others trunk deformities may not be as easy to answer, as no statistical difference was found in terms of age or instrumentation type (fusion versus growing rods) (Table 3). Perhaps the increase in Sacral Incidence seen immediately post-operative contributes to the likelihood of trunk tip or that subtle preoperative cervical kyphosis or post-operative head positioning contributes to later cervical kyphosis. Lack of adequate upright pre-operative cervical imaging for every patient leaves only conjecture. The authors had hypothesized that the forward tipped trunk may occur more readily in the presence of dislocated hips as the proximal migration of the femurs may act to over lengthen the hamstrings and gluteal muscles allowing the pelvis (and the attached, fused spine) to tip forward in response to the lordotic imbalance, however our limited sample size did not support this explanation.

The lack of uniform, adequate, upright lateral radiographs is the main limitation of this study. While being a tertiary referral center for these children provided the necessary patient volume to allow recognition of the clinical deformities; it complicates assuring that all patients have uniform imaging at their referring institutions and that all images ended up in our PACS for review. This was especially true over the study period as many institutions were transitioning from standard radiographs to digital imaging during this time frame (1993–2015). Furthermore, the underlying diagnosis also complicates standard imaging

as their overall weakness and spinal deformities can make upright radiographs for some impossible. Thus, while having full sets of pre-, post- and follow-up radiographs would have strengthened this study, the authors feel that the available radiographs were able to bring to light the ultimate sagittal deformities and highlights to others caring for these children the importance in obtaining AP and lateral upright sitting radiographs including the cervical spine and femurs before and after spinal surgery if possible. Furthermore, while the authors acknowledge that the lack of adequate radiographs severely limited our evaluation into the cause of the deformity, enough imaging was available to determine that at least 40% of our children with SMA type 2 and posterior instrumentation developed these deformities. As this prevalence was determined by taking those with an identified deformity and dividing that number by all the children with SMA type 2 and spinal instrumentation cared for at our institution (regardless of adequate films), additional adequate imaging would have only increased this prevalence, if more deformities had been identified.

Focusing on only SMA type 2 children may also be seen as a limitation of the current study. As children with SMA type 1 are unable to sit upright [44–46], and those with type 3 have less muscle weakness [47,48], these findings may be unique to SMA type 2 children. However, as more children are treated with the newer disease modifying drugs [49–53], the classic typing of SMA may become blurred as children become stronger [54,55]. Given the fact that the incidence of type 1 nearly doubles that of type 2 [56,57], we may find many more children with a phenotype similar to the classic SMA type 2 that may require spine surgery and develop compensatory deformities described in this study.

Exactly why these deformities develop and how best to prevent them was not completely answered in this study. It may be the result of a combination of several factors. First, the relative stiffness of the implants may result in a concentration of forces at the cephalad and caudal ends of the implants. Second, the overall muscle weakness from the disease itself results in the lack of muscular support for the unfused segments of the spine. A similar effect has been previously described as it relates to the collapse of the rib cage in children with spinal muscular atrophy known as the parasol rib deformity. [58,59] Finally, the crankshaft effect may develop with the continued growth of the anterior spinal column. Fujak et al. described the crankshaft phenomenon occurring in patients with SMA treated with telescopic rods and recommended definitive spinal fusion between the ages of 10–12 [60]. We have demonstrated safety and overall good results in these patients using standard distraction based growing rods [10,11]. While the authors would not suggest anterior spinal fusion in these children given their underlying pulmonary issues [4,20,61–63], surgical variables such as increased frequency of lengthening (using magnetically controlled devices) [64], three-column fixation (pedicle screws) [65–68] or stiffer instrumentation [69] might provide strategies to prevent the hypokyphosis from occurring. Thus, moving forward, it will be important for the spinal deformity surgeon to be aware of these potential sagittal compensations and to determine the best intervention to prevent them.

One final question that remains unanswered is the potential effect of recent disease modifying therapies on the development of the described sagittal plane deformities. The use of these therapies was not controlled for with this study as most of the study period predated the widespread use of these agents at our institution and could be the focus for future studies.

#### **5. Conclusions**

This single center, retrospective radiographic analysis demonstrated a 40% prevalence of sagittal deformities occurring above and below posterior instrumentation in SMA type 2 patients and provide radiographic parameters to assess for these deformities. While only correlative, these patients appear very sensitive to a lordotic imbalance that develops following posterior spinal instrumentation resulting in cervical kyphosis or anterior tipping (i.e., flexion) of the trunk. From this study, the authors would recommend all children with SMA being evaluated with scoliosis to have upright radiographs extending from the skull to the femurs, particularly in the lateral view, to allow for the detection of these deformities. Children with significant cervical kyphosis should be evaluated for signs of myelopathy [70], masked by their underlying neurologic pathology. The authors would also recommend caution in the complete correction or over-correction of thoracic kyphosis during spinal instrumentation. Furthermore, while only a correlative risk factor, the continued loss of thoracic kyphosis following instrumentation might be mitigated by (1) increased frequency of growing rod lengthening, (2) stiffer posterior spinal rods and (3) additional points of three-column fixation (pedicle screws), but the authors would caution against each of these interventions as they may have other unintended negative consequences. Further follow-up and studies are necessary to determine the long-term effects of these compensations and to identify strategies to avoid them.

**Author Contributions:** M.A.H.: Conceptualization, data curation, formal analysis, interpretation, writing—original draft preparation. R.H.: Data curation, investigation, formal analysis, writing original draft preparation. J.B.: Formal analysis, writing—review and editing. M.T.: Data curation. S.S.: Project administration, writing—review and editing. K.P.: Writing—review and editing. K.J.N.: Resources, writing—review and editing. M.S. (Meredith Schultz): Writing—review and editing. M.K.S.: Writing—review and editing. M.S. (Mark Sharafinski): Data curation, writing—original draft preparation. B.P.H.: Writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funding was provided by UW-ICTR.

**Institutional Review Board Statement:** Ethical review and approval were waived for this study by the institutional Review Board at the University of Wisconsin Madison (ID number 2018-0209), due to it being secondary research for which consent was not required.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article.

**Acknowledgments:** The authors would like to acknowledge support from the Cure SMA and UW-ICTR.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**

