Effect of EOS Severity and Treatment on Pulmonary Function Relative to Stature (as Represented by Pelvic Width) in Children with SMA
Michael J. Troy1,2, Patricia Miller1, Robert Graham1, Brian Snyder1
1Boston Children’s Hospital, Boston, MA, 2Marian University College of Osteopathic Medicine, Indianapolis, IN
INTRODUCTION: Spinal muscular atrophy (SMA) consists of a group of neuromuscular disorders characterized by degeneration of alpha motor neurons located in the anterior horn of the spinal cord, leading to a lower motor neuron-type syndrome. Degeneration occurs due to insufficient production of survival motor neuron (SMN) protein, induced by mutations or deletions in SMN1 gene (exon 7 and/or 8) on chromosome 5q, inherited in an autosomal recessive manner. The extent of clinical involvement depends on the number of copies of the SMN2 “rescue” gene, which produces mostly non-functional SMN protein (2 copies = more severe disease; 3-4 copies = milder disease). The clinical phenotype and musculoskeletal manifestations stem from an associated progressive muscle atrophy, weakness and paralysis1 SMA phenotypes are classified by the maximum motor function achieved: Type I are very weak infants unable to sit; Type II are non-ambulatory children who can sit independently; Type III ambulate as children, but may lose functional gait in adulthood; Type IV remain ambulatory as adults. Owing to severe global weakness, children afflicted with SMA are prone to Early Onset Scoliosis (EOS) and thoracic insufficiency. Pulmonary Function Tests normalized by height are difficult to evaluate in SMA patients because of associated skeletal deformities. Pelvic Inlet Width (PIW) has been established as a surrogate for stature & thoracic height in EOS that can be reliably measured on spine X-rays2. Our hypotheses are 1) that pulmonary function as represented by FVC varies proportionately with stature (thoracic height). Scoliosis affects this relationship: children with mild or moderate scoliosis (Cobb angle 50o) follow this relationship, while children with more severe scoliosis (Cobb angle >50o) do not; 2) correction of the scoliosis by dual posterior rod spinal instrumentation (SI) partially restores this relationship compared to use of a thoraco-lumbar spinal orthosis (TLSO). Therefore, we evaluated whether FVC varies proportionately with stature, as represented by PIW, and whether this relationship was affected by EOS severity or treatment.
METHODS: For this retrospective comparative study, sequential bedside FVC and sitting spine X-ray in/out of TLSO or after SI with either growing rods that allow for continued spine growth by periodic distraction of the rods or definitive spine fusion were evaluated bi-annually. Less severe curves were treated w/TLSO, whereas severe curves, unresponsive to TLSO, were treated by SI. Cobb, PIW and FVC at initiation of treatment (TLSO, SI) were compared to last f/u: 53 SMA pts., age < 18 yrs., types: 1 (2%), 2 (53%), 3 (45%) were analyzed over 5 yrs. (SD 2.6; range 1.1-11.6 yrs). Since 2017, nearly all received Nusinersen treatment (intrathecal administered mRNA that modulates alternate splicing of the SMN2 gene, functionally converting to SMN1 gene, thereby increasing levels of SMN protein in CSF). Due to the disparity in subgroup sample sizes, and non-normal outcome measures (change in curve and change in FVC/PIW) linear mixed modeling was performed to achieve a two-way type analysis with modifications for incomplete, unbalanced, nonparametric data.
RESULTS: At presentation (Fig. 1A), for children with Cobb 50o, there was strong correlation between FVC and PIW (r=0.86; p<0.001); however for Cobb >50o there was no correlation (p=0.27). These relationships were similar @ last f/u (Fig.1B), regardless of treatment, indicating that treatment did not restore proportionality between FVC and PIW for Cobb >50o, even if Cobb corrected 50o. For the change in FVC/PIW the available data is almost completely grouped between scoliosis >50 treated by SI versus scoliosis 50 treated by TLSO. There were some differences in outcomes across curve severity groups and across intervention groups (based on nonparametric testing). When FVC was normalized by stature (PIW), a detectable difference in FVC/PIW overtime across subjects with large curves (>50°) versus small curves (50°) (p=0.004) and across subjects treated with SI versus TLSO (p=0.009) was observed (Fig. 2A). It is unclear, if either of these associations are specifically due to treatment or initial curve severity. There was a detectable difference in curve magnitude overtime across subjects with large curves (>50°) versus small curves (50°) (p=0.001), but no difference was detected across subjects treated with SI versus TLSO (p=0.06) (Fig. 2B).
DISCUSSION: For SMA children with mild to moderate scoliosis (Cobb 50o), 50-75% of FVC variability is accounted by variability in PIW, indicating that pulmonary function changes proportionately with stature and growth. However, for severe spinal deformity (Cobb >50o), there was no relationship, indicating that for SMA children with severe scoliosis, pulmonary function was disproportionate to stature, thus accounting for the thoracic insufficiency observed in these children and their requirement for augmentative respiratory interventions. Nusinersen treatment did not delay the onset or ameliorate the progression of scoliosis and SI did not restore the proportional relationship between pulmonary function (FVC) and stature (PIW). These findings are consistent with our rabbit model of pulmonary hypoplasia associated with EOS that we produced by constricting the hemithorax in a growing rabbit to evaluate how surgical distraction of the thorax affects lung volume, respiratory function, alveolar morphology, and spine growth3. While surgical expansion of the thorax performed early or late altered the predicted trajectory of progressive spine deformity and thoracic volume reduction, pulmonary growth and respiratory function (FVC) remained below normal (Fig. 3). Partitioned respiratory mechanics indicated that the respiratory elastance was increased, exacerbated by increased thoracic stiffness, which was a consequence of altered ventilator mechanics induced by anatomic deformity of the spine and rib cage, tissue scarring and the rigidity of the surgical instrumentation, as well as failure of the surgical intervention to improve the projected surface area and excursion of the diaphragm, the primary muscle (“piston”) responsible for respiratory mechanics.
SIGNIFICANCE/CLINICAL RELEVANCE: We establish a direct relationship between pelvic inlet width (PIW), a radiographic surrogate for thoracic height and forced vital capacity (FVC), a measure of pulmonary function in SMA children, age <18 yrs, with moderate spinal deformity (Scoliosis 50o) indicating that for SMA children with moderate spinal deformity, FVC varies proportionately with stature. However there was no relationship for severe scoliosis, indicating disproportionate pulmonary function relative to stature = thoracic insufficiency. Surgical treatment using posterior instrumentation did not appear to modify or restore this relationship, possibly because surgical intervention occurred too late, after intrinsic (and irreversible) changes to the thorax and lung parenchyma transpired.
REFERENCES:Lunn, M.R. and C.H. Wang, Spinal muscular atrophy. Lancet, 2008. 371(9630):2120-33.  Glotzbecker et al, Is there a relationship between thoracic dimensions and pulmonary functionin early onset scolisos? Spine, 2014/39(19):1590-5  Olson et al, Expansion Thoracoplasty in Rabbit Model: Effect of Timing on Preserving Pulmonary Growth and Corecting Spine Deformity. Spine, 43(15) E877-884