As the population continues to grow older and the incidence of adult spinal deformity (ASD) has become more prevalent, scientific interest in this condition has increased. Patients with ASD undergoing complex thoracolumbar surgical procedures requiring aggressive osteotomies are at an increased risk for cervical cord injury (if preexisting cord compression is present) due to trauma from intraoperative positioning or reduced cord perfusion, e.g., intraoperative hypotension due to acute blood loss.1 In addition, concurrent cervical spinal cord compression (CSCC) in patients with ASD can potentially increase the risk of progression of myelopathy or cervical cord injury. However, clinical signs and symptoms of mild myelopathy due to CSCC, especially gait disturbance or hyperreflexia, might be subtle and masked by the clinical manifestations of thoracolumbar deformity, including spinal imbalance or simultaneous lumbar stenosis. Therefore, making a diagnosis based on patients’ history and physical examination can sometimes be problematic, especially in elderly patients with severe thoracolumbar deformity.
Some reports have described the relationship between radiographic cervical degeneration and thoracolumbar spinal alignment in patients with ASD.2,3 Fujimori et al reported that the prevalence of degenerative changes in the cervical spine was higher in patients with ASD than those without ASD, and the severity of degenerative change in the cervical spine varied depending on the type of thoracolumbar deformity: flat-back syndrome induced compensatory kyphotic changes in cervical alignment, showing significant degenerative changes on radiography.3 However, to our knowledge, magnetic resonance imaging (MRI)-based prevalence of CSCC in patients with ASD has not been previously studied.
Preoperatively, spinal deformity surgeons should be aware that CSCC can be present in patients with ASD who are undergoing major thoracolumbar realignment surgery to minimize the chance of perioperative cervical cord injury. This study seeks to identify the prevalence and characteristics of CSCC in patients with ASD based on MRIs. We hypothesized that the incidence of CSCC is relatively high in patients with ASD and its severity differs depending on the extent of sagittal malalignment of thoracolumbar deformity.
MATERIALS AND METHODS
This study was conducted at a single institution and approved by the institutional review board prior to data collection. A review of the medical records of patients with ASD who were indicated for major thoracolumbar realignment surgery at our institution between September 2015 and September 2017 was carried out. All ASD patients routinely undergo preoperative cervical MRI evaluation at our institution. Preoperative assessment consisted of routine neurological examination, including thorough motor/sensory, deep tendon reflex, Hoffman reflex, and gait testing. Overall, 121 patients during this period met the inclusion and exclusion criteria. Inclusion criteria were patients (1) aged > 20, (2) indicated for >5 level arthrodesis surgery, and (3) who underwent full-body stereoradiography (EOS imaging system; EOS Imaging SA, Paris, France) evaluation. Exclusion criteria were (1) cervical fusion surgery prior to cervical MRI evaluation and (2) neuromuscular or connective tissue disorders; four patients were excluded from the analysis because of prior cervical surgery to decompress the spinal cord.
Patient demographics including age, gender, diagnosis, and body mass index (BMI) were recorded. Radiographic evaluation was performed based on the full-body stereoradipgraph. Spinopelvic parameters were measured using the validated software (KEOPS, SMAIO, Lyon, France), including pelvic incidence (PI), lumbar lordosis (LL), PI-LL mismatch, thoracic kyphosis (T2-T2, T5-12, and T5-12 Cobb's angles), thoracolumbar kyphosis (T10-L2 Cobb's angle), pelvic tilt (PT), and sacral slope (SS). The interobserver reproducibility and intraobserver repeatability of the software have been previously shown as intraclass correlation coefficients (ICCs) of 0.9960 and rc = 0.9872, respectively.4 Parameters that could not be measured with the software were measured on the picture archiving and communication system (PACS) by two independent spine surgeons (TS and SP), and the interrater reliability (ICCs) was assessed. These included the cervical spine and sagittal balance parameters. Cervical parameters were C0-C2 angle (C0-2, occiput to C2 Cobb's angle), C2-C7 lordosis (C2-7L: C2-C7 Cobb's angle), C2-C7 SVA (cSVA: horizontal distance between a plumb line dropped from C2 to the posterosuperior corner of C7), and T1 slope (T1S: angle between the superior endplate of T1 and a horizontal reference line). The spinopelvic balance was evaluated using C7 sagittal vertical axis (C7SVA) and T1 pelvic angle5 (TPA: angle between the line from the femoral head axis to the center of the T1 vertebra and the line from the femoral head axis to the middle of the S1 superior endplate). The global balance parameters included the global sagittal axis6 (GSA: a global alignment angle measured from the line from the midpoint of the two distal femoral condyles to the center of the C7 vertebra and the line from the center of the two distal femoral condyles to the posterosuperior corner of the S1 endplate). The Torg-Pavlov ratio7 at the narrowest cervical segment was measured in each patient according to the previously published method. The presence of ossification of the posterior longitudinal ligament (OPLL) was also evaluated based on computed tomography (CT) if available or cervical radiography in cases in which CT was not performed.
CSCC was diagnosed based on the sagittal and axial views of 1.5T T2-weighted MRI and also graded according to its severity by two independent spine surgeons. CSCC grade was determined using the modified Cord Compression Index8 (CCI, Table 1): the sum of the anterior 0 to 3 points and the posterior 0 to 3 points at each segment (C2/3, C3/4, C4/5, C5/6, C6/7, and C7/T1) according to the compressed intervertebral disc or ligamentum flavum. The CCI has three grades: Grade 1 (compression of 0–2 points), Grade 2 (3–4 points), and Grade 3 (5–6 points). Significant CSCC was defined as Grade >2. The presence of cord signal change on T2-weighted imaging was also noted.
In the study cohort, the prevalence of CSCC was calculated according to the CCI. Patients with significant CSCC (Grade >2) were further analyzed in terms of (1) the distribution of cord compression level, (2) the number of compressed spinal cord segments, and (3) prevalence according to age by decade. A multivariate linear regression analysis was performed to identify the predictive factors for CSCC (CCI score). Patients with OPLL and Torg-Pavlov ratio < 0.8 were excluded in the regression analysis as these are different etiologies from degenerative diseases.
Data are presented as mean ± standard deviation, unless specified otherwise. The interobserver reliability was calculated according to the ICC and Cohen's κ value for continuous variables and categorical variables (CCI), respectively, and classified as poor (0–0.39), moderate (0.4–0.74), or excellent (0.75–1). A multivariate regression analysis was performed with variables that were selected using a stepwise method from all the demographic and radiographic variables. JMP version 13 (SAS, Cary, NC) was used for all analyses. The statistical significance was set at P < 0.05.
Demographics and Radiographic Analysis
Overall, the inter-rater reliability for radiographic measurements was excellent (ICCs 0.880–0.981).
Table 2 shows the patient characteristics and radiographic parameters. The mean age was 55.3 ± 14.1 (range, 23–82) years, with BMI of 24.4 ± 4.5 kg/m2, and female participants were the dominant group (M/F, 22:99). Etiologies were adult idiopathic (56.1%), degenerative kypho-scoliosis (23.9%), fixed sagittal imbalance syndrome (15.7%), Scheuermann's kyphosis (1.6%), postfracture kyphosis (1.6%), and ankylosing spondylitis (0.8%). Radiographic parameters demonstrated significant thoracic and thoracolumbar curvature in the coronal plane (main thoracic, 40.3 ± 27.1°; Lumbar/Thoracolumbar, 45.0 ± 25.1°) and slight sagittal malalignment (PI-LL mismatch, 8.5 ± 23.7°; C7SVA, 46.0 ± 89.7 mm; TPA, 21.7 ± 14.8°; and GSA, 3.5 ± 5.9°). The Torg-Pavlov ratio at the narrowest level was 1.0 ± 0.1, and three patients had <0.8 ratio. OPLL was observed in three patients (2.4%) with Grade 2 CCI.
Prevalence of Cervical Cord Compression
The inter-rater reliability for grading CCI was moderate (Cohen's κ = 0.701, 95% CI [0.679–0.721]). Table 3 shows the overall incidence of CSCC. Of 121 patients with ASD, 41 patients (33.8%) demonstrated significant CSCC (32 with Grade 2 and 9 with Grade 3). Cord signal change (myelomalacia) on T2-weighted imaging was present in eight patients (6.6%). Significant CSCC was most commonly observed at the C4/5 level (Figure 1). Among the 41 patients with significant CSCC, 22 patients had one-level cord compression, while the remaining 19 patients had more than two-level cord compressions (Figure 2). In addition, 35 of 41 patients (85.3%) were asymptomatic or with myelopathy that is difficult to detect (e.g., subtle Hoffman reflex), while 6 of 41 patients were myelopathic (e.g., hyperreflexia of lower extremities and weakness/atrophy in upper extremities).
Table 4 shows the distribution of CSCC according to each age generation. Significant CSCC was commonly seen even in patients in their 30s (37.5%) and significantly prevalent in those aged over 70s. Prior to the thoracolumbar reconstruction surgery, four patients (3.3%) underwent cervical decompression and fusion; in these patients, three had Grade 3 CSCC, of which two patients showed cord signal change (Figure 3A–E), whereas the other one had Grade 2 CSCC with OPLL. The multivariate regression analysis revealed that age (P = 0.034), BMI (P < 0.01), and PI-LL mismatch (P < 0.01) independently predicted the CSCC grade (Table 5).
In this study, we report the prevalence of and predictive factors for cervical spinal cord compression based on MRI in patients with ASD undergoing major thoracolumbar reconstruction surgery. Among the patients, 33.8% (41/121) showed significant CSCC (CCI Grade>2). Age, BMI, and PI-LL mismatch independently predicted the severity of CSCC. Clinically, we had noticed a high rate of cervical cord compression findings both by careful physical examination and history in this patient population; and thus began obtaining routine c-spine MRI as a screening for compression. As our results show, there is a high rate of cervical spinal cord compression in thoracolumbar deformity patients.
Simultaneous stenosis of the cervical and lumbar spine has been well known as “tandem stenosis,” which was first reported by Dagi et al9 in 1987. The prevalence of tandem stenosis ranges from 5% to 24% depending on the modality used to diagnose it.8–12 Lee et al8 reported that the incidence of tandem stenosis based on sagittal MRI was 23.7%. However, only a few studies focused on the association of cervical stenosis with the extent of thoracolumbar deformity. Schairer et al2 demonstrated that 31.0% of patients with thoracolumbar deformity had a diagnosis of cervical spondylosis without detailed evaluation on MRI. Our study presented a more reliable, MRI-based prevalence of cervical cord compression in patients with ASD.
In the intraoperative phase of complex ASD surgeries, various factors, such as prolonged positioning of the cervical spine in hyperextension and suboptimal mean blood pressure due to blood loss and anesthetic agents, could increase and/or induce the risk of cervical cord injury at the compressed levels.1 In addition, fusion of the thoracolumbar spine may postoperatively worsen the cervical symptoms due to spontaneous reciprocal changes in the cervical alignment. Therefore, it is crucial for deformity surgeons to evaluate the cervical spine preoperatively. Physical examination alone is not reliable for detecting cervical myelopathy. In fact, 85.3% of patients with significant CSCC (Grade>2 CCI) were asymptomatic or with myelopathy that is difficult to detect in our series. Rhee et al13 described that one-fifth of patients with cervical myelopathy were negative for myelopathic signs. In these patients, diagnostic imaging study, most commonly MRI, should be considered a preoperative assessment, in addition to a thorough neurologic examination. Nevertheless, it may not be cost-effective to perform preoperative MRI in all patients undergoing ASD surgery. Our study demonstrated older age, BMI, and PI-LL mismatch as predictive factors for significant CSCC. We suggest performing MRI on these patients prior to thoracolumbar reconstruction. Age and obesity have been shown to significantly influence the presence of disc degeneration in the cervical spine as well as the lumbar spine.14 Determining the cut-off age to perform preoperative MRI is not easy; however, our results imply that patients aged over 70 years should undergo MRI due to the high prevalence of CSCC in this age group (over 50%). Also, it should be noted that even those patients in their 30s had a 37.5% rate of Grade>2 CCI. Teraguchi et al15 reported that patients with obesity (BMI > 27 kg/m2) had an odds ratio of 1.60 for the presence of cervical disc degeneration. The data in the present study are not able to provide a pathophysiologic explanation for the association between increased BMI and CSCC, we suspect metabolic derangement or potentially microvascular changes related to diabetes may play a role. Further study would be needed to confirm this association and to explore potential pathophysiologic explanations. The relationship between cervical degeneration and thoracolumbar alignment has rarely been reported. Fujimori et al3 described that radiographic degeneration score correlated with loss of cervical lordosis in the setting of compensatory hypothoracic kyphosis induced by degenerative flat back (that is, high PI-LL mismatch). These patients were commonly seen in the current series. Conversely, in decompensated thoracolumbar malalignment with high PI-LL mismatch, hyperlordosis of the cervical spine can occur.16 In either of the two situations, high PI-LL mismatch results in nonphysiologic hyper-/hypo-lordosis, which can accelerate degenerative change of disc and ligaments.
In this series of patients with ASD, four patients (3.3%) underwent cervical decompression and fusion prior to thoracolumbar reconstruction. In patients presenting with apparent myelopathic signs or symptoms, most surgeons would recommend cervical spine surgery before thoracolumbar reconstruction. On the contrary, in the case of an asymptomatic but significant cord compression on MRI (Figure 4A–F), prophylactic decompression prior to thoracolumbar deformity operation may be controversial. On the basis of the current literature, the occurrence of cord injury after minor trauma in patients with asymptomatic cervical cord compression is likely uncommon and even intramedullary T2 hyper-intensity cannot be a pure predictor for myelopathy development.17,18 However, it is difficult to completely exclude the presence of cervical myelopathy with physical examination alone, especially in patients with significant thoracolumbar deformity. Treating these patients without knowing the presence of coexisting cervical cord compression could theoretically lead to a devastating cord injury. In addition, we recommend intraoperative neuromonitoring with recording from both upper and lower extremities and keeping the neck in gentle neutral position using halo-ring or Gardner–Wells tong traction throughout the operation.
The segment in which significant CSCC was most frequently observed was the mid-cervical levels (C4/5 and C5/6). In addition, almost half of the patients with significant CSCC showed more than two-level compression. These findings are in agreement with previous findings that demonstrated a high prevalence of degenerative changes of the cervical spine in those segments.15,19 Surgeons should expect that multiple compressions commonly coexist in those segments in patients with thoracolumbar deformity.
Some limitations of this study should be noted. Due to the nature of the study, some patients referred from other hospitals had already undergone MRI and been diagnosed with cervical stenosis. This might have caused bias to the neurological evaluation at initial visit. Second, cervical radiculopathy was not assessed in this study on every patient unless there was something in the patient's history which prompted a formal upper extremity examination. Concurrent radiculopathy can also be worsened after a long thoracolumbar reconstruction operation. Future studies are planned with prospectively collected data, including postoperative progression of clinical signs and comparison between patients with and without cervical spine surgery prior to thoracolumbar realignment surgery.
The prevalence of concurrent cervical spinal cord compression in patients with adult spinal deformity is relatively high at 33.8% and, among them, 85.3% were asymptomatic or with myelopathy that is difficult to detect. Even those patients in their 30s had a 37.5% rate of significant cervical cord compression. Preoperative evaluation of cervical MRIs and examinations for signs/symptoms of myelopathy are essential for (1) older patients, (2) patients having increased BMI, and (3) patients with high PI-LL mismatch to avoid progressive myelopathy or cord injury during ASD surgery.
1. Ahn H, Fehlings MG. Prevention, identification, and treatment of perioperative spinal cord injury. Neurosurg Focus
2. Schairer WW, Carrer A, Lu M, et al. The increased prevalence
of cervical spondylosis in patients with adult
thoracolumbar spinal deformity
. J Spinal Disord Tech
3. Fujimori T, Le H, Schairer W, et al. The relationship between cervical degeneration and global spinal alignment in patients with adult spinal deformity
. Clin Spine Surg
4. Maillot C, Ferrero E, Fort D, et al. Reproducibility and repeatability of a new computerized software for sagittal spinopelvic and scoliosis curvature radiologic measurements: Keops((R)). Eur Spine J
5. Protopsaltis T, Schwab F, Bronsard N, et al. TheT1 pelvic angle, a novel radiographic measure of global sagittal deformity, accounts for both spinal inclination and pelvic tilt and correlates with health-related quality of life. J Bone Joint Surg Am
6. Diebo BG, Oren JH, Challier V, et al. Global sagittal axis: a step toward full-body assessment of sagittal plane deformity in the human body. J Neurosurg Spine
7. Pavlov H, Torg JS, Robie B, et al. Cervical spinal stenosis: determination with vertebral body ratio method. Radiology
8. Lee SH, Kim KT, Suk KS, et al. Asymptomatic cervical cord compression
in lumbar spinal stenosis patients: a whole spine magnetic resonance imaging study. Spine (Phila Pa 1976)
9. Dagi TF, Tarkington MA, Leech JJ. Tandem lumbar and cervical spinal stenosis. Natural history, prognostic indices, and results after surgical decompression. J Neurosurg
10. Epstein NE, Epstein JA, Carras R, et al. Coexisting cervical and lumbar spinal stenosis: diagnosis and management. Neurosurgery
11. LaBan MM, Green ML. Concurrent
(tandem) cervical and lumbar spinal stenosis: a 10-yr review of 54 hospitalized patients. Am J Phys Med Rehabil
12. Vogt MT, Cawthon PM, Kang JD, et al. Prevalence
of symptoms of cervical and lumbar stenosis among participants in the Osteoporotic Fractures in Men Study. Spine (Phila Pa 1976)
13. Rhee JM, Heflin JA, Hamasaki T, et al. Prevalence
of physical signs in cervical myelopathy: a prospective, controlled study. Spine (Phila Pa 1976)
14. Wang YX, Griffith JF, Ma HT, et al. Relationship between gender, bone mineral density, and disc degeneration in the lumbar spine: a study in elderly subjects using an eight-level MRI-based disc degeneration grading system. Osteoporos Int
15. Teraguchi M, Yoshimura N, Hashizume H, et al. Prevalence
and distribution of intervertebral disc degeneration over the entire spine in a population-based cohort: the Wakayama Spine Study. Osteoarthritis Cartilage
16. Diebo BG, Challier V, Henry JK, et al. Predicting cervical alignment required to maintain horizontal gaze based on global spinal alignment. Spine (Phila Pa 1976)
17. Wilson JR, Barry S, Fischer DJ, et al. Frequency, timing, and predictors
of neurological dysfunction in the nonmyelopathic patient with cervical spinal cord compression
, canal stenosis, and/or ossification of the posterior longitudinal ligament. Spine (Phila Pa 1976)
18. Chang V, Ellingson BM, Salamon N, et al. The risk of acute spinal cord injury after minor trauma in patients with preexisting cervical stenosis. Neurosurgery
19. Matsumoto M, Fujimura Y, Suzuki N, et al. MRI of cervical intervertebral discs in asymptomatic subjects. J Bone Joint Surg Br