Analysis of Dynamic Factors and Spinal Sagittal Alignment in Patients with Thoracic Spondylotic Myelopathy

Abstract

Introduction

Although thoracic spondylotic myelopathy (TSM) without ossification or disc disorder has been associated with some dynamic factors in the thoracolumbar area, a detailed investigation is yet to be published. Thus, in this study, we investigated the segmental motion and sagittal alignment of the thoracolumbar area in patients with and without TSM.

Methods

Patients with TSM who were treated from 2013 to 2020 were enrolled in this study. The non-TSM group consisted of sex- and age-matched patients with spinal disorders other than TSM. Segmental mobility from T10-L2 during passive maximum flexion and extension following myelography and the sagittal cobb angles of T10-L1 and L1-L5 in the standing position were measured using multidetector computed tomography (CT). The mobility of each segment was set as the difference in the angles between the two positions.

Results

In total, 10 patients (8 males and 2 females, mean age 65.8 years) with TSM and 20 without TSM were enrolled. The most stenotic level was observed at T10-T11 in four cases and T11-T12 in six. The average mobility at this segment in the TSM group (5.8°) was significantly greater than that in the non-TSM group (2.1°) (p<0.001). In the TSM group, the cobb angles of T10-L1 and L1-L5 were 2.3° and 17.4° of lordosis, respectively, which differed significantly from those in the non-TSM group, which were 8° of kyphosis and 32.2° of lordosis, respectively (p<0.001 and p=0.001, respectively).

Conclusions

Compared with those without TSM, patients with TSM were found to have greater segmental mobility at the most stenotic level, thoracolumbar lordosis, and decreased lumbar lordosis.

Introduction

Thoracic myelopathy has been most commonly attributed to the ossification of the posterior longitudinal ligament (OPLL)^1-5) or the yellow ligament (OYL), and/or thoracic disc herniation^6,7). However, it can also occur without any ossification or disc herniation, a condition known as thoracic spondylotic myelopathy (TSM). In the authors' experience, TSM is noted to usually occur at the thoracolumbar region. No studies have elucidated on the pathogenesis of TSM. We hypothesize that the pathogenesis of this disease is most likely related to spondylosis-related instability of the thoracic spine, along with age-related sagittal malalignment of the spinal column.

A standardized evaluation for the instability or flexibility of the thoracic spine has not been validated yet, perhaps because the thoracic spine is within a semi-rigid thorax, in which dynamic factors are unlikely to affect spinal cord function. In previous reports, Stanley et al.⁸⁾ investigated thoracolumbar mobility in vitro. Morita et al.^9,10) conducted an in vivo study using computed tomography (CT) images taken with the thoracic spine bent backward and found that the normal thoracic spine has some natural mobility at each level, but it is much smaller than that of the lumbar spine. To our knowledge, no reports have described the instability of the thoracic spine in patients with TSM.

Thus, in this current study, we investigated the role that dynamic factors and sagittal spinal alignment play in the pathogenesis of TSM.

Materials and Methods

Patient population and study design

This study was approved by the Institutional Review Board of the authors' affiliated institution. Patients with TSM who subsequently underwent decompression and fusion from 2013 to 2020 were deemed eligible. TSM was defined as the condition in which ligament thickening was the primary cause of scoliosis on MRI and CT (Fig. 1, 2). Patients with OPLL and OYL were excluded. The non-TSM group consisted of sex- and age-matched patients (±5 years) with spinal disorders other than TSM who underwent myelography with CT for preoperative evaluation. Because of the small sample size, the size of the non-TSM group was set to double that of the TSM cases for greater analytical precision. All CT scans were performed on a multidetector CT (120 kV, 22-120 mA, 1-mm slice; Aquilion; Toshiba Co., Tokyo, Japan). The variables measured were the following: (1) The clinical evaluation based on the modified Japanese Orthopedic Association (JOA) score (evaluation other than the upper extremities; full mark 11 points) and 3 months pre- and postoperatively; (2) the most stenotic level in the TSM group; and (3) segmental mobility from T10 to L2 measured on a multidetector-row CT in maximum passive flexion and extension following myelography (Fig. 3, 4) and the sagittal cobb angles from T10 to L1 and L1 to L5 measured in a standing position. The mobility of each segment was set as the difference between the angles in flexion and extension.

Figure 1.

T2WI sagittal magnetic resonance images (A). Sagittal reconstructed computed tomography images after myelogram in supine extension (B) and flexion (C).

Figure 2.

Post-myelogram axial computed tomography images at the most stenotic level in flexion (A) and extension (B).

Figure 3.

Posture during computed tomography after myelogram in flexion (A) and extension (B).

Figure 4.

Global sagittal balance in the standing position of a patient with TSM (A) and a patient without TSM (B). Arrow indicates hyperextension at the thoracolumbar junction in the patient with TSM.

Statistical analysis

Continuous variables were expressed as mean±standard deviation, whereas categorical variables were expressed as frequencies and percentages. The unpaired t-test and Fisher's exact test were employed for continuous and categorical variables, respectively, for comparisons between the TSM and non-TSM groups. All analyses were performed using SAS9.4 (SAS Institute Inc., NC). A p-value <0.05 was considered to be statistically significant.

Results

In total, 10 subjects with TSM aged 58-76 years (mean age, 65.8 years) were enrolled in this study (Table 1). Twenty sex- and age- matched subjects aged 57-74 years (mean age, 64.5 years) were selected from a preoperative investigation database of our section as a control group. All subjects were followed up for at least 3 months after surgical treatment. The most stenotic level was at T10-T11 in four cases and T11-T12 in six cases. All cases had thoracic myelopathy, as manifested in their lower extremities and bladder and bowel dysfunction. Their preoperative cervical JOA score, excluding upper extremity function, was an average of 4.8 out of 11 total possible points. In the non-TSM group, 14 subjects were found to have lumbar stenosis, whereas the other 6 had cervical spondylotic myelopathy. Although the average segmental mobility from T10 to L2 did not vary significantly between the TSM and non-TSM groups, the mobility at the most stenotic level in the TSM group was 5.8 degrees, whereas that in the non-TSM group was significantly lower at 2.1 degrees (p<0.01, Table 2). The sagittal cobb angle from T10 to L1 and L1 to L5 in the TSM group were 2.3 degrees of lordosis and 17.4 degrees of lordosis, respectively, whereas those in the non-TSM group were noted to vary significantly at 8 degrees of kyphosis and 32.2 degrees of lordosis, respectively (p<0.01). All TSM cases were treated surgically (Fig. 1), and all cases showed improvement at 3 months postoperatively. The average postoperative JOA score was 8.1 points, an improvement from the preoperative average of 4.8 points.

Table 1.

Patient’s Baseline Demographic Characteristics in Two Groups.

	TSM group	Control group
Male/female	8:2 cases	16:4 cases
Average age	65.8±9.1	64.5±8.8
Most stenotic level	Th10/11 4 cases Th11/12 6 cases
JOA score (full mark 11)	4.8±1.5
Primary disease	TSM	LSCS 14 cases CSM 6 cases

TSM: Thoracic spondylotic myelopathy

JOA: Japanese Orthopaedic Association

LSCS: Lumbar spinal canal stenosis

CSM: Cervical spondylotic myelopathy

Table 2.

Results of Each Segment Mobility and Spinal Sagittal Cobb Angle in the Two Groups.

	TSM group (n=10)	Control group (n=20)	p-value
Average segment mobility
Th10/11	2.75±2.14	2.34±0.92	0.67
Th11/12	3.55±2.29	2.18±1.12	0.15
Th12/L1	3.50±0.19	2.45±0.71	0.16
L1/L2	3.87±1.56	3.18±1.27	0.33
Average segment mobility at the most stenotic level
	5.82±0.56	2.08±0.97	<0.01
Sagittal Cobb angle
T10-L1	−7.5±3.2	2.33±3.63	<0.01
L1-L5	17.4±5.6	32.2±8.4	<0.01

TSM: thoracic spondylotic myelopathy

Sagittal Cobb angle; minus means kyphosis, plus means lordosis

Discussion

In this study, we investigated the mobility at each thoracolumbar segment, as well as the sagittal alignment at the thoracolumbar junction, in patients with TSM and in age- and sex-matched controls to help elucidate the factors associated with TSM. We found that the segmental mobility at the most stenotic level in patients with TSM was significantly greater than that in the non-TSM group. Moreover, the sagittal cobb angles from T10 to L1 were more lordotic, whereas those in L1 to L5 were significantly less lordotic in the TSM group than in the non-TSM group. In fact, the patients with TSM maintained a hyperextended posture at the thoracolumbar junction when in the standing position (Fig. 4).

Few articles have described the mobility of the thoracic spine^9,10). Morita et al. investigated the range of motion and dynamic changes in the cross-sectional area of the dural sac in the thoracic spine based on multidetector CT scans in the extension and flexion positions. They concluded that the segmental range of motion of the thoracic spine was an average of 2-4 degrees and increased toward the lumbar spine. Furthermore, the anteroposterior diameter, as well as the cross-sectional area of the dural sac, decreased when patients assumed an extension position. They have also found that the mobility of the thoracic spine decreased among patients older than 55 years, secondary to degeneration. In contrast, the patients with TSM in our study, with a mean age of 64 years, showed an average of 5.8 degrees of motion at the most stenotic segment, as measured using the same method as Morita's. The segmental mobility at the most stenotic level in the TSM group was significantly greater than that in the non-TSM group. The compressive element included a thickened yellow ligament, which occurs with spinal hypermobility.

We have also measured the sagittal cobb angles at the thoracolumbar and lumbar sections to identify underlying factors associated with the observed hypermobility of the thoracic segment. We found that the lumbar spine was significantly less lordotic, whereas the thoracolumbar alignment was more lordotic in the TSM group than in the non-TSM group. Whether the decreased lordosis in the lumbar spine compensated for a more lordotic thoracolumbar alignment or vice versa is impossible to determine. We have also failed to determine whether the decreased lordosis of the lumbar spine or the increased lordosis of the thoracolumbar spine in the TSM group caused thoracic myelopathy and was not an epiphenomenon secondary to the myelopathy. For now, we can only report the association between the condition and the disorder, the cause and effect of which is indeterminate. To clarify the cause and effect, a long-term longitudinal study involving a large number of subjects with thoracolumbar degeneration is warranted.

We excluded patients with OPLL or OYL that caused stenosis but did not exclude patients who had such conditions at non-stenotic levels. The authors assumed that ankylosing conditions, including OPLL, OYL, or diffuse idiopathic skeletal hyperostosis at levels adjacent to the stenotic level, would strongly impact the pathogenesis of TSM at the start of this study^11,12). However, no patients who exhibited the most stenotic level next to an ankylosed spinal region were included in this case series. Given that an ankylosed spine has been relatively common in the elderly and that the average age of our TSM group was 64 years old, future studies involving larger study populations may help clarify the association between TSM and ankylosed spines.

The authors elected to perform thoracic decompression and posterior instrumented arthrodesis for all patients with TSM in this series (Fig. 5) because we have had several recurrent cases of TSM at the same level after decompression surgery alone in the past. None of our patients who underwent decompression and instrumented arthrodesis had had recurrent TSM at the index level. When performing instrumented arthrodesis for these patients, the authors always take caution to avoid kyphosis. In the lumbar spine, it has been reported that fusion alignment correlates with adjacent intervertebral disorders in long-term postoperative outcomes^13,14). In the thoracic spine, the association between fusion alignment and adjacent intervertebral disorders has not been reported, but we are currently following this point in the long-term.

Figure 5.

Anteroposterior (A) and lateral (B) radiographs at 3 months postoperatively.

This present study has several potential limitations. First, the study population was small. However, the patients with TSM not due to herniated discs or ossifying conditions, such as OPLL or OYL, were few. A multi-center study involving a larger number of patients is necessary to fully elucidate the etiology. Nevertheless, our study had an adequate sample size for determining a statistically significant association between TSM and three factors: increased mobility at the index segment, a lordotic thoracolumbar segment, and decreased lordosis at the lumbar segment. Second, other parameters, including pelvic and lower extremity alignment, were not evaluated. Degeneration in the lower extremities can influence the lumbar spine; however, assessing the lower extremity alignment in patients with paralysis is impossible. Third, the non-TSM group should have consisted of healthy subjects. However, due to the lacking data on healthy subjects, we were constrained to include only patients with lumbar spinal canal stenosis. The results of this study showed that a certain amount of bias may have occurred in the selection of the non-TSM group. Fourth, other confounding factors other than spinal sagittal balance were not considered. Conclusively, although we found that increased segmental mobility, thoracolumbar lordosis, and decreased lumbar lordosis were associated with TSM, we could not identify the cause and effect. Is it TSM that causes these conditions, or do these conditions contribute to the pathophysiology of TSM? Only a long-term study of a large group of patients who have these conditions but have not yet developed TSM can provide further clarification on this aspect. Given the relatively rare incidence of TSM, however, it may be difficult, if not practically impossible, to perform such a study. Fifth, the accuracy of the angle measurement may need to be verified for inter- and intra-observer errors.

Conclusions

The authors compared the segmental mobility and sagittal alignment of the thoracolumbar spine between a TSM group and a non-TSM control group to identify factors associated with TSM. The segmental mobility at the most stenotic level in the TSM group was significantly greater than that in the non-TSM group. In the TSM group, the sagittal cobb angle from T10 to L1 was found to be more lordotic, whereas that from L1 to L5 was less lordotic. Patients with TSM presented with a hyperextended thoracolumbar spine when in the standing position. Further studies are warranted to determine the cause and effect of these associated findings.

Conflicts of Interest: The authors declare that there are no relevant conflicts of interest. The manuscript submitted does not contain information about medical device(s)/drug(s).

Sources of Funding: This study was supported by JSPS KAKENHI grant number 15K10423 (2015) and 18K09086 (2018).

Author Contributions: N.W: patient registration and manuscript drafting

Y.S, N.O, T.S: supervising the study design and manuscript drafting

H.I, Y.O: measurement of radiographic information

A.H: patient registration.

Ethical Approval: The protocol for this study was approved by the Institutional Review Board of our institution (2016-H128).

Informed Consent: Informed consent was obtained in all cases.