Are there direct relationships in canal dimension and canal-body ratio between cervical and lumbar spine?

A known prevalence of concurrent cervical and lumbar spinal stenosis was shown to be 5%-25%, but there is a lack of evidence regarding direct relationships in canal dimension and canal-body ratio between cervical and lumbar spine. Total 247 patients (mean age: 61 years, male: 135) with cervical and lumbar computed tomography scans were retrospectively reviewed. Midsagittal vertebral body and canal diameters in reconstructed images were measured at all cervical and lumbar vertebrae, and canal-body ratios were calculated. The canal diameter and ratio were also compared according to the gender and age, and correlation analysis was performed for each value. There were signicant correlations between cervical (C3-C7) and lumbar (L1-L5) canal dimension (p < .001). C5 canal diameter was most signicantly correlated with L4 canal diameter (r = .435, p < .001). Cervical canal-body ratios (C3-C7) were also correlated with those of lumbar spine (L1-L5) (p < .001). The canal-body ratio of C3 was most highly correlated with L3 (r = 0.477, p < .001). Meanwhile, mean canal-body ratios of C3 and L3 were signicantly smaller in male patients than female (p = .038 and p < .001) and patient’s age was inversely correlated with C5 canal diameter (r=-.223, p < .001) and C3 canal-body ratio (r=-.224, p < .001). Spinal canal dimension and canal-body ratio have direct relationships between cervical and lumbar spine. Physicians can explain that if the cervical canal is narrow, the lumbar canal is also likely to be narrow.


Introduction
Tandem spinal stenosis (TSS) is caused by the simultaneous involvement of the cervical and lumbar spines. Since Dagi et al. rst described TSS, there has been growing interest in the relationship between the cervical and lumbar spine and proper evaluation and treatment of the TSS. [1][2][3] The TSS is known to occur in 5%~25% in previous literature. 4,5 Because the patient with tandem spinal stenosis shows cervical and lumbar stenotic symptoms simultaneously, physicians should distinguish between the symptoms of cervical spinal problems and lumbar spinal problems and apply the proper remedy. [6][7][8] Bajwa et al. 4 and Lee et al. 9 proved that congenital stenosis of the cervical spine was associated with congenital stenosis of the lumbar spine, using adult skeletal specimens. However, although they reported that about 15 -30 percentage of patients showed combined cervical and lumbar spinal stenosis, they did not show extent of the relationships in canal dimension and canal-body ratio between cervical and lumbar spine and special features regarding their relationships.
Until now, previous studies for the TSS mainly focused on the prevalence and appropriate management of the patients' problems 2,10-12 , but there are few reports for assessing the degree of cervical and lumbar spinal interrelationship. The aim of this study is to verify direct relationships in canal dimension and canal-body ratio between cervical and lumbar spine and to evaluate the characteristics according to the patients' age and gender, using cervical and lumbar computed tomography (CT) scans.

Patient populations
A total of 284 patients that visited our institution between Jan. 2013 and Apr. 2017 and had simultaneous cervical and lumbar CT scans regardless of their diagnosis were retrospectively reviewed.
Younger patients (< 20 years old) and patients with de nite ossi cation of posterior longitudinal ligament, spondylolysis, deformity of vertebral body, spinal fracture, or in ammatory disease such as ankylosing spondylitis, were excluded from this study. Also, patients with lms done outside our institution were not included. Finally, 247 patients (mean age: 61 (21-82), male: 135) were included in this study.
For evaluation of canal stenosis, vertebral body and canal diameters were measured at the cervical (C3-7) and the lumbar (L1-5) vertebrae, and spinal canal to vertebral body ratios were calculated at all levels.
The canal diameter and ratio were also analyzed according to the patient's age and gender, and correlation analysis was performed for each value.

Radiographic measurements
In this study, reformatted images of CT scans were used instead of conventional lateral radiographs. The following measurements were taken for the C3-C7 and L1-5 levels on sagittal views of the reformatted images of CT scans: the midsagittal diameter of the vertebral body and the canal diameter. We measured the midsagittal diameter of the vertebral body from the midpoints of the anterior surface to the posterior surface, and the canal diameter between the midpoint of the posterior surface of the vertebral body and the nearest part of the opposite lamina ( Figure 1). 13 To minimize inter-and intra-observer errors, two independent orthopedic surgeons evaluated the digital radiographs, which were uniformly magni ed twice. Inter-and intra-observer intra-class correlation coe cients (ICCs) were assessed for the vertebral body and canal diameters.

Statistical analysis
Statistical analysis was performed by a professional medical statistical con sultant using SPSS version 19.0 statisti cal software (IBM Corp, Armonk, New York). Values were recorded as mean ± standard deviation. Depending on the normality of the data, correla tions among the measured variables were analyzed by Pearson's product-moment or Spearman's rank correlation coe cient. An independentsamples t-test or Mann-Whitney U-test was used to compare pa rameters. Signi cance was accepted for a p-value of less than 0.05.

Ethical consideration and approval
This study and patient consent exemption have been approved by the institutional review boards at Kyung Hee University hospital (KHUH 2017-05-077-003) and all patients' data were made anonymous and kept con dential. All procedures were indicated and performed in compliance with our department's standards and the Declaration of Helsinki.

Results
Inter-and intra-observer reproducibilities were high for vertebral body and canal diameter measurement. Inter-observer ICCs for vertebral body and canal diameter measurement were 0.821 and 0.879, respectively, and the corresponding intra-observer ICCs were 0.912 and 0.920, respectively.
The cervical canal diameters (C3-C7) were signi cantly correlated with lumbar canal diameters at all segments (L1-L5) (p < 0.001). The C5 cervical canal diameter showed the highest signi cant correlation with the L4 canal diameter (r = 0.435, p < 0.001) ( Table 1).   Table 2). Table 2 Correlations between cervical and lumbar spinal canal-body ratios. All cervical canal-body ratios from C3 to C7 were correlated with lumbar canal-body ratios at all segments. Particularly, the C3 canal-body ratio showed highest correlation with L3 canal-body ratio.  (Fig. 2). Uniquely, the patient's age was inversely correlated with the C5 canal diameter (r=-0.201, p = 0.002) and the C3 canal-body ratio (r=-0.243, p < 0.001), but did not show a signi cant relationship with the L4 canal diameter or the canalbody ratio of the L3 ( Table 3). The spinal canal stenosis is more remarkable in male and elderly patients than in female and young patients. In previous studies, mainly the prevalence of tandem stenosis and its treatment were discussed. 14  In our study, the degree of correlations between actual cervical and lumbar canal stenosis was evaluated. The canal diameters and spinal canal to vertebral body ratios between cervical and lumbar spinal canal showed a moderate degrees of correlation. The highest correlation coe cient was 0.435 between the C5 and L4 canal diameter (p<0.001) and 0.477 between the C3 and L3 canal to body ratio (p<0.001), respectively. Particularly, our results show the most relevant segment of cervical and lumbar spine and provide bene cial information to help choose the representative segment to predict the extent of tandem spinal stenosis. Meanwhile, patient's age had a negative correlation with the cervical canal diameter and canal to body ratio, not with the lumbar canal diameter and canal to body ratio. Also, mean canal to body ratio was signi cantly lower in the male patients than female patients. Our results will be helpful to predict a possibility of the patients' spinal stenosis in different spinal parts without direct images.

Points of Difference
We used the spinal canal to vertebral body ratio 17,18 to identify a correlation between cervical and lumbar canal diameters and the degree of its relationship. The value is calculated by dividing the mid-vertebral sagittal diameter of the cervical spinal canal by the sagittal diameter of the vertebral body. Conventionally, the Torg-Pavlov ratio is measured using conventional lateral radiographs in cervical spine, but a lumbar spinal canal diameter cannot be measured exactly, due to the overlapping shadows of the pedicle and lamina and osteophyte formation, and ossi cation around the vertebral body may impede the accurate measurement of the canal diameter. The authors thought that the reformatted sagittal images of CT scans could reduce errors and provide more accurate measurements than x-rays or magnetic resonance images.
Interestingly, there were signi cant differences in spinal canal diameter and Torg-Pavlov ratio according to the patients' ages and genders. A patient's age was inversely correlated with the cervical canal diameter and canal to body ratio, but not with the lumbar canal or ratio. Meanwhile, male patients showed signi cantly lower canal to body ratios than female patients. With these results, although the exact causes could not be explained, the authors carefully concluded that elderly male patients were vulnerable to cervical canal stenosis. To clarify the cause of these results, a large-scale epidemiological study is necessary.
Although there are studies on the importance of the cervical-pelvic relationship, few papers have found its de nite association. 19 In this study, we have found that the size of the cervical and lumbar canal and canal-body ratio are inherently related. This result is considered to be important evidence not only for tandem spinal stenosis between the cervical and lumbar spines but for explaining similar characteristics of the cervical and lumbar spine, even if there are no additional images of other spinal department. Embryologically, the cervical and lumbar spines begin similarly, as secondary curves are developed. Therefore, cervical and lumbar spinal association is a matter of course. Until now, there has been a lack of evidence for the direct relationship between cervical and lumbar spinal features. In this sense, the authors' result would be an important evidence for explaining the interrelationship between the cervical and lumbar spinal problems.

Study limitations
There were some weaknesses in this study. Firstly, this study evaluates only radiographic results, not including the data related to clinical symptoms. However, there were some studies for symptomatic tandem spinal stenosis and in this study, we have focused spinal canal and body diameter without signi cant degeneration and its ratio with using the reformatted images of CT scans. Although the CT scan don't reveal disc degeneration or ligament hypertrophy associated with clinical symptoms, it can accurately measure spinal canal and vertebral body, compared to using the x-ray or MRI 20 . Secondly, all the cases of this study have visited our hospital regardless of the patients' diagnosis and had both cervical and lumbar CT scans. There can be a bias in patients' selection. However, unlike previous studies, the purpose of this study is not to show a prevalence of tandem spinal stenosis compared to normality data, but to analyze degree of the relationships in canal dimension and canal-body ration between cervical and lumbar spine and to verify its differences according to age and gender. Therefore, the authors thought that this error would not have signi cant in uence on the results.
In conclusion, the authors con rmed that cervical canal diameter and canal-body ratio are signi cantly associated with lumbar canal diameter and ratio. Particularly, male patients showed lower canal-body ratios in the cervical and lumbar spines than female patients, and patient age was negatively correlated with canal diameter and canal-body ratio in the cervical spine. A prospective study with larger and standardly selected patients will be warranted.