Cervical laminoplasty (CLP), which allows posterior decompression of the spinal cord in the subaxial cervical spine, was introduced in Japan during the 1970s1,2 and is presently one of the standard procedures for treating patients with cervical spondylotic myelopathy (CSM). Compared to cervical laminectomy, CLP enables further preservation of the posterior elements while providing multilevel decompression, resulting in a lower risk of postoperative kyphotic changes of the cervical spine.3,4 Additionally, various countermeasures to prevent postoperative kyphotic change, including the modification of postoperative care and operative methods, have been introduced.5–8 However, kyphotic deformity after CLP cannot be completely prevented.
Kyphotic deformity after CLP, following the loss of cervical lordosis (CL), can lead to unfavorable sequelae, which are broadly classified into two issues. First, when a kyphotic deformity exists, the spinal cord is forced against the osteophytes of the posterior vertebral body or bulging disc, and its decompression effect is lost, thus resulting in poor neurological recovery, even though the spinal canal is surgically expanded.9,10 Second, when the kyphotic change progresses and sagittal balance in the cervical spine is disrupted, not only does neck pain worsen, but also health-related quality of life deteriorates.11,12 Although severe sagittal imbalance after CLP is very rare, the so-called dropped head syndrome, which markedly complicates daily living, after CLP has been reported in the literature.13,14 For these reasons, researchers have investigated the risk factors associated with the loss of CL after CLP, so that appropriate patients can be selected to attain a favorable postoperative outcome.15–18
Two recent studies have demonstrated that cervical range of motion (ROM) was closely associated with the loss of CL after CLP. One study demonstrated that a greater flexion ROM (fROM) was a risk factor for kyphotic change after CLP.19 Another study revealed that a smaller extension ROM (eROM) was the definitive predictor.20 These studies ultimately presented different conclusions but may reflect a similar phenomenon: the imbalance that is characterized by a greater fROM and smaller eROM for the total ROM can result in a higher risk of kyphotic change after CLP. In reality, however, fROM and eROM do not necessarily have a complementary relationship because the total cervical ROM varies individually. The indicator, which can relatively represent fROM and eROM by itself, may be even more effective in predicting CL loss after surgery than fROM or eROM alone.
On the basis of these two studies,19,20 we developed a novel indicator—the gap between fROM and eROM (gROM), which independently represents a greater fROM and/or smaller eROM—and we hypothesized that it was a highly useful indicator for predicting the loss of CL after CLP. The aim of the present study was to investigate the risk factors for CL loss after CLP in patients with CSM, including gROM as a potential predictor.
Methods
Subjects
We retrospectively reviewed the records of consecutive patients who had undergone CLP, ranging from vertebrae C3 to C7, for CSM between January 2013 and September 2019 at our institution and an affiliated hospital. One of the authors was the chief surgeon in all of these operations. All patients engaged in routine follow-up until at least 1 year postoperatively, and the subsequent follow-up was dictated by the attending surgeon after considering the neurological recovery and radiographic findings of the patients. Inclusion criteria were as follows: cases without a history of previous cervical surgery, cases without neuromuscular disease, cases without congenital anomalies in the cervical spine, and cases that did not undergo surgery on the thoracolumbar spine or lower extremities, which could influence cervical alignment, during the follow-up period. Exclusion criteria were cases in which the radiographic measurement (later described) could not be conducted owing to the difficulty in identifying bony landmarks because of poor image quality or overlap with a bony structure in the shoulders. The protocol of the present study was approved by the ethics review board of each study site.
Surgical Technique and Postoperative Care
All patients had undergone double-door CLP using suture anchors, which has been described elsewhere.21,22 During the exposure, detachment of the paravertebral muscles from the posterior bony structures was kept to a minimum, and the attachment of the semispinalis cervicis (SSC) muscle to the C2 spinous process was preserved in all cases. When performing the decompressive procedures from C3 to the caudal levels, laminectomy was performed at C3 in most cases unless the C3 lamina could be fully opened. This is done because, in cases in which laminoplasty is performed at C3 while preserving the SSC muscle, the opened C3 lamina often interferes with its muscle course, which can lead to an increase in postoperative axial pain,8 lamina closure,22 or kyphotic change in the cervical spine.5 After performing midline osteotomy and placing hinges at the bilateral laminofacet junctions using a high-speed drill, self-tapping titanium suture anchors were embedded in the lateral masses. The bilateral split laminae were stabilized in the maximum open position using suture anchors. Decompression of the spinal cord was confirmed macroscopically and ultrasonographically.23
From the next postoperative day, patients were permitted to stand and walk while wearing a soft collar. The soft collar was removed within 2 weeks after surgery.
Demographic and Surgical Variables
Demographic characteristics, age at surgery, sex, BMI, the morbidity of diabetes mellitus, and smoking history were collected. Regarding the surgical variables, surgical time, estimated blood loss, use of concomitant foraminotomy, and number of operated levels were noted. Furthermore, the levels of operation in terms of the most cephalad and most caudal levels were recorded. The most cephalad level was noted as C3 or C4 and the most caudal level as C6/upper or C7.
Radiographic Variables
On preoperative lateral radiographs of the cervical spine in the neutral position, the following parameters were measured: C2–7 angle, the Cobb angle between the inferior endplate of C2 and the superior endplate of C7; cervical sagittal vertical axis (cSVA), the distance between the plumb line from the center of the C2 vertebral body and the posterosuperior aspect of C7; C7 slope, the angle between the superior endplate of C7 and a horizontal line; and C2 slope, the angle between the inferior endplate of C2 and a horizontal line (Fig. 1A). The C2–7 angles were also assessed using radiographs in the flexion and extension positions, in which fROM indicates the difference between C2–7 angles in the neutral and flexion positions and eROM indicates the difference between C2–7 angles in the neutral and extension positions. The gROM was defined as follows: gROM (°) = fROM − eROM. An example of the computation for gROM is illustrated in Fig. 1B. Spinal instability was examined using radiographs of the flexion and extension positions; anteroposterior translation ≥ 3 mm on the radiographs showing flexion and extension was defined as instability based on previous studies.24,25 On the final follow-up radiograph, the C2–7 angle in the neutral position was measured in each case. The radiographic measurements were conducted using dedicated and validated software (Surgimap, Nemaris Inc.).26,27 As for the C2–7 angle, a positive value indicated lordosis, and a negative value indicated kyphosis.
A: Radiographic measurements: C2–7 angle (a), cSVA (b), C7 slope (c), and C2 slope (d). B: An example of measurements of fROM, eROM, and gROM. In this case, the C2–7 angle was 8° in the neutral position, −24° in the flexion position, and 17° in the extension position. Hence, fROM and eROM were 32° and 9°, respectively, resulting in 23° of gROM (fROM − eROM). Figure is available in color online only.
MRI Variables
The cross-sectional area (CSA) of the posterior paravertebral muscle was assessed using MRI. The CSA of the multifidus (MF)–SSC muscle complex (MF-SSC CSA) was examined28 since two of these muscles specifically act to stabilize and adjust the cervical sagittal alignment.29,30 The measurements were performed according to the technique that has been previously reported to be reliable for assessing the CSA of the posterior paravertebral muscles in the cervical spine:28,31 the axial view on preoperative T2-weighted MRI parallel to the intervertebral disc space was analyzed using ImageJ imaging software (version 1.43, National Institutes of Health). The MF-SSC CSA was measured separately at the C3–4 and C6–7 intervertebral disc levels in each case.
The first observer (T.F.) performed all measurements. Images from 20 patients were assessed by the first observer with an interval of more than 4 weeks from the initial measurements to test intraobserver reliability. Measurements were then obtained by the second observer (T.O.), who was blinded to this study, to test interobserver reliability.
Clinical Outcomes
The Japanese Orthopaedic Association (JOA) score for cervical myelopathy (from −2 [worst] to 17 points [best]) was collected to evaluate the severity of myelopathic symptoms, and the recovery rate was examined.3
Data Analysis and Statistics
The patients were divided into two cohorts based on previous reports,19,20 that is, cases with and cases without more than a 10° loss of CL, represented by the C2–7 angle, at the final follow-up period compared to preoperative measurements (CL loss [CLL] group and no CLL [NCLL] group, respectively). Demographic, surgical, and preoperative radiographic and MRI variables were compared between the CLL and NCLL groups by using univariate analysis. Regarding the significant variables, a receiver operating characteristic (ROC) curve was constructed, and an area under the ROC curve (AUC) was determined to evaluate the capacity for discriminating between the CLL and NCLL groups. Additionally, we examined whether the most significant variable in the above analysis affected neurological recovery, represented by the recovery rate of the JOA score.
In univariate analyses, comparisons of continuous variables were performed using the Wilcoxon signed-rank test, and comparisons of nominal variables were performed using Pearson’s chi-square test or Fisher’s exact test. The degree of correlation was assessed using the Spearman correlation coefficient (rS), and an absolute rS value < 0.2 was interpreted as almost negligible, 0.2 to 0.4 as low, 0.4 to 0.7 as moderate, 0.7 to 0.9 as high, and > 0.9 as a very high correlation.32 Intra- and interobserver reliability of the measurements for MF-SSC CSA was assessed using intraclass correlation coefficients (ICCs).33 The predictive capacity based on the AUC was judged as high when it was greater than 0.9, as moderate when 0.7–0.9, and as low when 0.5–0.7,34 and the method described by DeLong et al. was used to test the statistical significance among AUCs.35 Statistical analyses were performed using JMP 14 software (SAS Institute Inc.) and IBM SPSS version 25 (IBM Corp.). A probability value < 0.05 was considered statistically significant.
Results
Study Population
A total of 135 patients met the inclusion criteria, and 120 patients completed more than 1 year of follow-up. Among the latter group, we excluded 1 case because of radiographs that could not be used for measurements and 8 cases because radiographs were not taken at the final follow-up visit 1 year postoperatively. As a result, 111 cases (82.2%) were included in the analysis.
Table 1 shows the demographics and clinical outcomes of this study cohort, which consisted of 68 men and 43 women with a mean age at surgery of 68.3 years. The mean follow-up period was 18.2 months after CLP. Regarding the operated levels, the procedure was most frequently performed from C3 to C6 (60 cases), followed by C3–7 (18 cases) and C4–6 (13 cases). Ninety-three patients underwent decompressive procedures starting from C3 to caudal levels; C3 laminoplasty was performed in 6 cases and C3 laminectomy in 87 cases. The mean JOA score changed from 10.5 preoperatively to 13.8 at the final follow-up period, with a significant improvement (p < 0.001). The average recovery rate was 51.3%.
Demographics and clinical outcomes of the study population
| Variable | Value |
|---|---|
| No. of patients | 111 |
| Age at surgery in yrs (range) | 68.3 (28−89) |
| Sex | |
| Male | 68 |
| Female | 43 |
| BMI in kg/m2 (range) | 24.0 (17.1−46.0) |
| FU period in mos (range) | 18.2 (12−36) |
| JOA score (range) | |
| Baseline | 10.5 (0.5−16.5) |
| FU | 13.8 (7−17) |
| Recovery rate in % | 51.3 (0−100) |
FU = follow-up.
Group Division
Among the 111 cases in the study population, there were no cases in which the C2–7 angle was less than −10° preoperatively; the mean C2–7 angle of the total study population was 10.0° preoperatively and 8.6° at the final follow-up, which did not significantly differ (p = 0.538). A more than 10° loss in the C2–7 angle at the final follow-up compared to the preoperative measurement was identified in 10 cases (9.0%), which were assigned to the CLL group, and the remaining 101 cases (91.0%) were assigned to the NCLL group. For the CLL and NCLL groups, the mean changes in the C2–7 angle were −26.7° and 1.1°, respectively, and the mean C2–7 angles at the final follow-up were −16.5° and 11.1°, respectively. These values were significantly different between the two groups (both p < 0.001; Table 2).
Change in C2–7 angle in CLL and NCLL groups
| Variable | Total | CLL Group | NCLL Group | p Value |
|---|---|---|---|---|
| No. of patients | 111 | 10 | 101 | |
| Change in C2–7 angle (°) | −1.4 ± 9.9 | −26.7 ± 13.4 | 1.1 ± 4.7 | <0.001 |
| C2–7 angle during FU period (°) | 8.6 ± 12.9 | −16.5 ± 14.8 | 11.1 ± 9.7 | <0.001 |
CLL group = CL loss group; NCLL group = no CL loss group. Values are presented as the mean ± standard deviation, unless indicated otherwise. Boldface type indicates statistical significance.
When analyzing 93 cases that underwent decompressive procedures involving C3, C2–7 angles preoperatively and at the final follow-up did not significantly differ for 6 cases with C3 laminoplasty and 87 cases with C3 laminectomy (12.4° vs 9.8°, p = 0.595; 9.6° vs 8.8°, p = 0.981, respectively). Similarly, there was no significant difference in the mean change in the C2–7 angle between the cases with C3 laminoplasty and those with C3 laminectomy (−2.8° vs −1.0°, p = 0.214). All 6 cases with C3 laminoplasty belonged to the NCLL group.
Investigation of Risk Factors for CL Loss After CLP
Table 3 shows comparisons of collected variables between the CLL and NCLL groups. Demographic variables did not differ between the two groups (p = 0.536, p > 0.999, p = 0.568, p = 0.108, and p = 0.737 for age, sex, BMI, morbidity of diabetes mellitus, and smoking history, respectively). Regarding the surgical variables, there were no significant differences in estimated blood loss, use of concomitant foraminotomy, number of operated levels, or most cranial and caudal levels operated upon (p = 0.215, p = 0.594, p = 0.299, p = 0.664, and p = 0.118 respectively). The surgical time was significantly shorter in the CLL group than in the NCLL group (119.4 vs 141.1 minutes, p = 0.047).
Comparison of variables between CLL and NCLL groups
| Variable | Total | CLL Group | NCLL Group | p Value |
|---|---|---|---|---|
| No. of patients | 111 | 10 | 101 | |
| Demographic variable | ||||
| Age at surgery in yrs | 68.3 ± 11.7 | 71.0 ± 9.7 | 68.1 ± 11.9 | 0.536 |
| Male sex | 61.3 (68) | 60.0 (6) | 61.4 (62) | >0.999 |
| BMI in kg/m2 | 24.0 ± 3.9 | 26.7 ± 8.8 | 23.8 ± 3.1 | 0.568 |
| Diabetes mellitus | 19.8 (22) | 40.0 (4) | 17.8 (18) | 0.108 |
| Smoking history | 39.6 (44) | 30.0 (3) | 40.6 (41) | 0.737 |
| Surgical variable | ||||
| Surgical time in mins | 139.1 ± 38.7 | 119.4 ± 40.2 | 141.1 ± 38.2 | 0.047 |
| Estimated blood loss in ml | 40.1 ± 40.8 | 35.0 ± 61.6 | 40.6 ± 38.5 | 0.215 |
| Foraminotomy | 9.9 (11) | 0 (0) | 10.9 (11) | 0.594 |
| No. of operated levels | 3.8 ± 0.8 | 3.6 ± 0.7 | 3.9 ± 0.8 | 0.299 |
| Most cranial level operated on | ||||
| C3 | 83.8 (93) | 80.0 (8) | 84.2 (85) | 0.664 |
| C4 | 16.2 (18) | 20.0 (2) | 15.8 (16) | |
| Most caudal level operated on | ||||
| C6/upper | 79.3 (88) | 100 (10) | 77.2 (78) | 0.118 |
| C7 | 20.7 (23) | 0 (0) | 22.8 (23) | |
| Preop radiographic variable | ||||
| C2–7 angle in ° | 10.0 ± 9.4 | 10.2 ± 8.0 | 10.0 ± 9.6 | 0.885 |
| cSVA in mm | 25.8 ± 13.5 | 23.5 ± 15.4 | 26.0 ± 13.4 | 0.338 |
| C7 slope in ° | 26.7 ± 9.0 | 23.8 ± 6.8 | 26.9 ± 9.1 | 0.370 |
| C2 slope in ° | 16.6 ± 7.9 | 13.6 ± 8.8 | 16.9 ± 7.7 | 0.222 |
| fROM in ° | 27.8 ± 10.3 | 40.2 ± 8.8 | 26.6 ± 9.6 | <0.001 |
| eROM in ° | 12.0 ± 7.4 | 8.6 ± 4.8 | 12.4 ± 7.5 | 0.114 |
| gROM in ° | 15.8 ± 11.7 | 31.6 ± 7.7 | 14.3 ± 10.9 | <0.001 |
| Instability | ||||
| Nix | 86.5 (96) | 90.0 (9) | 86.1 (87) | 0.338 |
| C3–4 | 8.1 (9) | 0 (0) | 8.9 (9) | |
| C4–5 | 2.7 (3) | 0 (0) | 3.0 (3) | |
| C5–6 | 2.7 (3) | 10.0 (1) | 2.0 (2) | |
| MRI variable | ||||
| C3–4 MF-SSC CSA in mm2 | 423.9 ± 117.0 | 436.0 ± 114.2 | 422.7 ± 117.7 | 0.703 |
| C6–7 MF-SSC CSA in mm2 | 808.9 ± 213.3 | 811.6 ± 268.0 | 808.6 ± 208.7 | 0.730 |
Values are presented as the mean ± standard deviation or percentage (number of cases), unless indicated otherwise. Boldface type indicates statistical significance.
Comparisons of the preoperative radiographic variables showed no significant differences in C2–7 angle, cSVA, C7 slope, C2 slope, eROM, and spinal instability between the CLL and NCLL groups (p = 0.885, p = 0.338, p = 0.370, p = 0.222, p = 0.114, and p = 0.338, respectively). In contrast, the CLL group had a significantly greater fROM than the NCLL group (40.2° vs 26.6°, p < 0.001).
The mean gROM for the entire study cohort was 15.8°. The gROM exhibited a highly positive correlation with fROM and a moderately negative correlation with eROM (rS = 0.793, p < 0.001; rS = −0.431, p < 0.001, respectively; Fig. 2A). When comparing gROM between the two groups, the CLL group had a significantly greater gROM than the NCLL group (31.6° vs 14.3°, p < 0.001).
Scatter diagrams according to gROM. The charts for fROM and eROM (A) and for the recovery rate of the JOA score (B). Figure is available in color online only.
As for the measurements of MRI variables, the intra- and interobserver reliabilities were acceptable, with ICC values of 0.931 and 0.917 for C3–4 MF-SSC CSA and 0.900 and 0.841 for C6–7 MF-SSC CSA, respectively. Significant differences in MRI variables between the CLL and NCLL groups were not identified (p = 0.703 and 0.730 for C3–4 and C6–7 MF-SSC CSAs, respectively).
In the univariate analyses, surgical time, fROM, and gROM significantly differed between the CLL and NCLL groups. Of these variables, fROM and gROM were studied using ROC curve analyses to evaluate their ability for discriminating between the two groups, as shown in Fig. 3. The discriminating capacity of fROM was moderate (AUC = 0.860, SE = 0.069, p < 0.001, 95% CI 0.666−0.950), and that of gROM was high (AUC = 0.906, SE = 0.048, p < 0.001, 95% CI 0.762−0.967). The AUCs for fROM and gROM did not statistically differ; however, gROM tended to have a higher AUC than fROM (p = 0.094). The optimal cutoff point of gROM was 27° with 80.0% sensitivity and 89.1% specificity.
ROC curves demonstrating the accuracy of fROM and gROM for discriminating the CLL and NCLL groups. The AUC for gROM was likely to be greater than the one for fROM (p = 0.094). Data within parentheses represent the 95% confidence intervals of the AUCs. Figure is available in color online only.
Relationship of Clinical Outcomes With gROM
The results of the above analyses indicated that gROM was the most significant predictor of postoperative CL loss; thus, its relationship with neurological recovery, represented by the recovery rate of the JOA score, was investigated.
Figure 2B shows the distribution of the recovery rate of the JOA score according to gROM, for which a significant correlation was not identified (rS = 0.083, p = 0.387). The optimal cutoff value of gROM for distinguishing CLL and NCLL groups was 27°. Hence, we compared the recovery rate between the cases with and the cases without a preoperative gROM ≥ 30°; however, there was no significant difference between them (53.9% vs 50.9%, respectively; p = 0.826; Table 4). In contrast, the 7 cases with ≤ −10° of C2–7 angle in the follow-up period, all of which belonged to the CLL group, tended to have a poorer recovery rate of the JOA score than the 104 cases without (35.2% vs 52.4%, respectively; p = 0.098).
Recovery rate of JOA score according to gROM and C2–7 angle at follow-up
| Variable | Recovery Rate (%) | p Value |
|---|---|---|
| gROM | ||
| ≥30° (n = 15) | 53.9 ± 29.0 | 0.826 |
| <30° (n = 96) | 50.9 ± 27.7 | |
| C2–7 angle during FU period | ||
| ≤−10° (n = 7) | 35.2 ± 21.1 | 0.098 |
| >−10° (n = 104) | 52.4 ± 27.9 |
Values are presented as the mean ± standard deviation, unless indicated otherwise.
Representative cases with greater and smaller gROMs are shown in the upper and lower sections of Fig. 4, respectively.
Pre- and postoperative radiographs obtained in a 66-year-old man with a greater gROM, who had undergone C3 laminectomy and C4−6 laminoplasty. In this case, the C2–7 angle in the neutral position was 23° preoperatively. In flexion and extension positions, the C2–7 angles were −29° and 27°, respectively. Consequently, fROM and eROM were 52° and 4°, respectively, indicating that gROM was 48° (A). During the patient’s postoperative course, his radiographs revealed the progression of kyphotic change in the cervical spine. At 1 year postoperatively, he showed good neurological recovery, as determined on the basis of JOA score (preoperatively, 12 points; 1 year postoperatively, 15.5 points; recovery rate, 70.0%); however, he suffered from severe neck pain with a C2–7 angle of −24° in the radiograph of neutral position (B). Pre- and postoperative images obtained in a 78-year-old woman with a smaller gROM, who had undergone C3 laminectomy and C4−6 laminoplasty. In the preoperative radiographs (C), C2–7 angles were 13°, −26°, and 40° in the neutral, flexion, and extension positions, respectively. Thus, fROM and eROM were 39°and 27°, resulting in 12° for gROM. Her preoperative JOA score was 11 points. One year after surgery, the C2–7 angle was 14° in the radiograph in the neutral position (D), indicating well maintained CL, and she had no complaint of neck pain. At that time, the JOA score was 15.5 points, for a 75.0% recovery rate. Figure is available in color online only.
Discussion
The present study demonstrated that the novel indicator, gROM, representing the gap between fROM and eROM, was highly useful in predicting a marked loss of CL after CLP.
Kyphotic deformity resulting from the loss of CL is one of the potential complications after CLP, which can cause loss of the decompression effect in the spinal cord or cervical sagittal imbalance. However, the prevalence of kyphotic change, which becomes a problem in clinical practice, is certainly not high. Recent studies have emphasized that cervical alignment can be sufficiently maintained in the majority of cases after CLP for CSM,19,36,37 and this finding is consistent with the present study, which showed that there was no significant difference between preoperative and postoperative CL, represented by C2–7 angles, of the entire study population (from 10.0° to 8.6°, p = 0.538). The present study also revealed that the potential surgical invasiveness values, based on surgical variables, including estimated blood loss, use of concomitant foraminotomy, and spinal levels operated on, were similar between the CLL and NCLL groups. Furthermore, surgical time was significantly shorter in the CLL group than in the NCLL group (Table 3). These findings could imply that the surgical invasiveness of CLP to the posterior cervical structure is not the direct cause of the postoperative development of kyphotic change, but that the cases with marked postoperative CL loss have some shared characteristics.
To date, several studies have investigated the risk factors for kyphotic change after CLP and have suggested that it is associated with a patient’s age, preoperative sagittal alignment, and the CSA of posterior paraspinal muscles.15–18 Two recent studies have indicated a significant relationship between ROM in the cervical spine and kyphotic change after CLP. Fujishiro et al. showed that a greater fROM was strongly associated with kyphotic change after CLP.19 They speculated that, in patients with a smaller fROM, the cervical spine is unlikely to move to the kyphotic position intrinsically because motion in the flexional direction is blocked by degenerative structures, such as vertebral osteophytes, intervertebral disc degeneration, and muscular-ligament contracture. Conversely, in patients with a greater fROM, such structural forces restricting motion toward the kyphotic position are weaker, and the burden in the posterior neck muscular-ligament complex (PMLC) to maintain cervical lordotic alignment is greater than in patients with a smaller fROM; hence, when the PMLC is surgically invaded, the equilibrium necessary to maintain cervical balance would be prone to disruption, resulting in a higher prevalence of a marked loss of CL. In parallel, Lee et al. demonstrated that a smaller eROM is a risk factor for the development of kyphotic change after CLP.20 They surmised that a greater eROM interrelates with a larger PMLC constriction reservoir; therefore, patients with a smaller eROM have less capacity to maintain cervical lordotic posture and are more likely to lose CL postoperatively.
The gROM, calculated by subtracting eROM from fROM, had a significant positive correlation with fROM and a negative correlation with eROM (Fig. 2A). More specifically, it is the indicator that interrelates with a greater fROM and smaller eROM. In the present study, we demonstrated that gROM can predict the loss of CL after CLP more accurately than fROM or eROM alone, as we had hypothesized. Its greater value, denoting greater fROM and/or smaller eROM, can signify a smaller structural stabilizing effect to restrain the cervical spine from lapsing into kyphotic alignment and/or a smaller capacity of the PMLC to maintain the lordotic alignment, indicating a higher risk of CL loss after CLP, and vice versa.
Clinically, the conception of gROM seems to be very practical in several respects. First is its simplicity in terms of measurement. Lateral radiographs of the cervical spine, including the flexion-extension position, are generally obtained preoperatively. The gROM can be calculated simply by measuring C2–7 Cobb angles on three lateral radiographs (in the neutral, flexion, and extension positions), without much difficulty in identifying bony landmarks. Second, it is accurate in predicting whether CL is markedly lost after surgery, as the capacity of gROM to discriminate between the CLL and NCLL groups, based on the AUC, was high (Fig. 3). The optimal cutoff value of gROM to discriminate between the CLL and NCLL groups was 27°; for patients with a preoperative gROM exceeding 30°, early removal of the cervical collar and the subsequent cervical ROM exercises should be encouraged. Otherwise, a concomitant fusion procedure or anterior cervical surgery should be optionally considered.
Lee et al. showed that a smaller eROM was a definite risk factor for kyphotic change after CLP.20 In the present study, the CLL group was likely to exhibit a smaller eROM than the NCLL group; however, the difference did not reach statistical significance (8.6° vs 12.4°, p = 0.114; Table 3). Note that there were some discrepancies in the view concerning eROM between the Lee et al. study and our present study. In the study by Lee et al., patients with ossification of the posterior longitudinal ligament (OPLL) accounted for more than half of the study population and were analyzed together with those with CSM.20 It has been reported that, in some patients with OPLL, cervical vertebrae fuse with the ossification mass; therefore, the cervical alignment changes after CLP in patients with and those without OPLL were different.38 The present study cohort was limited to CSM patients, and the inconsistency regarding eROM between the two studies may be derived from a difference in the etiology of the study populations.
Several studies have demonstrated that the postoperative kyphotic alignment of the cervical spine results in poor neurological recovery. Suda et al. reported that patients with a signal intensity change of the spinal cord on MRI exhibited a poor neurological recovery after CLP when cervical kyphosis exceeded 5°, whereas those without the change had a poor neurological recovery after CLP when cervical kyphosis exceeded 13°.10 In the present study, cases with a kyphotic C2–7 angle of more than 10° tended to have a poorer JOA score recovery rate than those without (Table 4), which is in accordance with previous studies. Nonetheless, gROM, which is a predictive indicator of postoperative kyphotic change, did not directly affect clinical outcome (Table 4 and Fig. 2B). There are several possible reasons for this result. First, the preoperative extent of CL varies individually. Therefore, gROM can predict whether CL is markedly lost after CLP, but estimating the degree of postoperative kyphotic deformity (mild or severe kyphotic deformity) is beyond its capacity. Another reason may be that the neurological recovery after CLP for degenerative cervical myelopathy does not only depend on the postoperative cervical alignment and is influenced by multiple factors, which has been well documented in previous studies.39–42
It is important to recognize the limitations of this study. First, this study was retrospective in design. Furthermore, the number of cases assigned to the CLL group was low, which might have led to low statistical power. This is likely because we took countermeasures to prevent kyphotic change in all patients, including preservation of the attachment of the SSC muscle to the C2 spinous process,43 selection of the cephalad vertebral level undergoing laminoplasty,5 and the choice of a shorter period of postoperative collar fixation,6,7 as well as the rarity of the marked loss of CL after CLP. To overcome these shortcomings and verify the results of the present study, prospective studies with larger sample sizes may be necessary. Second, the follow-up period in the present study was relatively short. However, Choi et al. showed that, after CLP, cervical sagittal curvature generally reached a plateau at 6 months postoperatively in most cases but did not change thereafter.15 Hence, the assessment of CL at 1 or more years after surgery, which was done in the present study, would be reasonable to investigate the risk factors for developing kyphotic change during the postoperative course. Finally, this study did not include measurements for neck pain and health-related quality of life, although their relationships with sagittal balance of the cervical spine have been well investigated.11,12
Conclusions
The gROM is a highly useful indicator for predicting whether CL is markedly lost after CLP. A greater gROM value, denoting greater fROM and/or smaller eROM, can signify a smaller structural stabilizing effect to prevent the cervical spine from lapsing into kyphotic alignment and/or a reduced capacity of the PMLC to maintain lordotic alignment, indicating a higher risk for CL loss after CLP. For CSM patients with a preoperative gROM exceeding 30°, CLP should be carefully considered since kyphotic change is likely to develop after surgery.
Acknowledgments
We sincerely appreciate Hiromitsu Moriuchi, MD, PhD, of the Department of Orthopedic Surgery, First Towakai Hospital, for collaboration during data collection.
Disclosures
The authors report no conflicts of interest concerning the materials or methods used in this study or the findings specified in this paper.
Author Contributions
Conception and design: Fujishiro, Hayama, Nakaya. Acquisition of data: Fujishiro, Obo. Analysis and interpretation of data: Fujishiro. Drafting the article: Fujishiro. Critically revising the article: Hayama, Obo, Nakaya, Nakano, Usami, Nozawa, Neo. Reviewed submitted version of manuscript: Hayama, Obo, Nakaya, Nakano, Usami, Nozawa, Baba, Neo. Statistical analysis: Fujishiro. Study supervision: Fujishiro, Neo.
References
- 1↑
Hirabayashi K. Expansive open-door laminoplasty for cervical spondylotic myelopathy. Article in Japanese. Shujutsu. 1978;32(11):1159–1163.
- 2↑
Oyama M, Hattori S, Morikawa N, et al.. A new method of cervical laminectomy. Article in Japanese. Cent Jpn J Orthop Traumatol. 1973;16(3):792–794.
- 3↑
Hirabayashi K, Watanabe K, Wakano K, et al.. Expansive open-door laminoplasty for cervical spinal stenotic myelopathy. Spine (Phila Pa 1976).1983;8(7):693–699.
- 4↑
Matsunaga S, Sakou T, Nakanisi K. Analysis of the cervical spine alignment following laminoplasty and laminectomy. Spinal Cord. 1999;37(1):20–24.
- 5↑
Michael KW, Neustein TM, Rhee JM. Where should a laminoplasty start? The effect of the proximal level on post-laminoplasty loss of lordosis. Spine J. 2016;16(6):737–741.
- 6↑
Cheung JPY, Cheung PWH, Law K, et al.. Postoperative rigid cervical collar leads to less axial neck pain in the early stage after open-door laminoplasty—a single-blinded randomized controlled trial. Neurosurgery. 2019;85(3):325–334.
- 7↑
Hida T, Sakai Y, Ito K, et al.. Collar fixation is not mandatory after cervical laminoplasty: a randomized controlled trial. Spine (Phila Pa 1976).2017;42(5):E253–E259.
- 8↑
Takeuchi K, Yokoyama T, Aburakawa S, et al.. Axial symptoms after cervical laminoplasty with C3 laminectomy compared with conventional C3-C7 laminoplasty: a modified laminoplasty preserving the semispinalis cervicis inserted into axis. Spine (Phila Pa 1976).2005;30(22):2544–2549.
- 9↑
Kimura I, Shingu H, Nasu Y. Long-term follow-up of cervical spondylotic myelopathy treated by canal-expansive laminoplasty. J Bone Joint Surg Br. 1995;77(6):956–961.
- 10↑
Suda K, Abumi K, Ito M, et al.. Local kyphosis reduces surgical outcomes of expansive open-door laminoplasty for cervical spondylotic myelopathy. Spine (Phila Pa 1976).2003;28(12):1258–1262.
- 11↑
Lafage R, Challier V, Liabaud B, et al.. Natural head posture in the setting of sagittal spinal deformity: validation of chin-brow vertical angle, slope of line of sight, and McGregor’s slope with health-related quality of life. Neurosurgery. 2016;79(1):108–115.
- 12↑
Tang JA, Scheer JK, Smith JS, et al.. The impact of standing regional cervical sagittal alignment on outcomes in posterior cervical fusion surgery. Neurosurgery. 2012;71(3):662–669.
- 13↑
Koda M, Furuya T, Kinoshita T, et al.. Dropped head syndrome after cervical laminoplasty: a case control study. J Clin Neurosci. 2016;32:88–90.
- 14↑
Odate S, Shikata J, Soeda T. Catastrophic dropped head syndrome requiring multiple reconstruction surgeries after cervical laminoplasty. Spine Surg Relat Res. 2018;2(3):243–247.
- 15↑
Choi I, Roh SW, Rhim SC, Jeon SR. The time course of cervical alignment after cervical expansive laminoplasty: determining optimal cut-off preoperative angle for predicting postoperative kyphosis. Medicine (Baltimore). 2018;97(47):e13335.
- 16
Cao J, Zhang J, Yang D, et al.. Multivariate analysis of factors associated with kyphotic deformity after laminoplasty in cervical spondylotic myelopathy patients without preoperative kyphotic alignment. Sci Rep. 2017;7(1):43443.
- 17
Lee BJ, Park JH, Jeon SR, et al.. Importance of the preoperative cross-sectional area of the semispinalis cervicis as a risk factor for loss of lordosis after laminoplasty in patients with cervical spondylotic myelopathy. Eur Spine J. 2018;27(11):2720–2728.
- 18
Sakai K, Yoshii T, Hirai T, et al.. Cervical sagittal imbalance is a predictor of kyphotic deformity after laminoplasty in cervical spondylotic myelopathy patients without preoperative kyphotic alignment. Spine (Phila Pa 1976).2016;41(4):299–305.
- 19↑
Fujishiro T, Nakano A, Yano T, et al.. Significance of flexion range of motion as a risk factor for kyphotic change after cervical laminoplasty. J Clin Neurosci. 2020;76:100–106.
- 20↑
Lee SH, Son DW, Lee JS, et al.. Does extension dysfunction affect postoperative loss of cervical lordosis in patients who undergo laminoplasty? Spine (Phila Pa 1976).2019;44(8):E456–E464.
- 21↑
Usami Y, Nakaya Y, Hayama S, et al.. Impact of multifidus muscle swelling on C5 palsy after cervical laminoplasty. Spine (Phila Pa 1976).2020;45(1):E10–E17.
- 22↑
Fujishiro T, Nakano A, Baba I, et al.. Double-door cervical laminoplasty with suture anchors: evaluation of the clinical performance of the constructs. Eur Spine J. 2017;26(4):1121–1128.
- 23↑
Nakaya Y, Nakano A, Fujiwara K, et al.. Percutaneous ultrasonographic evaluation of the spinal cord after cervical laminoplasty: time-dependent changes. Eur Spine J. 2018;27(11):2763–2771.
- 24↑
Oichi T, Oshima Y, Taniguchi Y, et al.. Cervical anterolisthesis: a predictor of poor neurological outcomes in cervical spondylotic myelopathy patients after cervical laminoplasty. Spine (Phila Pa 1976).2016;41(8):E467–E473.
- 25↑
Suzuki A, Daubs MD, Inoue H, et al.. Prevalence and motion characteristics of degenerative cervical spondylolisthesis in the symptomatic adult. Spine (Phila Pa 1976).2013;38(17):E1115–E1120.
- 26↑
Lafage R, Ferrero E, Henry JK, et al.. Validation of a new computer-assisted tool to measure spino-pelvic parameters. Spine J. 2015;15(12):2493–2502.
- 27↑
Iyer S, Lenke LG, Nemani VM, et al.. Variations in occipitocervical and cervicothoracic alignment parameters based on age: a prospective study of asymptomatic volunteers using full-body radiographs. Spine (Phila Pa 1976).2016;41(23):1837–1844.
- 28↑
Fortin M, Dobrescu O, Jarzem P, et al.. Quantitative magnetic resonance imaging analysis of the cervical spine extensor muscles: intrarater and interrater reliability of a novice and an experienced rater. Asian Spine J. 2018;12(1):94–102.
- 29↑
Vasavada A, Ward S, Delp AL, Lieber RL. Architectural design and function of human back muscles. In: Pepper D, LoGiudice B, eds.The Spine. 6th ed. Elsevier;2011:54–69.
- 30↑
Nolan JP Jr, Sherk HH. Biomechanical evaluation of the extensor musculature of the cervical spine. Spine (Phila Pa 1976).1988;13(1):9–11.
- 31↑
Hou X, Lu S, Wang B, et al.. Morphologic characteristics of the deep cervical paraspinal muscles in patients with single-level cervical spondylotic myelopathy. World Neurosurg. 2020;134:e166–e171.
- 33↑
Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res. 2005;19(1):231–240.
- 35↑
DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837–845.
- 36↑
Machino M, Ando K, Kobayashi K, et al.. Postoperative kyphosis in cervical spondylotic myelopathy: cut-off preoperative angle for predicting the postlaminoplasty kyphosis. Spine (Phila Pa 1976).2020;45(10):641–648.
- 37↑
Tamai K, Suzuki A, Yabu A, et al.. Clinical impact of cervical imbalance on surgical outcomes of laminoplasty: a propensity score-matching analysis. Clin Spine Surg. 2020;33(1):E1–E7.
- 38↑
Sakaura H, Ohnishi A, Yamagishi A, Ohwada T. Differences in postoperative changes of cervical sagittal alignment and balance after laminoplasty between cervical spondylotic myelopathy and cervical ossification of the posterior longitudinal ligament. Global Spine J. 2019;9(3):266–271.
- 39
Lee TT, Manzano GR, Green BA. Modified open-door cervical expansive laminoplasty for spondylotic myelopathy: operative technique, outcome, and predictors for gait improvement. J Neurosurg. 1997;86(1):64–68.
- 40
Kawaguchi Y, Matsui H, Ishihara H, et al.. Surgical outcome of cervical expansive laminoplasty in patients with diabetes mellitus. Spine (Phila Pa 1976).2000;25(5):551–555.
- 41
Handa Y, Kubota T, Ishii H, et al.. Evaluation of prognostic factors and clinical outcome in elderly patients in whom expansive laminoplasty is performed for cervical myelopathy due to multisegmental spondylotic canal stenosis. A retrospective comparison with younger patients. J Neurosurg. 2002;96(2)(suppl):173–179.
- 42
Kim HJ, Moon SH, Kim HS, et al.. Diabetes and smoking as prognostic factors after cervical laminoplasty. J Bone Joint Surg Br. 2008;90(11):1468–1472.
- 43↑
Iizuka H, Shimizu T, Tateno K, et al.. Extensor musculature of the cervical spine after laminoplasty: morphologic evaluation by coronal view of the magnetic resonance image. Spine (Phila Pa 1976).2001;26(20):2220–2226.
