Preoperative Disc Height as a Prognostic Factor in Laminoplasty for Cervical Spondylotic Myelopathy with Degenerative Spondylolisthesis

Abstract

Introduction

Laminoplasty (LP) is widely used to treat cervical spondylotic myelopathy (CSM) with degenerative spondylolisthesis (DS), but predictors of postoperative outcomes remain unclear. This study aimed to identify factors associated with neurological recovery after LP in patients with CSM and DS.

Methods

We retrospectively reviewed 136 patients who underwent LP for CSM between 2005 and 2020. Of these, 29 patients (20 men, 9 women) with DS at the responsible level were included. The responsible level was defined as the segment showing intramedullary signal change or cerebrospinal fluid space obliteration on magnetic resonance imaging. The mean age was 71.7 years (range: 50-87), with a mean follow-up of 5.5 years (range: 2-12). Patients were grouped on the basis of achievement of a ≥50% recovery rate (RR) in the Japanese Orthopaedic Association (JOA) score, defined as the minimum clinically important difference (MCID). Radiographic parameters, including C2-7 alignment, C2-7 range of motion (ROM), C7 slope, slip distance, segmental ROM, and disc height at the responsible level, in addition to Pfirrmann grade, were compared.

Results

The mean JOA RR was 50.2%, with 66% achieving MCID. No statistically significant differences were found between the two groups in terms of C2-7 alignment, C2-7 ROM, C7 slope, slip distance, segmental ROM, or Pfirrmann grade. However, preoperative disc height was significantly greater in the Good Improvement group (5.9 mm) than in the Poor Improvement group (4.8 mm) (p=0.03).

Conclusions

Greater preoperative disc height was associated with better neurological recovery, suggesting its potential utility as a simple prognostic marker in CSM with DS.

Introduction

Cervical spondylotic myelopathy (CSM), the most common cause of spinal cord dysfunction in adults, often progresses to neurological deterioration if untreated. When CSM is accompanied by degenerative spondylolisthesis (DS), it introduces additional mechanical instability, complicating surgical planning and making outcome prediction challenging. Although posterior fusion is often considered in such cases, several studies have reported postoperative stabilization at the slipped segment after laminoplasty (LP) alone¹⁾, and thus, LP without fusion remains a common surgical approach. Several radiological predictors of postoperative outcomes have been proposed. For example, Oichi et al.²⁾ reported that anterolisthesis ≥3 mm was an independent predictor of poor neurological recovery, whereas Sakai et al.³⁾ found that posterior spondylolisthesis was more problematic than anterior slips. Ninomiya et al.⁴⁾ indicated that patients with anterolisthesis and kyphosis had significantly lower recovery rates (RR) than did those with lordotic alignment. Moreover, sagittal imbalance, particularly a high T1 slope, has also been associated with poor outcomes due to compensatory mechanical strain on cervical segments⁵⁾.

In the lumbar spine, the degenerative cascade model suggests that advanced disc degeneration leads to disc height loss and segmental restabilization⁶⁾. However, in the cervical spine, the prognostic significance of preoperative disc height has not been thoroughly investigated. Given its critical role in intervertebral mobility, motion-induced anterior spinal cord compression, and overall cervical spine biomechanics, identifying the prognostic value of preoperative disc height may enhance patient selection and guide surgical planning. We hypothesized that preoperative disc height may serve as a simple and accessible radiographic parameter to support surgical decision-making in patients with CSM complicated by DS. This study aimed to determine whether preoperative disc height at the responsible level can serve as a predictor of functional recovery after LP in patients with CSM complicated by DS.

Materials and Methods

This retrospective study analyzed patients who underwent LP for CSM at our institution between 2005 and 2020. A total of 136 patients were initially identified. Patients were excluded if they had a follow-up period of less than 2 years, if pre- or postoperative Japanese Orthopaedic Association (JOA) scores were unavailable, if they had no DS, or if the level of spondylolisthesis did not correspond to the responsible segment. Ultimately, 29 patients (20 men, 9 women) were included in the final analysis (Figure 1).

Figure 1.

Patient selection flow diagram.

The responsible level was determined using T2-weighted magnetic resonance imaging (MRI), defined as the intervertebral segment showing intramedullary high signal intensity or complete obliteration of the cerebrospinal fluid space. Only cases in which the level of DS matched this responsible segment were included^7,8). Spondylolisthesis was defined as anterior or posterior slippage of 2 mm or more observed in at least one of the three dynamic radiographic views (neutral, flexion, or extension), as previously reported²⁾.

All patients underwent double-door LP using a modified technique originally described by Kirita and Miyazaki⁹⁾. A midline split was first performed along the central axis of each lamina using a high-speed air burr drill. Bilateral grooves were then created on both sides of the split. The laminae were symmetrically opened and maintained in the expanded position by suturing them to the deep fascia of the paraspinal muscles. Unlike the original technique, no local bone grafts were used to support the opened laminae. The operated levels varied by case: C3-6 in 19 patients, C4-6 in 5 patients, and other cervical levels in 15 patients. The semispinalis muscles of C2 and C7 were preserved as much as possible in all cases. The relationship between the responsible levels and the operated levels is summarized in Table 1.

Table 1.

Relationship between the Responsible Level and the Operated Levels.

Operated level Responsible level	C1-7	C2-6	C3-5	C3-6	C3-7	C4-6	C4-7
C3/4	0	1	1	6	0	0	0
C4/5	0	0	0	10	1	2	1
C5/6	1	0	0	1	0	3	0
C6/7	0	0	0	2	0	0	0

Radiographic parameters were evaluated preoperatively and postoperatively using lateral cervical spine radiographs. These parameters included C2-7 alignment, C2-7 range of motion (ROM), and segmental ROM, which was defined as the angular difference between flexion and extension. Slip distance at the responsible level was assessed in all three positions (neutral, flexion, and extension), and the maximum value among these positions was used for analysis. The C7 slope was measured in the neutral position both before and after surgery. Disc height was defined as the perpendicular distance from the center of the superior endplate of the lower vertebral body to the point of intersection with the inferior endplate of the upper vertebral body. It was measured at the responsible intervertebral level on midsagittal T2-weighted MRI images (Figure 2). All radiographic parameters were measured twice by two independent observers. Intra-observer reliability was assessed by calculating the intraclass correlation coefficient (ICC) between the first and second measurements of each observer. Inter-observer reliability was assessed by calculating the ICC using the mean values of the two measurements obtained by each observer. The final values used for the analysis were based on the mean. For qualitative grading of intervertebral discs, Pfirrmann grading¹⁰⁾ was also evaluated at the responsible intervertebral level on sagittal T2-weighted MRI images.

Figure 2.

Method of disc height measurement: Disc height was defined as the perpendicular distance from the center of the superior endplate of the lower vertebral body to the point of intersection with the inferior endplate of the upper vertebral body.

Clinical outcomes included the preoperative JOA score, postoperative JOA score, and the JOA score RR, which was calculated using the following formula:

(Postoperative JOA score−Preoperative JOA score)×100/(17−Preoperative JOA score). According to Kato et al.¹¹⁾, the minimum clinically important difference (MCID) for the JOA score in CSM is 52.8%. On the basis of this finding, patients with a RR of ≥50% were classified as the Good Improvement group and those with <50% as the Poor Improvement group. Clinical and radiographic parameters were compared in the two groups.

Statistical analysis

Continuous variables were compared using the Student t-test, and categorical variables were analyzed using the Chi-square test. Pearson's correlation coefficient was used to assess a linear relationship between continuous variables. A p-value <0.05 was considered statistically significant. All statistical analyses were performed using JMP software (version 17.0; SAS Institute, Cary, NC, USA). Continuous data are presented as mean±standard error.

Results

The mean JOA score RR was 50.2%. On the basis of the achievement of the MCID, defined as a JOA RR of 50% or higher, 66% of patients (19 of 29) were classified into the Good Improvement group, whereas the remaining 10 patients were categorized as the Poor Improvement group. The mean ages of the Good Improvement group and the Poor Improvement group were 71.3 and 72.5 years, respectively (p=0.76). The male-to-female ratios were 14:5 and 6:4 (p=0.45), and the mean follow-up durations were 5.3 and 6.0 years (p=0.56), respectively. There were no significant differences between the two groups in these baseline characteristics. The preoperative JOA score was significantly lower in the Good Improvement group than in the Poor Improvement group (10.2±2.3 vs. 11.7±1.5, p<0.05). The postoperative JOA score was significantly higher in the Good Improvement group than in the Poor Improvement group (14.7±1.2 vs. 12.7±2.3, p=0.02). Notably, the RR was remarkably better in the Good Improvement group, with a mean of 65.9%±14.8% compared with 20.4%±27.1% in the Poor Improvement group (p<0.001) (Table 2).

Table 2.

Intergroup Comparison of Demographic Data and Clinical Outcome.

Clinical outcome	Good Improvement group (n=19)	Poor Improvement group (n=10)	p-Value
Age (y)	71.3±2.2	72.5±3.0	0.76
Gender (M:F)	14:5	6:4	0.45
Follow-up periods (y)	5.3±0.7	6.0±0.9	0.56
Preoperative JOA score	10.2±2.3	11.7±1.5	<0.05*
Postoperative JOA score	14.7±1.2	12.7±2.3	0.02*
JOA recovery rate	65.9±14.8	20.4±27.1	<0.001*

F: female; JOA: Japanese Orthopaedic Association; M: male

*Statistically significant differences were observed in the preoperative JOA score, postoperative JOA score, and JOA recovery rate.

Radiographic parameters were compared in the two groups. The preoperative disc height at the responsible level was significantly greater in the Good Improvement group than in the Poor Improvement group (5.9±0.3 mm vs. 4.8±0.4 mm, p=0.03). However, no significant differences were observed in other radiographic measurements, including direction of spondylolisthesis (anterolisthesis vs. retrolisthesis), preoperative C2-7 alignment (14.6°±1.8° vs. 10.1°±6.1°, p=0.43), postoperative C2-7 alignment (8.9°±2.3° vs. 2.5°±8.4°, p=0.48), preoperative C7 slope (24.7°±2.0° vs. 23.8°±3.8°, p=0.84), postoperative C7 slope (25.8°±1.8° vs. 26.8°±3.3°, p=0.81), preoperative C2-7 ROM (39.8°±3.7° vs. 43.8°±4.3°, p=0.49), postoperative C2-7 ROM (33.5°±2.4° vs. 32.1°±5.5°, p=0.81), preoperative slip distance at the responsible level (3.1±0.2 mm vs. 3.5±0.2 mm, p=0.14), and postoperative slip distance at the responsible level (3.4±0.2 mm vs. 3.3±0.5 mm, p=0.88), preoperative segmental ROM at the responsible level (10.9°±1.3° vs. 7.9°±1.4°, p=0.14), postoperative segmental ROM at the responsible level (9.8°±1.6° vs. 7.0°±2.2°, p=0.32), and preoperative Pfirrmann grade (p=0.50), respectively (Table 3). There was no significant association between the surgical levels and either clinical outcomes or postoperative radiographic parameters. Furthermore, a moderate positive correlation was found between preoperative disc height and JOA RR (r=0.4, p=0.03), supporting a potential association between these two variables (Figure 3). These findings suggest that greater preoperative disc height might be associated with better postoperative functional outcomes. In contrast, other commonly evaluated sagittal alignment and mobility parameters showed no significant correlation. The intra-observer ICCs for radiographic and MRI measurements were 0.86 for examiner A and 0.89 for examiner B. The inter-observer ICC between examiners A and B was 0.85, indicating good reliability of the measurements.

Table 3.

Intergroup Comparison of Radiographic Parameters.

Clinical outcome	Good Improvement group (n=19)	Poor Improvement group (n=10)	p-Value
Preoperative disc height at responsible level (mm)	5.9±0.3	4.8±0.4	0.03*
Direction of spondylolisthesis (anterolisthesis:retrolisthesis)	11:8	5:5	0.68
Preoperative C2-7 alignment (°)	14.6±1.8	10.1±6.1	0.43
Postoperative C2-7 alignment (°)	8.9±2.3	2.5±8.4	0.48
Preoperative C7 slope (°)	24.7±2.0	23.8±3.8	0.84
Postoperative C7 slope (°)	25.8±1.8	26.8±3.3	0.81
Preoperative C2-7 range of motion (°)	39.8±3.7	43.8±4.3	0.49
Postoperative C2-7 range of motion (°)	33.5±2.4	32.1±5.5	0.81
Preoperative slip distance (mm)	3.1±0.2	3.5±0.2	0.14
Postoperative slip distance (mm)	3.4±0.2	3.3±0.5	0.88
Preoperative segmental range of motion at the responsible level (°)	10.9±1.3	7.9±1.4	0.14
Postoperative segmental range of motion at the responsible level (°)	9.8±1.6	7.0±2.2	0.32
Preoperative Pfirrmann grade (Grade 3:Grade 4:Grade 5)	5:11:1	2:6:2	0.50

*A significant difference was observed in the preoperative disc height at the responsible level.

Figure 3.

Correlation between preoperative disc height and JOA RR.

JOA: Japanese Orthopaedic Association; RR: recovery rate

Discussion

Greater preoperative disc height may predict better functional recovery after LP for CSM with DS. To the best of our knowledge, no previous studies have reported similar findings. Although an intergroup difference of approximately 1 mm may appear minimal, our intra- and inter-observer analyses indicated ICC values exceeded 0.8, indicating that the measurement method was sufficiently reliable. In addition, recent biomechanical studies have highlighted that even small changes in cervical disc height can substantially influence spinal mechanics. Wang et al.¹²⁾, in an in vitro investigation of total cervical disc replacement, showed that altering implant height by as little as 1 mm above the native disc height led to a measurable increase in facet joint pressures and changes in motion behavior. Notably, implants that were 2 mm taller caused marked elevation of facet joint loading and significant reduction in range of motion. Given that the physiological cervical disc height itself measures only a few millimeters, these results indicate that a difference as small as 1 mm can meaningfully modify biomechanical stresses affecting neural elements, facet joints, and intraspinal pressure. Accordingly, the 1-mm difference in preoperative disc height observed in our cohort may indeed exert a biologically relevant effect on the degree of neural decompression achieved and the subsequent neurological recovery―an interpretation supported by existing biomechanical evidence.

In the lumbar spine, according to the concept of the “degenerative cascade” proposed by Kirkaldy-Willis⁶⁾, progressive loss of disc height eventually leads to a stage of “restabilization.” Indeed, previous studies¹³⁾ have shown that patients with reduced disc height tend to exhibit minimal postoperative slip progression and generally achieve relatively favorable clinical outcomes. In contrast, whether the cervical spine follows a similar course remains controversial. Aoyama et al.¹⁴⁾ reported that cervical disc height narrowing does not necessarily suppress dynamic instability, suggesting that the mechanism of “disc collapse leading to stabilization,” commonly observed in the lumbar spine, may not be directly applicable to the cervical region. In the present study, we found that patients in the Good Improvement group showed greater preoperative disc height. These findings suggest that although the cervical spine shares certain degenerative features with the lumbar spine―such as intervertebral disc degeneration and spondylolisthesis―the factors contributing to surgical outcomes may differ. Cervical-specific pathological features, including dynamic spinal cord compression and preserved segmental mobility, may play a more significant role than slip stabilization alone.

One possible explanation is that preserved disc height may indicate not only structural integrity but also residual segmental mobility and the potential reversibility of spinal cord compression. Prior studies have shown that greater disc height correlates with dynamic instability and increased intervertebral motion¹⁵⁾, which can lead to intermittent anterior cord impingement, particularly during cervical extension. This dynamic compression may be amenable to posterior decompression, contributing to neurological improvement after laminoplasty. Kinematic MRI is effective in detecting such dynamic lesions, as revealed by Joaquim et al.¹⁶⁾ and Liu et al.¹⁷⁾, who reported that cervical extension MRI can reveal cord impingement not apparent in the neutral position. Lao et al.¹⁸⁾ further showed that greater disc height is associated with enhanced segmental mobility and more pronounced anterior bulging during extension. Consistent with these findings, our cohort exhibited a trend toward greater ROM at the responsible level in patients with higher preoperative disc height, although the difference was not statistically significant. In contrast, severely collapsed discs may indicate fibrotic or ankylosed segments with fixed deformity, when long-standing static compression could contribute to irreversible cord damage. Such cases may be less responsive to decompression. Taken together, these results support the interpretation that preserved disc height may serve as an indirect marker of dynamic anterior compression―a potentially reversible mechanism that may explain better surgical outcomes after laminoplasty in selected patients.

A statistically significant positive correlation was observed between preoperative disc height and JOA RR (r=0.40, p=0.03). However, the strength of the correlation was moderate, suggesting that disc height alone may not be a robust predictor of postoperative outcomes. Although this finding supports the group-based results, it should be interpreted with caution. Disc height may contribute to surgical outcomes as part of a multifactorial framework rather than as an isolated factor. Standardized preoperative examination protocols that include disc height measurement may help clinicians better identify patients likely to benefit from LP alone, without the need for additional fusion.

We also evaluated prognostic factors previously reported in the literature. Oichi et al.²⁾ identified anterior spondylolisthesis ≥3 mm as an independent predictor of poor outcomes; however, our analysis did not reproduce this association. This may be due to the presence of favorable factors such as preserved disc height in some anterior slip cases, variability in surgical indication, or differences in compression mechanisms. These findings suggest that the amount of anterior slippage alone may not consistently predict outcomes. Similarly, no significant difference in clinical outcomes was observed between anterior and posterior spondylolisthesis, in contrast to findings by Sakai et al.³⁾ and Ninomiya et al.⁴⁾, who reported poorer recovery in patients with anterior slippage―particularly when combined with kyphotic alignment. Given that only 4 cases of kyphosis alignment were included in our analysis, these findings may reflect insufficient statistical power. In contrast to prior predictors that vary by the direction of spondylolisthesis, our findings suggest that disc height at the responsible level may function as a direction-independent prognostic factor, potentially providing a more stable and reliable parameter for preoperative risk stratification.

This study has several limitations. First, the retrospective design introduces risks of selection and information bias. Although inclusion criteria were predefined, unmeasured confounders may have affected outcomes. Furthermore, given all patients underwent LP alone, those with greater instability or deformity may have been preferentially excluded, limiting generalizability. Second, the small sample size reduced statistical power and precluded multivariate regression analysis, limiting adjustment for potential confounders such as age, neurological severity, alignment, and comorbidities. Third, postoperative MRI was not routinely obtained in this study. As a result, we were unable to directly assess the degree of spinal cord decompression, posterior shift, or the presence of residual or recurrent compression. In addition, postoperative disc height could not be measured, limiting our ability to evaluate structural changes after surgery. These limitations may have influenced our interpretation of radiological and clinical correlations. Finally, the follow-up period varied widely from 2 to 12 years. This heterogeneity in follow-up duration raises the possibility that long-term neurological outcomes may have been influenced by factors unrelated to CSM, such as age-related decrease, comorbid conditions, or unrelated neurological disorders. However, because the follow-up duration did not differ significantly between the groups, its potential impact on group comparisons may be limited. Although dynamic changes in alignment could be assessed with flexion-extension radiographs, the behavior of the spinal cord and discs under dynamic conditions could not be evaluated owing to the absence of kinematic MRI. Future prospective studies with larger cohorts, standardized surgical indications, and dynamic imaging techniques such as kinematic MRI or upright MRI are needed to validate whether preoperative disc height independently predicts postoperative outcomes in patients with CSM complicated by DS.

Conclusions

This study highlights the potential of preoperative disc height as a reliable predictor of favorable surgical outcomes in patients with CSM with DS who undergo LP. Incorporating disc height into preoperative evaluation may refine surgical decision-making and improve outcome prediction. Although promising, these findings require validation in larger, multicenter cohorts to confirm their prognostic value and guide treatment strategies.

Author Contributions: Y.D.: Data curation, Data analysis, Writing - original draft. K.T.: Conceptualization, Methods, Formal analysis, Writing - review and editing. K.Y.: Supervision, Validation. K.Ko.: Supervision, Validation. H.S.: Supervision, Validation. K.Ka.: Supervision, Validation. Y.N.: Supervision, Validation.

Conflicts of Interest: The authors declare that there are no relevant conflicts of interest.

Ethical Approval: This study was approved by the institutional review board of Kyushu University Hospital (Approval Number 2023-78).

Informed Consent: Informed consent was obtained using an opt-out approach owing to the retrospective nature of the study.