INTRODUCTION
Ossification of posterior longitudinal ligament (OPLL) and degenerative spondylosis are common causes of myelopathy, and posterior cervical fusion (PCF) is one of the most widely performed procedures when multilevel treatment is necessary with laminoplasty and anterior cervical fusion surgery3,13,15). The reported incidence of pseudarthrosis in PCF ranges from 3.5% to 28%9,12). However, contrary to cases in lumbar fusion surgery and anterior cervical fusion surgery, the clinical significance of pseudarthrosis in PCF is frequently underrated. As interest in posterior cervicothoracic deformity surgery increases and with a better understanding of the PCF procedure, various factors such as bone mineral density (BMD) and several radiologic parameters that affect the fusion rate in PCF have been reported to date. OPLL is an abnormal lesion caused by heterotrophic OPLL and can lead to cord compression via narrowing of the spinal canal10,14). However, it is known that the genes involved in OPLL formation, such as COL6A1, bone morphogenetic protein (BMP), and transforming growth factor-β, also play key roles in bone formation and maintaining collagen structure and bone equilibrium5). Thus, in this study, we hypothesized that the presence of OPLL may positively support fusion in PCF and sought to determine if the presence of OPLL influences the fusion rate.
MATERIALS AND METHODS
1. Patient Selection
This research is a retrospective cohort study that was approved by the Institutional Review Board of our institution. Patients included in the study were those with myelopathy symptoms with modified Japanese Orthopaedic Association (mJOA) score of less than 17 and three or more lesion points in the cervical spine due to cervical OPLL or spondylosis, thus requiring PCF surgery. From 2011 to 2019, we examined patients who underwent lateral mass screw fixation (LMSF) C3 to C6 and decompression surgery. We also included patients who had radiographs (including dynamic X-ray and computed tomography [CT]) performed at 12 months post-surgery. Exclusion criteria were patients who had previously undergone other cervical surgeries; patients who received concurrent anterior cervical discectomy and fusion (ACDF); patients with cervical deformity (C2-7) sagittal vertical axis (SVA) >4 cm, chin-brow vertical angle <10° or >25°, T1-slope-cervical lordosis >15°, according to criteria by Ames et al.1), pathological, infectious, or traumatic conditions; patients with incomplete medical records; and patients who used fusion assist materials such as BMP or demineralized bone matrix.
2. Surgical Procedures
Surgeries were performed in a prone position under general anesthesia. From C3 to C6, the entry site was set 1 mm medial to the center of the lateral mass, and screws were inserted 25 degrees superolateral to the entry point using the Magerl technique. All decompressions were carried out with total laminectomy at the fixation points (C3-6), followed by fixation using rods and set screws. Posterolateral fusion was performed by decorticating the lateral mass and using only autologous chips obtained from laminectomy. Postoperatively, a semi-rigid neck collar was worn for three months.
3. Radiological Measurement
By examining medical records, patient information was obtained. OPLL was defined as heterotrophic OPLL thicker than 2 mm Anteroposterior, lateral, flexion-extension radiographs and CT scans were reviewed for radiological evaluation. Using imaging software (GE Medical Systems, Milwaukee, WI, USA; Centricity Enterprise Web V 3.0), all radiologic parameters were measured. The C2-7 angle was defined as the cervical lordosis angle (cervical lordosis, angle subtended by lines drawn along the posterior vertebral bodies of C2 and C7). Then, the C2-7 range of motion (ROM) was calculated as the difference between the angles formed by the C2 and C7 vertebrae during cervical extension and flexion. The distance between the C2 plumb line and the posterior C7 upper endplate was defined as the C2-7 SVA. Pseudarthrosis was defined as the absence of radiographic ossification of all fusion surfaces, a radiolucent gap around the hardware, or the presence of instrument failure in one year postoperative CT scan (Fig. 1, 2). Solid fusion was determined based on bone bridge formation and minimal segmental motion on dynamic X-ray (Fig. 3, 4).
4. Clinical Outcome
We measured the mJOA score, neck disability index (NDI) and visual analogue scale (VAS) score for neck pain before and at one year postoperatively.
5. Statistical Analysis
The radiologic and clinical data of the groups were compared using the student’s t-test and the χ2 test. The descriptive data are shown as the mean standard deviation. The Fisher's exact test was utilized to compare proportional data with small occurrences, such as disease type. All statistical analyses were performed using the IBM SPSS (version 25.0; IBM Corp., Armonk, NY, USA), with a significance level of p-value less than 0.05 for all analyses. In addition, the odds ratio (OR) was calculated for dichotomous variables to describe the strength of the relationship between the risk factors and pseudarthrosis. Multivariate logistic regression models for pseudarthrosis were developed to investigate the relative contributions of the risk factors.
RESULTS
1. Comparison of the Fusion and Pseudarthrosis Groups (Table 1-3)
Eighty patients were included in this study (39 patients with spondylosis, and 41 patients with OPLL (Table 1). There were no significant differences in sex, age, or BMD between the patients with spondylosis and the patients with OPLL. However, patients with OPLL have smaller ROM (15.9 ± 10.5°) than patients with spondylosis (23.5 ± 12.3°; p=0.004). One year after surgery, 18 patients (22.5%) showed pseudarthrosis and belonged to the pseudarthrosis group, while 62 belonged to the fusion group. At two years follow-up of the pseudarthrosis group, 3 patients showed fusion, 8 patients still showed pseudarthosis and 7 were lost for follow-up. The mean age was 67.2 ± 10.6 years in the pseudarthrosis group and 67.7 ± 9.7 years in the fusion group (p=0.855). However, BMD T-score was significantly lower in the pseudarthrosis group (-1.9 ± 0.7) compared to the fusion group (-0.6 ± 1.3; p=0.003) and the incidence of pseudarthrosis was lower in patients with OPLL than in patients with only spondylosis (OPLL vs. spondylosis, 12.2% vs. 33.3%; p=0.024) (Table 2). Segment by segment analysis of fusion rate in pseudarthrosis group showed that pseudarthrosis occurred mostly in the C5-6 segment (Table 3).
2. Comparison of Radiologic Parameters (Table 4)
The preoperative C2-7 ROM and change in cervical lordosis after surgery were greater in the pseudarthrosis group compared to the fusion group (pseudarthrosis vs. fusion, C2-7 ROM, 26.7 ± 13.3° vs. 17.6 ± 10.9°; p=0.004; change of cervical lordosis after surgery, -8.1 ± 7.9° vs. -2.7 ± 7.9°; p=0.013). There were no significant differences between the two groups in preoperative and postoperative cervical lordosis and C2-7 SVA.
3. Comparison of Clinical Outcomes (Table 4)
While preoperative VAS, NDI, and mJOA scores and 1-year postoperative NDI and mJOA scores did not differ significantly between the two groups, the 1-year postoperative neck VAS score was higher in the pseudarthrosis group (3.3 ± 0.9) than in the fusion group (2.2 ± 0.9; p<0.001). However, no patients required revision surgery due to pseudarthrosis.
4. Risk Factors for Pseudarthrosis (Table 5)
In the univariate logistic regression analysis to identify risk factors for pseudarthrosis, absence of OPLL (OR, 3.600 [1.142-11.348]; p=0.029), lower BMD (OR, 0.528 [0.312-0.894]; p=0.018), larger preoperative cervical ROM (OR, 1.069 [1.017-1.124]; p=0.009), and a larger decrease in postoperative cervical lordosis (OR, 1.097 [1.017-1.183]; p=0.017) were found to increase the OR. In the multivariate logistic regression analysis, three variables were found to significantly predict pseudarthrosis: Absence of OPLL (OR, 11.192 [1.114-112.451]; p=0.04), lower BMD (OR, 0.198 [0.064-0.618]; p=0.005), and a larger decrease in postoperative cervical lordosis (OR, 1.222 [1.009-1.480]; p=0.041).
DISCUSSION
In this study, pseudarthrosis occurred in 22.5% of cases. The reported rates of pseudarthrosis in PCF range from 3.5% to 28%, with an average of 4.8%. Our findings align with those in the literature4). Pseudarthrosis is a significant complication that affects clinical outcomes, such as the need for revision surgery. Associated risk factors of cervical fusion surgery reported to date include low BMD T-scores, a larger number of fused vertebrae, large preoperative C2-7 SVA, and low cervical lordosis10,25).
Previous studies by Sohn and Chung19), Yamauchi et al.22) have demonstrated that OPLL patients have a higher BMD than their counterparts. Furthermore, Hyun et al.6) reported that in patients who underwent laminoplasty for cervical myelopathy, patients with OPLL showed greater postoperative decrease in ROM, which may be attributed to greater spontaneous laminar fusion. Our study's primary objective was to determine whether the presence of OPLL itself influences the fusion rate. In our study, patients with OPLL showed a statistically significantly higher fusion rate compared to patients without OPLL.
High levels of high-sensitivity C-reactive protein in serum may indicate that inflammation can influence the progression of OPLL, according to a study by Kawaguchi et al.7). In addition, serum concentrations of the formation regulator, periostin, were found to be elevated and numerous investigations have reported results suggesting a genetic link to ectopic bone formation in ligaments17,21,24). We postulate that such factors may have contributed to a greater fusion rate in OPLL. In this study, we also found lower BMD as a significant predictor of pseudarthrosis using multivariate logistic regression. Despite the fact that OPLL patients tend to have higher BMD and lower pseudarthrosis rates, our study helped establish that the presence of OPLL may serve as an independent factor contributing to increased fusion rate in PCF8,16,22).
Larger preoperative ROM or larger decreases in cervical lordosis were also found to be associated with pseudarthrosis. Specifically, univariate and multivariate analyses revealed that these are risk factors for pseudarthrosis. Choi et al.2) reported that a substantial preoperative ROM is associated with pseudarthrosis in ACDF, it is less clear whether ROM is a risk factor in PCF2). We hypothesize that this is a result of the residual anterior disc motion that persists after fixation of the posterior element and the facet joint potentially interfering with fusion. In our study, patients with OPLL had a smaller ROM compared to patients without OPLL. This may be because some patients experienced a reduction in ROM due to partial fusion caused by OPLL. However, in the multivariate logistic regression test, larger ROM was not an independent risk factor of the pseudarthrosis. Several studies have reported correlations between mechanical failure and cervical deformities in terms of T1 slope, cervical lordosis, and SVA10,18). A significant decrease in cervical lordosis was observed in the pseudarthrosis group, suggesting a correlation between cervical imbalance and pseudarthrosis due to mechanical failure. In our study, there was an increased risk of pseudarthrosis in cases where there is an absence of OPLL, a larger preoperative ROM in the cervical spine, an anticipated postoperative increase in the SVA, or low BMD. In such patients' surgeries, surgeons should consider rigorous osteoporosis treatment and additional surgeries such as ACDF or pedicle screw fixation rather than LMSF only. In clinical comparisons, we observed higher postoperative VAS scores for the neck in the pseudarthrosis group, suggesting that pseudarthrosis may affect patients' neck pain.
The retrospective design and limited sample size of this study may hinder the generalizability of its findings. Future study is required to evaluate the long-term clinical outcomes and the influence of interventions on these risk factors on the fusion rate. It has also been argued that OPLLs should be categorized into continuous and segmental types, depending on whether they are genetic or not21). Continuous (including mixed) type OPLL is reported to have higher expression of the endochondral ossification gene encoding osteocalcin, alkaline phosphatase, and type I collagen than segmental (including localized) type, so it grows faster and has a different age of onset11,20,23,24). Future studies on the fusion of OPLLs may yield more meaningful results if genetic factors are also considered.
CONLUCSION
In conclusion, our study found that an absence of OPLL, lower BMD, larger changes in cervical lordosis, and larger preoperative cervical ROM were significant predictors of pseudarthrosis. These findings may help in better patient selection and preoperative planning, as well as in providing appropriate patient counseling regarding the risks and benefits of the procedure.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
