Abstract
Study Design
Retrospective cohort study
Objectives
To describe the clinical characteristics and surgical outcomes of patients with multilevel-ossification of the posterior longitudinal ligament (mT-OPLL), and to identify risk factors for unfavorable outcomes.
Methods
Patients who were diagnosed with mT-OPLL and underwent one-stage thoracic posterior laminectomy combined with selective OPLL resection, spinal cord de-tension, and fusion surgery between August 2012 and October 2020 were recruited. Patients’ demographic-, surgical- and radiological-related parameters were collected and analyzed. Neurological status was evaluated with mJOA score, and recovery rate (RR) was calculated using the Hirabayashi formula. According to RR, patients were divided into a favorable outcome group (FOG, RR ≥50%) and an unfavorable outcome group (UOG, RR <50%). Univariate and multivariate analyses were used to compare the difference between the 2 groups and to identify risk factors for unfavorable outcomes.
Results
A total of 83 patients were included, with an average age of 50.6 ± 8.3 years. Cerebrospinal fluid leakage (60.2%) and transient neurological deterioration (9.6%) were the most common complications. The average mJOA score improved from preoperative 4.3 ± 2.2 to 9.0 ± 2.4 at the last follow-up, and the mean RR was 74.9 ± 26.3%. Disease duration, preoperative nonambulatory status, and the number of decompressed levels were identified as potential risk factors by Univariate analysis (all P < .05). Multivariate analysis showed that the preoperative disease duration and nonambulatory status were independent risk factors for unfavorable outcomes.
Conclusions
Long disease duration and nonambulatory status before surgery were independent risk factors for unfavorable outcomes.
Keywords
Introduction
Multilevel ossification of the posterior longitudinal ligament in the thoracic spine (mT-OPLL) is a special type of OPLL, which is particularly challenging to treat. It has been reported that the high number of OPLL levels is significantly associated with the long recovery period of postoperative motor palsy.1,2 Although various surgical methods, such as laminoplasty or laminectomy, 3 posterior decompression, and fusion with or without concomitant dekyphosis,4-7 anterior decompression,8-10 2-stage posterior and anterior decompression,11-14 and circumferential decompression,12,15 have been conducted to treat this disease; however, it is not always possible to achieve favorable results, which is mainly due to the high incidence of perioperative complications and the low neurological recovery rate (RR).1,16
Unlike the cervical or lumbar spine, the thoracic spine’s kyphosis limits the spinal cord’s posterior shift, and when mT-OPLL ventrally compresses the spinal cord, the longitudinal tension of the spinal cord increases and the spinal cord becomes more fragile. In such cases, posterior decompression and fusion combined with dekyphosis was considered to be the optimal choice. Accordingly, in our previous study, we introduced a new “de-tension” guide surgical strategy for mT-OPLL, achieving an average neurological RR of 71.3%. 17 However, there was still no improvement in some cases, or aggravation of symptoms was observed after surgery. To date, the factors affecting the surgical outcomes of mT-OPLL are still unclear.
Thus far, few studies have investigated the risk factors for unfavorable outcomes of T-OPLL (including single-level and multi-level lesion), where most large sample researches were multicenter studies. The bias caused by patient selection, surgical methods, and surgeon experience was inevitable, which may hinder reaching a consistent conclusion. Also, no studies have specifically focused on the risk factors for unfavorable outcomes of mT-OPLL. Therefore, the present study aimed to investigate the surgical results and complications of mT-OPLL patients who underwent one-stage thoracic posterior laminectomy combined with selective OPLL resection, spinal cord de-tension and fusion surgery (TPDF) in our single-center, and to identify factors affecting the surgical outcomes.
Methods
A total of 83 consecutive patients diagnosed with mT-OPLL who underwent TPDF surgery in our single center between August 2012 and October 2020 were included. Indications for surgery were progressive gait disturbance, lower limb muscle weakness, and sphincter dysfunction. The inclusion criteria were as follows: (1) OPLL involved 3 or more levels continuously; (2) TPDF surgery was performed; (3) medical records were completed; (4) patients were followed up for at least 2 years. This study was approved by the institutional review board of the Peking University Third Hospital (LM2022238). Informed consent was obtained from all patients.
Patient Demographics and Clinical Characteristics
The demographic data of patients, such as age, gender, disease duration (from the onset of the initial symptom to surgery), body mass index (BMI, kg/m2), and data on comorbidities such as hypertension, diabetes mellitus (DM), and history of cervical/thoracic/lumbar surgery were collected. Surgical details, including operative time, estimated blood loss, and blood transfusion, were recorded.
Radiologic Parameters
All radiologic parameters were determined based on computed tomography (CT) and magnetic resonance (MRI) images. Data were measured and calculated by 2 independent spine surgeons (CGH and FTQ). The number of OPLL levels, OPLL segment distribution (defined as upper thoracic spine (T1-T4), middle thoracic spine (T5-8), and lower thoracic spine (T9-12)), OPLL canal occupying ratio (measured at the maximal stenosis level), concomitant cervical OPLL (C-OPLL)/thoracic ossification of the ligamentum flavum (T-OLF), and decompressed and fused levels were recorded. Pre- and post-operative T1-12 thoracic kyphosis angle (TK) and segment kyphosis angle in the fusion area (FSK) were measured on CT images to calculate the degree of dekyphosis in the sagittal plane. The shortened distance (SD) of the posterior edge of the vertebral body in the circumferential decompression (CD) segment was measured according to our previous reports.
17
The configuration of OPLL was classified into two types according to the criteria proposed by Matsuyama et al.
18
ie, the beak type and the flat type (Figure 1). The presence of intramedullary high-intensity areas (HIA) was detected on T2-weighted MRI and was graded as 0/1/2. Two types of OPLL. (A) Flat type. (B) Beak type.
Neurological Function and Surgical Outcome Assessment
The preoperative manual muscle test (MMT) score for the lower extremities, numbness status, non-ambulatory status, and bladder and bowel disturbance status were recorded. Ambulatory status was defined as walking (unaided or with support), and nonambulatory status as described by Imagama et al. 13 Pre- and post-operative neurological status were evaluated according to a modified Japanese Orthopedic Association (mJOA) scale (maximum 11 points). The neurological recover rate (RR) was calculated using the Hirabayashi method (RR% = (postoperative JOA score – preoperative JOA score) × 100/(11- preoperative JOA score)). The surgical outcomes were assessed by RR and were graded as excellent (75% - 100%), good (50-74%), fair (25-49%), unchanged (0% - 24%), or deteriorated (<0%). Patients were divided into 2 groups according to RR, ie, the favorable outcome group (FOG, RR ≥50%) and the unfavorable outcome group (UOG, RR <50%).
Statistical Analysis
All statistical analyses were conducted using IBM SPSS Statistics 27.0 (Armonk, NY, USA). Descriptive statistics (mean ± standard deviation and frequency) were used to summarize demographic data. Two-tailed independent t test or Mann-Whiney U test was used for continuous variables, and the Chi-square test or Fisher exact test was used for categorical variables. The univariate and multivariate analyses of various clinical, surgical and radiological parameters were performed to identify factors affecting the surgical outcomes. Factors with P < .05 or with clinical significance were included in multivariate logistic regression analysis. Odds ratios and 95% confidence intervals (CI) were estimated for significant factors. A P < .05 was considered statically significant.
Results
Demographic Data
Baseline Demographic and Clinical Characteristics of mT-OPLL Patients.
BMI: body mass index, JOA: Japanese Orthopaedic Association, MMT: manual muscle testing, RR: recovery rate.
Radiologic Findings
Radiologic Characteristics of mT-OPLL patients.
OPLL: Ossification of the posterior longitudinal ligament, OLF: Ossification of the ligamentum flavum, HIA: High intense area, TK: thoracic kyphosis, SD: shortened distance of posterior edge of the vertebral body in circumferential decompression segment.
Complications
There were 76 perioperative complications in 58 cases, achieving a rate of 69.9% in all cases. These patients experienced 1 or more complications. The most common complication was CSF leakage, which occurred in 50 patients (60.2%). CSF pseudocysts, which were observed in 2 patients, were treated with revision surgery. Among 8 patients who suffered transient neurological deterioration after surgery, 7 recovered after conservative treatment, and 1 suffered permanent motor paralysis due to a second surgical site and pulmonary infection, even with debridement surgery. The neurological function of 4 patients did not change or even deteriorate after surgery, where 2 patients had severe thoracic myelopathy preoperatively and did not recover after surgery. Neurological function improved immediately in 1 patient after surgery but deteriorated after atlantoaxial surgery and debridement for Staphylococcus aureus infection (Figure 2), and 1 patient developed neurological deficits even though there was no evidence of spinal cord injury during surgery. One patient developed a distal fixed vertebral fracture and motor paralysis 3 weeks after the operation and underwent revision surgery to extend the fixed segment with good results (Figure 3). Another revision surgery was performed because of the distal fixed vertebral fracture 7 months after surgery. The remaining complications included epidural hematoma (1 case) and surgical site infection (2 cases) (Table 3). A 46-year-old female with T5-7 OPLL. (A) Preoperative X-ray of the whole spine. (B, C) Preoperative CT and MRI show T5-7 continuous OPLL and T2-12 multilevel OLF. (D) Postoperative CT shows T1-L2 expansive laminectomy and T6-7 OPLL resection. (E, F) Postoperative X-ray and CT at the 18-month follow-up showed that T6-7 was partially fused. (G, H) Lateral X-ray and CT of the cervical spine show atlantoaxial dislocation, continuous OPLL at C2-T1 level, and internal fixation in place. (I) A cervical MRI shows spinal cord compression at the occipitocervical junction. (J-K) Postoperative X-ray and CT show that partial atlas and odontoid processes were resected, the atlas and axis were restored, and the occipitocervical fixation was good. (L, M) Postoperative X-ray and CT at 65-month follow-up show a good fusion of the T6-7 segment. A 52-year-old male with T1-7 OPLL. (A) Preoperative X-ray of the thoracic spine. (B) Preoperative CT shows T1-7 continuous OPLL. (C) Preoperative MRI shows thoracic spinal stenosis and spinal cord compression. (D, E) Postoperative X-ray and CT shows that T3-4 OPLL was resected. (F) Postoperative CT at 3-week follow-up shows T7 vertebral fracture. (G) Axial CT shows T7 vertebral transverse fracture. (H, I) Postoperative X-ray and CT show that the screws of the T7 vertebral were removed, and the internal fixation was extended to T9. (J-K) Postoperative X-ray and CT at 59-month follow-up showed good fusion of T3-4 and T7 without internal fixation failure. Perioperative complications. CSF: cerebrospinal fluid.
Surgical Outcomes
All patients were followed up for at least 2 years (mean, 40.2 ± 11.7 months). The mJOA score improved from 4.3 ± 2.2 before surgery to 9.0 ± 2.4 at the last follow-up, and the mean RR was 74.9 ± 26.3. According to the Hirabayashi classification, 48 cases were rated as excellent, 29 as good, 1 as fair, 2 as unchanged, and 3 as deteriorated. There were 77 patients in FOG and 6 patients in UOG.
Univariate and Multivariate Analysis of Factors Affecting Surgical Outcomes
Analysis of factors affecting the surgical outcomes.
BMI: Body mass index, JOA: Japanese Orthopaedic Association, MMT: Manual muscle testing, OPLL: Ossification of the posterior longitudinal ligament, OLF: Ossification of the ligamentum flavum, HIA: High intense area, TK: thoracic kyphosis, SD: shortened distance of the posterior edge of vertebral body in circumferential decompression segment, RR: recovery rate, CSF: Cerebrospinal fluid. DISH: Diffuse idiopathic skeletal hyperostosis.
aStudent t test.
bFisher exact test.
cMann-Whitney test.
d P value less than .05.
Multivariate logistic regression analysis of risk factors affecting the surgical outcomes.
BMI: Body mass index, MMT: Manual muscle testing, JOA: Japanese Orthopaedic Association, OPLL: Ossification of the posterior longitudinal ligament
aP value less than .05.
Discussion
MT-OPLL is a special type of OPLL that compresses the spinal cord from the ventral side at multiple levels. It has been identified as the main cause of thoracic myelopathy. The treatment of such a disease is quite challenging, considering that the thoracic spine is naturally kyphotic and the spinal cord at this site is fragile due to its avascularity. Previously, various surgical procedures have been employed to manage thoracic OPLL; however, favorable outcomes cannot always be achieved, and no consistent surgical protocol has been established, especially for mT-OPLL. To the best of our knowledge, only a few studies have specifically addressed this disease, 2 probably because it is difficult for single-center to collect many surgical cases for such a rare disease. While multicenter studies or systematic reviews provide solutions for such conditions, it is difficult to draw consistent conclusions because of the inconsistency in surgical methods, surgeon experience, and surgical indications. In a multicenter study by Matsumoto et al. 19 the surgical outcomes of various surgical methods were compared, revealing that the surgical outcomes of circumferential decompression tended to be more favorable than other methods. Nevertheless, there are still different opinions on the choice of surgical methods that could balance adequate nerve decompression and increased surgical complications.4,20-23 To address this issue, we have previously introduced a new “de-tension” guided surgical strategy for mT-OPLL, and good results have been achieved in most cases, but the neurological function remains unchanged or even worsens in some cases. 17 The reason for the unfavorable outcomes remains unclear. In this study, we utilized the data of 83 consecutive patients to elucidate the clinical characteristics of mT-OPLL and surgical outcomes of patients undergoing TPDF surgery and to identify the risk factors affecting the surgical outcomes. The reported results further the understanding of this rare disease and could help surgeons to make appropriate surgical plans when dealing with this disease, thereby improving the prognosis of mT-OPLL patients.
Clinical Characteristics of mT-OPLL Patients
In the present study, we found that the number of female patients undergoing surgery was much higher than that of male patients (65:18). Although previous studies have reported a higher prevalence of thoracic OPLL in female patients, 24 such a large gender ratio difference in patients undergoing surgery has not been explored. It remains to be further clarified whether mT-OPLL is more common in women and whether female patients are more prone to develop symptoms and require surgical intervention after suffering mT-OPLL. Moreover, 39.8% of patients had C-OPLL, and 78.3% had OLF, which is consistent with previous studies. Liang et al. 25 reported that 38% of TOPLL had C-OPLL, and 53% of TOPLL had T-OLF. In another study, Fujimori et al. 26 reported that more than half of individuals with TOPLL had C-OPLL, and 46% had OLF. These results indicate that tandem ossification was common in patients with mT-OPLL and that great attention should be paid to the whole spine to address the possibility of ossification of other ligaments when dealing with this disease.
Complications
The treatment of mT-OPLL was associated with a high incidence of complications. In this case series, nearly 69.9% of patients had at least one complication, and the overall rate was higher than in most previous studies. In a nationwide multicenter study, Imagama et al. 1 reported 88 perioperative complications in 59 cases, resulting in a rate of 51.3% in all cases. Xu et al. 16 Conducted a systematic review of the complications of OPLL in thoracic spine surgery and found that the average incidence was 39.4%. The reason for the high incidence of complications in this study was that 50 patients experienced CSF leakage, which accounted for the majority of complications (60.2%). Still, the CSF leakage rate was high, partly due to dural ossification, resulting in a dural tear during the removal of the ossification mass. The other reason may be that the dural injury was caused by unintended durotomy, in which no CSF leakage was observed during the operation, but a large amount of pale blood was drained after the operation.
In addition to CSF leakage, transient neurological deterioration was the most common complication. It was detected in 8 patients, resulting in an incidence of 9.64%, which was much lower than previous reports. E.g., Imagama et al. 1 reported that 26.3% of patients suffered neurological deterioration just after surgery, while Hu et al. 2 Found 34.6% neurological deterioration in mT-OPLL patients. Although the neurological function of most patients improved during the final follow-up, the high incidence of this complication indicated that the surgery for mT-OPLL was a relatively high-risk task. Some other complications, such as CSF pseudocyst, pulmonary infection, and SSI, have also been observed. A common feature of these complications was that they were all accompanied by CSF leakage. Although we could not conclude from the later univariate analysis that CSF does affect the prognosis of patients with mT-OPLL, it is an objective fact that patients with CSF leakage are prone to many postoperative adverse events. These results highlight the importance of proper management of dura tears during operation. Another interesting finding was that 2 patients underwent revision surgery to extend the fixed segment because of the distal fixed vertebral fracture, one of whom also suffered motor paralysis 3 weeks after surgery. It is possible that when long-segment laminectomy and internal fixation fusion were performed, the increased stress on the distal fixed vertebra led to vertebral fracture, which resulted in spine column instability, repeated spinal cord irritation, and eventually motor paralysis. Saiwai et al. 21 reported that a lower number of instrumented spinal fusion levels than decompression levels were exacerbating factors for the neurological improvement in T-OPLL surgery. Although favorable outcomes were achieved in both cases, the results suggested that instrumentation length should be equal to or one vertebral longer than the decompression segments, which was necessary for good outcomes.
Surgical Outcomes
Since previous studies reported that most neurological recovery peaked within 1-2 years after surgery, 1 we mainly assessed the neurological function of patients who were followed up for at least 2 years. Our results showed that the mJOA score significantly improved from 4.3 ± 2.2 before surgery to 9.0 ± 2.4 at the last follow-up, and the mean RR was 74.9 ± 26.3. Unfortunately, as only a few studies have specifically investigated the surgical outcomes of mT-OPLL, it was not easy to make an effective comparison with previous studies. Nevertheless, the results of this study were superior to most of them, as many previous studies reported RR ranging from 24.7% to 77.6%,2,6,16 regardless of the type of OPLL and surgical methods. However, there were still 6 patients with RR <50%, 2 with no change in neurological function, and 3 with a deteriorated condition. Nevertheless, these results demonstrated the critical importance of exploring factors that could influence surgical outcomes and making evidence-based decisions to reduce such complications.
Factors Affecting the Surgical Outcomes
To date, factors affecting the surgical outcomes of mT-OPLL have not been studied. Although various factors, such as longer preoperative duration of symptoms, 20 lower JOA score, 20 major intraoperative blood loss,2,20 concomitant OLF,16,20 use of instrumentation, 19 beak-type OPLL, 18 and large kyphosis angle on preoperative MRI 27 have been reported to be associated with unfavorable outcomes, there are still no consistent conclusions. The main reason is that mT-OPLL is a relatively rare disease, and there is no consistent surgical protocol to guide the surgery. In addition, the surgical experience of surgeons from different institutions may also hinder drawing convincing conclusions. To our knowledge, this is the first study that systematically compared the demographic, surgical, and radiological characteristics of mT-OPLL patients between FOG and UOG. Disease duration, preoperative nonambulatory status, and the number of decompressed/fused levels were identified as potential risk factors for poor prognosis. Still, when these variables and some other clinically significant factors, such as the number of OPLL levels, preoperative JOA score, and preoperative MMT in lower extremities, were included in the multivariate analysis, only the disease duration and preoperative nonambulatory status reached statistical significance, and were determined as independent risk factors for unfavorable outcomes.
OPLL levels have been reported to be associated with poor surgical outcomes.1,2 However, in the present study, no significant difference was observed in OPLL levels between FOG and UOG, which may be because, in previous studies, mT-OPLL was mainly treated by either indirect decompression or multilevel circumferential decompression via a posterior approach. The former method is less effective because mT-OPLL still compresses the spinal cord, while the latter is relatively invasive and has more complications. Therefore, these approaches led to surgical outcomes that were not always desirable. Takahata et al. 15 have reported that multilevel circumferential decompression of 5 or more vertebral levels is associated with unfavorable surgical outcomes. Our surgical strategy included a limited OPLL resection to reduce the surgical approach’s invasiveness. We also performed dekyphosis and spinal column shortening to reduce spinal cord tension. We believe that a balance between reducing surgical invasiveness and sufficient spinal cord decompression has been achieved.
Preoperative lower JOA score has been reported to be a risk factor for poor clinical outcomes. 20 Yet, in the present study, there was no significant difference in preoperative mJOA between the two groups. Moreover, preoperative MMT was not statistically significant. Interestingly, we found that many patients had very low preoperative MMT and mJOA scores in FOG but showed great improvement in postoperative neurological function. While some other patients in the UOG had the same preoperative neurological status, their postoperative neurological function did not change or even worsen. In addition to the secondary SSI and revision surgery, which may affect the neurological function, we speculate that the capacity of the patient’s neurological function may be an important factor affecting the surgical outcomes. As reported by Kim et al. 28 the decrease in functional recovery capacity is important in a poor clinical course. Therefore, we suggest that the disease duration and preoperative nonambulatory might be used as sensitive indicators reflecting the ability of neurological recovery.
Strengths and Limitations
To our knowledge, this is currently the largest single-center study that specifically characterized the clinical features of mT-OPLL. The surgical outcomes, perioperative complications, and risk factors for the unfavorable outcomes were discussed in detail. Nevertheless, there are some limitations. Firstly, as the number of patients in UOG (6 cases with RR <50%) was very small, the power of statistical analysis was limited, especially for the multivariate analysis. Secondly, although the bias caused by surgical methods and surgeon experience could be effectively avoided, the diversity and heterogeneity of patients cannot be guaranteed, which may have a certain impact on our analysis results. A future nationwide multicenter study with a consistent surgical strategy for mT-OPLL is needed to further confirm these findings.
Conclusions
The present study provides detailed information on the treatment of mT-OPLL with TPDF. Although favorable results were achieved in most cases, patients with long disease duration and nonambulatory status before surgery were prone to unfavorable outcomes. These results highlight the importance of early diagnosis and surgical timing for mT-OPLL. Future large-scale multicenter studies are needed to further confirm these findings.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Natural Science Foundation of China (82072479, 81772381, 81874031), Talent Incubation Foundation of Peking University Third Hospital (Y77477-03) and China Postdoctoral Science Foundation (2022M720293).
