Factors Predicting Poor Surgical Outcome in Patients with Thoracic Ossified Ligamentum Flavum – Analysis Of 106 Patients in a Tertiary Care Hospital in South India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.360929
Source of Support: None, Conflict of Interest: None
Keywords: Decompressive laminectomy, myelopathy, ossification of ligamentum flavum, thoracic spine
Ossification of the ligamentum flavum (OLF) has been recognized as a common cause of thoracic myelopathy; it leads to acquired thoracic spinal canal stenosis. It is comparatively more prevalent in the Japanese population. However, nowadays, the incidence of thoracic OLF (T-OLF) in the Indian population has been increasingly recognized.
Several surgical procedures have been developed for the management of OLF, but the ideal approach is posterior decompression by laminectomy. Surgical treatment of T-OLF remains one of the most challenging problems for spinal surgeons. The OLF compresses the spinal cord tightly, thereby rendering the surgical decompression risky. In severe cases, the dural ossification associated with the OLF makes surgery more difficult.
The recovery rate varies from 25% to 100% even with adequate posterior decompression in patients with T-OLF. Spinal surgeons remain unclear regarding surgical outcomes and their related factors in patients with T-OLF because of the very few clinical studies done on this condition.,, In patients with T-OLF surgery, prognostic factors are still unpredictable, and it is very difficult for the surgeon to predict postoperative recovery.
The purpose of the current study was to evaluate the effectiveness of posterior surgical decompression and to determine factors predicting surgical outcome in patients with T-OLF.
The present study aimed to examine the factors predicting poor surgical outcome and the effectiveness of posterior decompressive laminectomy and OLF resection in patients with T-OLF.
We included 106 patients having T-OLF, who underwent decompressive laminectomy and OLF resection from 2007 to 2018 at our institute in this retrospective study. All patients with dorsal OLF who underwent surgery were included; those who presented with thoracic myelopathy caused by other than OLF were excluded in this study.
Preoperative diagnosis was established based on clinical symptoms, physical examination, and imaging studies. Computed tomography (CT) scan was done to confirm the presence of OLF and the type of OLF to look for the ossification of the dura. Magnetic resonance imaging (MRI) was used to detect signal changes in the spinal cord.
The following parameters were studied: sex, age, duration of symptoms, preoperative mJOA score, number of levels involved, intramedullary signal change on MRI, dural ossification, OLF type based on Sato classification, the combined spinal disorders, Dural breach during surgery, and cerebrospinal fluid (CSF) leak. All the patients were followed up till 6 months during the post-operative period.
Surgery: In all patients, a posterior decompressive laminectomy was performed by a single experienced surgeon under microscopic guidance. Initially, the laminectomy was performed by using a high-speed drill or an ultrasonic Bone scalpel system (MISONIX) until the ossified part is shelled out. The presence of Dural ossification limits the resection of OLF, and the process of drilling should be limited until a fragment of floating ossified dura was found which was left behind. The intraoperative Dural tears were managed using various methods such as repair using fascia, Dural substitutes (Duragen Plus, Adhesion barrier matrix, and Integra), and fibrin glue (TISSEEL Lyo, two-component fibrin sealant).
The modified Japanese Orthopedic Association (mJOA) scores for thoracic myelopathy and modified Nuric grade were used in neurological assessment both preoperatively and during follow-up at the end of 6 months. The postoperative neurological recovery was estimated based on the recovery rate. RR = (postoperative mJOA–preoperative mJOA)/(11– preoperative mJOA) × 100%.
The results of continuous measurements were presented as the mean ± standard deviation. Data were analyzed using paired t-test and ANOVA. P values of <0.05 and <0.001 were considered statistically significant and highly significant, respectively. Multiple regression analysis was conducted to know the correlation between factors and surgical outcomes. Statistical analyses were performed using the SPSS 16.0 software (SPSS Inc., Chicago, IL, USA).
In total, 106 symptomatic T-OLF patients were retrospectively evaluated. The mean age of the 106 patients (55 male and 51 female, ratio = 1.07:1) was 47.37 ± 9.5 years. The mean preoperative duration of symptoms was 15.7 months. The location of the OLF was upper thoracic spine (T1-T4) in 12 (11.3%) patients, middle thoracic spine (T5-T8) in 10 (9.4%) patients, and lower thoracic spine (T9-T12) in 84 (79.2%) patients. Imaging studies (MRI and/or CT) showed combined spinal pathology in 11 patients (10.3%): nine patients had cervical OPLL, and two patients had lumbar OLF. Out of 106 patients, 64 (60.37%) patients had a single/two-level, 38 (35.85%) patients had three or four levels, four (3.77%) patients had more than five levels dorsal OLF. Dural Ossification (DO) occurred in 23 cases, and the incidence was 21.7%. The mean preoperative mJOA score was 5.67 ± 2.13. The mean JOA score at the end of follow-up was 7.50 ± 2.60. The recovery rate was 43.29 ± 30.55% [Table 1].
[Table 2] shows a comparison of preoperative and postoperative mJOA scores, modified Nurick score, Ashworth grade, and effectiveness of surgical decompression. The overall mean preoperative and postoperative mJOA scores and modified Nurick scores were (5.67 ± 2.13, 7.5 ± 2.60) and (3.86 ± 1.01, 1.97 ± 1.16), respectively. The mean preoperative and postoperative Ashworth grades were 3.15 ± 0.97 and 1.28 ± 0.47, respectively. After surgical decompression, there were statistically significant differences between the preoperative and postoperative mJOA score, modified Nurick score, and Ashworth grade (P < 0.001).
[Table 3] shows the relationships between the pre- and post-operative neurological status and various factors. Paired t-tests revealed that there were statistically significant differences between the pre and postoperative mJOA score and sex, age, and duration of symptoms (<6M, 6M–12M), Type of OLF (lateral type), involved levels (<3, 3–4 levels), lower T-OLF involvement, absence of combined spinal disorder, absence of dural tear, and absence of CSF leak (P < 0.05).
Patients in whom the longer duration of symptoms (13M–24M), ≥5 involved levels of OLF, Tuberous type of OLF, as well as middle thoracic level OLF patients had lower postoperative scores. There were no statistically significant differences between the pre and postoperative mJOA scores and these factors (P > 0.05).
[Table 4] shows the results of multiple regression analyses of various factors to predict surgical outcome. The post-op mJOA score at follow-up and recovery rate correlated significantly with the age of the patient, trauma, intramedullary signal change on T2WI, number of OLF levels, type of OLF (tuberous type)and preoperative severity of myelopathy, preoperative mJOA score, preoperative Nurick score (P < 0.05). No correlation was identified with the preoperative duration of symptoms, dural ossification, dural tear, and CSF leak (P > 0.05).
Pre- and post-operative images
[Figure 1]: Sato classification of OLF based on morphology.
[Figure 2]: Thoracic spine T2WI showed T10–T12 OLF. A postoperative CT scan showed sufficient decompression of the spinal cord.
The first report of OLF was done in the late 1920s. The prevalence of OLF is more common in Asians and reported mainly in Japanese patients.,, OLF is the major cause of thoracic myelopathy in our locality. Surgical risks and unpredictable surgical outcomes are the main problems in the management of T-OLF. Surgical outcomes and their related factors for T-OLF remain unclear because of the less published data on this condition. It is very important to determine the effectiveness of surgical decompression and to determine various factors that are most closely related to the postsurgical prognosis of T-OLF. To our knowledge, this is the largest study of surgery for T-OLF reported to date.
According to the published data, there are conflicting results on gender ratio. Very few studies showed female preponderance; most studies have shown a male preponderance., Males (55, 51.9%) were more commonly affected in our study. Most of our study population were residents of rural areas; heavier physical activity in them caused more stress on the ligamentum flavum. T-OLF and was predominantly seen in the age group of 31–70 years according to different study series., In our study, the mean age of the patient was 47.37 ± 9.5 years (range: 41–50 years). No case was reported below 30 years of age. Myelopathy caused by T-OLF progress slowly.,, In our study series, the mean preoperative duration to develop initial symptoms was approximately 2 years (15.77 ± 9.36 months). As most of our patients belonged to the rural background they appreciate the symptoms only when they become very severe.
In most of the studies, the most commonly affected segment was the lower thoracic spine (T10-T12), whereas the middle thoracic spine (T5-T8) was rarely affected. Location-wise, our cases demonstrated that the lower thoracic spine was involved in 84 (79.2%) cases, whereas the middle thoracic spine was the least affected (10 (9.4%) cases). Based on the morphology, Sato et al. classified OLF into the following types: lateral type, extended type, enlarged type, fused type, and tuberous type. In our study, the most common type of OLF was lateral type (50, 47.2%) followed by tuberous type (29, 27.4%). A high-intensity signal in the spinal cord onT2-weighted MRI suggests edema, demyelination, myelomalacia, cavitation, or necrosis, and is related to the severity of the myelopathy. Our results showed that 21 (19.8%) patients presented with a high-intensity signal. The dura may be ossified together with ligamentum flavum, and when it occurred, the difficulty and risks of surgery increased significantly. It indicates the importance of diagnosis of dural ossification, ranging from 11% to 62% in various studies., In this study, dural ossification was seen in 23 (21.7%) patients; our results are comparable with previous studies. Specifically, more advanced grades such as enlarged, fused, and tuberous types are frequently associated with dural ossification (DO). Sun et al. reported DO in 86% of the tuberous types. Our results (78.2%) were similar to that reported by Sun et al. OLF was frequently accompanied by the ossification of other spinal ligaments. Cervical ossified posterior longitudinal ligament (OPLL) can coexist with OLF, which further worsens myeloradiculopathy. Inamasu et al. found a 5% coexistence of OLF and OPLL in their study. In our series, cervical OPLL was seen in nine (8.4%) patients. It is wise to perform an MRI of the whole spine as surveillance before any surgical intervention.
In symptomatic OLF patients, conservative treatment had no role., As the T-OLF compresses the spinal cord posteriorly, surgery is the only treatment that can provide sufficient decompression by excision of the ossified segments. Various surgical procedures such as fenestration, laminoplasty, hemilaminectomy, open door laminectomy, en bloc laminectomy, and laminectomy with fusion have been described in the literature.,,, Among all, laminectomy provides direct access and complete resection of the offending pathology.,
Surgical outcomes for OLF were unclear because of the few studies on this condition. In our study, the mJOA scores of all patients improved from 5.67 ± 2.13 to 7.50 ± 2.60, and the mean JOA RR improved to 43.29 ± 30.5% after surgical decompression. Our results are comparable to Zhong et al. and Wang et al. In their studies, the mJOA scores improved from 6.67 ± 1.73 to 8.63 ± 1.81 and from 6.82 ± 1.81 to 8.58 ± 1.62, respectively, and the mean JOA RR improved to 50.45% ± 24.37%. Li et al. in their retrospective study analyzed 85 cases and found that weakness and spasticity showed the best improvement after surgical decompression. In our study, the lower limb motor function of modified Nurick grade reduced from 3.86 ± 1.01 to 1.97 ± 1.16, and Ashworth's grade reduced from 3.15 ± 0.97 to 1.28 ± 0.47. The results of our study indicate that surgical outcomes of OLF-induced thoracic myelopathy were significantly (P < 0.001) improved after decompression.
Until now, very few studies examined relationships between the pre- and post-operative neurological status and various factors. In addition to this, in our study, the results of these analyses demonstrated significant differences in the relationships between the pre- and post-operative neurological status and various clinical, radiological, and intraoperative factors. In our study, we found statistically highly significant differences between the pre and postoperative mJOA scores of patients with ≤50 years old, patients with a shorter duration of symptoms (<6, 6–12 months), patients with single- or dual-level lesions (<3, 3–4 levels), lateral type of OLF, lower thoracic (T9-T12) OLF involvement, absence of combined spinal disorder, absence of dural tear, and absence of CSF leak (P < 0.001). There were no statistically significant differences between the pre and postoperative mJOA scores of patients with longer duration of myelopathy (13–24 months), multiple levels of OLF (≥5 levels), tuberous type of OLF, and middle thoracic level OLF (P > 0.05). These patient factors tended to have lower postoperative scores and were unlikely to modify the outcome after surgery.
At present, very few studies have proposed various clinical and radiological factors such as gender, age, duration of symptoms, preoperative JOA score, intramedullary signal change on T2WI, number of levels involved, and OLF type in predicting the postsurgical prognosis of patients with T-OLF.,, All these studies examined the correlation between these factors and JOA RR only. In our study, in addition to the JOA RR, we also examined the correlation between these factors and the post-op mJOA score.
Multiple linear correlation analysis of our study revealed that age of the patient, patient with associated trauma, intramedullary signal change on T2WI, type of OLF (tuberous type), number of levels involved, preoperative severity of myelopathy, preoperative mJOA score, and preoperative Nurick score were predicting the postsurgical prognosis of patients with T-OLF. Our results were comparable with those of previous studies evaluating the prognostic factors of T-OLF.,,,
Preoperative severity of myelopathy and preoperative JOA score were considered important predictors of the postoperative outcome. In our study, preoperative severity of myelopathy, the preoperative mJOA scores, and preoperative Nurick score correlated with post-op mJOA score (−0.694, 0.735, and −0.530, respectively, P = 0.0001) and RR (−0.396, 0.414, and −0.359, respectively, P = 0.0001) in the correlation analysis. Signal intensity (SI) changes of the spinal cord on MRI reflect pathological changes in the spinal cord. Zhong et al. reported a correlation between signal intensity changes in MRI and surgical outcomes in patients with T-OLF. Our results are comparable with Zhong et al.; patients with no signal intensity changes on MRI correlated with high JOA RR (P = 0.0001). The morphology of OLF may affect the post-surgical neurological status of the patients., In our study, we found patients with tuberous type of OLF (0.684, 0.332, P = 0.000) had a significant correlation with post-op mJOA score and recovery rate. Factors associated with good outcomes included a single-level disease and unilateral lesions. In our study, patients with single-level disease had a significant correlation with postop mJOA score (−0.211, P = 0.03), which is comparable with previous studies.
There was no significant correlation between the duration of symptoms and the surgical recovery rate in our study; the reason for this contradiction may be that most of our patients came after a long-term compression of the spinal cord, thus leading to irreversible damage. Thus, the preoperative duration does not significantly affect surgical recovery. According to the published data, coexisting spinal disorders make the surgical decision more complicated and gives unpredictable surgical outcome. Similar to the results reported by Liao et al., we did not find any significant correlation between the coexisting spinal disorders and the postoperative JOA score and RR (P = 0.5). In our study, we observed no correlation between Dural ossification (DO) and postoperative neurological recovery (P = 0.3). Our results are comparable with Sun et al. A recent study showed that dural tear and CSF leak are the predictive factors of postoperative neurologic deterioration in T-OLF surgery. However, we did not find any correlation between the post-op mJOA score and RR with these factors. This may be due to that these complications are not only affected by the patient's clinical and imaging characteristics but also by intraoperative factors.
Most of the previous studies in the literature had small patient cohorts and failed to examine and characterize possible complications that occur following T-OLF surgery. The decompressive surgical procedures mend for T-OLF are technically demanding. Dural tears and CSF leakage are the most common complications of T-OLF surgery and can lead to CSF fistula formation, pseudomeningocele, meningitis, arachnoiditis, epidural abscess, and wound dehiscence. In our study of the 106 patients, dural tears (22, 20.7%), CSF leakage (2, 1.8%), pseudomeningocele (2, 1.8%), wound site infection (2, 1.8%), and meningitis (1, 0.9%) were seen. The incidence of dural tears and CSF leaks are comparable with previous studies, in which the reported incidence was 11%–32%.,,, Patients with dural tears most of them showed ossified dura. According to the surgical strategies used in the present study, a floating fragment adherent to the dura may be left behind after drilling and freeing the sides in case ossified dura was encountered, which was already described by Wang et al. No postoperative spinal instability was observed in our study series. We performed a procedure in which medial one-third of the facet was eliminated. To avoid more complications, we recommended that the entire decompression procedure should be carried out under microscopic magnification with utmost care.
Thoracic ossification of ligamentum flavum is a progressive disease that causes spinal canal stenosis and serious neurological sequelae. OLF is a relatively common disease in our locality. It is important to identify preventable risk factors for poor surgical outcomes for T-OLF. Age of the patient, associated trauma, OLF level, tuberous type OLF, intramedullary signal change on T2WI, preoperative severity of myelopathy, pre-op mJOA score, and pre-op Nurick grade were important predictors of surgical outcome in our study series. No correlation was identified with the preoperative duration of symptoms, dural ossification, dural tear, and CSF leak.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]