Multiple thoracic disc herniations are rare and there are few reports in the literature. Between December 1998 and July 2002, we operated on 12 patients with multiple thoracic disc herniations. All underwent an anterior decompression and fusion through a transthoracic approach. The clinical outcomes were assessed using the Frankel neurological classification and the Japanese Orthopaedic Association (JOA) score. Under the Frankel classification, two patients improved by two grades (C to E), one patient improved by one grade (C to D), while nine patients who had been classified as grade D did not change. The JOA scores improved significantly after surgery with a mean recovery rate of 44.8% ± 24.5%. Overall, clinical outcomes were excellent in two patients, good in two, fair in six and unchanged in two. Our results indicate that anterior decompression and fusion for multiple thoracic disc herniations through a transthoracic approach can provide satisfactory results.
The incidence of thoracic disc herniation is reported to be one per million per year.1 It represents between 0.5% and 4.5% of all disc herniations and between 0.15% and 1.8% of all surgically-treated herniations.2–,5 Recently, some authors, including the senior author of this report (KS), have reported the rarer clinical entity of multiple thoracic disc herniations.2,4–,12
Most thoracic disc herniations are found in the lower thoracic spine with more than 75% occurring below T8, mainly at T11/12.1,8–10,13–,20 The aetiology is thought to be weakness of the posterior longitudinal ligament and increased mobility of the lower thoracic spine.4,9,14,21 In addition, in the thoracic spine, the ratio of the diameter of the spinal cord to that of the spinal canal is large and the blood supply in this region is limited.2,3,17,21–,23 This makes the spinal cord vulnerable to compression from disc herniation. The diagnosis of a thoracic disc herniation can be difficult as there is no characteristic pattern of the onset of symptoms.1,3,6,8,11,13,20,21,24–,26
Although some researchers have shown conservative treatment to be effective for patients without long-tract clinical signs,14,27 there is no doubt that surgical treatment is required for patients with myelopathy.1,3,9–14,21,23,26 In 1996, we began treating patients with a single-level thoracic disc herniation associated with myelopathy surgically. We used a transthoracic, anterolateral approach, which allows a wide exposure for decompression. Subsequently, we began to use this same procedure for patients with multiple thoracic disc herniations. The purpose of this study is to assess the clinical outcome of anterior decompression and fusion through this approach for the treatment of multiple thoracic disc herniations.
Patients and Methods
Between December 1998 and July 2002, we treated 12 patients (seven men and five women) with a mean age of 63.0 years (47 to 75) with multiple thoracic disc herniations (Table I⇓). The mean length of time between the onset of symptoms and the diagnosis was 46.9 months (eight days to 120 months). The symptoms varied, but included back pain, sensory changes, weakness in both lower limbs and urinary symptoms. Myelography, MRI and CT myelography were used both for diagnosis and to determine the specific disc levels which required decompression. A mean of three (two to four) disc herniations compressed the spinal cord at levels which ranged from T3/4 to T12 and L1. The most commonly affected levels were T7/8 and T9/10 (six patients). All patients underwent anterior decompression and fusion with a rib graft which was inserted through a transthoracic approach. The mean post-operative follow-up period was 32.6 months ± 11.0 (17 to 50).
The level of the thoracotomy was decided pre-operatively. The patient was placed in the lateral decubitus position with the centre of the lesion convex. After thoracotomy, the disc herniations and the relevant posterior vertebral bodies were resected with the aid of an operating microscope. An anterior decompression was performed until the dura could be seen. The resected rib was cut into struts and inserted into the decompressed area. For the first two patients, we inserted the bone graft longitudinally and without the use of instrumentation. Both developed a kyphotic deformity because of fracture of the rib graft. Consequently, we used spinal instrumentation for all operations after September 1999 (ten patients). With this technique, the vertebral endplates and discs are resected and the rib grafts are inserted transversely. For these ten patients, a mean of four (two to eight) levels were instrumented. As disc herniation of the upper lumbar spine occurred in two patients, curettage of these levels was also performed. After surgery, a drain was placed in the extra-pleural space and the wound was closed. When the pleural effusion had reduced to less than 100 ml per day, the drain was removed and the patients allowed to walk while wearing a rigid corset.
Assessment of clinical outcome.
The magnitude of surgery was measured by the operating time and by blood loss.
Pre- and post-operative symptoms were assessed by the Frankel neurological classification28 and the Japanese Orthopaedic Association score for thoracic myelopathy (JOA score, Table II⇓).29,30 The JOA score measures three functions; motor function in the lower limbs, sensory (in the lower limbs and trunk) and bladder function.
The JOA score recovery rate was calculated using the method of Hirabayashi et al:31
The Wilcoxon signed-rank test was used to compare the pre- and post-operative values of each parameter and of the total points. The JOA score was calculated at final follow-up. The recovery rate and the final results were classified into five groups: excellent (75% to 100%), good (50% to 74%), fair (25% to 49%), unchanged (0% to 24%) and worse (less than 0%).29
Magnitude of surgery.
The mean operating time was 430 minutes (329 to 645). The mean blood loss was 850 ml (160 to 3223) (Table I⇑).
Frankel neurological classification
Using the Frankel neurological classification, two patients improved from grade C to E and one patient from grade C to D. Nine patients, who had initially been classified as grade D, had no change (Table III⇓).
For all patients, the total JOA score increased after surgery (Table IV⇓). The improvements in each individual parameter of the JOA score were statistically significant. The mean increase in lower limb motor function was 1.3 ± 1.3 points (0 to 4, p < 0.05). The mean increase in sensory function was 1.2 ± 1.1 points (0 to 3, p < 0.05), bladder function was 0.5 ± 0.7 points (0 to 7, p < 0.05) and total JOA score was 3.0 ± 2.0 points (1 to 8, p < 0.01). The mean recovery rate for the total JOA score was 45.7% ± 26.6% (14.3 to 100). The overall results were excellent in two patients, good in two, fair in six and unchanged in two. No patient was classified as worse. The mean recovery rate for the JOA score for sensory function was the lowest at 40.9%± 29.9% (0 to 75) as compared with lower limb motor function, bladder function and total score. For sensory function, no patient had a complete recovery.
Post-operatively, pneumonia occurred in one patient, chylothorax in another and pain was present at the site of the skin incision in two patients (Table III⇑). The patients with pneumonia and chylothorax recovered with conservative treatment, while an intercostal nerve block was used to treat those with pain at the site of the incision.
The clinical symptoms of thoracic disc herniation vary. Some patients are asymptomatic, while others have midline thoracic pain, girdle pain or myelopathy. Because there are no characteristic symptoms and thoracic spinal disc herniation is relatively rare, the diagnosis can be difficult.1,3,8,11,13,20–,26 Our patients had a variety of pre-operative symptoms and for some there was a lengthy period between the onset of symptoms and diagnosis.
We found acceptable clinical outcomes in 12 consecutive patients with multiple thoracic disc herniations treated by anterior decompression and fusion through a transthoracic approach (Fig. 1⇓).20 Using the Frankel neurological classification, the pre- and post-operative scores for nine of the 12 patients were unchanged. However, because of their slight neurological deficits, each of these nine patients had been classified pre-operatively as grade D. Post-operatively, each was reclassified as grade D because of slight residual symptomatology. This does not mean that the surgical procedure was ineffective. Using the JOA score, sensory function was the most resistant to improvement. Its recovery rate was lower than the other JOA variables (lower limb motor and bladder functions), and a degree of sensory disturbance remained after surgery in all patients. We believe these results are satisfactory, although the patients had not fully recovered at final follow-up, if account is taken of the poor pre-operative scores and long clinical course of the condition.
Fujimura et al29 reported the surgical outcome of 33 patients with thoracic myelopathy caused by ossification of the posterior longitudinal ligament which were treated by anterior decompression. They found significant negative correlations between the clinical outcome and the duration of symptoms.29 They also reported 33 patients with thoracic disc herniation which had been treated by a trans-thoracic approach and concluded that the indicators of a poor prognosis included an age of 60 or more years, a long duration of symptoms and severe pre-operative myelopathy.32 In some of our patients, the period between the onset of symptoms and the time of diagnosis was relatively long. Awareness of the possibility of single or multiple thoracic disc herniations, in combination with physical and radiological assessments, would enable earlier recognition and improve the outcome of treatment.
In order to accomplish our surgical goals when treating patients with multiple thoracic disc herniations, a trans-thoracic anterior approach is necessary for two reasons. First, it provides sufficient exposure to perform a safe and extensive decompression of the spinal cord with bone grafting and the use of spinal instrumentation.6,11,12,16,20,22,23,33–,35 Secondly, a thoracic herniation frequently extrudes or protrudes into the central area of the spinal canal,1 where the conventional posterior36,37 or transpedicular17,38,39 approaches may not allow sufficient decompression.6,16,19,23,33 However, the extent of the transthoracic anterior approach for the thoracic spine compared with the posterior or posterolateral approaches cannot be ignored.3,11,13,23 Complications such as respiratory failure or pneumonia have been reported using this approach.3,11,13,23
Our results confirm that this surgical strategy provides satisfactory outcomes. Nevertheless, some limitations exist in the interpretation of our results. First, the number of patients was small and they were assessed retrospectively. In order to obtain conclusive proof of the benefits of an anterior approach for this surgery, randomised controlled trials are necessary. In order to focus on disabilities in these patients, a widely-accepted measure for these disabilities, such as the Oswestry score,40 could have been used. Spinal cord monitoring was not performed in all our patients. The accumulation and analysis of cord monitoring data would have provided valuable evidence to assess the immediate effect of anterior decompression and obtain an accurate diagnosis of the affected levels. However, we have been able to show that anterior decompression and fusion for the treatment of multiple thoracic disc herniations can provide satisfactory results without major complications.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received May 27, 2004.
- Accepted August 3, 2004.
- © 2005 British Editorial Society of Bone and Joint Surgery