Efficacy of Surgical Decompression in Regard to Motor Recovery in the Setting of Conus Medullaris Injury

Abstract

Background/Objective:

An assessment of neurological improvement after surgical intervention in the setting of traumatic conus medullaris injury (CMI).

Methods:

A retrospective evaluation of a cohort of patients with a blunt traumatic CMI from T12 to L1. The neurologic and functional outcomes were recorded from the acute hospital admission to the most recent follow-up. Data collected included age, level of injury, neurologic examination according to the Frankel grading system and motor index score, and the mechanism and timing of CMI decompression.

Results:

A total of 24 patients with a mean age of 27 years (men, 87%) were identified. The most common level of bony injury was L1, and the most frequent mechanism of injury was a motor vehicle crash. Before surgical intervention, 16 of 24 patients (66.7%) had a complete neurological deficit below the level of injury. The median interval from injury to surgery was 6 days (range, 7 hours to 390 days). Decompression, fusion, and adjunctive internal fixation were the most common surgical procedures. Median length of follow-up was 32 months after surgery. Improvement in spinal cord and bladder function was seen in 41.6% and 63.6% of patients, respectively. Root recovery was seen in 83.3% of patients.

Conclusions:

In the setting of CMI, no correlation between the timing of surgical decompression and motor improvement was identified. Root recovery was more predictable than spinal cord and bladder recovery.

INTRODUCTION

Blunt spinal trauma complicated by injury to the conus medullaris is a devastating event on a personal and family level, as well as a tremendous financial burden to society because of its attendant morbidity, expense, and prolonged treatment requirements (1). Traumatic spinal injury occurs most frequently in the young men with an average age of 35 years. The lack of controlled, prospective, multicenter clinical studies has contributed to confusion regarding optimal treatment methods for patients with injuries to the thoracolumbar spine. The transitional anatomy of the thoracolumbar junction makes it vulnerable to injury resulting from high-energy motor vehicle collisions and falls (2–4). Approximately 40% of patients with traumatic spinal cord injury (SCI) present with a complete SCI, 40% with an incomplete injury, and 20% with either no cord injury or only root lesions (5). Formulation of a treatment plan for patients with injuries to the thoracolumbar spine depends on the presence and extent of neurologic injury and existing spinal stability. Both nonsurgical and surgical treatment options are available to achieve the goals of preservation of neurologic function and restoration of spinal stability (6). To date, the role of decompression in patients with incomplete SCI is only supported by class 3 and limited class 2 medical evidence (2). Because of the absence of scientific literature examining injuries specific to the conus medullaris, a retrospective pilot study was undertaken to access the efficacy and potential morbidities related to the surgical management (decompression and stabilization) of these injuries. This study will serve as a foundation for future prospective, multicenter studies evaluating the safety and efficacy of surgical intervention in neurologically and mechanically unstable injuries to the thoracolumbar junction.

METHODS

From October 1994 through March 2005, a total of 108 patients with blunt traumatic SCI were identified at a regional level 1 trauma center in southeastern Iran. Of these patients, a subset was identified in which (a) a neurological deficit was attributable to a traumatic conus medullaris injury (CMI), (b) follow-up was a minimum of 6 months (7), and (c) the CMI was caused by an acute nonpenetrating traumatic event with radiographically documented cord compression caused by cord encroachment by anterior vertebral body elements, disk material, or posterior vertebral elements as a result of fracture subluxation or dislocation.

Patients were excluded if (a) their neurologic deficit was associated with a preexisting spinal cord abnormality or disease process (eg, multiple sclerosis or preexisting myelopathy as a result of severe spondylosis without trauma), (b) they could not actively participate in the follow-up neurologic examination process, or (c) there were inadequate follow-up data available.

Data Collection

Data collected included patient age, sex, associated injuries, mechanism of injury, admitting and follow-up Frankel grade and motor index score (MIS), time interval from injury to surgical decompression and stabilization, imaging studies documenting the spinal injury, and the type of surgical procedure.

Neurologic Evaluation

Motor and sensory examinations (Frankel grade and MIS) were performed at admission, daily during the acute hospitalization, before and after surgery, and at all follow-up outpatient encounters. Patients were assigned an initial MIS that included manual muscle test scores of all key muscles, sensory examination (prick and touch), sacral and deep tendon reflexes, and muscle tone evaluation. Sensory level was recorded as the most caudal dermatomal level of bilateral intact sensation.

Treatment

Standard spinal immobilization and resuscitation were implemented by emergency medical personnel. All patients were prescribed intravenous (IV) methylprednisolone (30 mg/kg IV bolus over 15 minutes followed 45 minutes later by a 5.4-mg/kg/h IV infusion over 23 hours) if they arrived to the emergency room within 8 hours of the accident (8). All patients underwent preoperative myelography, computerized tomography, and/or magnetic resonance imaging. Patients with image-documented spinal cord compression [from vertebral bony elements, spinal malalignment (subluxation or dislocation), or epidural hematoma] underwent surgical decompression and spinal column stabilization. The surgical approach was determined by the location of cord compression and the type and degree of spinal instability. Adequacy of decompression was determined by postoperative computed tomography and magnetic resonance imaging (9). Statistical data analysis was performed using SPSS-11.5 software applications. A P value of less than 0.05 was considered statistically significant.

Outcome Assessment

Outcome assessment was performed using the method of Bohlman and Anderson (10). A patient was considered to have an excellent result if they became a household or community ambulator or had marked improvement in ambulator status. A good outcome was recorded if there was recovery of one or more motor-root levels in the lower extremities or partial recovery of multiple levels. A fair result was recorded if there was partial improvement of 1 or 2 motor-root levels, and a poor result was shown by no cord or root improvement.

RESULTS

Twenty-four patients satisfied the inclusion and exclusion criteria for this study (Table 1). Before treatment, 16 of 24 patients (66.7%) had a functionally complete (Frankel A) neurological deficit below the level of spinal injury (Table 2). Mean patient age was 26.7 ±8.6 years, and 87.5% of the patients were men (Table 3). The most frequent level of bony injury was L1, followed by T12, and the most frequent mechanism of injury was a motor vehicle crash. The median time interval from injury to surgery was 6 days (range, 7 hours to 390 days). The length of follow-up after surgery ranged from 8 months to 12 years, with a median time period of 32 months (Table 3). The primary indications for surgery were documented spinal cord compression in the setting of a spinal cord neurologic deficit associated with instability. Posterior transpedicular/extracavitary decompression, followed by an intertransverse process fusion without anterior strut graft placement, and instrumentation 2 levels above and below the injury level were the most common surgical procedures performed (Table 2).

Table 1.

Summary of Patient Data

Table 2.

Patient Data By Frequency

Table 3.

Patient Data: Mean, Median, and Range

Follow-up employment status (listed from most to less frequent job description) was unemployed, member of office staff, housewife, driver, and laborer. Spinal cord functional improvement and nerve root recovery was seen in 10 of 24 (41.6%) and 20 of 24 (83.3%) patients, respectively. Overall, there was an average improvement of 1.5 Frankel grades among the study population. Patients with an initial Frankel grade of A improved an average of 1.6 Frankel grades. No patient had an initial Frankel grade of B. Patients with an initial Frankel grade of C improved an average of 1.2 grades. All patients with an initial Frankel grade of D improved to a Frankel grade of E (Table 4). Overall, there was an average MIS improvement of 14.5 points in the study population. Patients with an initial Frankel grade of A improved an average of 15.1 motor points. Patients with an initial Frankel grade of C improved an average of 16 motor points, and patients with an initial Frankel grade of D improved an average of 6.5 points (Table 4).

Table 4.

Average Improvement of Neurologic Function According to the Frankel Scale and MIS

Bladder function improvement was seen in 7 of 11 (63.6%) patients (Table 5). Partial or complete bladder function was documented at follow-up in 5 of 8 (62.5%) patients with an initial Frankel grade of A, one-half (50%) of patients with an initial Frankel grade of C, and 1 of 1 (100%) patients with an initial grade of D. Bladder function outcome was not adequately documented in the remaining 13 patients.

Table 5.

Bladder Function Improvement According to the Frankel Scale

Additionally, there was no significant correlation between time to spinal cord decompression and spinal cord or nerve root improvement according to the Frankel grading system and the MIS (Table 6).

Table 6.

Correlation of Time to Decompression and Neurologic Outcome According to the Frankel Grading System and MIS

Complications were recorded in 10 patients, including 6 cases of delayed instrumentation failure, 2 extended urinary tract infections, 2 cases of symptomatic pressure ulcers, and 1 case of severe neuropathic pain syndrome. The cases with instrumentation failure were not associated with recurrent or progressive spinal deformity and did not result in symptomatic neural compression.

DISCUSSION

Our study showed that, in cases of blunt traumatic CMI, some degree of spinal cord recovery was seen in 41.6% of the patients who underwent surgical intervention. Root recovery was seen in 83% of the patients. A precise understanding of spinal cord vs peripheral nerve root recovery was difficult because of the close proximity of the conus medullaris to the cauda equina and the common occurrence of nerve root escape. This made assessment of bowel and bowel function extremely valuable in understanding improvement in spinal cord function. Of the 11 patients assessed, 7 (63.6%) were noted to have improvement in bowel and bladder function.

The efficacy or futility of surgical intervention could not be determined in the absence of a control (non-surgical) group because of the retrospective nature of this study. A retrospective review of our blunt trauma patient population did not produce any patient treated non-surgically with a CMI injury and evidence of spinal instability. Our data, however, did show a defined neurologic outcome that can be expected if a surgeon decides to surgically decompress a compromised spinal cord in a posttraumatic patient with a CMI.

The efficacy of decompression after SCI in enhancing neurological recovery in animal models has been widely shown (7, 11–20). There are 8 prospective nonrandomized case series (class 2 evidence) (21–28) and several retrospective case series with historical controls (class 3 evidence) that have addressed the role of spinal cord decompression in the setting of a contused and compressed spinal cord. None have shown an advantage to surgery in the setting of a complete SCI.

An extensive review of the literature did not discuss any clinical reviews of patients with SCIs specific to the conus medullaris. The majority of studies were retrospective reviews of a mixture of thoracic and lumbar injuries with various degrees of spinal cord and peripheral nerve root injuries. In general, patients with an incomplete neurologic deficit often showed improved lower extremity motor and/or bladder function with either non-operative or operative intervention (29–41).

Boerger et al (42) performed a meta-analysis of the world's literature in 2000 to evaluate the effectiveness of surgical decompression in the context of a neurological deficit associated with a thoracolumbar burst fracture. The design of 9 studies was sufficiently similar to allow pooling of their results, all of which failed to establish a significant advantage of surgical over nonsurgical treatment in regard to neurological improvement. They found that patients with an incomplete neurological deficit who had undergone surgical decompression and stabilization experienced a better neurological recovery compared with patients who had undergone nonsurgical treatment (43). Comparison of surgical intervention using Harrington instrumentation and recumbence therapy showed that neurological improvement was much more predictable with surgical intervention in the complete paraplegia group from T9 to L2, but there was no difference in neurological recovery between the 2 groups in patients with incomplete paraplegia (44). However, Waters et al (27) showed motor recovery did not significantly differ between patients categorized in various surgical subgroups or between those having surgery and those treated nonoperatively. Geisler et al (45) concluded that the sparseness of prospective data on the treatment of traumatic spinal cord injury at 28 centers in North America suggests that treatment guidelines have limited empirical support and should be made cautiously.

To date, no definitive data exist that correlate the timing of surgery with neurologic outcome (32, 46–49). Clohisy et al (50) observed that early anterior thoracolumbar decompression for traumatic injuries at the thoracolumbar junction was associated with improved rates of neurologic recovery compared with late decompression in the presence of an incomplete neurologic deficit.

Transfeldt et al (34) showed that late decompression of more than 3 months resulted in neurologic improvement in 46.5% of patients with an incomplete post-traumatic spinal injury. If the surgery was performed less than 2 years after injury, neurologic improvement occurred in 68%, with an improvement in Frankel grades of 32%. Bladder function improved overall in 27% of patients, and if decompression occurred less than 2 years after injury, improvement occurred in 43% of patients. Conus medullaris decompression resulted in a 50% improvement in bladder function. There was an 83% improvement in the pattern of pain after spinal decompression. Because of the retrospective nature of our study, a correlation between the timing of neurologic compression and neurologic improvement could not be determined.

An intrinsic problem with our study is the small number of cases, which decreases the power of our study and prevents us from employing any meaningful statistical analysis. A true understanding of the role of surgical intervention in the setting of traumatic thoracic spinal cord can only be determined through a randomized, multicenter controlled clinical trial.

CONCLUSIONS

Partial spinal cord and bowel and bladder recovery was identified in this patient cohort. In the setting of CMI, no correlation between the timing of surgical decompression and motor improvement according to the Frankel grading system and the MIS was identified. Root recovery was more predictable than spinal cord recovery. Clearly, to better define the role of surgery in the management of acute SCI, randomized, controlled prospective trials are required.