SURGERY VERSUS CONSERVATIVE MANAGEMENT: TIMING OF SURGERY

Knowledge of the natural history of CSM is critical in decision making for treatment of CSM. However, few studies and no Level I evidence studies are available. Early studies on the course of CSM suggest that most patients with myelopathy experience progressive neurological worsening, most commonly with episodic deterioration.^2–4 In contrast, another long-term study of myelopathic patients by Lees and Turner⁵ shows that a long period of nonprogression was the rule and progressive deterioration was the exception. A small cohort study of 24 patients, treated by collar immobilization and followed for a mean duration of 6.5 years, found that the conditions of approximately one third of patients improved, one third deteriorated, and one third were stable.⁶ Similar results were confirmed by Nurick.⁷ In the majority of CSM cases, there is an initial phase of deterioration, followed by a stable period lasting a number of years, during which the degree of disability does not change significantly for those mildly affected. In general, older patients and those with motor deficits are more likely to experience development of progressive deterioration. As a result, Nurick⁷ states that surgery should be reserved for those with progressive disability and those older than 60 years. Patients with milder disease may have a better prognosis.⁸ However, some authors state that patients treated conservatively show progressive neurologic deterioration.^9,10 In addition, patients with CSM may be at increased risk for spinal cord injury after minor trauma,¹¹ which supports early intervention for even mildly symptomatic patients. Few direct comparisons of conservative and operative treatment in patients with myelopathy exist even in the recent literature. A nonrandomized cohort study comparing medical and surgical treatment in 43 patients, 23 of whom were treated without surgery, reported significant worsening of their ability to perform activities of daily living, with worsening of neurologic symptoms.¹² In contrast, 20 surgically treated patients had a significant improvement in functional status and overall pain with improvement also observed in neurologic symptoms. Although this study demonstrates that the results of surgical treatment are better than those of conservative management, it lacks randomization and had significant treatment selection bias. In the 2002 Cochrane review on the role of surgery for cervical myelopathy,¹³ one small study reports 49 patients with mild or moderate myelopathy who were randomized to surgery versus conservative treatment. Although age and sex ratios were similar between the two groups, the conservative group had slightly better modified Japanese Orthopaedic Association (mJOA) scores, suggesting a possible bias in treatment allocation. At 6 months, mJOA scores and gait scores were better in the conservatively treated group. But no differences were reported at 2 years. In addition, a subgroup with severe disability improved after surgical intervention.¹³ A more recent 3-year, prospective, randomized study of 68 patients with mild-to-moderate nonprogressive CSM did not demonstrate a significant difference in outcomes (mJOA score and self-evaluation) between surgically and nonsurgically treated patients.¹⁴ In addition, timed 10-m walk in the nonsurgical group was significantly better than that in the surgical group. The authors conclude that their results could mean that the conservative approach can treat CSM with a degree of success similar to that of surgery for at least 3 years, supporting rather than proving the “wait and see” strategy.¹⁴ This study failed to prove that surgical intervention has any advantage over conservative management. However, the poor specificity of the mJOA scoring scale, the small number of randomized patients, and relatively short follow-up period for this disorder (3 years) may account for the apparent lack of any lasting beneficial effect of surgery on the natural history of cervical myelopathy. In addition, the potential risk for spinal cord injury after minor trauma, which is impossible to quantify, should be considered when conservative management is selected.

Kadanka and colleagues’¹⁵ more recent study of a subclass analysis demonstrated that the patients with a good outcome in the surgically treated group had a more serious clinical picture (expressed in mJOA score and slower walk). They conclude that surgery is more suitable for patients who are clinically worse and have a spinal canal transverse area of less than 70 mm².¹⁵ Once moderate signs and symptoms develop, patients are less likely to improve on their own and would likely benefit from surgical intervention.¹⁶

One crucial question is what group of patients will experience development of CSM and clinically pro-gress. Bednarik and investigators¹⁷ conducted a prospective cohort study of clinically asymptomatic cervical cord compression cases (66 cases). Patients were managed to determine which patients experienced development of clinical signs of cervical myelopathy. During a median 4-year follow-up period, clinical signs of myelopathy were detected in 13 patients (19.7%). These signs were also associated with symptomatic cervical radiculopathy, electromyographic evidence of an anterior horn lesion, and abnormal somatosensory-evoked potentials.

Little information is available on nonoperative treatment of cervical myelopathy. Cervical collars have been recommended for symptomatic relief but have no effect on long-term outcomes, including neurologic progression. Although a recent study demonstrated the effectiveness of “rigorous” conservative management including 3- or 4-hour cervical traction, cervical orthosis, drug therapy, and exercise therapy,¹⁸ manipulation and traction are generally considered to be contraindicated because of the potential for aggravation of neurologic symptoms.²

SELECTION OF THE SURGICAL APPROACH

Selection of the appropriate surgical approach is based on a complete understanding of the factors responsible for the cord dysfunction. Because abnormal radiographic findings are common even in asymptomatic patients,^17,19 clinicians need to be careful to correlate the patient’s complaints, physical examination, and imaging results to discern the precise diagnosis. The purpose of surgery is to decompress neural elements, restore lordosis, and stabilize the spine to prevent additional degeneration at the affected level. Surgery for CSM has been performed by anterior, posterior, and combined approaches; each has unique advantages and disadvantages. The decision of which surgical approach to use is based on multiple factors, including the source of spinal cord compression, the number of vertebral segments involved in the disease process, cervical alignment, the magnitude of coexisting neck pain, patient comorbidities, and the surgeon’s familiarity with various techniques. Therefore, the selection of surgical approach remains controversial, especially in patients with multilevel degenerative disease.

In general, primarily ventral pathology causing cord compression, particularly if it is focal rather than contiguous over multiple levels (Fig. 8-1), is best treated via an anterior approach.²⁰ An anterior approach provides for direct visualization and removal of the offending pathologic lesion without manipulation of the cord. When a neutral or kyphotic cervical sagittal alignment is present, anterior procedures may also serve to restore physiologic lordosis. Restoration of lordosis allows for shifting of the cord dorsally to diminish the effect of anterior compression. After anterior decompression, spinal column stability is restored through segmental arthrodesis, which may have the added benefit of eliminating painful motion from the spondylotic motion segment.²¹ In contrast, if the compression is posterior and related primarily to facet hypertrophy or buckling ligamentum flavum, posterior decompression should be considered.⁹ However, the optimal surgical approach for the treatment of cervical myelopathy resulting from stenosis at three or more levels remains controversial. In addition to the earlier discussion, some other principles can assist in the selection of the appropriate approach, although an element of the surgeon’s preference will also be involved.²² An anterior arthrodesis of three or more motion segments is associated with a greater incidence of nonunion and graft-related problems than one- or two-level procedures.^23–25 A posterior procedure, including laminectomy with or without fusion or laminoplasty, can achieve an indirect decompression and may be an excellent alternative to anterior decompression if the spine is lordotic²¹ (Fig. 8-2). If the spine is kyphotic, however, a posterior approach may be contraindicated because the spinal cord cannot displace dorsally from the anterior compressive structures, and a ventral approach is indicated. If the patient’s spine is straight, either procedure can be used.¹⁰

FIGURE 8-1 Sagittal T2-weighted magnetic resonance imaging scan shows a large herniated cervical disc at C6-C7 causing spinal cord compression.

FIGURE 8-2 Sagittal T2-weighted magnetic resonance imaging scan demonstrates multilevel degenerative disease with spondylotic changes spanning from C2 to T1 resulting in spinal canal stenosis.

Other considerations include patient’s age and overall medical condition. Anterior surgery is more prolonged, and patients with multiple levels will be more likely to have dysphagia and voice problems after surgery; therefore, posterior surgery may be preferable if the alignment is not kyphotic for an older or a high medical risk patient.²²

Although the optimal selection of surgical approach for CSM involving three or more motion segments in the presence of a lordotic sagittal alignment remains controversial (Fig. 8-3), few comparative studies are available. In 1985, one study compared the results of laminectomy versus corpectomy and anterior cervical discectomy and fusion (ACDF) for the treatment of multilevel CSM, and demonstrated that multilevel corpectomy has a significantly greater rate of recovery and decreased late neurologic deterioration compared with laminectomy.²⁶ These results disfavor laminectomy mainly because of the well-known sequelae, such as segmental instability, kyphosis, and perineural adhesions.^27–36 In Japan, relatively poor outcomes associated with cervical laminectomy led to the evolution of laminoplasty, which remarkably reduced the sequelae after laminectomy.^37–40 A radiographic analysis of cervical alignment after laminectomy and laminoplasty in humans demonstrated that segmental kyphosis was present in 33% of laminectomy patients and 6% of laminoplasty patients at an average of 6 years after surgery.⁴¹ A later study compared the results of laminoplasty with subtotal corpectomy⁴² and reported that there were no significant differences in either the maximum JOA recovery rate or the final recovery rate for the two procedures. However, a notable difference was the overall incidence rate of complications (29% for corpectomy and 7% for laminoplasty). A similar retrospective study⁴³ showed that both procedures had similar rates of maintained neurologic improvement, and the disadvantages noted for the two procedures were pseudarthrosis (26%) and asymptomatic adjacent segment degeneration in a majority of patients after corpectomy, and decreased range of motion and axial discomfort (40%) after laminoplasty. A more recent prospective cohort study demonstrated that patients undergoing both procedures enjoyed a similar degree of subjective and objective neurologic improvement, but the incidence of complications was significantly greater for patients in the corpectomy cohort, especially persistent dysphagia and dysphonia.⁴⁴ The authors conclude that laminoplasty may be the preferred method of treatment for multilevel cervical myelopathy in the absence of preoperative kyphosis. Another interesting observation of this study was that the laminoplasty cohort tended to require less pain medication at final follow-up than did the multilevel corpectomy cohort.⁴⁴

FIGURE 8-3 Sagittal T2-weighted magnetic resonance imaging scan shows multilevel degenerative disease with spondylotic changes at C3-C4 and from C5 to C7 causing spinal cord compression.

Surgical indications of the combined approach to CSM are limited. A combined anterior and posterior approach may be considered for the following patients: (1) those with both anterior and posterior compression, which is difficult to treat with a single approach; (2) those with a loss of lordosis (either straightening or kyphosis) in addition to multilevel (three or more disc spaces) severe anterior compression from an osteocartilaginous spur or ossification of posterior longitudinal ligament (OPLL)^45,46; or (3) those with severe osteoporosis, diabetes, or heavy nicotine use who have a poor rate of spinal fusion with ventral surgery alone.⁴⁶ Some studies demonstrate that a single-stage combined procedure maximizes the decompression and reduces the graft-related surgical complications, as well as reducing perioperative complications when compared with staged combined procedures.^45–47

ANTERIOR APPROACH

Anterior decompression and fusion has been widely applied to the treatment of cervical stenosis resulting from herniated discs, spondylosis, or OPLL. When the pathology occurs at the level of the intervertebral disc, ACDF provides sufficient access to the stenotic focus. Clinical studies have demonstrated that successful arthrodesis can be achieved in 92% to 96% of patients after single-level ACDF with satisfactory clinical results.^21,48–50 Selection of surgical procedure (multilevel discectomy vs. cervical corpectomy) in multilevel pathology is controversial.^51,52 Few studies directly compare multilevel ACDFs and corpectomy. In general, if compressive lesions are present at the level of the disc only, either a single-level or a multilevel ventral discectomy should be performed. However, if the lesion is behind the vertebral body, a corpectomy should be performed.

The advantage of performing a multilevel ACDF instead of a single-level or multilevel corpectomy and fusion is that with the multilevel discectomy, postdecompression segmental fixation can be achieved by placing screws into the intervening vertebral bodies. It is easier to restore sagittal balance after a multilevel ventral cervical discectomy as opposed to a cervical corpectomy by using this strategy.⁵³ However, the disadvantage of multilevel ACDFs is that the incidence of nonunion increases with the number of levels being fused, although the rate of neurologic improvement remains high for multilevel ACDFs.⁵⁴ The recently reported fusion rate for a one-level ACDF using autograft iliac crest without plating (two graft–host sites) was 96%.⁴⁹ This decreased to a 75% fusion rate when a two-level ACDF was performed and to a 56% fusion rate with a three-level ACDF (six graft–host fusion sites).⁵⁵

Corpectomy is an alternate option to improve the fusion rate after multilevel anterior decompression. Only two points must fuse in a corpectomy, as compared with multiple surfaces that must fuse with multilevel ACDFs. The additional advantage of a single-level or multilevel cervical corpectomy is the ability to decompress lesions behind the vertebral bodies.⁴⁶ The studies that compared the fusion rate after multilevel ACDF and corpectomy reported the nonunion rate in corpectomy was 7% to 10% compared with 34% to 36% for patients undergoing multilevel ACDFs.^23,25 Based on these factors, corpectomy may be considered preferable to multilevel ACDF, especially in higher-risk patients, such as smokers, patients with diabetes, or revision cases.²¹