Surgery for Cervical Spondylotic Myelopathy: Right Evidence or Evidence Right Now?

    loading  Checking for direct PDF access through Ovid


Cervical spondylotic myelopathy (CSM) is indeed a formidable adversary. Cloaked under the veil of the normal aging process, it can present abruptly or insidiously with disabling neurologic dysfunction of varying clinical manifestations and severity. Moreover, these symptoms can improve, stabilize with intervening phases of subtle deterioration, or worsen progressively. Many conservatively treated patients will ultimately require surgical intervention…but not all1.
Surgical intervention is often recommended for patients with CSM because of disease persistence, severity, or rapid progression. The reported outcomes following surgery have been variable2, and this has prompted clinicians to attempt to identify the patients with CSM who would most benefit from operative treatment. A number of clinical factors have been suggested with the hope that they have prognostic value. The best evidence to date, although not universally accepted, has implicated the patient’s age, spinal canal and cord dimensions, and duration and/or severity of symptoms at presentation as major predictors of surgical outcome3. What has been universally acknowledged is the 11% to 38% prevalence of complications typically associated with operative treatment of patients with CSM4,5.
The paper by Fehlings et al. constitutes a large, thoughtfully designed, and well-executed prospective multicenter clinical study that sought to validate the merits of surgery for patients with mild, moderate, and severe CSM. All study subjects had documented clinical and imaging evidence of CSM, were adult (at least eighteen years old), had objective signs of cervical cord compression on magnetic resonance imaging, had no prior surgical treatment for CSM, and had no symptomatic lumbar stenosis. The study enrolled patients with varying degrees of CSM severity and duration, and it utilized several validated outcome measures (mJOA [modified Japanese Orthopaedic Association] score, Nurick grade, NDI [Neck Disability Index], and SF-36 [Short Form-36] score) to characterize the study subjects prior to and at twelve months following surgical intervention.
However, the study lacks a control group, and it is unclear whether it represents a consecutive series. Although the authors acknowledge the potential logistic and ethical concerns associated with having a nonoperative control group in such patients with CSM, it is regrettable that a study with this methodology cannot provide Level-I evidence. This is particularly relevant for the issue being studied because patients with symptomatic CSM can stabilize with nonoperative treatment. Furthermore, as referenced by Fehlings et al., a prior prospective randomized study by Kadanka et al.6 (n = 68, with three-year follow-up utilizing the mJOA score) suggested that surgery was not superior to conservative treatment in patients with mild to moderate CSM. However, the characteristics of these two study cohorts do not entirely match, and this may account for the differences in the conclusions reached.
The duration of symptoms is considered by many to be critical in the CSM decision-making process, but the importance of this clinical parameter is not reflected by the existing literature. Unfortunately, the study by Fehlings et al. does not address this void. Although the mean baseline duration of CSM symptoms was documented preoperatively, according to the clinical level of severity, for all patients with CSM, its impact on the outcomes was not analyzed. This is especially warranted because of the wide range of symptom duration among the enrolled patients at the time of the preoperative baseline assessment.
This study has yielded a positive answer to the question of whether surgery for CSM can be effective; however, it has also raised other questions. The most important lingering question may be, who actually requires surgery? At this time, the precise operative indications for patients with CSM remain elusive.
    loading  Loading Related Articles