12 Cervical Spondylosis with Myelopathy
Cervical spondylosis with resultant myelopathy and/or radiculopathy is an increasingly prevalent disease in our growing aging population. As such, clinicians and surgeons alike must be well-equipped for the medical and surgical management of this clinical entity. This chapter examines the epidemiology of cervical spondylotic myelopathy and details the biomechanics, pathophysiology, and natural history of the disease process. The chapter explores the multitude of non-surgical and surgical treatment options and their respective risks and benefits, highlighting both the challenges inherent in decision-making and the evidence in the existing literature. After reading this chapter, the reader should have an understanding of cervical spondylotic myelopathy, should know how this disease behaves and progresses, and should have a foundation on which to make evidence-based management decisions.
Cervical spondylosis is the leading cause of spinal cord dysfunction in aging patients
Cervical spondylosis is a degenerative process resulting from static, dynamic, and histopathologic factors
Anywhere from 20 to 63% of patients with cervical spondylotic myelopathy will suffer worsening neurological function 3 to 6 years from diagnosis; however, it is still unclear which patients will deteriorate
With appropriate counseling, patients with mild CSM may be considered for nonoperative management
Patients with moderate-to-severe CSM generally should be offered surgical management
Surgical treatment for CSM involves decompression of the spinal cord with or without fixation (fusion) of the affected vertebral levels
Surgery may be performed from an anterior, posterior, or combined approach; surgical decision-making should be tailored to each individual case
Cervical spondylosis is a chronic, progressive degenerative disease of the spine and is the leading cause of spinal cord dysfunction in the aging population worldwide. 1 , 2 Spondylosis refers to the cascade of degenerative changes which happens throughout the aging spine. In contrast to inflammatory conditions, spondylosis is a naturally-occurring process as a result of chronic, normal compressive forces on the spine. The end result of these degenerative changes is narrowing of the spinal canal that can lead to compression and subsequent dysfunction of the cervical spinal cord (Fig. 12‑1a and b). The term cervical spondylotic myelopathy (CSM) is used specifically to describe the phenomenon of cervical spinal cord dysfunction due to the process of cervical spondylosis. CSM presents insidiously with symptoms, including gait instability, bladder dysfunction, and difficulty with finger fine motor tasks; classic signs are upper motor neuron dysfunction referable to the highest level of compression, which include motor weakness, hyperreflexia, and diminished proprioception sense. In order to objectively quantify the functional status of patients with CSM, a number of numeric scales have been created. Perhaps the most commonly used of these scales is the modified Japanese Orthopedic Association (mJOA) scale (Fig. 12‑2). The mJOA scale is an investigator-administered, 18 point scale which assesses motor dysfunction of the upper and lower extremities, sensory dysfunction of the upper extremities, and urinary sphincter dysfunction. A full 18 points indicates no neurologic dysfunction, and the score decreases with increasing severity of neurologic injury. In order to standardize the clinical assessment of CSM, the scale has been further divided into ranges for mild (15–17), moderate (12–14), and severe (0–11) myelopathy 3
The prevalence of spondylotic changes throughout the cervical spine has been reported in a number of studies. Gore et al 4 examined lateral cervical radiographs of 200 asymptomatic patients aged 60 to 65 and found evidence of degenerative changes in 95% of men and 70% of women. Boden et al 5 examined magnetic resonance imaging (MRI) studies in 63 asymptomatic patients and found evidence of disc degeneration in 25% of patients younger than 40 but in 60% of those aged older than 40. Ernst et al 6 reported a 73% prevalence of bulging discs and 37% prevalence of annular tears on MRI of 30 asymptomatic patients. In a larger study, Nakashima et al 7 prospectively examined MRI studies of 1,211 healthy volunteers aged 20 to 70 years. They reported evidence of disc bulge in 87.6% of patients, which significantly increased with age in terms of severity, frequency, and number of levels. In contrast, only 5.3% of asymptomatic subjects had evidence of spinal cord compression; however, this number also increased with age, particularly after age 50. A follow-up study by Kato et al 8 demonstrated a significant decrease in the diameter of the spinal canal, dural tube, and spinal cord in correlation with increasing age. These studies collectively support that cervical spondylosis is primarily a disease of the aging population.
The prevalence of CSM remains significantly more challenging to elucidate. While the vast majority of aged patients have been shown to have evidence of spondylotic changes on imaging studies, it is not entirely clear what fraction of these patients will progress to symptomatic myelopathy. The picture is further complicated by the lack of a universal classification of cervical degenerative disease throughout the literature. Specifically, there is considerable controversy as to whether ligamentous conditions such as ossification of the posterior longitudinal ligament (OPLL) should be included within the category of true CSM. Indeed, OPLL is a known etiology of cervical spinal cord compression and myelopathy, and for that reason it is sometimes included in epidemiologic studies of CSM. Recent studies 9 , 10 have even suggested a more overarching term–Degenerative Cervical Myelopathy (DCM) – to encompass the wide range of degenerative spinal disorders which result in cervical cord compression. Those authors define spondylotic changes (e.g. disc degeneration and facet arthropathy) and ligamentous aberrations (OPLL) as two separate categories, both of which fall within the overarching term “DCM”. The lack of consistency throughout previous literature partially explains the difficulty in knowing the true prevalence of CSM. For the purposes of this chapter, we define CSM as cervical myelopathy as a result of degenerative spondylosis, not including OPLL.
Despite these challenges, it is well known that CSM is the most common cause of spinal cord dysfunction in the aging population. In a study by Moore and Blumhardt, 11 CSM was the most common (23.6%) cause of nontraumatic paraparesis or tetraparesis in adults. It more commonly affects men than women, with a reported ratio of 2.7:1, and the average age of diagnosis has been reported as 64 years of age. 12 Some studies have attempted to estimate the prevalence and incidence of CSM from nontraumatic spinal cord injury (SCI). In a literature review, New et al 13 reported that degenerative diseases constituted 59% of nontraumatic SCI in Japan, 54% in the United States, and 31% in Europe. However, the authors noted a relative paucity of quality studies in this field, as well as a lack of consistency in classification of nontraumatic SCI. For instance, many studies included only paraplegia/tetraplegia as SCI, likely missing a great deal of patients with milder myelopathy from degenerative disease. Furthermore, these studies included all cases of nontraumatic SCI, not just cervical disease. Other studies have attempted to estimate epidemiological trends for CSM based on hospital admission rates. A study by Boogaarts and Bartels 14 estimated the prevalence of CSM as 1.6 per 100,000 based on the surgically-treated cases at their institution from 2009 to 2012. A retrospective, 12-year, nationwide database analysis by Wu et al 1 estimated an overall prevalence of CSM-related hospitalizations of 4.04 per 100,000; of note, this study included cases of OPLL, which is well-known to be more common in Asian populations. 2 While these are some of the only estimates reported, significant limitations must be considered. First, geographical restrictions limit the extrapolation of these estimates to a more global scale. Furthermore, estimating prevalence of CSM based on surgical treatment or hospitalization likely excludes a large portion of patients with milder, less-symptomatic CSM (i.e. those not requiring hospitalization or surgery); this almost certainly underestimates the true prevalence.
12.1.1 Natural History
When considering treatment options for CSM, it is essential that clinicians be familiar with the natural history of the disease. This knowledge allows for management of patient’s expectations, as well as an accurate assessment of the relative risks and benefits of existing treatment options. In regards to the management of CSM, in particular, knowledge of the natural history of progression is of critical importance in deciding between nonsurgical vs. surgical treatments. The natural history of CSM remains rather challenging to investigate for a number of reasons, including heterogeneity of populations and the subjectivity of questionnaires used to grade myelopathy and quality of life (QOL) outcomes. Furthermore, the natural history of CSM is only approximated by studies using various combinations of nonoperative interventions. Still, recent literature does provide us with some insight into the natural progression of CSM. In a prospective cohort study by Bednarik et al, 15 199 patients with clinically “silent” asymptomatic spondylotic cervical cord compression on MRI were followed for at least 2 years. In this study, 22.6% of patients progressed to symptomatic CSM within the follow-up period; 35.5% of these patients progressed within the first 12 months. In a more recent study by Sumi et al, 16 60 patients with mild CSM (JOA score 13) were followed prospectively for a mean of 78.9 months (records were available at follow up for 55 patients). Clinical deterioration (defined by a decline in JOA score below 13 with a decrease of at least 2 points) occurred in 25.5% of cases, while 74.5% remained stable without any deterioration over more than 5 years. A retrospective review by Oshima et al 17 investigated patients with mild myelopathy (defined by motor JOA score of ≥ 3 in both upper and lower extremities) in addition to cervical cord compression with increased T2 signal intensity on MRI. Of the 45 patients in this study, 16 (35.6%) deteriorated and underwent surgery, 2 (4.4%) worsened after minor trauma, while 27 (60%) remained neurologically stable. Shimomura et al 18 prospectively followed 70 patients with mild CSM for an average of 35.6 months; 56 patients were observed for the duration of the study. Eleven of these 56 patients (19.6%) deteriorated to moderate or severe myelopathy. A retrospective cohort study by Yoshimatsu et al 19 found a 62% rate of deterioration in patients with mild CSM who were managed conservatively. In a prospective series by Kadanka et al, 20 the authors reported a progressive worsening of activities of daily living (ADL) scores over time, with 56% of patients deteriorating after 10 years of follow-up.
Numerous studies have attempted to identify clinical or imaging predictors of progression in patients with CSM. A study by Shimomura et al 18 evaluated prognostic indicators of neurological deterioration and found only circumferential cord compression (as opposed to partial compression) as the only statistically significant predictor. They found no association with age, sex, developmental factor, dynamic factor, or high intensity signal on T2 MRI. Oshima et al 17 evaluated risk factors for patients with CSM treated nonoperatively to convert to surgical treatment. They reported that (a) total cervical range of motion (ROM) greater than 50º, (b) segmental kyphosis in the maximum compression segment, or (c), the presence of a local slip were associated with increased risk of requiring surgery. Age greater than 60, sex, C2-C7 alignment, spinal cord diameter less than 50%, developmental canal stenosis, and segmental ROM were not associated with increased risk of surgery.
In summary, the proportion of patients with CSM who deteriorate by at least one point on the JOA scale 3 to 6 years after initial diagnosis ranges from 20–62%, while the proportion of patients with worsening ADL scores may be as high as 56% at 10 years. 21 However, while there is some weak evidence supporting patient risk factors for deterioration, it is still mostly unclear which patients are more likely to progress.
12.2 Biomechanics and Pathophysiology
Cervical spondylotic myelopathy results from a combination of three factors—static, dynamic, and histopathologic—that contribute to compression and dysfunction of the cervical spinal cord. 2 , 21 , 22 , 23
12.2.1 Static Factors
The degenerative process of cervical spondylosis is thought to begin with desiccation, or loss of normal water content, of the nucleus pulposus of the intervertebral discs. Normally, the nucleus pulposus, comprised largely of proteoglycans, has a relatively high water content. This gives the nucleus pulposus a viscoelasticity which allows it to convert great axial load into hoop stress contained within the surrounding annulus fibrosus. 9 With aging, the water content of the nucleus pulposus decreases. This, in combination with the repetitive, chronic biomechanical forces of daily use, results in disc degeneration and flattening. Subsequently, the uncovertebral processes become flattened, and the load-bearing capacity of the vertebrae is altered. 2 , 23 As a result, there is increased stress on the articular cartilage endplates of the vertebral bodies, and hypermobility of the facet joints. It is thought that these structural changes result in uneven forces exerted on the vertebrae, leading to compensatory osteophyte formation in an attempt to stabilize the uneven segment. The loss of disc height is also thought to result in the buckling and hypertrophy of the ligamentum flavum, as well as in the straightening of cervical lordosis or even progression to kyphosis. The combination of these factors ultimately results in stenosis of the spinal canal and compression of the spinal cord.
12.2.2 Dynamic Factors
While static factors involved in the cervical degenerative process are the primary event in CSM, dynamic factors are also important to understand in the development of this disease. Because cervical spine flexion and extension affect the sagittal diameter of the spinal canal, they both can contribute significantly to CSM. Flexion can induce compression of the cervical cord against ventral bone spurs or bulging discs, a phenomenon that is compounded in the case of cervical kyphosis. 24 Extension can result in buckling of the ligamentum flavum, which leads to dorsal compression of the spinal cord. 22 , 24 , 25 Furthermore, instability of cervical segments from ligamentous laxity can result in subluxation during flexion or extension, leading to a pincer phenomenon causing more damage to the spinal cord. 22
12.2.3 Histopathologic Factors
As discussed above, the combination of static and dynamic mechanical factors results in narrowing of the spinal canal, followed by chronic compression and repetitive trauma to the cervical spinal cord. Secondary to this structural damage, the spinal cord has been shown to undergo multiple histopathological and vascular changes, resulting in ischemia/infarction and other cytotoxic processes. Much of our preliminary understanding of this process comes from extrapolation from models of acute SCI. However, the pathophysiologic process of CSM differs from traumatic SCI in that there is no acute mechanical insult; there is an absence of hemorrhagic necrosis of the cord; and the progressive nature of the disease results in compensatory changes within the cord. 26 Current knowledge suggests that progressive compression of the cervical spinal cord leads to a chronic hypoxic/ischemic insult due to compression of spinal arteries and resultant decrease in blood flow. 2 This ischemic state results in damage to oligodendrocytes and neurons, which elicits a unique immune response. Inflammation combined with the chronic hypoxemic state is thought to cause endothelial cell loss and compromise of the blood-spinal cord barrier, which results in edema and entry of neurotoxic substances. 2 , 23 , 26 In particular, glutamate neurotoxicity is thought to play a significant role in the pathophysiology of CSM by causing neuronal degeneration, as demonstrated by Karadimas et al in a study utilizing a novel animal model. 27 Other studies have added evidence that apoptosis of neurons and oligodendrocytes further contributes to the pathobiology of CSM. 28 Further research is necessary to more fully understand the molecular pathways and cytotoxic changes involved in CSM. A better understanding of the pathophysiologic processes could implicate potential pharmaceutical therapies to augment or delay the need for surgical treatment.
12.3 Treatment Options
12.3.1 Nonsurgical Treatment
Nonsurgical treatment options reported in the literature vary widely. These range from bedrest and discouragement of high-risk activities to cervical traction and spinal injections. Most of the nonsurgical treatment methods described in the literature involve a combination of techniques. Kadanka et al 20 , 29 utilized intermittent soft collar, anti-inflammatory medications, intermittent bedrest, and avoidance of high-risk activities and environments. Sampath et al 30 used pain medications (either narcotic or nonsteroidal), steroids, bed rest, home exercise, cervical traction, neck bracing, and various spinal injections (epidural, facet, nerve blocks). Yoshimatsu et al 19 utilized cervical traction (3–4hrs per day), immobilization with cervical orthosis, medications, exercise therapy, and thermal therapy.
12.3.2 Surgical Treatment
Surgical treatment strategies include anterior (ventral), posterior (dorsal), or combined approaches. The goal of surgical treatment is to decompress the spinal cord and stabilize the cervical spine if dynamic injury to the spinal cord is suspected. In deciding on a surgical approach, several variables must be considered by the surgeon, including the area of compression (ventral vs. dorsal), sagittal alignment, focal vs. diffuse disease, the presence of radiculopathy or axial pain, age, comorbidities, and surgeon familiarity with each specific procedure.
A multitude of techniques may be used via the anterior cervical approach. These include single or multiple discectomy or corpectomy (or a hybrid combination of the two), most often in conjunction with anterior plate fixation and fusion. Alternative anterior nonfusion options include cervical disk arthroplasty and oblique minimally destructive corpectomies. These techniques aim to avoid the complications associated with fusion, including adjacent segment disease and altered cervical alignment. Anterior techniques tend to be used when fewer levels are involved (focal disease), when compression is more ventral in nature, and in cases of significant kyphosis.
The two most common posterior cervical procedures are laminoplasty or laminectomy with fusion. While laminectomy alone was used more commonly in the past, accumulating evidence of postlaminectomy kyphosis (in the absence of fusion) has led to a general trend away from this procedure. In general, posterior decompression procedures are more commonly reserved for multi-level, diffuse disease or cord compression that is primarily dorsal.
12.4 Benefits and Risks
As explored in more detail below, the relative benefits and risks of treatment options for CSM depend heavily upon the patient’s presenting mJOA status, the treatment option selected, and patient-specific factors (e.g., age or co-morbidities). Just as it has been difficult to ascertain the true natural history of CSM due the heterogeneity inherent in both the disease and the patients affected, it is also challenging to clearly elucidate the real benefits attained or risks assumed for any given CSM treatment. Clinical decision-making for these patients must, therefore, involve a synthesis of 1) the data available from longitudinal (though retrospectively analyzed) clinical studies, 2) surgeons’ expertise and appraisal of the individual patient, and 3) patient preferences.
12.5 Pitfalls, Complications, and Avoidance
Though this chapter does not focus on specific surgical techniques, several general “pearls” can be offered. When deciding between nonoperative and operative treatment, patients with moderate or severe mJOA scores should, if possible, be offered surgery. Patients with asymptomatic or mild disease should be counseled on what is known of the natural history of CSM and offered the options of nonoperative vs. operative intervention. In managing patients’ expectations for surgery, it should be made clear that the primary goal of surgical intervention is prevention of worsening of neurological status, and that improvements in mJOA scores, though likely, are not guaranteed. When selecting the procedure for patients undergoing surgery, those with significant kyphosis and/or greater than 50% canal compromise from ventral disease are likely to benefit from an anterior approach, either alone or combination anterior-posterior. Finally, when CSM patients have an accompanying deformity, surgeons should consider obtaining standing full spine X-rays to assess and document sagittal plane derangements both for the cervical region and globally.
Evidence for outcomes of treatment for CSM centers mostly around comparative studies (Table 12‑1). Outcomes are challenging to assess, given the heterogeneity of populations and treatment modalities, but the literature does provide some insight.
12.6.1 Nonoperative Treatment vs. Surgery
Kadanka et al 20 followed patients with mild CSM (mJOA score > 12) in a randomized controlled trial comparing nonoperative treatment (mentioned above) with surgery (anterior discectomy, corpectomy, or laminoplasty) over a 10-year period. The authors evaluated four primary outcomes: mJOA score, timed 10-minute walk, and ADL as scored by the clinician and the patient. There was no statistically significant difference in any of these outcome measures over the 10-year follow-up period, suggesting surgical treatment is no more effective than conservative therapies in patients with mild CSM. However, it should be noted that within both groups there were patients who declined, and others who improved. In another study by Kadanka et al examining the same study population, 31 the authors attempted to identify predictive factors for outcomes after both conservative and surgical treatments. They found that positive response to conservative treatment after 3 years was more likely in patients with older age, lower body height, higher anterior-posterior spinal cord diameter, lower entry mJOA score, and normal central motor conduction time. Positive response to surgical treatment was associated with slower 10-minute walk time, and lower entry mJOA score. It is important to mention here that both studies are somewhat underpowered to draw any significant conclusions.
Two cohort studies compared nonoperative treatment with surgical treatment in patients with moderate to severe CSM. A prospective study by Sampath et al 30 compared outcomes in non-operative and surgical treatment groups with a 1-year follow-up. They found statistically significant improvement in pain and functional status, and no improvement in neurological symptoms in the surgical cohort. On the other hand, they found significant worsening in ability to perform ADL and nonsignificant worsening of neurological symptoms in the conservative cohort in comparison to baseline. Although no direct comparisons were made between cohorts, the results suggest that surgery may be more effective in treating CSM. It should be noted that the surgical cohort had more severe myelopathy at baseline compared to the conservative cohort, which may have made this group more likely to improve. A retrospective study by Yoshimatsu et al 19 compared outcomes in 32 patients who elected to undergo surgical treatment for CSM with 69 patients who elected for nonoperative care. Patients who chose surgery had more severe CSM (mJOA 9.1) compared with those who chose conservative management (mJOA>12). The authors found that 78% of patients who underwent surgery had improved mJOA scores at a mean follow-up of 29 months, compared to only 23% in the nonoperative group. These results again suggest that surgical treatment may be superior to conservative therapy but it remains unclear for which subgroups of CSM patients (i.e., mJOA strata).