In 1956, cervical myelopathy was recognized by Clarke and Robinson as a potentially devastating and irreversible neurologic condition.⁵ Their study of 120 patients indicated that once symptoms developed, patients were unable to return to a neurologically normal state. Although half of the study subjects experienced some improvement in their symptomatology with conservative therapy, the remaining 50% of patients progressed in a slow, steady manner with short episodes of rapid worsening. Several subsequent studies have corroborated the ominous nature of this disease and noted the need for aggressive intervention.⁶^–⁸

A review of the Joint Guidelines Committee of the American Association of Neurological Surgeons and the Congress of Neurological Surgeons demonstrates the relative lack of data regarding the natural history of cervical spondylotic myelopathy, especially in the context of modern imaging, objective myelopathy scoring, and blinded outcome.⁹ Indeed the authors’ key issues statement summarizes the problem nicely:

It is evident that there is a need for a randomized clinical trial examining patients with mild CSM (cervical spondylotic myelopathy) and comparing operative and nonoperative treatment. There is sufficient evidence indicating that patients with severe symptoms and a long duration of symptoms will generally not improve, and attempts at randomizing these patients to different treatment groups may not be ethically feasible.

The same committee suggested an appropriately powered prospective randomized trial; however, to date, such a trial has yet to be performed. Hence, this systematic review, unfortunately, does not include a more recent study for which to base clinical decision making with respect to patients with minimal myelopathy and cervical stenosis.

Best Evidence to Date

Recently, Bednarík et al.¹⁰ reported their findings in a group of 199 patients with “presymptomatic spondylotic cervical cord compression” who were followed for a minimum of 2 years. Forty-five (22.6%) of the patients developed clinical evidence of myelopathy and 18 (9.0%) were offered surgical intervention. Significantly, slightly more than one third of these patients who developed clinical myelopathy did so within the first year of observation. The best predictors of early progression to myelopathy were symptomatic cervical radiculopathy and abnormal motor-evoked potentials (MEPs) or somatosensory-evoked potentials (SEPs). The best predictor of late progression to myelopathy was MRI hyperintensity. It is clear that most patients with MRI evidence of cord compression can likely be followed expectantly. It appears that at least about 20% of patients with no or minimal myelopathic signs or symptoms but with MRI evidence of cervical cord compression or those with only mild CSM will progress to some degree during a 2-year follow-up period of observation. The next task is to identify factors that predict probable worsening of cervical spondylotic myelopathy.

Surgical Decision Making for Patients with Presymptomatic Spinal Cord Compression

Given the current lack of data regarding the relative benefits of surgical versus nonsurgical treatment for minimal myelopathy, it becomes very difficult to counsel patients. The main question remains, in what setting is operative treatment best? Knowing when to intervene prevents unnecessary surgery in those whose symptoms are unlikely to progress and, conversely, treats appropriate patients prior to a point of significant debilitation. Attempting to identify CSM patients would be a negligible exercise if a successful treatment option was not available. It is clear, however, that surgical decompression does produce favorable results ¹¹^–¹³ and should be used early in the disease course. Failing to aggressively treat this group of patients with operative intervention will result in clinical worsening, which at some point will become irreversible.

Risk Factors for Progression of Neurologic Dysfunction

Numerous factors in the radiologic, clinical, electrophysiologic, and genetic realms have been considered in attempting to predict which patients will develop progressive cervical spondylotic myelopathy. Some of the positive predictive factors in neurologic worsening are considered in the following sections.

Cervical Spinal Cord Diameter

There is a strong correlation between the severity of cervical spondylotic spinal cord compression and the likelihood of progressive clinical myelopathy.

The extent of the spatial compromise of the cervical spinal canal and the consequent severity of spinal cord compression correlates with the severity of the initial clinical presentation of CSM. A congenitally narrow spinal canal results in the development of CSM in 40% to 47% of these patients, which is a much higher rate than is seen in the general population.^14,¹⁵ Long periods of severe cervical stenosis over many years result in demyelination of the spinal cord’s white matter and in necrosis of both the gray and white matter, leading to potentially irreversible neurologic deficits.⁹ In patients with more severe MRI evidence of cord compression and/or MRI evidence of spinal cord atrophy, functional neurologic deficits with lower postoperative modified Japanese Orthopaedic Association (mJOA) scores were seen more consistently.^16,¹⁷ In a prospective, randomized 3-year clinical study of stable or very slowly progressing mild and moderate clinical SCM (mJOA ≥12), several factors were associated with good outcome in conservatively treated, nonoperative patients. Patients demonstrating lower rates of stenosis of the spinal canal in the anteroposterior (AP) dimension were less likely to show progression of their myelopathy. Also, a slightly larger spinal cord transverse area was found to have a positive predictive power for favorable response to conservative treatment.¹⁸ Although the correlation between the severity of cervical spondylotic cord compression and the likelihood of rapidly progressive myelopathy appears strong, not every study correlates this finding.¹⁹

A spinal cord transverse area of less than 50 mm ² at the level of maximum compression significantly correlated with clinical symptoms by mJOA.²⁰ In the same study, this correlation was strengthened if these areas of maximal spinal cord compression were associated with hyperintensities on T2-weighted MR imaging within the spinal cord. Other studies examining the AP dimensions of the cervical canal have demonstrated significant differences in functional neurologic status; patients with a 7.1-mm AP canal diameter were more likely to have clinical myelopathy, whereas patients with a 10.8-mm canal were more likely to be nonmyelopathic.²¹ In a study of mild and moderate CSM (mJOA ≥12), Kadanka et al. randomized patients to either early surgery or nonoperative management and found that patients with less significant spinal cord compression responded more favorably to surgery.¹⁸ Considering this potential postoperative improvement, if cervical stenosis is permitted to progress preoperatively, is there a point where surgical results will be suboptimal? If so, and if that point can be identified, then treatment prior to that time may prevent the occurrence of irreversible symptoms. Several groups have attempted to identify the exact degree of spinal cord compression that will best predict the optimal timing of surgery, but despite promising data, these results have rarely been consistent.²²^–²⁴ It has been demonstrated, however, that spinal cord architecture appears to be well restored by decompression in these patients with mild-to-moderate cervical myelopathy.²³

Intramedullary Spinal Cord Signal Intensity

Cervical intramedullary signal changes, especially when accompanied by significant stenosis, can help identify a patient who will have a favorable surgical outcome. Once these changes span multiple levels or include hypointensity on T1-weighted MR imaging, however, the prognosis worsens.

The observation that severe cervical stenosis and cord compression often result in intramedullary spinal cord signal intensities in CSM patients has prompted the evaluation of a possible relation between these signal changes and the severity of cervical myelopathy. Although the results of such studies have produced promising results, the exact correlation and the strength of its predictive value remain unclear.

Bednarík et al. suggested that MRI hyperintensity, when present in patients with asymptomatic cervical spondylosis, may predict a delayed (>12 months) progression to symptomatic CSM.²⁵ The influence on surgical outcome of these signal intensities has become an area of immense interest. Overall, there appears to be a significant correlation between surgical result and the preoperative presence of intramedullary signal alteration.^20,²⁶^–³⁰ The best results occur in those patients with either no signal changes or hyperintensity on T2-weighted MRI.²⁶^–³⁰ When T2 hyperintensity occurs at a single segment, this single finding does not portend a poor surgical outcome; these changes are reversible.²⁷^–³¹ When hyperintensity onT2-weighted imaging spans multiple levels or is coupled with T1-weighted hypointensity, however, the surgical prognosis is significantly less favorable.²⁷^–³¹ Unlike T2-weighted changes, T1-weighted hypointensities are not reversible,²⁸ but they never exist as an isolated finding (i.e., without accompanying T2-weighted hyperintensities).³⁰ There does not appear, however, to be a relationship between the duration of observed signal intensity and outcome.²⁸

As mentioned previously, a review of the cross-sectional transverse area and signal change on MRI found that a T2-weighted MRI of a cross-sectional area of the spinal cord between 50 and 60 mm ² in association with hyperintensities was a critical point of significant disability as identified by the mJOA score.²⁰ In addition, differentiating between mild and pronounced spinal cord hyperintensities produced a significant difference in mJOA scores of patients with CSM.²⁰ Although decompressive surgery can result in the rapid resolution of these imaging abnormalities, clinical improvement does not always accompany these radiologic changes.³¹ Suri et al., however, were able to demonstrate that patients who experienced regression of any intramedullary signal changes were statistically likely to experience better surgical outcomes.³⁰

Electrophysiologic Abnormalities

Electrophysiologic examination may be helpful in identifying candidates for early surgery. Surgical results will be best in those with normal or mildly abnormal studies. In the setting of cervical stenosis, an abnormal electromyelogram (EMG) or clinical radiculopathy should prompt strong consideration for decompression due to the ominous association with the development of symptomatic myelopathy.

Evaluation of patients with cervical spondylotic myelopathy in various electrophysiologic (EP) studies has yielded promising results. In general, when these results are normal, a more favorable clinical course is likely and these patients will likely respond well to nonsurgical management.^10,¹⁸ In contrast, in a study of 30 asymptomatic patients with spondylotic cervical cord compression, one-third of those demonstrating an initial EP abnormality eventually developed CSM during the study’s 2-year follow-up.¹⁰ In a subsequent study analyzing the utility of SEPs and MEPs in predicting and monitoring the effect of therapy in spondylotic cervical myelopathy, however, the same authors were unable to demonstrate as strong a correlation between these EP studies and the clinical status.³²

In their 2004 study involving a 2-year prospective follow-up of 66 patients with MRI-demonstrated spondylotic cervical cord compression but without clinical myelopathy, Bednarík et al. identified several variables significantly associated with the development of myelopathy. Among these were the presence of symptomatic cervical radiculopathy, which was present in 92% of patients eventually developing CSM versus 24% who did not develop myelopathy (P = .0001), electromyelographic signs of an anterior horn lesion (61% vs. 11.3%, P < .01), and abnormal SEPs (38.5% vs. 9.4%, P < .02).¹⁹ These authors also noted that the presence of symptomatic cervical radiculopathy and electrophysiologic abnormalities of cervical cord dysfunction detected by SEPs or MEPs were associated with the early development (<12 months) of CSM.²⁵ Using the combination of these two parameters, radiculopathy, and EP abnormalities, early progression could be predicted accurately in over 80% of patients.²⁵ In fact, in a recent meta-analysis of the CSM literature, class I evidence suggested that in the setting of cervical stenosis without myelopathy but with either an abnormal EMG or clinical radiculopathy, decompression should be strongly considered due to the association with the development of symptomatic myelopathy.⁹

The value of additional electrophysiologic parameters for the detection, evaluation, and treatment of CSM has been examined.³³^–³⁵ In general, these studies demonstrated a correlation between the severity of the cervical myelopathy and the degree of EP abnormalities. This relationship indicates that as EP studies worsen, surgical intervention should be more readily employed as a treatment option to prevent further myelopathic progression.³⁴ There may be, however, a point of diminishing return in which the EP abnormalities are so severe that a patient may not improve clinically despite the treatment modality, including surgery. Both Hu et al. and Yu et al. have shown that the absence of any SEP waveform is correlated with impairment that is least likely to improve with operative intervention.^36,³⁷

Symptom Severity and Duration

Patients with no or mild myelopathic symptoms will respond better to surgery than those with more severe disease. Once permitted to progress, cervical myelopathy becomes more refractory to surgical intervention. Adjunctive tests (e.g., 30-meter timed walk test, nine-hole-peg test) may help objectify clinical symptoms and can be used to follow patients treated operatively or nonoperatively.

In identifying those patients with no or mild CSM symptoms who will benefit from early surgery, both the severity and the duration of symptoms must be identified. Perhaps the most important predictive factor for surgical intervention is the pretreatment neurologic condition.^38,³⁹ Whereas 70% of patients with mild myelopathy (mJOA ≥12) had good outcomes with surgery, only 55% of those with more severe symptoms (mJOA <7) fared as well.³⁸^–⁴⁰ Additionally, multiple reports have suggested the negative prognostic impact of surgical treatment performed “too late.”^* There is a significant correlation between the duration of symptoms and neurologic condition after surgery, that is, a longer duration of symptoms portends a poorer prognosis. The point at which patients will begin to experience this negative effect is unclear. Some studies have suggested that those patients who have had preoperative myelopathic symptoms for more than 12 months have a worse postoperative prognosis.^28,⁴⁰ Others, however, found a significant difference only at the 2-year time point.^11,³⁰ Regardless, it is clear that patients have a better outcome when treated earlier in the disease course.