Axial Pain

Neck pain is a common presenting chief complaint seen by the general practitioner. Contributing anatomic sources of neck pain are multiple and include neck musculature, tendons, ligaments, facet joints, intervertebral discs, and cervical vasculature. Referred pain can be seen with shoulder and temporomandibular joint pathology as well. The intervertebral disc is innervated ventrally by branches from the sympathetic plexus and dorsally by the sinuvertebral nerve, which arises from the ventral nerve root. The sinuvertebral nerve also innervates the PLL, dura, and a substantial portion of the vertebral body periosteum. Cervical facet joints are innervated by branches arising from the dorsal ramus.

The most common cause of nondegenerative isolated neck pain is cervical strain—as frequently seen with whiplash injury—resulting from injury to neck muscles, tendons, and ligaments. Beliefs about the anatomic source of isolated neck pain in patients with cervical spondylosis vary. The intervertebral disc is commonly cited as the source of axial pain. Tears in the annulus may stimulate the sinuvertebral nerve and injection of local anesthetic in the disc space can temporarily relieve pain in some patients. The facet joints are another potential source of axial pain. Stimulation of subaxial facet joints generates reproducible neck pain patterns in normal volunteers. However, facet steroid injections and percutaneous radiofrequency neurotomy have demonstrated mediocre results.

Isolated axial pain that fails to respond to initial conservative therapy can be further evaluated with cervical radiographs. Cervical spondylotic changes on radiograph are ubiquitous in the aging population and include loss of disc height, osteophyte formation, kyphosis, and subluxation. Flexion-extension cervical films greatly help in ruling out instability or motion that may be a source of significant pain.

In appropriately selected patients, several studies have demonstrated good results in operative management of axial pain. Neck pain is common in rheumatoid arthritis and can be secondary to instability or from basilar invagination, and surgery is commonly employed in this population. One must always be alert to the possibility of a C3-4 radiculopathy as a source of axial pain. Unilateral pain should alert the practitioner to look for sensory alterations, ask about paresthesias in this distribution, and look for a positive Spurling sign. C3-4 radiculopathy that causes axial pain generally responds very well to surgical decompression. Pseudarthrosis from previously attempted fusion can also lead to significant axial pain with or without radiculopathy and is a condition that also responds well to reoperation.

Acute neck pain deserves a trial of nonsteroidal anti-inflammatory drugs (NSAIDs) and short-term muscle relaxants if needed. A temporary soft neck collar can provide comfort as well. Chronic neck pain can be managed with analgesia and physical therapy exercises to strengthen the cervical musculature. Surgery should be reserved for patients with well-accepted indications.

Radiculopathy

Cervical radiculopathy results from compression of an exiting cervical nerve root. This often results from uncovertebral and facet osteophyte formation extending into the neural foramen. Patients often describe a sharp or burning radiating pain in a dermatomal distribution. Nerve compression can also result in paresthesias or impaired sensation in a dermatomal distribution or weakness in the respective myotome. Physical examination is often significant for a positive Spurling sign: axial compression with lateral bending to the ipsilateral side reproduces the radicular pain. The abduction relief sign—relief of radicular pain by abducting the ipsilateral arm and putting the hand on the head—can help clinicians to differentiate radiculopathy from thoracic outlet syndrome or shoulder pathology. One must carefully evaluate the radiculopathic complaint and consider alternative etiologies such as peripheral entrapment syndromes, thoracic outlet syndrome, brachial neuritis, shoulder pathology, reflex sympathetic dystrophy, and even angina.

MRI has become the standard for evaluating the neural foramina for radiculopathy. With existing hardware, computed tomography (CT) myelography is more useful. In attempting to sort out radiculopathy from peripheral syndromes, electrodiagnostic studies such as electromyography and nerve conduction studies are routinely used.

Radiculopathy without significant weakness deserves an appropriate trial of conservative therapy. This includes analgesia, NSAIDs, and possibly anticonvulsant therapy. Although not approved by the U.S. Food and Drug Administration (FDA), anticonvulsant therapy with gabapentin or pregabalin, which has demonstrated benefits in diabetic neuropathy, is now frequently used for radiculopathic pain with good anecdotal results. Epidural steroid injection or localized nerve blocks can provide therapeutic relief, and the latter can help to confirm diagnostic hypotheses.

When conservative therapy fails and the diagnosis of cervical root compression is certain, surgical decompression provides good results. When alignment is well maintained, a minimally destabilizing approach includes dorsal laminoforaminotomy. When fusion is needed, either anterior cervical discectomy and fusion (ACDF) or dorsal decompression with fusion provides good results in class III evidence. Persson and associates randomized 81 patients with cervical spondylotic radiculopathy to ACDF, physical therapy, or cervical collar immobilization. Evaluation at 3 to 4 months revealed improved pain scores (using a visual analog scale) and motor and sensory improvements with surgery compared to nonoperative alternatives. This effect dissipated at the 12-month follow-up; however, a disability rating index showed improved return to work and dressing ability at 12 months with surgery.

Myelopathy

Patients with myelopathy commonly present with unsteady gait and difficulty with fine motor coordination in the hands. Physical examination may demonstrate hyperreflexia below the level of compression, increased muscle tone, clonus, the Babinski sign, the Hoffman sign, and the finger escape sign. Some patients may describe the Lhermitte sign (electric shock sensations traveling down the spine with flexion), which is thought to be due to stimulation of the dorsal columns. Hands may demonstrate intrinsic muscle atrophy, which is a classic sign in myelopathy. Some patients may complain of urinary retention or spastic detrusor activity leading to frequent urges with or without incontinence. Additional localizing upper motor signs include pectoral muscle reflex, which is suggestive of compression at or above C2-4, and the jaw jerk, which if present suggests a lesion above the foramen magnum. Patients with severe cervical spondylosis with canal stenosis can experience central cord syndrome with even minor trauma, particularly in hyperextension injury. Greater motor impairment is seen in the upper extremities and is often accompanied by urinary retention. Burning hands have been described in football injuries and are thought to be a variant of central cord syndrome in patients with congenital canal stenosis. The differential diagnosis for CSM is broad and includes multiple sclerosis, syringomyelia, atrophic lateral sclerosis, subacute combined degeneration, intraspinal tumor, spinal arteriovenous malformation, epidural abscess, Chiari malformation, ossification of the posterior longitudinal ligament, normal pressure hydrocephalus, tabes dorsalis, hereditary spastic paraplegia, and tropical spastic paraparesis. Several grading systems have been developed to classify the severity of CSM in an objective, reliable, and valid assessment that can also be used to measure responsiveness to therapeutic interventions. The Japanese Orthopaedic Association (JOA) scale and the modified version by Benzel and colleagues are the two most widely used systems and have demonstrated good interobserver and intraobserver reliability. Other accepted systems include gait analysis and the Short Form-36 (SF-36).

The gold standard for imaging in CSM has become MRI because it provides the best view of the spinal cord, exiting nerve roots, and cerebrospinal fluid (CSF) signal. CT myelography may be more useful in cases of previous surgery because it is superior to MRI in viewing residual bony anatomy and produces less artifact with existing hardware. The examiner must be aware that the degree of stenosis on imaging frequently does not correlate with clinical impairment. In one study of asymptomatic elderly patients, 26% had some degree of spinal cord impingement on MRI. Multiple studies have attempted to correlate spinal cord signal changes on MRI with neurologic recovery after decompression. Several class III studies demonstrate that T2 hyperintensity at a single segment does not predict outcome, but when present at multiple levels or in combination with T1 hypointensity, it does correlate with poor neurologic recovery after surgery. Other studies have attempted to correlate the degree of canal stenosis with neurologic recovery after surgery. Most studies demonstrate poorer neurologic recovery in patients with greater radiographic canal stenosis, with most studies using a canal area of 30 to 45 mm ² as the cutoff to dichotomize groups. One study did not corroborate these findings.

Although electrodiagnostic studies are not necessary for diagnosis of CSM, a class I study by Bednarik and coworkers followed 66 patients (average age 50 years) with radiographic spinal cord compression from cervical spondylosis without clinical myelopathy. These patients were followed for an average of 4 years, during which 19.7% developed CSM. Bednarik and colleagues found that electromyography and sensory-evoked potential abnormalities and, additionally, clinical radiculopathy, when present initially, predicted the development of CSM.

Traditional teaching portrays the natural history of CSM as progressive stepwise neurologic deterioration. However, after initial presentation of neurologic impairment, the natural history is mixed. Some patients remain neurologically stable for long periods of time, with some even improving; others will continue to accrue additional deficits. Many class III studies have tracked the natural history of CSM. One of the initial studies by Clarke and Robinson in 1956 retrospectively reviewed 120 patients with CSM, 26 who never underwent surgery plus the preoperative course of 94 patients who eventually underwent surgery. They found that the majority (75%) of patients experienced episodes of neurologic deterioration with intervening periods of stability. Of the smaller cohort that did not undergo surgery, half experienced some degree of neurologic improvement with conservative management. Another study by Nurick in 1972 found that most patients remained neurologically stable after initial deficits, and he advocated surgery for those with progressive symptoms and those older than 60 years of age. One of the only class I studies, by Kadanka and associates, demonstrated that 80% of patients younger than 75 years of age with 1 year of mild CSM (defined as modified JOA > 12) remained neurologically stable with conservative management (NSAIDs, rest, cervical immobilization) over 2 to 3 years as measured by the modified JOA scale, a timed 10-meter walk, and a video evaluation of activities of daily living (ADL) performance. A 2002 Cochrane review of CSM concluded that there is no clear evidence to support the idea that CSM patients inevitably deteriorate neurologically. However, all of these studies of conservative management excluded patients who underwent early surgery, likely because of more severe or progressive forms of CSM. Therefore, these studies have a selection bias toward patients with a more benign natural history and cannot be generalized.

Furthering the decision maker’s dilemma, numerous studies demonstrate that increased symptom duration—most studies using between 12 and 24 months as the cutoff—portends worse neurologic recovery. Therefore, it appears that in mild CSM, a conservative management trial is reasonable, but patients with unacceptable neurologic deficits and those with progressive symptoms should be considered for early decompression.

Multiple class III studies demonstrate that the majority of patients either improve or remain neurologically stable, by JOA or Nurick scores, after surgical decompression using both ventral and dorsal approaches.

Dorsal Approach

For isolated radiculopathy without myelopathy, dorsal laminoforaminotomy provides an effective alternative to decompress the exiting nerve root. The goal of foraminotomy is to provide additional space to the exiting root without necessarily resecting the offending osteophyte.

Laminectomy is frequently used for multilevel pathology, including multilevel cervical spondylosis, congenital canal stenosis, and dorsal compressive pathology such as ligamentum flavum hypertrophy or ossification. Laminectomy has also been successful for treatment of ossification of the posterior longitudinal ligament (OPLL). Laminectomy alone is better suited for the straight or lordotic spine but not the kyphotic spine. Long-term studies show that the rate of postoperative kyphosis after isolated laminectomy ranges from 14% to 47%. The incidence of postoperative kyphosis increases with loss of lordosis on preoperative radiographs. Particularly concerning is that when multilevel ventral pathology exists, increasing kyphosis may result in further draping of the cord over ventral osteophytes. Laminectomy alone also results in a higher rate of kyphosis than laminoplasty. Numerous studies highlight an increased risk of late neurologic deterioration with laminectomy alone compared to ventral or dorsal decompression with fusion.

Laminectomy can be supplemented with arthrodesis when there is concern for the development of kyphosis. Laminoplasty using either a French door or an open-door technique has also been employed with success. There currently is no class I or II evidence to suggest superiority among laminoplasty, laminectomy with arthrodesis, anterior cervical corpectomy and fusion (ACCF), or anterior cervical discectomy and fusion (ACDF).

Ventral Approach

When the offending compressive elements are ventral, a ventral approach allows better access for direct decompression. When there are three or fewer diseased levels, ACDF or ACCF may be used. For longer segments, either a dorsal or a combined approach is utilized. Ventral plate fixation and instrumentation have become commonplace but should not be a substitute for good graft technique. Kaiser and colleagues retrospectively compared 251 patients with ACDF with plate fixation showing a 96% fusion rate for single-level ACDF and a 90% fusion rate for two-level ACDF compared to historical fusion rates for single-level ACDF (91%) and two-level ACDF (72%). Additionally, graft complications with plate fixation were reduced from 6% to 1.3%. A large retrospective review by Caspar and coworkers found that the reoperation rate for pseudarthrosis was 4.8% for ACDF and 0.7% for ACDF with plate fixation. However, Resnick and Trost performed a systemic review of randomized trials that showed no clear benefit for ventral plate fixation in single-level ACDF. From a biomechanical perspective, plate fixation results in greater preservation of lordosis. Troyanovich and associates calculated that lordosis at the fused segment decreased by 2.5 degrees in ACDF but increased by 5.7 degrees in ACDF with plate fixation.

ACCF is an alternative to ACDF. Traditionally, ACCF demonstrated higher rates of fusion than ACDF without plate fixation. However, ACCF appears to yield results equivalent to those of ACDF with plate fixation. A pooled analysis of 2682 patients by Fraser and Hartl found that two-level ACDF with plate fixation yielded fusion rates (> 90%) similar to those of ACCF. For three-level disease, ACDF with plate fixation yielded significantly lower fusion rates (82.5%) than ACCF with plate fixation (96.2%).

Ventral plates vary among manufacturers, and some have more recently produced dynamic plates that allow for motion. Class III studies show no difference in fusion rates between dynamic and rigid fixation plates. However, one study found a higher screw failure rate with rigid fixation but increased dysphagia with dynamic plates.

Cervical disc arthroplasty presents an alternative to fusion with the theoretic benefit of motion sparing at the treated level and the hope of decreasing adjacent segment disease. It appears that cervical disc arthroplasty is at least as good as ACDF. However, long-term neurologic outcome and safety are still to be determined.