J Neurosurg 74:163-170, 1991
Central corpectomy for cervical spondylotic myelopathy: a consecutive series with long-term follow-up evaluation RICHARD L. SAUNDERS, M.D., PHILIP M. BERNINI, M.D., THOMAS G. SHIRREFFS, JR., M.D., AND ALEXANDER G. REEVES, M.D.
Sections of Neurosurgery and Orthopedics, Department of Surgery, and Section of Neurology, Department of Medicine, Dartmouth-Hitchcock Medical Center, Hanover, New Hampshire
V Since 1984, a consecutive series of patients with cervical spondylotic myelopathy has been treated by central corpectomy and strut grafting. This report focuses on 40 cases operated on between 1984 and 1987 and followed from 2 to 5 years. The perioperative complication rate was 47.5%, with a 7.5% incidence of persistent sequelae: severe C-5 radiculopathy in one patient, swallowing dysfunction in one, and hypoglossal nerve palsy in one. No single factor (age, duration of symptoms, or severity of myelopathy) was absolutely predictive of outcome; however, syndromes of short duration had the best likelihood of cure. Similar outcomes were associated, individually, with long duration of symptoms, age over 70 years, and severe myelopathy. After factoring a 5% regression of improvement, the long-term cure rate was 57.5% and the failure rate was 15%. Myelopathy worsening was not documented. KEY WORDS
cervical spine • myelopathy • spondylosis • corpectomy •
spinal cord
ENTRAL
corpectomy is reportedly a more effective treatment for cervical spondylotic myelopathy than laminectomy and discectomy/interbody fusion. 4,6.21,27,31,42,45,48.51 The logic of corpectomy lies in an aggressive direct approach to the spondylotic
process (the anterior spine) by fully decompressing the anterior spinal canal and eliminating multisegmental spinal mobility. Oddly, neurosurgeons seem to have neglected this strategy; the reported experience since 1976 is orthopedic, except for that of Kojima, et a1., 31 in 1989 (Table 1). In 1984, we initiated the use of central corpectomy for treatment of a consecutive unselected series of patients with cervical spondylotic myelopathy at our institution; we describe our surgical method and longterm outcome in 40 patients, followed now for 2 to 5 years. Our observations suggest that the benefits of this approach for this disorder far exceed those of more commonly performed procedures. Summary of Cases
Since 1984, a consecutive series of patients diagnosed as having cervical spondylotic myelopathy has been evaluated for surgical treatment. By prospective plan, the only surgical treatment offered was that of central corpectomy and strut grafting. Of 94 patients, 82 have J. Neurosurg. / Volume 74 / February, 1991
undergone surgery to date. The median age of the group was 64 years. Among the patients not yet operated on, there are no severe or progressive syndromes. From 1984 until 1987, 40 patients (22 males and 18 females) were operated on and form the basis of this report.
Diagnosis The diagnosis of cervical spondylotic myelopathy was based on clinical presentation and radiographic
documentation of spinal cord compression by spondylotic disease. Two patients had spondylosis superimposed on congenital stenosis. Myelography alone or combined with computerized tomography or, more recently, magnetic resonance (MR) imaging was used for imaging the spinal pathology. Many patients have had all three studies in the course of investigation.
Preoperative Symptoms Virtually all patients had either the transverse syndrome or the central syndrome according to the myelopathic classification of Crandall and Batzdorf. 9 Patients with transverse syndromes had varying degrees of spasticity, gait disturbance, and posterior column dysfunction. The gait disorder was graded by the classification of Nurick 37 (Table 2). Patients with either a Brown-Séquard syndrome or episodes of transient pa163
R. L. Saunders, et al. TABLE I Published corpectorny series for cervical spondyloirc myelopathy* Authors & Year Whitecloud & LaRocca, 1976 Boni, et al., 1984 Hukuda, or al., 1985 Yonenobu, of al., 1985 Senegas, of al., 1985 Hanai, or al., 1986 Bernard & Whitecloud, 1987 Zdeblick & Bohlman, 1989 Kojima, of al., 1989
No. of Cases 9 29 12 21 45 30 21 8 19
Outcome Age (yrs) 45 (mean)
53 (mean) 56 (mean) 54 (mean) 46 (mean) 55 (mean)
Improved
Worse
100% 97%
0
73% 100% 85%t 100% 87%
0 0
0 0 0
Follow-Up Period (mos) 10-41 6-150 31 (mean) 30 (mean) 48 (mean) 36 (mean) 32 (mean)
0
*— = Data not available. f At 1 year follow-up review this dropped to 76%.
TABLE 2
TABLE 3
Classification of disability in spondylotic myelopathy*
Characterization of central syndrome in 12 cases
Definition Grade 0 root signs & symptoms, no evidence of cord involvement 1 sign of cord involvement, normal gait 2 mild gait impairment, able to be employed 3 gait abnormality prevents employment able to ambulate only with assistance 4 5 chair bound or bedridden * Classification according to Nurick. 37
ralysis were also included in the transverse-syndrome group. Those with central syndromes had upper-extremity disturbances of various degrees of severity, characterized by sensory and motor symptoms, and/ or signs (Table 3), Most of the minor or subjective myelopathies were grouped into the central category, as well. Twenty-eight of the 40 cases suffered from transverse syndromes, 19 of which were considered severe, having a gait disturbance of at least Grade 3, an associated central syndrome, a Brown-S6quard syndrome, or resolving acute quadriplegia. Five of the 12 cases of central syndromes were considered severe, due to the patients' hyperpathic useless hands or associated gait disturbance. Symptoms were considered arbitrarily of short duration if present for less than 1 year, and long if otherwise. In recent years, four-level corpectomies have been performed; however, in this series the longest decompression was a three-level corpectomy, performed in 22 of the 40 cases; there were 15 two-level and three one-level corpectomies.
Surgical Procedure The procedure is carried out under nonparalytic anesthesia if feasible and prophylactic antibiotics are used. The patient is positioned supine; the head and neck are placed in a neutral position, with 15 lbs of Gardner tong traction. One iliac crest and lower leg (fibula) are prepared and draped, and the surgery is commenced at the neck. 164
Signs & Symptoms
No. of Cases
weak hands useless hands hand paresthesiae hand dysesthesiae hand hyperpathia Lhermitte's sign bladder dysfunction spastic paraparesis
1 5 1 5 2 6 1 3
In the short fat neck, a transverse incision is made in the appropriate crease, without great concern for the exact level. In the long thin neck, an incision paralleling the anterior border of the right sternocleidomastoid muscle is made, curving onto the muscle just below the pinna. Through either incision, the anterior border of the sternocleidomastoid muscle is mobilized throughout as long an extent as possible to allow for ultimate full soft-tissue relaxation. The avascular plane between the lateral carotid sheath and the midline strap muscles is widened, and the prevertebral space entered. The omohyoid muscle is transected and not repaired. A lateral spine radiograph is taken upon access to the prevertebral space. The bony ligamentous keel lying on the midline of C-2 is a reliable landmark for further tailoring the exposure. After the spine film confirms the exact level of surgery, a table-fixed retraction device is positioned. The midline structures are retracted just beyond the midpoint of the left longus colli muscle using a wide retractor blade perpendicular to the muscle. If further cephalad exposure for resection of the C-3 vertebra is required, two long narrow blades are angled under the mandible from a rod extension placed over the face from either the left or right side. Right lateral retraction is accomplished with a Richardson-like blade cephalad and an angled shallow blade caudally, placing these lateral retractors in as low a position as possible to J. Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy minimize the depth of the prevertebral exposure. The longus colli muscles are then resected about 5 mm to either side to expose a swath down the midline of the vertebral bodies to be resected, measuring about 1.5 to 1.8 cm. The pressure from the left-sided retractor will retract the left longus colli muscle adequately. Stay sutures with elastic bands are placed on the right longus colli muscle for further retraction. The intervertebral spaces of the segments to be removed are then evacuated of disc material, facilitated by a sharp air-driven burr if the spaces are substantially narrowed. Spreaders are not used. The uncinate processes on both the right and left are defined at each interspace to define a central approach to the spinal canal. With the use of a Lexell rongeur and the highspeed air-driven cutting burr, the vertebral bodies are then removed, down to the posterior cortical bone overlying the posterior longitudinal ligament. The width of the anterior bone trough measures about 1.5 cm, and flares in a trapezoid-like fashion toward the anterior spinal canal to a point measuring approximately 1.8 cm at the canal itself. Dissection to this point has been carried out under loupe magnification; with the appearance of islands of the posterior longitudinal ligament through the posterior vertebral cortex, the operating microscope is first used, and the air-driven burr is changed to a diamond burr. All bone is then removed from the posterior longitudinal ligament under the operating microscope, and the oppposing cartilaginous end plates of the remaining vertebral bodies are curetted away. A No. 15 Bard Parker blade or an arachnoid knife is used to incise the posterior longitudinal ligament as caudally as possible in the midline. The underlying, dura is readily apparent in most cases due to the appearance of a scant amount of moisture and the shining dural surface. Some delicate adhesions of the ligament to the dura are not unusual, and do not ordinarily interfere with the ligament's definition from the dura. With the dural plane identified, the ligament is lifted cephalad with a surgical clamp, and lateral gutter attachments are sharply cut, using a blade, a sharp curette, or a small angled Kerrison rongeur. The ligament is then transected at the uppermost point close to the remaining vertebral segment. The remaining ligament in the midline caudally is removed with an angled Kerrison ronguer. The ligament in the lateral gutters will ordinarily thin out, and its removal is unnecessary; however, in some cases the liagment will be heavy in the lateral gutters, and removal is elected. There are large epidural veins in the lateral gutter and in the substance of the lateral ligament itself, which can pose a problem. If these cannot be coagulated with bipolar forceps set at a 5- or 10-W setting, Gelfoam held in place with a surgical patty is usually effective. Any minor spondylotic ridge at the caudal remaining vertebra is removed using the high-speed diamond burr. A spondylotic ridge at the remaining cephalad vertebral segment is an indication for removal of that segment J. Neurosurg. / Volume 74 / February, 1991
FIG. 1. Operative view of a corpectorny "trough" with wide exposure of the longus colli muscles. Note the characteristic bulging of the dura with excision of the posterior longitudinal ligament.
if it is of any magnitude at all. This strategy avoids excessive reduction in height of the cephalad vertebral segment resulting from the angle of the interspace. Removal of a spondylotic ridge at the upper vertebral segment may reduce the anterior vertebral height precariously, risking fracture of the vertebra. With removal of the posterior longitudinal ligament, the dura characteristically bulges into the decompression (Fig. 1), and the spinal cord may be seen through the translucent dura. If the uncinate processes have been used to centralize the decompression, ultrasound or x-ray imaging are probably not necessary. By actual measurement, the decompression at the anterior spinal canal is at least 1.8 cm. After mortices are created in the opposing surfaces of the remaining vertebral bodies, the orthopedic team harvests the appropriate bone and tailors it to the decompression site. The fit of the graft is reasonably secure without added traction, using a so-called "keystone" graft. The 15 lbs of tong traction is removed, and a lateral spine radiograph is taken to determine whether the graft is acceptable in its alignment and relationships to the anterior and posterior vertebral body planes. The intraoperative film can demonstrate a misaligned axis of the graft, incomplete seating of the end of the graft in the mortice, or excessively thick posterior mortice lips at the expense of proper depth of the graft. If any inadequacy such as these is noted, the graft is revised. A later film taken in the recovery room is used for comparison, to assess any displacement of the graft with reversal of anesthesia. The surgical wounds are then irrigated and closed over suction drains. The orthotic appliance of the orthopedist's choice is placed. This is usually either a cervicothoracic jacket or a Philadelphia collar. Halo orthotic systems are rarely used, and are predicated on the determination of pos165
R. L. Saunders, et al. TABLE 4 ,S'umrnary ref perioperative complications
FIG. 2. Magnetic resonance image in a patient suffering return of myelopathy 36 months after "cure by multilevel corpectomy. Note the inadequate longitudinal decompression and characteristic bulging of suharachnoid space with decompression.
terior spinal column instability. Such a situation might exist in patients with a previous laminectomy. The patient is mobilized the following day, and the cervical drain is removed. Daily radiographs are obtained until the graft is believed to be stable, and the patient is discharged as soon as the discomfort at the hip or leg graft donor site allows. Outpatient postoperative follow-up evaluations are scheduled for 2 weeks, 6 weeks, 3 months, 6 months, and then yearly, with plain x-ray films taken at each visit. If an iliac crest graft is used, the orthotic device is usually discontinued at approximately 2 months. After the orthotic system is discontinued, isometric and stretching exercises are initiated to promote neck flexibility. Radiographs are taken approximately 2 weeks after orthotic discontinuance. If residual symptoms or myelopathy are present, an MR image is obtained during the first months after surgery. If the patient has had resolution of myelopathic symptoms, it is advised that MR imaging be performed sometime during the 1st postoperative year for decompression documentation.
Outcome Improvement was usually apparent during the 1st postoperative week or weeks; however, three patients in our series experienced a slow recovery, with improvement not being apparent for 6 months. Gait disturbance, bladder dysfunction, and severe hyperpathia were the earliest complaints most likely to be relieved. The outcome was defined as a cure if the patient was satisfied, had no complaints, and there were no signs other than active reflexes or residual minor hand intrinsic atrophy. Improvement was defined as a change of at least two grades in the Nurick classification (Table 2), and/or resolution of hyperpathia and dysesthesia. Less change in grade or regression of improvement was considered a failure. 166
Perioperative Complications C-5 radiculitis graft mortise fracture graft displacement requiring surgery protracted swallowing dysfunction hoarseness soft-tissue hematorna requiring surgery stridor, reintubated hyperventilation escalation of hiatal hernia symptoms pneumonia incarcerated hernia at iliac crest osteomvelitis at donor site hyponatremia perforated sigmoid diverticulum
No. of Cases
5 1 1 1 1 2 1 1 1 1 1 1
By these criteria, 25 patients (62.5%) were considered cured by corpectomy; however, two (8%) of these patients had regression of improvement, resulting in a 57.5% cure rate at the time of this report, The two patients who regressed, at 18 months and 36 months, had exhibited posterior column signs preoperatively, which ultimately recurred. Repeat MR imaging showed inadequate decompression (Fig. 2). If the six patients with minor central myelopathies (five of whom were cured) are eliminated as unrepresentative cases because the symptoms were largely subjective, there are 18 continuing cures in 34 long-term patients, for a 53% cure rate. Equal numbers of the 40 patients had symptoms of long and short duration, with 45% and 70% cure rates, respectively. Of 24 patients preoperatively considered severely impaired, 50% were cured. Four of nine severely impaired patients with symptoms of long duration were cured. Of the 14 patients aged 70 years or older, six were cured. Of the 28 patients with transverse syndromes, 13 had persistent cures, six of whom had been severely impaired. All four patients with central
syndromes were cured, three of whom had had symptoms of short duration. The failure group comprised 15% of the series. This included two cases of cure regression, one case of error in diagnosis (multiple sclerosis), and three patients with severe syndromes of long duration with only minimal improvement. Perioperative Complications
The perioperative complication rate was 47.5%, the specifics of which are enumerated in Table 4. The most common problem was a C-5 radiculitis. One bone graft displacement occurred immediately after surgery and was revised. One patient with mortice fracture was treated with bedrest and cervical traction. Morbidity
There was one error in diagnosis, which ultimately Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy TABLE 5 Long-term sequelae of corpectomy Complication
No. of Cases
permanent (7.5%) C-5 radiculopathy swallowing dysfunction hypoglossal nerve palsy resolved (5%) recurrent laryngeal nerve palsy pseudarthrosis (? fracture)
proved to be multiple sclerosis, as suspected preoperatively. There were five patients with long-term sequelae of corpectomy (Table 5); three of these persist at the time of this report but do not include myelopathic worsening. Mortality
One patient, who was quadriplegic preoperatively, had subtle improvement postoperatively, but remained unable to breathe without assistance; by her own choice she was sedated and allowed to die. One patient with a severe myelopathy improved insignificantly after corpectomy, but died of unrelated causes 10 months postsurgery. Discussion The poor surgical results in patients with cervical spondylotic myelopathy are often ascribed to degenerative or ischemic processes of the spinal cord compounding spondylotic compression. 3 .''' These poor results are striking by comparison with those of other forms of benign spinal cord compression, such as extraaxial tumor and ossification of the posterior longitudinal ligament, even though the myelopathies are fundamentally similar. 20,22,33,35,36,43 The basis for the role of ischemia in this disorder lies not in pathological findings of vascular disease in cervical spondylotic myelopathy,'""° but the not-surprising experimental observations that ischemia will compound a compressive mydopathy. 17,18,26,28,47 An equally compelling explanation for surgical failure in cervical spondylotic myelopathy lies in the possibility that conventional surgical treatment for cervical spondylotic myelopathy is different, or less definitive, than that for extra-axial tumors and ossification of the posterior longitudinal ligament.
Pathogenesis of Myelopathy Spondylotic myelopathy is thought to stem from at least three treatable factors: spinal stenosis;"'"'"'" anterior osteophytesP" and relatively excessive spinal mobility.'" 2 Adams and Logue' first reported that those patients with best results after laminectomy or conservative methods had relatively stiff spines in comparison with those patients with poor results. Experimental and clinical evidence suggests that a myelopathy can be ameliorated by spinal immobilization. 5 . 11 '"'" J. Neurosurg. / Volume 74 / February, 1991
Surgical Approaches The commonly performed surgical procedures for cervical spondylotic myelopathy, laminectomy, and discectomy/interbody fusion seemingly ignore the multifactorial causes of cervical spondylotic myelopathy. 10,23.44 The cure rate with laminectomy is no more than 20% (CA Fager, unpublished data, 1989), and the overall improvement rate is under 70%. 8 . 19 Conventional anterior procedures have a cure rate of 33%, and an overall incidence of beneficial outcome of less than 75%." Corpectomy with strut grafting, which treats simultaneously several factors causative in cervical spondylotic myelopathy, has a success rate of 73% to 100% (Table 1). Surgical Outcome The 2- to 5-year follow-up period in this small group has not yet documented progression of osteophyte formation above and below the immobilized cervical segment. With longer follow-up monitoring, the progression of spondylosis may demonstrate inevitable recurrent spinal cord compression, although this arthritic progression may be an overestimated concern. 4' 12 The presumption that certain conditions must have a specific treatment (such as laminectomy for congenital stenosisi has not been borne out in this series (Fig. 3). The presumed need for posterior decompression for posterior canal encroachment' also has not been corroborated by our experience (Fig. 4). Since there were no selection criteria in this consecutive series, corpectomy was performed in some frail elderly patients; more than 25 % of the patients were older than 70 years. In our experience, cardiovascular risk factors and osteoporosis, and not age in itself, are critical factors with corpectomy. Fifty percent of the patients over 70 years of age were cured, as were nearly 50% of the patients with severe myelopathies. Long duration of severe symptoms prior to treatment emerged as the single complex most predictive of outcome; nevertheless, three such patients had long-term cures. Length of Decompression Our series suggests that patients with resection of the most segments have the best outcomes, favoring the choice of longer longitudinal decompression using fibular grafts if necessary. We prefer autogenous iliac crest as strut graft material rather than fibula, and accordingly have restricted the extent of longitudinal decompression to three segments. The use of iliac bone for a longer graft is simply not feasible, at least in our hands. Iliac crest graft, involving heavily cancellous bone of substantial diameters, has been preferred for its early incorporation. This graft, however, should be cautiously selected in patients requiring a strut of greater than two levels, especially in cases of osteoporosis, as two such grafts have fractured in our series. Yonenobu, et al.," have cautioned that patients requiring a vertebrectomy involving more than four 167
R. L. Saunders, et al.
FIG. 3. Left: Preoperative magnetic resonance image showing congenital canal stenosis with superimposed spondylosis. Right: Postoperative image from the same patient after a four-level decompression and fibular graft placement. The canal is now capacious.
segments be managed by laminectomy. The highly stressed, if not mobile, motion segment between the cervical spine and the relatively immobile thoracic spine, if crossed with a bone strut would result in a lever arm of such length that it would readily fracture a mortice. This would indicate that the most radical cervical corpectomy feasible is the removal of a lower portion of C-2 through C-6, with grafting from C-2 to C-7. (Resection of 3 mm of the upper body of C-7 might safely address a C6-7 ridge.) Some range of motion will be lost with such a multisegmental strut, but it is practically insignificant, since the degenerative joint disease of the older patient's neck limits flexion, extension, and rotation to the occiput, C-1, and C-2 levels. Short cervical grafts at the cervicothoracic junc-
FIG. 4. Left: Preoperative myelogram showing marked spondylotic ridging at C3-4 and C4-5, in particular. There is apparent posterior cord encroachment. Right: Myelogram in the same patient after C3-5 corpectomy. The canal is now capacious anteriorly as well as posteriorly.
1 68
tion are probably not as dangerous; we have treated one patient who had an uncomplicated strut graft from C6 to T-2, and Yoshizu, et at," described a graft from C-6 to T-10 without apparent consequence. Choice of Bone Graft We use autogenous bone and avoid implanted hardware to minimize the possibility of long-term complications. We recognize that the use of plates and screws, at least in the short term, would lessen the risk of graft displacement' and the use of allograft bone would eliminate donor site complications. Generally, our perioperative complications have decreased with practice. However, in our ongoing series we continue to have about a 10% incidence of graft fractures and/or displacements, not all necessarily requiring surgical revision. The preference for using fibula to graft more than two segments' should be questioned in light of its limitations in certain patients. The fibula, being mostly cortical, is incorporated very slowly. The shortest period reported for the uniting of fibula with vertebra is 6 months; 4 ' the time for complete incorporation of the fibula graft is more likely closer to 1 year. In the patient with osteoporosis, the fibula probably should not be used, unless protected with a halo vest. The dense character of fibula used in the osteoporotic spine creates the phenomenon of "pistoning" (Fig. 5), which risks fracturing off the front of the vertebral body as the fibula graft settles through its mortices. Although we have recognized but one pseudarthrosis, which occurred after substantial neck stress and could be considered a graft fracture, this concern might warrant a more liberal use of fibula. There is no reported pseudarthrosis with fibular grafting, whereas the incidence with iliac crest may be as high as 20% (T Whitecloud, unpublished data, 1978). Of our 40 patients with long-term follow-up monitoring, 10% have .1. Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy
FIG. 5. Studies of a recently performed procedure not included in this long-term series. Left: X-ray film of the fibular graft obtained immediately postoperatively in a osteoporotic individual. Right: X-ray film in the same patient 3 days later showing that the dense graft material has penetrated the end plate. This is an example of "pistoning. -
neck pain, raising the question of possibly undiagnosed pseudarthrosis. Perioperative Complications The etiology of the C-5 radiculitis seen in our series (Table 4) is as yet uncertain, and has not been described by others. N Epstein (personal communication, 1989) has implicated the unusual length of the C-5 root, noting that C-5 complications occur in posterior decompression as well as conventional anterior surgery. The C-5 root may be uniquely sensitive to inadequate foraminal decompression, inherent in the central corpectomy procedure. Alternatively, this may be an unusual expression of a segmental myelopathy or a surgical myelopathic complication (JN Abramowitz, unpublished data, 1989). That this is not a myelopathic effect is supported by the delayed onset, often on the 2nd day after surgery, and the complete resolution of the syndrome in most patients within several months. Most recently, by avoiding a paralytic anesthesia technique at Sypert's suggestion (G Sypert, personal cornmunication, 1989), we have noted the C-5 segment to be most sensitive not to instrumental compression but to the heat of either unipolar or bipolar coagulation. As low a bipolar setting as 10 W, when applied to the posterior longitudinal ligament, may evoke a major deltoid twitch. It could be that we have inadvertently caused heat injuries to a uniquely sensitive root. Our long-term cure rate of greater than 50% is unusual in the treatment of cervical spondylotic myelopathy 19 and places the substantial perioperative complications in perspective. Other corpectomy series do not describe any mortality, and considerably less morbidity; however, graft problems are reported in 3% to 21% of cases ,46,2 1,23 a favorable comparison with our 5%. Despite our high perioperative complication rate, as in other corpectomy series we have noted no myelopathic morbidity. This underscores one of the most important aspects of corpectomy when considered against the significant incidence of postoperative worsJ. Neurosurg. / Volume 74 / February, 1991
ening with cervical laminectomy and discectomy/interbody fusion.' Regression of improvement after corpectomy deserves more attention, and warrants long-term followup evaluation of series prior to reporting. Of the corpectomy series (Table 1), only that of Bernard and Whitecloud 4 includes two patients who deteriorated 1 year after initial improvement, a regression rate of 11%, with which our rate of 5% compares favorably. Regression after laminectomy is reported as between 19% and 53% and is in large part due to instability and/or kyphosis. 4 Galera and Tovi,' discussing long-term follow-up results after the Cloward procedure for cervical spondylotic myelopathy, described most of their patients as ultimately deteriorating; the report of Lunsford, et al.,' was nearly as pessimistic. In our two patients with long-term deterioration, follow-up MR images showed that their longitudinal decompressions were inadequate (Fig. 3). Furthermore, as noted earlier, longer decompressions are related positively to outcome. In conclusion, the logic of a multifactorial surgical approach to cervical spondylotic myelopathy is supported by the results of this unselected series. A cure rate in excess of 50%, a regression rate of 5%, and nonexistent myelopathic morbidity associated with central corpectomy are suggestive evidence that the generally accepted surgical results in cervical spondylotic myelopathy should be reconsidered, especially if our observations are confirmed by a randomized trial. References 1. Adams CBT, Logue V: Studies in cervical spondylotic myelopathy. II. The movement and contour of the spine in relation to the neural complications of cervical spandylosis. Brain 94:569-586, 1971 2. Barnes MP, Saunders M: The effect of cervical mobility on the natural history of cervical spondylotic myelopathy. J Neurol Neurosurg Psychiatry 47:17-20, 1984 3. Bedford PD, Bosanquet FD, Russell WR: Degeneration of the spinal cord associated with cervical spondylosis. Lancet 2:55-59, 1952 4. Bernard TN, Whitecloud TS: Cervical spondylotic myelopathy and myeloradiculopathy: anterior decompression and stabilization with autogenous fibula strut graft. Clin Orthop 221:149-157, 1987 5. Bohlman HH: Cervical spondylosis with moderate to severe myelopathy. A report of seventeen cases treated by Robinson anterior cervical discectomy and fusion. Spine 2:151-162, 1977 6. Boni M, Cherubino P, Denaro V, et al: Multiple subtotal somatectomy. Spine 9:358-362,1984 7. Brown JA, Havel P, Ebraheim N, et al: Cervical stabilization by plate and bone fusion. Spine 13:236-240, 1988 8. Carol MP, Ducker TB: Cervical spondylotic myelopathies: surgical treatment. J Spinal Dis 1:59-65, 1988 9. Crandall PH, Batzdorf U: Cervical spondylotic myelopathy. J Neurosurg 25:57-66,1966 10. Cusick JF, Myklebust JB: Biomechanics of cervical spondylotic myelopathy. Contemp Neurosurg 9(5):1-8,1987 11. Cusick JF, Steiner RE, Berns T: Total stabilization of the
1 69
R. L. Saunders, et al. cervical spine in patients with cervical spondylotic myelopathy, Neurosurgery 18:491-495, 1986 12. DePalma AF, Rothman RH, Lewinnek GE, et al: Anterior interbody fusion for severe cervical disc degeneration. Surg Gynecol Obstet 134:755-758, 1972 13. Ducker TB: Experimental injury of the spinal cord, in Vinken PJ, Bruyn GW (eds): Handbook of Clinical Neurology. Injuries of the Spine and Spinal Cord and Column.
Amsterdam: North-Holland, 1976, Vol 25, pp 9-26 14. Epstein JA, Carras R, Lavine LS, et al: The importance of removing osteophytes as part of the surgical treatment of myeloradiculopathy in cervical spondylosis. J Neurosurg 30:219-226, 1969 15. Galera R, Tovi D: Anterior disc excision with interbody fusion in cervical spondylotic myelopathy and rhizopathy. J Neurosurg 28:305-310,1968 16. Gillilan LA: Veins of the spinal cord. Anatomic details: suggested clinical applications. Neurology 20:860-868, 1970 17. Gooding MR, Wilson CB, Hoff JT: Experimental cervical myelopathy: autoradiographic studies of spinal cord blood flow patterns. Surg Neurol 5:233-239,1976 18. Gooding MR, Wilson CB, Hoff JT: Experimental cervical myelopathy. Effects of ischemia and compression of canine cervical spinal cord. J Neurosurg 43:9-17, 1975 19. Gorier K: Influence of laminectorny on the course of cervical myelopathy. Acta Neurochir 33:265-281, 1976 20. Hanai K, Adachi H, Ogasawara H: Axial transverse tomography of the cervical spine narrowed by ossification of the posterior longitudinal ligament. J Bone Joint Surg (Br) 59:481-484, 1977 21. Hanai K, Fuminori F, Kunitaka K: Subtotal vertebrectomy and spinal fusion for cervical spondylotic myelopathy. Spine 11:310-315, 1986 22. Harsh GR IV, Sypert OW, Weinstein PR, et al: Cervical spine stenosis secondary to ossification of the posterior longitudinal ligament. J Neurosurg 67:349-357, 1987 23. Herkowitz HN: The surgical management of cervical spondylotic radiculopathy and myelopathy. Clin Orthop 239:94-108, 1989 24. Hinck VC, Sachdev NS: Developmental stenosis of the cervical spinal canal. Brain 89:27-36, 1966 25. Hughes JT, Brownell B: Spinal cord ischemia due to arteriosclerosis. Arch Neurol 15:189-202, 1966 26. Hukuda S, Mochizuki T, Ogata M, et al: A comparison of the results of anterior and posterior procedures. J Bone Joint Surg (Br) 67:609-615, 1985 27. Hukuda S, Ogata M, Katsuura A: Experimental study on acute aggravating factors of cervical spondylotic myelography. Spine 13:15-20, 1988 28. Hukuda 5, Wilson CB: Experimental cervical myelopathy: effects of compression and ischemia on the canine cervical cord. J Neurosurg 37:631-652,1972 29. Idwards WC, LaRocca H: The developmental segmental sagittal diameter of the cervical spinal canal in patients with cervical spondylosis. Spine 8:20-27, 1983 30. Jellinger K: Spinal cord arteriosclerosis and progressive vascular myelopathy, J Neurol Neurosurg Psychiatry 30: 195-206,1967 31. Kojima T, Waga S, Kubo Y, et al: Anterior cervical vertebrectomy and interbody fusion for multi-level spondylosis and ossification of the posterior longitudinal ligament. Neurosurgery 24:864-872, 1989 32. Lunsford LD, Bissonette DJ, Zorub DS: Anterior surgery for cervical disc disease. Part 2: Treatment of cervical spondylotic myelopathy in 32 cases. J Neurosurg 53: 12-19,1980 33. Manabe S, Nomura S: [Anterior decompression for ossi170
fication of the posterior longitudinal ligament of the cervical spine.] Neurol Surg 5:1253-1259,1977 (Jpn) 34, Miller CA: Shallow cervical canal: recognition, clinical symptoms, and treatment. Contemp Neurosurg 7(7):1-6, 1985 35. Murakami N, Muroga T, Sobue I: Cervical myelopathy due to ossification of the posterior longitudinal ligament. A clinicopathologic study. Arch Neurol 35:33-36,1978 36. Nagashima C: Cervical myelopathy due to ossification of the posterior longitudinal ligament. J Neurosurg 37: 653-660,1972 37. Nurick S: The natural history and the results of surgical treatment of the spinal cord disorder associated with cervical spondylosis. Brain 95:101-108, 1972 38. Nurick S: The pathogenesis of the spinal cord disorder associated with cervical spondylosis. Brain 95:87-100, 1972 39. Payne EE, Spillane JD: The cervical spine. An anatomicopathological study of 70 specimens (using a special technique) with particular reference to the problem of cervical spondylosis. Brain 80:571-596, 1957 40. Roberts AH: Myelopathy due to cervical spondylosis treated by collar immobilization. Neurology 16:951-954, 1966 41. Rossier AB, Hussey RW, Kenzora JE: Anterior fibula interbody fusion in the treatment of cervical spinal cord injuries. Surg Neurol 7:55-60,1977 42. Senegas J, Guerin J, Vital JM, et al: Decompression medullaire etendue par voie anterieure dans le traitement des myelopathies par cervicarthrose. Rev Chir Orthop 71: 291-300,1985 43. Stookey B: Compression of the spinal cord due to ventral extradural chondromas. Arch Neurol Psychiatry 20: 275-280,1928 44. Whitecloud TS: Anterior surgery for cervical spondylotic myelopathy. Smith-Robinson, Cloward, and vertebrectomy. Spine 13:861-863, 1988 45, Whitecloud TS, LaRocca H: Fibular strut graft in reconstructive surgery of the cervical spine. Spine 1:33-43, 1976 46. Wilkinson M: The morbid anatomy of cervical spondylosis and myelopathy. Brain 83:589-617,1960 47. Wilson CB, Landry RNA: Experimental cervical myelopathy: 1. Blood supply of the canine cervical spine cord. Neurology 14:809-8 14, 1964 48. Yonenobu K, Fuji T, Ono K, et al: Choice of surgical treatment for multisegmental cervical spondylotic myelopathy. Spine 10:710-716, 1985 49. Yoshizu T, Hara T, Honma T, et al: Experience with anterior spinal fusion by vascularized fibula as a strut graft in kyphosis and kyphoscoliosis. J Microsurg 3: 123-124,1981 (Abstract) 50. Yu YL, Jones SJ: Somatosensory evoked potentials in cervical spondylosis. Correlation of median, ulnar and posterior tibial nerve responses with clinical and radiological findings. Brain 108:273-300, 1985 51. Zdeblick TA, Bohiman HH: Cervical kyphosis and myelopathy. Treatment by anterior corpectomy and strutgrafting. J Bone Joint Surg (Am) 71:170-182,1989 Manuscript received January 11,1990. Accepted in final form June 29,1990. This paper was presented in part at the Annual Meeting of the American Association of Neurological Surgeons, Dallas, Texas, on May 4,1987. Address reprint requests to: Richard L. Saunders, M.D., Section of Neurosurgery, Dartmouth-Hitchcock Medical Center, 2 Maynard Street, Hanover, New Hampshire 03756.
J. Neurosurg. Volume 74 / February, 1991
Central corpectomy for cervical spondylotic myelopathy: a consecutive series with long-term follow-up evaluation RICHARD L. SAUNDERS, M.D., PHILIP M. BERNINI, M.D., THOMAS G. SHIRREFFS, JR., M.D., AND ALEXANDER G. REEVES, M.D.
Sections of Neurosurgery and Orthopedics, Department of Surgery, and Section of Neurology, Department of Medicine, Dartmouth-Hitchcock Medical Center, Hanover, New Hampshire
V Since 1984, a consecutive series of patients with cervical spondylotic myelopathy has been treated by central corpectomy and strut grafting. This report focuses on 40 cases operated on between 1984 and 1987 and followed from 2 to 5 years. The perioperative complication rate was 47.5%, with a 7.5% incidence of persistent sequelae: severe C-5 radiculopathy in one patient, swallowing dysfunction in one, and hypoglossal nerve palsy in one. No single factor (age, duration of symptoms, or severity of myelopathy) was absolutely predictive of outcome; however, syndromes of short duration had the best likelihood of cure. Similar outcomes were associated, individually, with long duration of symptoms, age over 70 years, and severe myelopathy. After factoring a 5% regression of improvement, the long-term cure rate was 57.5% and the failure rate was 15%. Myelopathy worsening was not documented. KEY WORDS
cervical spine • myelopathy • spondylosis • corpectomy •
spinal cord
ENTRAL
corpectomy is reportedly a more effective treatment for cervical spondylotic myelopathy than laminectomy and discectomy/interbody fusion. 4,6.21,27,31,42,45,48.51 The logic of corpectomy lies in an aggressive direct approach to the spondylotic
process (the anterior spine) by fully decompressing the anterior spinal canal and eliminating multisegmental spinal mobility. Oddly, neurosurgeons seem to have neglected this strategy; the reported experience since 1976 is orthopedic, except for that of Kojima, et a1., 31 in 1989 (Table 1). In 1984, we initiated the use of central corpectomy for treatment of a consecutive unselected series of patients with cervical spondylotic myelopathy at our institution; we describe our surgical method and longterm outcome in 40 patients, followed now for 2 to 5 years. Our observations suggest that the benefits of this approach for this disorder far exceed those of more commonly performed procedures. Summary of Cases
Since 1984, a consecutive series of patients diagnosed as having cervical spondylotic myelopathy has been evaluated for surgical treatment. By prospective plan, the only surgical treatment offered was that of central corpectomy and strut grafting. Of 94 patients, 82 have J. Neurosurg. / Volume 74 / February, 1991
undergone surgery to date. The median age of the group was 64 years. Among the patients not yet operated on, there are no severe or progressive syndromes. From 1984 until 1987, 40 patients (22 males and 18 females) were operated on and form the basis of this report.
Diagnosis The diagnosis of cervical spondylotic myelopathy was based on clinical presentation and radiographic
documentation of spinal cord compression by spondylotic disease. Two patients had spondylosis superimposed on congenital stenosis. Myelography alone or combined with computerized tomography or, more recently, magnetic resonance (MR) imaging was used for imaging the spinal pathology. Many patients have had all three studies in the course of investigation.
Preoperative Symptoms Virtually all patients had either the transverse syndrome or the central syndrome according to the myelopathic classification of Crandall and Batzdorf. 9 Patients with transverse syndromes had varying degrees of spasticity, gait disturbance, and posterior column dysfunction. The gait disorder was graded by the classification of Nurick 37 (Table 2). Patients with either a Brown-Séquard syndrome or episodes of transient pa163
R. L. Saunders, et al. TABLE I Published corpectorny series for cervical spondyloirc myelopathy* Authors & Year Whitecloud & LaRocca, 1976 Boni, et al., 1984 Hukuda, or al., 1985 Yonenobu, of al., 1985 Senegas, of al., 1985 Hanai, or al., 1986 Bernard & Whitecloud, 1987 Zdeblick & Bohlman, 1989 Kojima, of al., 1989
No. of Cases 9 29 12 21 45 30 21 8 19
Outcome Age (yrs) 45 (mean)
53 (mean) 56 (mean) 54 (mean) 46 (mean) 55 (mean)
Improved
Worse
100% 97%
0
73% 100% 85%t 100% 87%
0 0
0 0 0
Follow-Up Period (mos) 10-41 6-150 31 (mean) 30 (mean) 48 (mean) 36 (mean) 32 (mean)
0
*— = Data not available. f At 1 year follow-up review this dropped to 76%.
TABLE 2
TABLE 3
Classification of disability in spondylotic myelopathy*
Characterization of central syndrome in 12 cases
Definition Grade 0 root signs & symptoms, no evidence of cord involvement 1 sign of cord involvement, normal gait 2 mild gait impairment, able to be employed 3 gait abnormality prevents employment able to ambulate only with assistance 4 5 chair bound or bedridden * Classification according to Nurick. 37
ralysis were also included in the transverse-syndrome group. Those with central syndromes had upper-extremity disturbances of various degrees of severity, characterized by sensory and motor symptoms, and/ or signs (Table 3), Most of the minor or subjective myelopathies were grouped into the central category, as well. Twenty-eight of the 40 cases suffered from transverse syndromes, 19 of which were considered severe, having a gait disturbance of at least Grade 3, an associated central syndrome, a Brown-S6quard syndrome, or resolving acute quadriplegia. Five of the 12 cases of central syndromes were considered severe, due to the patients' hyperpathic useless hands or associated gait disturbance. Symptoms were considered arbitrarily of short duration if present for less than 1 year, and long if otherwise. In recent years, four-level corpectomies have been performed; however, in this series the longest decompression was a three-level corpectomy, performed in 22 of the 40 cases; there were 15 two-level and three one-level corpectomies.
Surgical Procedure The procedure is carried out under nonparalytic anesthesia if feasible and prophylactic antibiotics are used. The patient is positioned supine; the head and neck are placed in a neutral position, with 15 lbs of Gardner tong traction. One iliac crest and lower leg (fibula) are prepared and draped, and the surgery is commenced at the neck. 164
Signs & Symptoms
No. of Cases
weak hands useless hands hand paresthesiae hand dysesthesiae hand hyperpathia Lhermitte's sign bladder dysfunction spastic paraparesis
1 5 1 5 2 6 1 3
In the short fat neck, a transverse incision is made in the appropriate crease, without great concern for the exact level. In the long thin neck, an incision paralleling the anterior border of the right sternocleidomastoid muscle is made, curving onto the muscle just below the pinna. Through either incision, the anterior border of the sternocleidomastoid muscle is mobilized throughout as long an extent as possible to allow for ultimate full soft-tissue relaxation. The avascular plane between the lateral carotid sheath and the midline strap muscles is widened, and the prevertebral space entered. The omohyoid muscle is transected and not repaired. A lateral spine radiograph is taken upon access to the prevertebral space. The bony ligamentous keel lying on the midline of C-2 is a reliable landmark for further tailoring the exposure. After the spine film confirms the exact level of surgery, a table-fixed retraction device is positioned. The midline structures are retracted just beyond the midpoint of the left longus colli muscle using a wide retractor blade perpendicular to the muscle. If further cephalad exposure for resection of the C-3 vertebra is required, two long narrow blades are angled under the mandible from a rod extension placed over the face from either the left or right side. Right lateral retraction is accomplished with a Richardson-like blade cephalad and an angled shallow blade caudally, placing these lateral retractors in as low a position as possible to J. Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy minimize the depth of the prevertebral exposure. The longus colli muscles are then resected about 5 mm to either side to expose a swath down the midline of the vertebral bodies to be resected, measuring about 1.5 to 1.8 cm. The pressure from the left-sided retractor will retract the left longus colli muscle adequately. Stay sutures with elastic bands are placed on the right longus colli muscle for further retraction. The intervertebral spaces of the segments to be removed are then evacuated of disc material, facilitated by a sharp air-driven burr if the spaces are substantially narrowed. Spreaders are not used. The uncinate processes on both the right and left are defined at each interspace to define a central approach to the spinal canal. With the use of a Lexell rongeur and the highspeed air-driven cutting burr, the vertebral bodies are then removed, down to the posterior cortical bone overlying the posterior longitudinal ligament. The width of the anterior bone trough measures about 1.5 cm, and flares in a trapezoid-like fashion toward the anterior spinal canal to a point measuring approximately 1.8 cm at the canal itself. Dissection to this point has been carried out under loupe magnification; with the appearance of islands of the posterior longitudinal ligament through the posterior vertebral cortex, the operating microscope is first used, and the air-driven burr is changed to a diamond burr. All bone is then removed from the posterior longitudinal ligament under the operating microscope, and the oppposing cartilaginous end plates of the remaining vertebral bodies are curetted away. A No. 15 Bard Parker blade or an arachnoid knife is used to incise the posterior longitudinal ligament as caudally as possible in the midline. The underlying, dura is readily apparent in most cases due to the appearance of a scant amount of moisture and the shining dural surface. Some delicate adhesions of the ligament to the dura are not unusual, and do not ordinarily interfere with the ligament's definition from the dura. With the dural plane identified, the ligament is lifted cephalad with a surgical clamp, and lateral gutter attachments are sharply cut, using a blade, a sharp curette, or a small angled Kerrison rongeur. The ligament is then transected at the uppermost point close to the remaining vertebral segment. The remaining ligament in the midline caudally is removed with an angled Kerrison ronguer. The ligament in the lateral gutters will ordinarily thin out, and its removal is unnecessary; however, in some cases the liagment will be heavy in the lateral gutters, and removal is elected. There are large epidural veins in the lateral gutter and in the substance of the lateral ligament itself, which can pose a problem. If these cannot be coagulated with bipolar forceps set at a 5- or 10-W setting, Gelfoam held in place with a surgical patty is usually effective. Any minor spondylotic ridge at the caudal remaining vertebra is removed using the high-speed diamond burr. A spondylotic ridge at the remaining cephalad vertebral segment is an indication for removal of that segment J. Neurosurg. / Volume 74 / February, 1991
FIG. 1. Operative view of a corpectorny "trough" with wide exposure of the longus colli muscles. Note the characteristic bulging of the dura with excision of the posterior longitudinal ligament.
if it is of any magnitude at all. This strategy avoids excessive reduction in height of the cephalad vertebral segment resulting from the angle of the interspace. Removal of a spondylotic ridge at the upper vertebral segment may reduce the anterior vertebral height precariously, risking fracture of the vertebra. With removal of the posterior longitudinal ligament, the dura characteristically bulges into the decompression (Fig. 1), and the spinal cord may be seen through the translucent dura. If the uncinate processes have been used to centralize the decompression, ultrasound or x-ray imaging are probably not necessary. By actual measurement, the decompression at the anterior spinal canal is at least 1.8 cm. After mortices are created in the opposing surfaces of the remaining vertebral bodies, the orthopedic team harvests the appropriate bone and tailors it to the decompression site. The fit of the graft is reasonably secure without added traction, using a so-called "keystone" graft. The 15 lbs of tong traction is removed, and a lateral spine radiograph is taken to determine whether the graft is acceptable in its alignment and relationships to the anterior and posterior vertebral body planes. The intraoperative film can demonstrate a misaligned axis of the graft, incomplete seating of the end of the graft in the mortice, or excessively thick posterior mortice lips at the expense of proper depth of the graft. If any inadequacy such as these is noted, the graft is revised. A later film taken in the recovery room is used for comparison, to assess any displacement of the graft with reversal of anesthesia. The surgical wounds are then irrigated and closed over suction drains. The orthotic appliance of the orthopedist's choice is placed. This is usually either a cervicothoracic jacket or a Philadelphia collar. Halo orthotic systems are rarely used, and are predicated on the determination of pos165
R. L. Saunders, et al. TABLE 4 ,S'umrnary ref perioperative complications
FIG. 2. Magnetic resonance image in a patient suffering return of myelopathy 36 months after "cure by multilevel corpectomy. Note the inadequate longitudinal decompression and characteristic bulging of suharachnoid space with decompression.
terior spinal column instability. Such a situation might exist in patients with a previous laminectomy. The patient is mobilized the following day, and the cervical drain is removed. Daily radiographs are obtained until the graft is believed to be stable, and the patient is discharged as soon as the discomfort at the hip or leg graft donor site allows. Outpatient postoperative follow-up evaluations are scheduled for 2 weeks, 6 weeks, 3 months, 6 months, and then yearly, with plain x-ray films taken at each visit. If an iliac crest graft is used, the orthotic device is usually discontinued at approximately 2 months. After the orthotic system is discontinued, isometric and stretching exercises are initiated to promote neck flexibility. Radiographs are taken approximately 2 weeks after orthotic discontinuance. If residual symptoms or myelopathy are present, an MR image is obtained during the first months after surgery. If the patient has had resolution of myelopathic symptoms, it is advised that MR imaging be performed sometime during the 1st postoperative year for decompression documentation.
Outcome Improvement was usually apparent during the 1st postoperative week or weeks; however, three patients in our series experienced a slow recovery, with improvement not being apparent for 6 months. Gait disturbance, bladder dysfunction, and severe hyperpathia were the earliest complaints most likely to be relieved. The outcome was defined as a cure if the patient was satisfied, had no complaints, and there were no signs other than active reflexes or residual minor hand intrinsic atrophy. Improvement was defined as a change of at least two grades in the Nurick classification (Table 2), and/or resolution of hyperpathia and dysesthesia. Less change in grade or regression of improvement was considered a failure. 166
Perioperative Complications C-5 radiculitis graft mortise fracture graft displacement requiring surgery protracted swallowing dysfunction hoarseness soft-tissue hematorna requiring surgery stridor, reintubated hyperventilation escalation of hiatal hernia symptoms pneumonia incarcerated hernia at iliac crest osteomvelitis at donor site hyponatremia perforated sigmoid diverticulum
No. of Cases
5 1 1 1 1 2 1 1 1 1 1 1
By these criteria, 25 patients (62.5%) were considered cured by corpectomy; however, two (8%) of these patients had regression of improvement, resulting in a 57.5% cure rate at the time of this report, The two patients who regressed, at 18 months and 36 months, had exhibited posterior column signs preoperatively, which ultimately recurred. Repeat MR imaging showed inadequate decompression (Fig. 2). If the six patients with minor central myelopathies (five of whom were cured) are eliminated as unrepresentative cases because the symptoms were largely subjective, there are 18 continuing cures in 34 long-term patients, for a 53% cure rate. Equal numbers of the 40 patients had symptoms of long and short duration, with 45% and 70% cure rates, respectively. Of 24 patients preoperatively considered severely impaired, 50% were cured. Four of nine severely impaired patients with symptoms of long duration were cured. Of the 14 patients aged 70 years or older, six were cured. Of the 28 patients with transverse syndromes, 13 had persistent cures, six of whom had been severely impaired. All four patients with central
syndromes were cured, three of whom had had symptoms of short duration. The failure group comprised 15% of the series. This included two cases of cure regression, one case of error in diagnosis (multiple sclerosis), and three patients with severe syndromes of long duration with only minimal improvement. Perioperative Complications
The perioperative complication rate was 47.5%, the specifics of which are enumerated in Table 4. The most common problem was a C-5 radiculitis. One bone graft displacement occurred immediately after surgery and was revised. One patient with mortice fracture was treated with bedrest and cervical traction. Morbidity
There was one error in diagnosis, which ultimately Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy TABLE 5 Long-term sequelae of corpectomy Complication
No. of Cases
permanent (7.5%) C-5 radiculopathy swallowing dysfunction hypoglossal nerve palsy resolved (5%) recurrent laryngeal nerve palsy pseudarthrosis (? fracture)
proved to be multiple sclerosis, as suspected preoperatively. There were five patients with long-term sequelae of corpectomy (Table 5); three of these persist at the time of this report but do not include myelopathic worsening. Mortality
One patient, who was quadriplegic preoperatively, had subtle improvement postoperatively, but remained unable to breathe without assistance; by her own choice she was sedated and allowed to die. One patient with a severe myelopathy improved insignificantly after corpectomy, but died of unrelated causes 10 months postsurgery. Discussion The poor surgical results in patients with cervical spondylotic myelopathy are often ascribed to degenerative or ischemic processes of the spinal cord compounding spondylotic compression. 3 .''' These poor results are striking by comparison with those of other forms of benign spinal cord compression, such as extraaxial tumor and ossification of the posterior longitudinal ligament, even though the myelopathies are fundamentally similar. 20,22,33,35,36,43 The basis for the role of ischemia in this disorder lies not in pathological findings of vascular disease in cervical spondylotic myelopathy,'""° but the not-surprising experimental observations that ischemia will compound a compressive mydopathy. 17,18,26,28,47 An equally compelling explanation for surgical failure in cervical spondylotic myelopathy lies in the possibility that conventional surgical treatment for cervical spondylotic myelopathy is different, or less definitive, than that for extra-axial tumors and ossification of the posterior longitudinal ligament.
Pathogenesis of Myelopathy Spondylotic myelopathy is thought to stem from at least three treatable factors: spinal stenosis;"'"'"'" anterior osteophytesP" and relatively excessive spinal mobility.'" 2 Adams and Logue' first reported that those patients with best results after laminectomy or conservative methods had relatively stiff spines in comparison with those patients with poor results. Experimental and clinical evidence suggests that a myelopathy can be ameliorated by spinal immobilization. 5 . 11 '"'" J. Neurosurg. / Volume 74 / February, 1991
Surgical Approaches The commonly performed surgical procedures for cervical spondylotic myelopathy, laminectomy, and discectomy/interbody fusion seemingly ignore the multifactorial causes of cervical spondylotic myelopathy. 10,23.44 The cure rate with laminectomy is no more than 20% (CA Fager, unpublished data, 1989), and the overall improvement rate is under 70%. 8 . 19 Conventional anterior procedures have a cure rate of 33%, and an overall incidence of beneficial outcome of less than 75%." Corpectomy with strut grafting, which treats simultaneously several factors causative in cervical spondylotic myelopathy, has a success rate of 73% to 100% (Table 1). Surgical Outcome The 2- to 5-year follow-up period in this small group has not yet documented progression of osteophyte formation above and below the immobilized cervical segment. With longer follow-up monitoring, the progression of spondylosis may demonstrate inevitable recurrent spinal cord compression, although this arthritic progression may be an overestimated concern. 4' 12 The presumption that certain conditions must have a specific treatment (such as laminectomy for congenital stenosisi has not been borne out in this series (Fig. 3). The presumed need for posterior decompression for posterior canal encroachment' also has not been corroborated by our experience (Fig. 4). Since there were no selection criteria in this consecutive series, corpectomy was performed in some frail elderly patients; more than 25 % of the patients were older than 70 years. In our experience, cardiovascular risk factors and osteoporosis, and not age in itself, are critical factors with corpectomy. Fifty percent of the patients over 70 years of age were cured, as were nearly 50% of the patients with severe myelopathies. Long duration of severe symptoms prior to treatment emerged as the single complex most predictive of outcome; nevertheless, three such patients had long-term cures. Length of Decompression Our series suggests that patients with resection of the most segments have the best outcomes, favoring the choice of longer longitudinal decompression using fibular grafts if necessary. We prefer autogenous iliac crest as strut graft material rather than fibula, and accordingly have restricted the extent of longitudinal decompression to three segments. The use of iliac bone for a longer graft is simply not feasible, at least in our hands. Iliac crest graft, involving heavily cancellous bone of substantial diameters, has been preferred for its early incorporation. This graft, however, should be cautiously selected in patients requiring a strut of greater than two levels, especially in cases of osteoporosis, as two such grafts have fractured in our series. Yonenobu, et al.," have cautioned that patients requiring a vertebrectomy involving more than four 167
R. L. Saunders, et al.
FIG. 3. Left: Preoperative magnetic resonance image showing congenital canal stenosis with superimposed spondylosis. Right: Postoperative image from the same patient after a four-level decompression and fibular graft placement. The canal is now capacious.
segments be managed by laminectomy. The highly stressed, if not mobile, motion segment between the cervical spine and the relatively immobile thoracic spine, if crossed with a bone strut would result in a lever arm of such length that it would readily fracture a mortice. This would indicate that the most radical cervical corpectomy feasible is the removal of a lower portion of C-2 through C-6, with grafting from C-2 to C-7. (Resection of 3 mm of the upper body of C-7 might safely address a C6-7 ridge.) Some range of motion will be lost with such a multisegmental strut, but it is practically insignificant, since the degenerative joint disease of the older patient's neck limits flexion, extension, and rotation to the occiput, C-1, and C-2 levels. Short cervical grafts at the cervicothoracic junc-
FIG. 4. Left: Preoperative myelogram showing marked spondylotic ridging at C3-4 and C4-5, in particular. There is apparent posterior cord encroachment. Right: Myelogram in the same patient after C3-5 corpectomy. The canal is now capacious anteriorly as well as posteriorly.
1 68
tion are probably not as dangerous; we have treated one patient who had an uncomplicated strut graft from C6 to T-2, and Yoshizu, et at," described a graft from C-6 to T-10 without apparent consequence. Choice of Bone Graft We use autogenous bone and avoid implanted hardware to minimize the possibility of long-term complications. We recognize that the use of plates and screws, at least in the short term, would lessen the risk of graft displacement' and the use of allograft bone would eliminate donor site complications. Generally, our perioperative complications have decreased with practice. However, in our ongoing series we continue to have about a 10% incidence of graft fractures and/or displacements, not all necessarily requiring surgical revision. The preference for using fibula to graft more than two segments' should be questioned in light of its limitations in certain patients. The fibula, being mostly cortical, is incorporated very slowly. The shortest period reported for the uniting of fibula with vertebra is 6 months; 4 ' the time for complete incorporation of the fibula graft is more likely closer to 1 year. In the patient with osteoporosis, the fibula probably should not be used, unless protected with a halo vest. The dense character of fibula used in the osteoporotic spine creates the phenomenon of "pistoning" (Fig. 5), which risks fracturing off the front of the vertebral body as the fibula graft settles through its mortices. Although we have recognized but one pseudarthrosis, which occurred after substantial neck stress and could be considered a graft fracture, this concern might warrant a more liberal use of fibula. There is no reported pseudarthrosis with fibular grafting, whereas the incidence with iliac crest may be as high as 20% (T Whitecloud, unpublished data, 1978). Of our 40 patients with long-term follow-up monitoring, 10% have .1. Neurosurg. / Volume 74 / February, 1991
Central corpectomy for cervical myelopathy
FIG. 5. Studies of a recently performed procedure not included in this long-term series. Left: X-ray film of the fibular graft obtained immediately postoperatively in a osteoporotic individual. Right: X-ray film in the same patient 3 days later showing that the dense graft material has penetrated the end plate. This is an example of "pistoning. -
neck pain, raising the question of possibly undiagnosed pseudarthrosis. Perioperative Complications The etiology of the C-5 radiculitis seen in our series (Table 4) is as yet uncertain, and has not been described by others. N Epstein (personal communication, 1989) has implicated the unusual length of the C-5 root, noting that C-5 complications occur in posterior decompression as well as conventional anterior surgery. The C-5 root may be uniquely sensitive to inadequate foraminal decompression, inherent in the central corpectomy procedure. Alternatively, this may be an unusual expression of a segmental myelopathy or a surgical myelopathic complication (JN Abramowitz, unpublished data, 1989). That this is not a myelopathic effect is supported by the delayed onset, often on the 2nd day after surgery, and the complete resolution of the syndrome in most patients within several months. Most recently, by avoiding a paralytic anesthesia technique at Sypert's suggestion (G Sypert, personal cornmunication, 1989), we have noted the C-5 segment to be most sensitive not to instrumental compression but to the heat of either unipolar or bipolar coagulation. As low a bipolar setting as 10 W, when applied to the posterior longitudinal ligament, may evoke a major deltoid twitch. It could be that we have inadvertently caused heat injuries to a uniquely sensitive root. Our long-term cure rate of greater than 50% is unusual in the treatment of cervical spondylotic myelopathy 19 and places the substantial perioperative complications in perspective. Other corpectomy series do not describe any mortality, and considerably less morbidity; however, graft problems are reported in 3% to 21% of cases ,46,2 1,23 a favorable comparison with our 5%. Despite our high perioperative complication rate, as in other corpectomy series we have noted no myelopathic morbidity. This underscores one of the most important aspects of corpectomy when considered against the significant incidence of postoperative worsJ. Neurosurg. / Volume 74 / February, 1991
ening with cervical laminectomy and discectomy/interbody fusion.' Regression of improvement after corpectomy deserves more attention, and warrants long-term followup evaluation of series prior to reporting. Of the corpectomy series (Table 1), only that of Bernard and Whitecloud 4 includes two patients who deteriorated 1 year after initial improvement, a regression rate of 11%, with which our rate of 5% compares favorably. Regression after laminectomy is reported as between 19% and 53% and is in large part due to instability and/or kyphosis. 4 Galera and Tovi,' discussing long-term follow-up results after the Cloward procedure for cervical spondylotic myelopathy, described most of their patients as ultimately deteriorating; the report of Lunsford, et al.,' was nearly as pessimistic. In our two patients with long-term deterioration, follow-up MR images showed that their longitudinal decompressions were inadequate (Fig. 3). Furthermore, as noted earlier, longer decompressions are related positively to outcome. In conclusion, the logic of a multifactorial surgical approach to cervical spondylotic myelopathy is supported by the results of this unselected series. A cure rate in excess of 50%, a regression rate of 5%, and nonexistent myelopathic morbidity associated with central corpectomy are suggestive evidence that the generally accepted surgical results in cervical spondylotic myelopathy should be reconsidered, especially if our observations are confirmed by a randomized trial. References 1. Adams CBT, Logue V: Studies in cervical spondylotic myelopathy. II. The movement and contour of the spine in relation to the neural complications of cervical spandylosis. Brain 94:569-586, 1971 2. Barnes MP, Saunders M: The effect of cervical mobility on the natural history of cervical spondylotic myelopathy. J Neurol Neurosurg Psychiatry 47:17-20, 1984 3. Bedford PD, Bosanquet FD, Russell WR: Degeneration of the spinal cord associated with cervical spondylosis. Lancet 2:55-59, 1952 4. Bernard TN, Whitecloud TS: Cervical spondylotic myelopathy and myeloradiculopathy: anterior decompression and stabilization with autogenous fibula strut graft. Clin Orthop 221:149-157, 1987 5. Bohlman HH: Cervical spondylosis with moderate to severe myelopathy. A report of seventeen cases treated by Robinson anterior cervical discectomy and fusion. Spine 2:151-162, 1977 6. Boni M, Cherubino P, Denaro V, et al: Multiple subtotal somatectomy. Spine 9:358-362,1984 7. Brown JA, Havel P, Ebraheim N, et al: Cervical stabilization by plate and bone fusion. Spine 13:236-240, 1988 8. Carol MP, Ducker TB: Cervical spondylotic myelopathies: surgical treatment. J Spinal Dis 1:59-65, 1988 9. Crandall PH, Batzdorf U: Cervical spondylotic myelopathy. J Neurosurg 25:57-66,1966 10. Cusick JF, Myklebust JB: Biomechanics of cervical spondylotic myelopathy. Contemp Neurosurg 9(5):1-8,1987 11. Cusick JF, Steiner RE, Berns T: Total stabilization of the
1 69
R. L. Saunders, et al. cervical spine in patients with cervical spondylotic myelopathy, Neurosurgery 18:491-495, 1986 12. DePalma AF, Rothman RH, Lewinnek GE, et al: Anterior interbody fusion for severe cervical disc degeneration. Surg Gynecol Obstet 134:755-758, 1972 13. Ducker TB: Experimental injury of the spinal cord, in Vinken PJ, Bruyn GW (eds): Handbook of Clinical Neurology. Injuries of the Spine and Spinal Cord and Column.
Amsterdam: North-Holland, 1976, Vol 25, pp 9-26 14. Epstein JA, Carras R, Lavine LS, et al: The importance of removing osteophytes as part of the surgical treatment of myeloradiculopathy in cervical spondylosis. J Neurosurg 30:219-226, 1969 15. Galera R, Tovi D: Anterior disc excision with interbody fusion in cervical spondylotic myelopathy and rhizopathy. J Neurosurg 28:305-310,1968 16. Gillilan LA: Veins of the spinal cord. Anatomic details: suggested clinical applications. Neurology 20:860-868, 1970 17. Gooding MR, Wilson CB, Hoff JT: Experimental cervical myelopathy: autoradiographic studies of spinal cord blood flow patterns. Surg Neurol 5:233-239,1976 18. Gooding MR, Wilson CB, Hoff JT: Experimental cervical myelopathy. Effects of ischemia and compression of canine cervical spinal cord. J Neurosurg 43:9-17, 1975 19. Gorier K: Influence of laminectorny on the course of cervical myelopathy. Acta Neurochir 33:265-281, 1976 20. Hanai K, Adachi H, Ogasawara H: Axial transverse tomography of the cervical spine narrowed by ossification of the posterior longitudinal ligament. J Bone Joint Surg (Br) 59:481-484, 1977 21. Hanai K, Fuminori F, Kunitaka K: Subtotal vertebrectomy and spinal fusion for cervical spondylotic myelopathy. Spine 11:310-315, 1986 22. Harsh GR IV, Sypert OW, Weinstein PR, et al: Cervical spine stenosis secondary to ossification of the posterior longitudinal ligament. J Neurosurg 67:349-357, 1987 23. Herkowitz HN: The surgical management of cervical spondylotic radiculopathy and myelopathy. Clin Orthop 239:94-108, 1989 24. Hinck VC, Sachdev NS: Developmental stenosis of the cervical spinal canal. Brain 89:27-36, 1966 25. Hughes JT, Brownell B: Spinal cord ischemia due to arteriosclerosis. Arch Neurol 15:189-202, 1966 26. Hukuda S, Mochizuki T, Ogata M, et al: A comparison of the results of anterior and posterior procedures. J Bone Joint Surg (Br) 67:609-615, 1985 27. Hukuda S, Ogata M, Katsuura A: Experimental study on acute aggravating factors of cervical spondylotic myelography. Spine 13:15-20, 1988 28. Hukuda 5, Wilson CB: Experimental cervical myelopathy: effects of compression and ischemia on the canine cervical cord. J Neurosurg 37:631-652,1972 29. Idwards WC, LaRocca H: The developmental segmental sagittal diameter of the cervical spinal canal in patients with cervical spondylosis. Spine 8:20-27, 1983 30. Jellinger K: Spinal cord arteriosclerosis and progressive vascular myelopathy, J Neurol Neurosurg Psychiatry 30: 195-206,1967 31. Kojima T, Waga S, Kubo Y, et al: Anterior cervical vertebrectomy and interbody fusion for multi-level spondylosis and ossification of the posterior longitudinal ligament. Neurosurgery 24:864-872, 1989 32. Lunsford LD, Bissonette DJ, Zorub DS: Anterior surgery for cervical disc disease. Part 2: Treatment of cervical spondylotic myelopathy in 32 cases. J Neurosurg 53: 12-19,1980 33. Manabe S, Nomura S: [Anterior decompression for ossi170
fication of the posterior longitudinal ligament of the cervical spine.] Neurol Surg 5:1253-1259,1977 (Jpn) 34, Miller CA: Shallow cervical canal: recognition, clinical symptoms, and treatment. Contemp Neurosurg 7(7):1-6, 1985 35. Murakami N, Muroga T, Sobue I: Cervical myelopathy due to ossification of the posterior longitudinal ligament. A clinicopathologic study. Arch Neurol 35:33-36,1978 36. Nagashima C: Cervical myelopathy due to ossification of the posterior longitudinal ligament. J Neurosurg 37: 653-660,1972 37. Nurick S: The natural history and the results of surgical treatment of the spinal cord disorder associated with cervical spondylosis. Brain 95:101-108, 1972 38. Nurick S: The pathogenesis of the spinal cord disorder associated with cervical spondylosis. Brain 95:87-100, 1972 39. Payne EE, Spillane JD: The cervical spine. An anatomicopathological study of 70 specimens (using a special technique) with particular reference to the problem of cervical spondylosis. Brain 80:571-596, 1957 40. Roberts AH: Myelopathy due to cervical spondylosis treated by collar immobilization. Neurology 16:951-954, 1966 41. Rossier AB, Hussey RW, Kenzora JE: Anterior fibula interbody fusion in the treatment of cervical spinal cord injuries. Surg Neurol 7:55-60,1977 42. Senegas J, Guerin J, Vital JM, et al: Decompression medullaire etendue par voie anterieure dans le traitement des myelopathies par cervicarthrose. Rev Chir Orthop 71: 291-300,1985 43. Stookey B: Compression of the spinal cord due to ventral extradural chondromas. Arch Neurol Psychiatry 20: 275-280,1928 44. Whitecloud TS: Anterior surgery for cervical spondylotic myelopathy. Smith-Robinson, Cloward, and vertebrectomy. Spine 13:861-863, 1988 45, Whitecloud TS, LaRocca H: Fibular strut graft in reconstructive surgery of the cervical spine. Spine 1:33-43, 1976 46. Wilkinson M: The morbid anatomy of cervical spondylosis and myelopathy. Brain 83:589-617,1960 47. Wilson CB, Landry RNA: Experimental cervical myelopathy: 1. Blood supply of the canine cervical spine cord. Neurology 14:809-8 14, 1964 48. Yonenobu K, Fuji T, Ono K, et al: Choice of surgical treatment for multisegmental cervical spondylotic myelopathy. Spine 10:710-716, 1985 49. Yoshizu T, Hara T, Honma T, et al: Experience with anterior spinal fusion by vascularized fibula as a strut graft in kyphosis and kyphoscoliosis. J Microsurg 3: 123-124,1981 (Abstract) 50. Yu YL, Jones SJ: Somatosensory evoked potentials in cervical spondylosis. Correlation of median, ulnar and posterior tibial nerve responses with clinical and radiological findings. Brain 108:273-300, 1985 51. Zdeblick TA, Bohiman HH: Cervical kyphosis and myelopathy. Treatment by anterior corpectomy and strutgrafting. J Bone Joint Surg (Am) 71:170-182,1989 Manuscript received January 11,1990. Accepted in final form June 29,1990. This paper was presented in part at the Annual Meeting of the American Association of Neurological Surgeons, Dallas, Texas, on May 4,1987. Address reprint requests to: Richard L. Saunders, M.D., Section of Neurosurgery, Dartmouth-Hitchcock Medical Center, 2 Maynard Street, Hanover, New Hampshire 03756.
J. Neurosurg. Volume 74 / February, 1991
