Laminoforaminotomy

In 1946, Scoville presented the so-called keyhole foraminotomy technique at the annual Harvey Cushing Meeting. This was followed in 1951 by his report on 115 patients treated using this technique.¹ Despite its being developed several years before the now more common anterior cervical discectomy and fusion operations (Cloward^2,3 and Smith-Robinson types⁴), there are clear indications for and advantages to this approach.

The ideal patient for a posterior laminoforaminotomy is a patient with a single-level radicular syndrome, corresponding neurologic findings, and a concordant preoperative imaging study demonstrating foraminal stenosis. The nature of the foraminal stenosis can be adequately determined using a combination of plain radiographs, CT, and MRI. It can be due to a single or a combination of factors. The nerve root may be impinged by a soft disc herniation, typically seen as an extruded fragment that is hyperintense on T2-weighted imaging. This is usually the case in younger individuals. In older patients, the foramen is typically compromised by a disc-osteophyte complex that is a combination of disc material and uncovertebral osteophytes (hard disc). Foraminal stenosis, due to loss of disc height at the same level and osteophytes arising from the facet joint, may also coexist.

The keyhole foraminotomy typically provides a 3- to 5-mm exposure of the sensory root, which can then be retracted rostrally, thus facilitating the removal of the soft disc through the root’s axilla.⁵ It is difficult and risky to attempt removal of anterior uncovertebral osteophytes through this approach. In addition, this approach does not directly address foraminal stenosis due to a collapse of the disc space. However, as discussed later, numerous studies have shown excellent results in patients with hard discs. Webb et al. have described removing the superomedial aspect of the inferior pedicle using a drill to increase the foraminal volume in selected cases.⁶

The laminoforaminotomy has several advantages over the anterior discectomy and fusion. It preserves motion of the segment and therefore avoids the potential pitfalls associated with arthrodesis, namely, hardware complications, pseudarthrosis, graft subsidence, graft harvest complications, and the theoretical increased chance of accelerated adjacent-level degeneration. There are no issues with postoperative dysphagia or hoarseness and the decompression may be easily performed at multiple levels and bilaterally.

Because of the procedure’s motion-sparing advantage, spondylotic changes may continue and further surgery may be necessary for recurrent or new symptoms at the same level. Clarke et al. calculated the 5- and 10-year risk rates for development of same-segment disease as 3.2% and 5.0%, respectively.⁷ As previously mentioned, many patients have nerve root compromise as a result of bony growth into the foramen from the uncovertebral joint. The laminoforaminotomy allows an indirect decompression of the root but does not allow a direct decompression with removal of the anteriorly situated osteophytes, as would an anterior approach. Midline or collapsed disc pathology, kyphotic deformity, and disc herniation associated with unstable or traumatic cervical spine injury and the presence of significant axial neck pain are better treated with an anterior approach. The incision is typically more painful than a ventral approach for several days to weeks, and uses a midline approach and subperiosteal dissection of the paraspinal cervical musculature and lateral retraction. This may be improved with a paramedian muscle-splitting tubular approach.^8,9 No more than 50% of the facet should be resected to prevent instability and kyphosis.^10–18 Preoperative kyphosis at the symptomatic level, therefore, is a contraindication to this procedure.

A review of the recent literature is provided in Table 157-1. Typically, at least 80% to 90% of patients have good to excellent results. In the largest study, Henderson et al. reported resolution of radicular symptoms in 96% of 846 procedures performed in 736 patients, with 91.5% reporting good or excellent results.¹⁹ Jagannathan et al. recently published their experience in 162 patients who underwent a unilateral, single-level posterior cervical foraminotomy with a minimum 5-year clinical and radiographic follow-up.²⁰ Resolution of radiculopathy and improvement in the Neck Disability Index occurred in 95% and 93% of patients, respectively. There were no statistically significant changes in focal or overall cervical kyphosis with time. Thirty patients (20%) experienced postoperative loss of lordosis (defined as a Cobb angle <10 degrees measured from C2 to C7). Factors found to be associated with worsening sagittal alignment included age greater than 60 years, the presence of a preoperative cervical lordosis of less than 10 degrees, and the need for posterior surgery after the initial foraminotomy.

TABLE 157-1 Summary of Recent Literature on Posterior Cervical Laminoforaminotomy

Ventrolateral Foraminotomy

In 1996, Jho described a modified ventrolateral approach to the cervical foramen (ventral foraminotomy) to treat radiculopathy.²¹ It was based on the previously described and more extensive lateral approach of Verbiest²² and the transuncodiscal approach of Hakuba.²³ The key step in the procedure described by Jho is drilling and removal of the lateral uncovertebral joint (measuring approximately 5 to 8 mm transversely by 7 to 10 mm vertically) to expose the exiting nerve root as it passes under the vertebral artery. The surgical technique emphasizes preservation of the intervertebral disc as a functioning motion segment, while allowing direct access to the offending lesion (posterolateral disc herniation or uncovertebral osteophyte). Others, however, appear to have been using a similar technique before Jho’s publication.²⁴ As in the posterior laminoforaminotomy, the ideal patient is one with one or more ipsilateral radiculopathies due to disc herniation or uncovertebral osteophytes. It cannot address foraminal stenosis due to loss of disc space height and would achieve only an indirect decompression of the root if the stenosis was due to facet arthropathy.

The ventrolateral approach has not found widespread acceptance among spine surgeons, with most publications coming from Europe and few centers in the United States. The reasons for this include fear of unnecessarily injuring the vertebral artery and sympathetic chain (Horner syndrome), inability to address bilateral symptoms, and the widespread use and teaching of the more common anterior (fusion and now arthroplasty) and posterior (laminoforaminotomy) approaches. Furthermore, a cadaveric study demonstrated a significant alteration in the mobility of the segment after unilateral uncoforaminotomy, with an increase in the range of motion in lateral bending and axial rotation.²⁵ The clinical implication is that with time this may lead to further disc degeneration and its sequelae (disc collapse, reherniation, development of osteophytes), requiring a reoperation. For example, a recent study by Yi et al. comparing 15 patients who underwent a total disc replacement with 13 patients who had an anterior foraminotomy found good clinical results in both groups, but the anterior cervical foraminotomy caused a significant loss of disc height after surgery.²⁶ Hacker and Miller found an unacceptably high revision rate of 30% (7 patients) in their series of 23 patients, with good or better outcomes using Odom’s criteria in only 52% (12 patients).²⁷ The revisions were performed between 2 weeks and 14 months from the index procedure and there were two patterns of failure that led to reoperation: recurrent radiculopathy or intractable neck pain. Six patients underwent anterior cervical discectomy and fusion, and one patient required a corpectomy. Nonetheless, multiple publications have indicated that anterior foraminotomy improves radicular symptoms and neck pain, with good or better outcomes using Odom’s criteria in more than 80% of patients.²⁸ Some of this literature is summarized in Table 157-2. One of the largest series, by Jho et al., was published in 2002.²⁹ Of the more than 400 surgeries performed by the author and his team, 104 met the inclusion criteria for their study. Six weeks after surgery, 83 patients (79.8%) demonstrated excellent results, 20 patients (19.2%) demonstrated good results, and 1 patient (1%) experienced a fair outcome. No patient experienced a poor outcome or unchanged status. One patient required a reoperation to remove a persistent disc fragment, one patient developed discitis and had a spontaneous fusion with a slightly kyphotic angulation, one patient developed hemiparesis that resolved after 6 weeks (which the author believed was due to extended neck posture during surgery), and two patients developed transient Horner syndrome that resolved within 6 weeks. Dynamic cervical radiographs were obtained in 59 patients and all showed preservation of the motion segment.

TABLE 157-2 Summary of Literature on Anterior Foraminotomy

A few surgeons, including Jho, have expanded the indications for the approach by treating myelopathic patients with partial oblique corpectomies while still preserving much of the disc space.^30–35 These results are not reviewed here, but good outcomes have been reported and a sheep cadaveric model has shown that the ventral oblique corpectomy does not result in instability.³⁶

Laminectomy

Historically, cervical laminectomy has been a valuable motion-sparing, nonimplant surgery performed for the management of congenital or acquired multilevel cervical spinal stenosis associated with spondylosis, ossification of the posterior longitudinal ligament, and ossification or hypertrophy of the ligamentum flavum. The key factors in determining the optimal surgical candidate are the extent and nature of compression and the overall sagittal balance of the cervical spine.

Upright and dynamic (flexion-extension) MRI and CT are often recommended as preoperative imaging studies. The normal anteroposterior (AP) diameter of the spinal canal is approximately 17 mm. Stenosis is defined as a sagittal diameter of less than 12 to 13 mm. Congenital stenosis usually affects multiple levels and is readily apparent on a sagittal T2-weighted MRI. The Torg or Pavlov ratio, which is the AP diameter of the spinal canal divided by the AP diameter of the vertebral body at the same level, can be used as a guide in diagnosing congenital stenosis.³⁷ Stenosis is usually present if the ratio is less than 0.8. Another quick method to determine the presence of congenital stenosis is by assessment of plain lateral radiographs. If minimal space is present between the spinolaminar line and a line connecting the dorsal margin of the facet joints, significant canal stenosis is present.

Static and dynamic radiographs may provide information regarding instability. Concerning features, according to White and Panjabi,³⁸ include sagittal plane listhesis of greater than 3.5 mm on resting or dynamic radiographs, a sagittal plane rotation of greater than 20 degrees on flexion-extension radiographs, and a relative sagittal plane angulation of greater than 11 degrees on resting radiographs. It is critical to assess the overall sagittal balance of the cervical spine. Optimally, there should be approximately 14 degrees of lordosis from C2 to C7.

The optimal patient is one with single-level or multilevel circumferential central stenosis (i.e., congenital stenosis), or primarily dorsal impingement with preserved cervical lordosis. Preservation of lordosis is critical because maximal dorsal migration of the spinal cord takes place if the patient has an adequate cervical lordosis, typically 10 or more degrees, and 7 mm or less of ventral compression.³⁹ Those patients with a straight cervical spine or, of more concern, those with a kyphotic curve, may not be adequately decompressed through the dorsal approach and are at higher risk for development of a kyphotic deformity postoperatively. Decompression should be carried past the areas of focal stenosis to ensure that the transition areas from nonstenotic to stenotic regions do not themselves become compressed as the cord migrates dorsally. For example, if there is stenosis from C3 to C6, then the lamina of C2 should be undercut and the top of the C7 lamina should be resected.

Many reports since the 1970s have demonstrated the efficacy of laminectomy for the treatment of myelopathy. Ryken et al. recently provided an exhaustive review of this literature.⁴⁰ The surgery has numerous advantages; it can be performed relatively quickly and can address multiple levels, and it avoids the risks associated with anterior approaches, namely, hoarseness and dysphagia, which are significant risks in patients with advanced age and multilevel disease. There are disadvantages, however. The already compromised spinal cord may be damaged if instruments are placed under the midline lamina too aggressively. This has led some to recommend starting the decompression bilaterally at the laminofacet junction and then removing the central dorsal elements en bloc.⁴¹ Postoperative C5 palsy is a complication associated with multilevel dorsal decompressive surgery (laminectomy or laminoplasty), occurring with a frequency of about 5% to 10%.⁴² Its etiology continues to be a mystery and intraoperative monitoring has not been shown to prevent it.^43,44 Many believe that it may be due to the dorsal shift of the spinal cord, which is often maximal at the C5 level because it is usually at the midpoint in the laminectomy, causing tethering of the already short, less oblique C5 roots.^45,46 Some have advocated prophylactically performing foraminotomies at C4-5.^47,48 Recently, Sieh et al. have found that an average Pavlov ratio of less than 0.65 and compression at the C3-4 segment on preoperative MRI were reliable preoperative predictors for the development of this problem.⁴⁹

The development of postlaminectomy spinal instability manifested by kyphosis is a concern that has garnered much attention and has led to the development of laminoplasty as an alternative technique (see next section). In their review, Ryken et al. found the risk ranged from 14% to 47%.⁴⁰ However, as stated by the authors, no study has clearly demonstrated a relationship between postlaminectomy kyphosis and deterioration in the patient’s quality of life.

The trigger for the development of postoperative kyphosis is the denervation and weakening of the paraspinal musculature as a result of tissue dissection and bone removal. This alone may be enough to lead to kyphosis, such as in patients who already have a kyphotic cervical sagittal balance. On the other hand, in other patients, such as those with less cervical lordosis or straight spines, it may require the cumulative effect of other risk factors in addition to weakened muscles and lack of a bony scaffold to give rise to the kyphotic deformity. Not much can be done during surgery to limit the damage to the paraspinal musculature (other than possibly reducing the amount of electrocautery used and time under retraction), so it is the avoidance or limitation of other, more controllable factors that is of utmost importance in preventing this complication. One cannot overstress the importance of careful study of preoperative imaging, focusing on the presence of underlying instability and overall sagittal balance. The width and length of the laminectomy are important. If done for the treatment or prevention of myelopathy, the laminectomy should be taken to the lateral extent of the dural sac, which corresponds to the laminofacet junction. If further resection of the facets is required for decompression of the sac or individual nerve roots, this should be limited to 25% to 50% because numerous clinical, cadaveric, and finite element model studies have demonstrated that progressive facet resection produces increases in angular rotation and intervertebral disc stresses, leading to segmental hypermobility.^10–18 Laminectomies should be avoided at C2 and C7 because these laminae have been found to be relatively high load-bearing structures compared with C3 to C6, and the spinous process of C2 needs to be preserved because it is the insertion site of several key erector spinae muscles that are an integral part of the dorsal tension band.

The risk of postoperative kyphosis is greater in children for a number of anatomic reasons.⁵⁰ The child’s cervical spine is hypermobile because the muscles and supporting ligamentous structures are less well developed, the orientation of the facets is more horizontal, the vertebral bodies are wedge shaped, and the head is disproportionately larger compared with the spine.⁵¹