Pathology

Ossification of the posterior longitudinal ligament (OPLL) begins with hypervascular fibrosis and hypertrophy of the posterior longitudinal ligament, followed by cartilaginous proliferation, lamellar bone formation, and Haversian canal production. OPLL mineralization progressively narrows the anteroposterior (AP) diameter of the cervical canal by 0.4 mm a year, while lengthening by 0.67 mm longitudinally. Once the average canal diameter is reduced by more than 30 to 50 percent (less than an average of 9.4 mm), myelopathy typically appears. The fact that there is a greater incidence of neuronal damage in the posterolateral gray matter than in the white matter of the cervical cord confirms the idea that the predominant mechanism of spinal injury is vascular rather than compressive.

Incidence of OPLL

OPLL, which can be seen on plain cervical x-ray films accounts for 20 to 25 percent of cervical myelopathy. In terms of spinal location, 70 percent of OPLL is found in the cervical, 15 percent in the thoracic (T4-T6), and 15 percent in the lumbar (L1-L3) regions of the spinal canal. Unlike typical cervical spondylosis and disc disease, which originate at the C5-C6 and C6-C7 levels, OPLL usually originates proximally (C3-C4 to C4-C5).

Magnetic Resonance Imaging in OPLL

0PLL is rarely visualized on plain x-ray films and must attain a minimum of 5.4 mm in AP diameter before it can be visualized on magnetic resonance imaging (MRI) scans. On MRI, it appears as a mass of hypointensity separating the vertebral marrow anteriorly from the isointense spinal cord (on T1-weighted images) and hyperintense thecal sac (on T2-weighted images) posteriorly. Pathognomonic for OPLL is the demonstration of new bone marrow production, which can be found in 11 percent of individuals with segmental OPLL and in 56 percent of those with continuous OPLL. Sagittal T2-weighted images also provide longitudinal overviews of OPLL, particularly at the cervicothoracic junction.

Computed Tomographic Definition of OPLL Types

Direct demonstration of the calcification in OPLL by means of noncontrast computed tomography (CT), CT myelography, and 3-D CT scans is the best way to define the full extent and type of OPLL. Hirabayashi et al. used this modality as the basis for a classification of OPLL into four types: segmental (39 percent of cases), continuous (27 percent), mixed (29 percent), and other (5 percent). Segmental OPLL is located behind the vertebral bodies and not at the disc spaces, whereas continuous OPLL extends from body to body. Mixed OPLL has both segmental and continuous components, and OPLL classified as "other" is confined to the disc spaces alone. OPLL in evolution (OEV), an early variant of OPLL, is defined as a posterior longitudinal ligament (PLL) that shows focal hypertrophy and may also have punctate calcification at the interspaces. As the calcification typically covers several levels. OEV may be easily confused with multilevel disc disease.

Ranawat Classification of OPLL Patients

The surgical outcomes for OPLL patients treated with anterior cervical corpectomy and fusion (AVF), anterior cervical discectomy and fusion (ADF) laminectomy (LAM) and laminoplasty (LOP) are easily compared using the Ranawat neurological classes and grades. Class I patients exhibit no neurological deficits and are only seen postoperatively: class II patients show radiculopathy or mild myelopathy: class IIIA patients exhibit moderate to severe myelopathy: and class IIIB patients are severely myelopathic or frankly quadriplegic. The postoperative Ranawat grades (0 through 4), which are calculated as the preoperative Ranawat class minus the postoperative Ranawat class, provide a quantitative measure of relative outcome. For example. a patient who deteriorates from preoperative class II to postoperative class IIIA has a grade of - I. whereas one who improves from class IIIA to class I has a grade of + 2. Averaging grades across the different surgical categories facilitates comparison of outcomes between series.

Clinical Presentation

OPLL patients are Typically men in their mid-fifties (average ages 49-61) and usually present with a 1 to 2-year history of severe, progressive myelopathy. Patients with OEV average a decade younger (mean age, 45 years) and have less severe radicular findings.

Management Alternatives

Although many asymptomatic class I OPLL patients can be managed conservatively without surgery, others with radiographic or physiologic evidence of severe cord compromise may merit prophylactic decompression. The latter group includes patients whose T2-weighted MRI scans show hyperintense signal in the cord, whose CT studies show marked cord compression or who show severe delay in somatosensory evoked potentials (SSEPs).

Patients in classes II through IIIB. who exhibit progressive radicular and myelopathic deficits. require increasing degrees of surgical attention. Whereas surgery can be delayed for class II patients whose root or cord signs evolve slowly, others with a more precipitous, acute course of deterioration warrant more immediate attention. Similarly, patients with class IIIA disease and slowly progressive myelopathy can be treated electively, whereas those with more rapid functional loss warrant more urgent intervention. The most severely compromised class IIIB patients. with their severe preoperative myelopathy and near quadriplegia. are the least likely to benefit from either immediate or delayed surgical intervention. Certainly, for any patient with OPLL, the longer-standing and more severe the preoperative deficits, the more limited the probable outcome of surgery.

Surgical Techniques

Anterior Corpectomy and Fusion

Myelography followed by CT is the best way to determine the full extent of OPLL in cases that require multilevel anterior corpectomies with reversed iliac crest strut fusions. We prefer iliac crest auto grafts over iliac allografts or fibular auto grafts or allografts, because they are incorporated the most quickly and with the least morbidity (i.e., fracture). Occasionally, a halo brace, posterior wiring and fusion or an anterior or posterior plating system may be required to stabilize a graft, to promote fusion, or to treat a pseudarthrosis.

An extensive AVF or multilevel ADF is started with an appropriate left-sided transverse skin incision. When the prevertebral compartment is located. hand-held Cloward retractors are quickly replaced by two medial/lateral self-retaining Caspar retractors (smooth or mildly serrated blades only) placed superiorly and inferiorly in the wound. Caspar vertebral body screws are placed in the most cephalad and caudad vertebral bodies to facilitate longitudinal soft tissue retraction. Free-standing Caspar blades or Caspar blades secured to the Grossman retractor are placed behind these screws to enhance soft tissue dissection.

The ventral canal is exposed without distraction by performing partial or complete corpectomies. Disc space spreaders are not applied to the Caspar vertebral body screws prior to OPLL resection, even over single segments, because even minimal distraction can result in profound suppression or loss of SSEPs. Significant depression or loss of SSEPs occurring during the placement of Smith-Robinson or iliac crest strut fusions indicates over-distraction of the segment and mandates removal and revision of the graft.

Under magnified vision,  air drill is used to remove three-quarters of the depth of the vertebral body while creating a transverse corpectomy trough averaging 16 to 20 mm in diameter. This trough has to be wide enough to resect the lateral extent of the OPLL while avoiding the vertebral arteries. Under the operating microscope, removal of the remaining one-quarter of the cancellous portion of the vertebral body down to the posterior cortex. The diamond burr, micro Kerrison punches, micro bayoneted 3-0 and 4-0 curettes, micro nerve hooks, and other microinstruments are then employed to remove the residual posterior cortical margin and the OPLL. The dissection should be accomplished in a deliberate cephalad-to-caudad fashion, leaving the inferoposterior cortical margin intact to protect the more inferior dura and cord as dissection proceeds inferiorly.

The extent to which ossification involves the posterior longitudinal ligament and dura varies both between and within individual patients. Sometimes the dura is left completely intact. In other cases or at other levels in the same patient, neither the dura nor residual intact PLL can be distinguished from the lesion. This situation contributes significantly to the complexity of OPLL surgery.

Routine use of the operating microscope is essential to permit adequate resection of OPLL while limiting morbidity, including neurological injury and CSF fistula formation. Certainly the 16 to 25 percent incidence of CSF fistula reported in conjunction with OPLL surgery may largely be avoided by leaving occasional thin shelves of OPLL adherent to atretic dura. So long as these remaining OPLL fragments have been freed from the bony perimeter of the OPLL, they float on the pulsating dura in the decompression site and do not contribute to ongoing cord or root compromise. These maneuvers may avoid the need for lumbar drains or lumbo­peritoneal shunts.

Anterior Discectomy and Fusion

A modified anterior discectomy and fusion (ADF) is used for OPLL patients whose disease involves the disc spaces and contiguous end plates. This modification includes the additional removal of the cephalad and caudad vertebral margins and the placement of somewhat larger Smith-Robinson fusions. Vertebral body screws placed in the upper half of the most proximal vertebral body and the lower half of the most distal vertebral body provide adequate soft tissue retraction while not interfering with visualization or manipulation under the operating microscope. Distraction is used only after OPLL resection for graft impaction where tolerated. (As mentioned earlier, distraction is never used before OPLL resection.) Anterior plate systems are not routinely employed to supplement ADF.

Occasionally, a combined vertebral strut fusions (AVFs) with the ADF procedure, with single or multiple intervening skipped levels is performed. Also, some individuals with OPLL at adjacent interspaces are best managed with single AVF struts rather than tandem ADF grafts, for example where the cortical end plates of the intervening vertebral body have been extensively dissected. In such a case, a single graft could be placed and the intervening posterior cortical vertebral margin left alone.

Risks of Anterior Surgery 

Complications of OPLL surgery in particular, and of cervical surgery in general, include increased myelopathy and radiculopathy. In one series of 19 patients undergoing OPLL surgery, 2 (10.5 percent) developed postoperative quadriplegia. McAfee et al. reported an even higher 23 percent incidence (3 patients) of cord dysfunction following cervical surgery in 13 spondylotic patients, whereas Saunders et al. encountered a 2.2 percent incidence of cord injury following 90 multilevel corpectomies. Of interest, Saunders and associates also reported a 17 percent frequency of postoperative C5 root injuries.

Routine use of the operating microscope makes it possible to avoid producing CSF fistulae; graft fracture and extrusion, and pseudarthrosis can be minimized by using iliac crest ADF or AVF grafts. Judicious application of hard collars and extension braces, occasional halos braces, and rare posterior wiring and fusion may also be called for. Some surgeons also routinely or intermittently use an anterior or posterior plating system to supplement fusion constructs. However, these devices may create complications of their own, such as plate extrusion or plate migration resulting in esophageal erosion, Horner's syndrome, hoarseness, or vertebral artery compromise.

Laminectomy

Laminectomies may be performed in patients with extensive OPLL involving three or more levels, and in patients unlikely to tolerate a multilevel anterior procedure, such as patients over 70 years. A prophylactic laminectomy, such as performed by Itoh and Tsuji before carrying out a multilevel AVF in patients with severe underlying cervical stenosis, may also be considered.

A LAM is initiated using a high-speed air drill to shave down the lamina in the lateral gutters. Once the ventral bony cortex or ligamentum flavum is exposed laterally, filed-down micro Kerrison punches, and microcurettes are used to complete the laminectomies and foraminotomies. Only very selective removal of discs, spurs, or lateral OPLL is performed, because these measures increase the risk of neural injury.

Although some surgeons, such as Nakano, have concluded that the long-term results of LAM are comparable to those of LOP, others still choose LOP, which offers an average 4.1 mm enlargement of the spinal canal while in theory enhancing spinal stability and avoiding kyphosis.

Risks of Laminectomy

LAM performed to decompress OPLL from behind may result in irreversible intraoperative cord or root injury while failing to relieve ventral cord or root compression. Radicular injuries. particularly those involving the C5 root, may be attributed to the progressive tethering of nerve roots over persisting ventral OPLL masses. A laminectomy that results in swan neck deformity, kyphosis, compressive laminectomy membranes, and more rapid OPLL progression may contribute to cord and root compromise.

Laminoplasty (LOP)

LOPs, such as Hukuda's French window laminoplasty, are better preserving the lordotic curvature and avoiding cervical instability than LAMs, and are therefore less likely than LAMs to exacerbate the postoperative progression of OPLL. Good outcomes were reported for 63 percent of the patients reported by Hirabayashi et al. who underwent expansive open door laminoplasties., and of Kawai's patients who underwent Hattori laminoplasties., 31 percent had an excellent and 53 percent a good postoperative recovery. However, complications unique to LOP included the closing of the "door," and a higher incidence of postoperative C5 root injuries.

Bone Grafts

Anterior cervical grafts fuse in an average of 3.4 months (range, 2.5 to 5 months). Although 80 percent of two-level fusions were complete by 3 months. the rate of pseudarthrosis rose dramatically when three or more levels were involved. A higher fusion rate and lesser inci­dence of graft collapse were observed where an iliac crest autograft was used, rather than an iliac crest allograft or fibular autograft or allograft. Vascularized fibular strut grafts were also an alternative that could be used to promote rapid fusion over multiple levels. Anterior plate instrumentation can also reduce the rate of pseudarthrosis, but it can increase the rate of soft tissue complications.

Outcomes

The Ranawat classes and grades were used to assess the outcome from 112 patients from six published surgical series. The procedures used for this group were 68 AVFs. 14 ADFs, and 30 LAM/LOPs. Although the average postoperative grade for all six series was 1.08, in individual series this varied from 0.4 to 1.8. Neither age nor sex factors nor the choice of anterior or posterior surgery was uniformly associated with the best postoperative outcome.

The highest average postoperative grade of 1.8 came from the series reported by Itoh and Tsuji of 13 patients who underwent LOPs. Taken by itself, this result would seem to indicate that posterior techniques are best for OPLL surgery. In the same vein, a greater morbidity for anterior surgery might be indicated by the series reported by Harsh et al. of 19 patients, all of whom had anterior procedures and who yielded the lowest average grade of only 0.4.5 This disparity in grade could not be attributed to differences in the severity of preoperative neurological dysfunction, because both groups had similar initial deficits: Three of Harsh's 19, and four of Itoh's 13 patients, were in class IIIb preoperatively, and nine and seven, respectively, were in class IIIA5.8 However, in the other four studies, a better postoperative grade and outcome was achieved in patients who had anterior surgery either alone (two series) or in combination with posterior surgery. Evaluation of 43 OPLL patients revealed that the best surgical results followed anterior procedures (average grade 1.6 for AVF; average grade 1.54 for ADF), even though the AVF patients had on average the worst preoperative deficits. In contrast, the patients who had a laminectomy showed the worst average outcome (grade 0.9) although they had only intermediate preoperative dysfunction. McAfee et al., similarly observed that the 6 of their 13 patients who had an anterior rather than a posterior procedure for OPLL had a better average postoperative Ranawat grade.

Intraoperative SSEP Monitoring

Awake nasotracheal intubation, awake positioning, and continuous intraoperative SSEP monitoring may limit the morbidity associated with OPLL surgery. A significant deterioration in the amplitude or latency of median (N20) or posterior tibial nerve (N38) SSEPs, detected over 100-sec recording intervals from all four extremities, may warn of impending neurological injury. A significant SSEP deterioration (defined as an increase of over 10 percent in latency or a decrease of more than 50 percent in amplitude that is replicated over 100 sec intervals) may indicate the need for prompt resuscitative measures, which may reverse the SSEP abnormalities and prevent neurological injury. Nonsurgical measures include reversing hypotension, reducing or eliminating inhalation anaesthetics, warming irrigating solutions, and administering high-dose methylprednisolone. Surgical manoeuvres include ceasing manipulation, releasing distraction, and, occasionally, removing the graft.

SSEP monitoring in OPLL surgery is feasible even for the most myelopathic individuals. Half of the absent preoperative SSEPs-typically involving no more than two out of eight potentials in any given patient-recovered intraoperatively. Additionally, in half of those patients exhibiting significant intraoperative SSEP deterioration, the SSEP recover intraoperatively, and in the other half, the potentials begin to return to base­line prior to closing. Furthermore, significant improvement in one or more SSEPs is observed in up to 75 percent of the OPLL patients.

Posterior tibial responses proved more sensitive than median potentials to impending cord injury, whereas significant changes in amplitude and in latency are nearly equally common. Significant changes are considered to be true rather than false positives, that is, to reflect real neurological injury successfully avoided by appropriate resuscitative techniques. It is assumed that the adoption of strict significance criteria helped to avoid false negatives or misses. Using the SSEP monitoring protocol described, it may be possible to reduce the incidence of quadriplegia and the frequency of root injuries described in some series of patients treated surgically for cervical spondylosis and OPLL.

Summary

OPLL is now recognized to be present in nearly 25 percent of cervical myelopathy. An accurate determination of the full extent of OPLL by means of CT is critical to surgical planning. Although there is still controversy regarding whether anterior procedures (AVF, ADF) or posterior procedures (LAM, LOP) are best for the surgical management of OPLL, the results of anterior techniques are increasingly promising. Routine intraoperative SSEP monitoring appears to limit morbidity.