Summary of Key Points
Compared to virgin spine operations, reoperations require more extensive exposures. Therefore, the risk of spine instability, neural damage, and infection increases.
Techniques used in the virgin operation are the same as those used in reoperations, but limitations created by scar tissue and the loss of bone and ligamentous structures that aid in spine stability are present. When undertaking any reoperation on the spine, the surgeon must keep in mind that with each succeeding operation, the challenge is increased and the chance of a good result is reduced.
Pain alone is not an indication for reoperation. Correctable anatomic abnormalities must be present to warrant repeat surgery.
It is always worthwhile to remember that the most common cause of spine surgery failure is not a technical error or complication but failure of appropriate patient selection. When patients have been improperly chosen for surgery in the first place, it is not likely that repair of an unintended consequence of the first operation will be beneficial. On the other hand, it is unfair to leave patients with an uncorrected abnormality that is producing symptoms. Repair of the demonstrated problem in such patients is reasonable.
All patients with failed spine surgery should, however, be carefully assessed for the presence of important comorbidities that may exaggerate the complaint of pain. These should be addressed simultaneously with reparative surgery.
Preventing repeat spine surgery is an important goal for surgeons and their patients. Reoperation is generally an undesirable outcome, implying persistent symptoms, progression of the underlying disease, or complications related to the initial operation. A higher risk of reoperation was observed among patients covered by workers’ compensation insurance compared with those with other types of insurance. Patients under age 60 were more likely than those age 60 years and older to have second operations. Males had a slightly lower risk of reoperation than females, and having any comorbidity resulted in a higher risk of reoperation.
Often reoperations on the spinal column are more technically difficult than the index procedure. Moreover, the risk of surgical complications is potentially greater. In addition to the technical problems of reoperation, clinical and radiographic evaluation of the patient is more difficult. Because normal anatomic relationships and normal tissue planes have been altered, imaging is less accurate and the surgical pathology is more difficult to recognize.
Patients undergoing surgery for degenerative spine disease may require further surgery for a variety of reasons such as progression of disease at the original operative level or at adjacent levels or for the development of instability or deformity typically at adjacent levels. Reoperation has proved to be much less effective than initial surgery, and it is estimated that only 30% to 50% of patients benefit from this second surgical procedure. The reoperation rate varies with the region of the spinal column, type of disease, and type of previous surgery. Reoperations are performed at the rate of 2.5% per year at the cervical spine level and range from 8.9% to 10.2% at the lumbar level. Reoperations are more expensive; one study found that the average hospital charge for a cervical spine reoperation is $57,205. Identifying modifiable factors, such as the choice of approach and gaining adequate stabilization of the spine, might reduce the need for spine reoperations and might improve public health and curb health care expenditures.
Usually, reoperations on the spinal column are performed for the following reasons: (1) recurrent or persistent neural compression, (2) development of new or persistent instability, (3) development of a deformity, (4) cerebrospinal fluid (CSF) leak, (5) postoperative hematomas, and (6) postoperative infection.
Role of Neuronavigation in Revision Spine Surgery
Since the 1990s, intraoperative neuronavigation, particularly computed tomography image-guided surgery (CT-IGS), has emerged as an alternative to fluoroscopy-based techniques. The argument for image-guided surgery has been most compelling in cases of spinal tumors, deformity surgery, and revision fusion surgery. In each of these instances, anatomy is significantly altered and difficulties are compounded by limitations in imaging visualization of bony landmarks. Intraoperative imaging during the revision surgery is important for accurate localization of instrumentation, its relationship to vital neurovascular structures, the achievement of adequate decompression, and maintaining structural integrity within the spinal column. Currently, many centers routinely use the fluoroscopy; however, fluoroscopy is limited in resolution and depth of field. Improved CT-IGS navigational functions of systems like the O-arm provide 2D and 3D resolution of structures in real time to ensure proper instrument placement and alignment. Neuronavigation reduces the risk that tactile feedback cues from previously violated or weakened bone or dense scarring will mislead the surgeon; it also provides the option of acquiring an intraoperative postinstrumentation CT image to confirm correct placement of screws.
A retrospective study found that CT-guided navigation provided safe screw placement in both revision and primary spine surgeries. This was true in nearly all locations of the spine, with all levels exceeding 96% in safe placement in the cervical, thoracic, and lumbar spine. Known limitations of CT-IGS are well documented and include concerns of cost, time, and radiation exposure. Revision surgery, however, presents its own set of concerns. In cases in which the spinous process or other rigid bony structure is unavailable, it may become a challenge to find a purchase in which to adequately affix the reference frame. There are several alternatives besides attachment of the reference frame to the spinous process. The more common solutions to cervical and high thoracic spine registration may be accomplished by attachment of the frame to a Mayfield head holder, whereas lumbar and sacral registration may be accomplished by frame attachment to a pelvic pin through the posterior-superior iliac spine.
Reoperations for Cervical Spine
Recurrent Cervical Radiculopathy or Myelopathy
Persistent radicular pain in the early postoperative period after anterior cervical discectomy and fusion may be due to retained disc fragment, especially when disc fragment has penetrated the posterior longitudinal ligament (PLL). Magnetic resonance imaging (MRI) or computed tomography (CT) myelography can help to diagnose the retained disc fragment. If a retained disc fragment is the suspected cause of persistent symptoms, reexploration and removal of the retained disc is advised. If the PLL was not opened completely at first surgery, it must be opened and inspected for any compression.
Late onset radicular symptoms may be due to symptomatic adjacent segment disease. Adjacent-level disease is defined as the development of a new radiculopathy or myelopathy referable to a segment adjacent to a previously fused level in the cervical spine. In up to 25% of patients, degenerative disease may develop at adjacent segments within 10 years of initial surgery, and 7% to 15% of these patients have been reported to require reoperation. Hilibrand and associates estimated that more than a quarter of all patients would develop radiographic degeneration of spinal elements adjacent to the index surgery, and two thirds of those would go on to require additional surgery. Indeed, the advent of total disc arthroplasty is predicated on the assumption that restoring normal spinal kinematics would alleviate further spinal degeneration and development of adjacent segment disease (ASD). Nonetheless, despite a number of studies looking at the effectiveness of total disc arthroplasty, to date there exists no level I evidence that restoration of spinal biomechanics using disc replacements prevents or even abrogates the incidence of ASD development, highlighting the need for a better understanding of this pathophysiologic process. The likelihood of developing moderate to severe adjacent level ossification was high when the anterior cervical plates were less than 5 mm from adjacent disc spaces. Decisions to use the same side approach must be considered carefully, if clinical and radiologic findings confirm symptomatic adjacent level disease. Otolaryngology evaluation of vocal cord motility should be done, if a contralateral approach is considered. If a vocal cord is noted to be paretic from a previous approach, then the repeat surgery should be same sided to avoid contralateral recurrent laryngeal nerve injury. Previous cervical plate system removal may be required to allow proper positioning of the plate for the adjacent level ( Fig. 210-1 ). Cage staple hybrids, threaded cages, are the alternatives that would obviate the need for previous plate removal ( Fig. 210-2 ). Patients undergoing an anterior revision surgery for adjacent segment disease after anterior cervical discectomy and fusion (ACDF) have higher rates of postoperative radiculopathy and redevelopment of ASD when compared with posteriorly approached patients. Patients receiving posterior revision surgery had greater intraoperative blood loss, hospitalizations, and postoperative complications such as wound infections and discharge to rehabilitation, but they had a statistically lower chance of redeveloping ASD requiring secondary revision surgery. This may be due to the fact that posterior revision surgeries involved more levels fused.
Vertebral osteophytes are a common radiologic finding affecting 20% to 30% of the elderly population. A number of factors are responsible for the local osteogenesis, notably mechanical factors. Both ACDF and posterior cervical foraminotomy represent surgical options to treat cervical radiculopathy. Certain cervical pathologies may benefit more from ACDF, such as central-lateral disc herniation, bilateral foraminal stenosis, osteophyte anterior to the nerve root, or facet hypertrophy. In a patient with a previous ventral cervical discectomy and fusion and with a persistent large osteophyte in the neural foramen, correction may be accomplished by performing a simple cervical foraminotomy from a dorsal approach, with or without drilling off the osteophyte. This procedure is probably easier than reoperating from the ventral approach and drilling out the previous fusion and decompressing the foramen. If, however, the osteophyte is ventral and medial in location and cannot be decompressed adequately from a dorsal approach, a reoperation from the ventral approach should be performed. Soft tissue planes may be scarred, but the tissue plane between the carotid sheath and the esophagus and trachea is usually maintained and easily dissected. If the soft tissue scarring is due to previous infection or radiation therapy, the operation may be simplified by operating from the opposite side. An operation being performed from the ventral approach for inadequate neural decompression requires increased bone resection, at least a minicorpectomy to allow definitive visualization of both nerve roots with magnification. This is often best accomplished by using a high-speed drill and an operating microscope. If the problem is a persistent central osteophyte or ossification of the posterior longitudinal ligament, corpectomy with interbody fusion is the safest ventral approach, allowing complete decompression of the spinal cord.
Posterior cervical foraminotomy (PCF) is a procedure that addresses foraminal nerve root compression, and it is considered to be a safe procedure with a low complication rate; however, radiculopathy symptoms may recur, in some cases necessitating reoperation. The reoperation rate with PCF is 5% to 6.6% with an average time to reoperation of 2.4 years after the first operation. The rate of reoperation at the index level was significantly higher than reoperation at adjacent or distant segments. Patients with preoperative neck pain had a higher risk for reoperation and a shorter time to reoperation.
Postoperative Cervical Instability
Extrusion of a bone graft, failure of fusion, iatrogenic instability, and failure of instrumentation are indications for reoperation. The graft extrusion is more common in C2-3, C6-7, and C7-T1 levels due to the anatomic variation in the bony spine. Extrusion of a strut graft associated with a corpectomy may be due to a poorly fitting graft or to fracture of the vertebral body into which the graft is fixed. This allows the caudal portion of the graft to extrude ventrally. The dislocation of the graft may be associated with collapse of the disc spaces, which, in turn, may cause nerve root irritation and pain. Furthermore, the extruded graft may result in adjacent structure compression, causing symptoms such as dysphagia and hoarseness of voice.
Extruded bone fragment is removed, the graft site is freshened, usually by use of a high-speed drill to accomplish good preparation of the end plates, and a new, appropriate size graft is inserted. A ventral plate-and-screw construct provides further assurance of retention of the bone graft. Other techniques for improved stability include hybrid constructs combining corpectomies and discectomies, which allow for more fixation points using an anterior plate, anterior-posterior fusion for greater stability, and multilevel interbody grafts as opposed to long corpectomies. High reoperation rates for extruded grafts and symptomatic pseudarthrosis have been associated with nonplated two-level anterior discectomy and fusion (ADF) and single-level anterior corpectomy with fusion (ACF) procedures. However, comparison of single-level ACF procedures performed with and without plates showed that plating did not appear to reduce pseudarthrosis or graft extrusion rates.
Iatrogenic Instability in Cervical Spine
Kyphosis may develop in up to 21% of patients who have undergone laminectomy for cervical spondylotic myelopathy. Progression of the deformity is likely if preoperative radiologic studies demonstrate loss of lordosis. The incidence of progressive deformity and instability after cervical laminectomy was highest in the pediatric population and in those with a malignant intramedullary lesion, adjuvant radiotherapy, or preoperative findings of kyphosis or instability. Other intraoperative factors, including multilevel laminectomies and removal of the facet joint, can also increase the risk for a subsequent development of deformity, instability, and neurologic sequelae. Although cervical laminoplasty has been proposed to reduce the risk of postsurgical deformity, reports suggest no significant reduction in the incidence of spine deformity, especially in the setting of intradural spinal tumor resection. Postoperative kyphosis is another cause or indication for revision cervical spine reconstruction. Postlaminectomy kyphosis involves the triad of deformity, incomplete soft tissue restraints, and neurologic compromise, all of which have to be addressed for the patient to obtain an adequate clinical improvement. After a destabilizing procedure, there is loss of sagittal alignment shifting the weight-bearing axis anteriorly, which leads to muscular fatigue and pain. Often, in the end stage, this results in the spinal cord being draped over the vertebral bodies anteriorly with subsequent neurologic compromise. Surgical strategies include lengthening the anterior column and shortening the posterior column. Anterior-only procedures tend to be inadequate because of the instability, with increased graft problems of extrusion, collapse, and failure of fusion. Anterior-posterior procedures lead to increased construct stability and the decreased need for external immobilization.
Failure of Fusion: Cervical Spine
The ultimate goal of spinal hardware is to provide temporary stability allowing for bony fusion, usually requiring 6 to 9 months. Failure of fusion and the development of pseudarthrosis or fibrous union are the sequelae of ongoing low-grade mobility. Pseudarthrosis is defined as an absence of bridging bone between grafted bone and vertebral bodies and the presence of a radiolucent defect, a halo sign, or a loss of grafted bone. Despite advances in the technologies and instrumentation of spine surgery, pseudarthrosis still occurs in 10% to 15% of all patients. Pseudarthrosis itself can be a source of pain, or it may provide a lead point for ongoing mobility leading to increased stress on hardware and inevitable failure, one of the indications for reoperation. Revision spinal arthrodesis for pseudarthrosis and loose instrumentation with widely dilated screw tracts is a difficult clinical problem.
Single-level ACDF is a highly successful procedure yielding high reported fusion rates, ranging from 83% to 97% for autograft and 82% to 94% for allograft, respectively. However, in multilevel ACDF, as the number of grafts increases, the cervical spine is predisposed to decreased fusion rates as contact stress increases between the graft-body interfaces, further contributing to unacceptable micromotion. Pseudarthrosis after ACDF has been recognized as a cause of continued cervical pain and unsatisfactory outcomes ( Fig. 210-3 ). Debate continues as to whether a revision ventral approach or a dorsal fusion procedure is the best treatment for symptomatic cervical pseudarthrosis. An outcomes analysis of the treatment of cervical pseudarthrosis with posterior fusion by Kuhns and coworkers demonstrated that 52% of such patients reported none or mild pain at follow-up, whereas 20% reported “discomforting” pain, and 28% reported continuous moderate-to-severe pain. The rationale for the dorsal approach includes the advantages of avoiding the scar tissue and potentially difficult tissue planes encountered in a revision ventral approach, as well as encountering a fresh fusion bed when a dorsal approach is used. Contraindications to the dorsal approach include those problems that can only be addressed through a ventral approach, such as graft migration or kyphosis. Advocates for a revision ventral approach suggest that patients experience more stiffness and pain after a dorsal approach secondary to disruption of the dorsal musculature. There is no level I evidence in the literature to support the superiority of an anterior or posterior approach. A meta-analysis concluded that anterior and posterior approaches appear to be viable surgical alternatives. Valid conclusions regarding the superiority of one approach over another are debatable; however, the limited data suggest that the posterior approach may have a greater potential for solid arthrodesis and clinical improvement.
Reoperation: Thoracolumbar Spine
Reoperation for Residual Thoracic Disc Herniation
Between 7% and 15% of patients with a thoracic disc herniation are asymptomatic; symptomatic herniation is uncommon, accounting for only 0.25% to 0.57% of herniated discs in the entire spine. The resection of herniated thoracic discs can present formidable technical difficulties, especially if the discs are calcified, large, centrally located, broad based, or extend transdurally. Residual herniated discs that cause neurologic symptoms secondary to spinal cord and nerve root compression or cause progressive neurologic deficit should be considered for reoperation. Two of the most common reasons for failure to remove a thoracic disc were mislocalization of the disc level and inadequate visualization of the pathologic entity. The latter principally reflects choosing the wrong route for operative exposure. Transpedicular or posterolateral approaches may be indicated for the treatment of a small, laterally positioned thoracic disc herniation. An anterior approach is indicated for the safe microsurgical removal of large, calcified, broad-based, midline, and transdural discs.
Mislocalization is a problem inherent to operating in the thoracic spine. The traditional method of marking the thoracic vertebrae believed to be involved by placing percutaneous sterile needles near the spinous process in a sequential fashion at every 3 or 4 levels starting at the sacrum is an accepted strategy. Fluoroscopy has been used to count the needles and the vertebral levels from the sacrum to the target level. Both methods are time consuming. In the past, preoperative skin surface localization with halibut liver oil surface markers filled with radiopaque material was used. Such can be visualized on MRI and CT to localize the involved thoracic spine level. This was a simple method, but it is problematic in patients with scoliosis, spinal deformity, obesity, and heavy skin folding. The skin markers may also shift during positioning. Prepreoperative CT-guided percutaneous fiducial screws, bone cement injection, and wire localization also described in the literature with good success.
Inadequate visualization of the herniated disc material can be prevented by selecting an approach that provides a full view of the ventral dura and herniated disc material and by removing enough of the vertebral body to ensure full visualization of the spinal cord. Stabilization was considered when a corpectomy was performed or when extensive bone was resected circumferentially at the same level during the multiple operative procedures. In contrast, fusion is rarely required after a first operation for resection of a herniated thoracic disc.
Persistent or Recurrent Neural Compressionin Lumbar Spine
Of all the indications for reoperation for neural compression, recurrent or persistent radiculopathy secondary to disc or scar tissue is by far the most common. Persistent symptoms with neural compression are seen in patients with a recurrent disc herniation, large foraminal osteophyte, inadequate decompression of thickened ligamentum flavum, facet joint hypertrophy causing root compression and inadequate decompression of the spinal cord or cauda equina in spinal stenosis, or recurrence of neoplasia.
The reported incidence of symptomatic recurrent disc herniation after lumbar discectomy varies between 3% and 18% in retrospective studies. Subjects with larger annular defects and those in whom a smaller proportion of disc volume was removed during the first surgery were associated with an increased risk of symptomatic recurrent disc herniation. Carragee and associates demonstrated that the reherniation rate varied from 1.1% with small fissure-like annular defects to 27.3% for large open annular defects. Recurrent disc herniation or progressive disc space loss after discectomy often leads to increased pain and disability, which necessitates repeat surgery. Revision surgery, however, does not always improve symptoms. Differentiation of a recurrent disc herniation from an epidural scar presents a dilemma. Characteristics associated with recurrent disc herniation include a nonenhanced or rim-enhanced abnormality surrounding a low-signal-intensity lesion on MRI and extension of contrast into the epidural space and an enhancing abnormality on CT/discography ( Fig. 210-4 ). Discovery of a focal mass of scar that is obviously compressing a nerve root may still be an indication for surgery; however, a diffuse epidural scar without nerve root compression is not. (FIG)
Patients who present with a recurrent lumbar disc herniation should initially be treated with a nonoperative approach, just like those who present with primary disc herniation. Patients who failed an adequate conservative treatment should be considered for repeat discectomy. Indications for more urgent surgical treatment include cauda equina syndrome (CES), progressive neurologic deficit, and severe intractable radicular pain. Reoperation for same site disc herniation is technically more demanding than the primary operation. Dissection begins in a normal area to reduce the chances of iatrogenic nerve root injury or dural tear. This is accomplished by enlarging the previous laminotomy and may include partial facetectomy to adequately visualize the lateral border of the nerve root.
Postoperative Cauda Equina Syndrome
Postoperative CES is a rare but troublesome complication of uneventful lumbar surgery, with an incidence ranging from 0.08 to 1.2%. Several theories have been proposed to explain the pathogenesis responsible for acute postoperative CES, such as underlying spinal canal stenosis along with low-lying conus medullaris; incarceration of the cauda equina through small and initially unrecognized dural defects, the effect of anesthetic agents, epidural abscess, retained surgical sponge, placement of excessive amounts of hemostatic agents, epidural hematoma, and sudden alteration of the vascular supply of the cauda equina. The venous congestion theory has gained popularity among spine surgeons because it is the only one that has some experimental support. In postoperative CES, the imaging studies are essential for revealing a compressive origin, even though, in the majority of postoperative CES cases, the imaging studies obtained revealed no evidence of compression. Postoperative CES associated with compressive pathology is commonly accepted as an indication for emergency surgery. However, controversy exists regarding the management of postoperative CES without a compressive pathology. Meta-analysis of surgically treated CES suggests benefit, if decompression is undertaken within 48 hours from symptom onset. So it is reasonable to recommend repeated exploration and decompression of cauda equina in any patient with postoperative CES and in whom there is no reasonable explanation for existing neurologic deficits. The prognosis for CES seems to be significantly affected by a patient’s preoperative neurologic condition. Patients with severe motor deficit prior to decompression tend to have less chance of motor recovery after decompression compared with patients who have minor deficits. Nevertheless, motor function is much more likely to return after surgery than is bladder function. Recovery of urinary retention and bowel incontinence is much more variable but occurs frequently in the perioperative period.
Iatrogenic Instability: Thoracic and Lumbar Spine
Revision lumbar decompression and extension of fusion remains the gold standard surgical treatment for patients with persistent, recurrent back and leg pain following a previous decompression and fusion. Compared with those for surgery for de novo low back pain (LBP), success rates for reoperations in elderly patients have been highly variable, ranging from 40% to 94%. Revision decompression and fusion provided a significant gain in health state utility for elderly patients with symptomatic pseudarthrosis, same-level recurrent stenosis, or ASD, with a mean 2-year cost of $80,594 per quality-adjusted life-years gained. When indicated, revision surgery for symptomatic ASD, same-level recurrent stenosis, and pseudarthrosis is a valuable treatment option for elderly patients experiencing persistent back and leg pain. Late failures in lumbar spinal stenosis may be due to persistent or acquired instability, recurrence of stenosis at operated levels, new stenosis at adjacent levels ( Fig. 210-5 ), epidural fibrosis, or arachnoiditis. Degenerative discogenic pain and reports of narrowing of the intervertebral disc space, reactive changes in adjacent vertebral bodies, vacuum disc phenomenon, spondylolisthesis (ventral or dorsal), and abnormal motion of 3 mm or more on flexion-extension radiographs all indicate lumbar instability. The development of postoperative intraspinal facet cysts has been related to the presence of postoperative segmental spinal instability, including a progression of spondylolisthesis and disc degeneration. Pedicle screw fixation fusion with interbody fusion gives better results in postoperative spinal instability.