Introduction
Despite excellent outcomes for carefully selected patients undergoing surgical management of lumbar spine pathology, a small but significant subset of patients will have persistent or recurrent symptomatology. Unfortunately, some of these patients may require revision lumbar spine surgery. Given the high patient, provider, and socioeconomic burden associated with revision spine surgery, the treating surgeon must carefully plan all revision interventions, including instrumentation choices. Considering the quickly expanding number of instrumentation options and approaches available to spine surgeons for use in the lumbar spine, a thorough understanding of each is paramount for a successful outcome. The purpose of this chapter is to provide an overview of instrumentation options for use in the revision setting in the lumbar spine.
Although these will be described in more detail in other chapters of this text, a brief overview of the indications for revision lumbar spine fusion is warranted in a discussion of instrumentation options. Patients may require a revision surgery for symptomatic adjacent segment disease, pseudarthrosis, previous instrumentation failure (which is sometimes associated with pseudarthrosis), infection, iatrogenic or unaddressed spinal deformity, and persistent or recurrent neural element compression. Accurate identification of the cause of persistent or recurrent patient symptoms is necessary for successful instrumentation selection in revision surgery. In particular, selection of the appropriate instrumentation may assist the treating surgeon addressing persistent or recurrent stenosis, pseudarthrosis for any reason, and correcting iatrogenic or unaddressed lumbar deformity.
Considerations
Need for Decompression
Addressing persistent or recurrent neural element compression may be a frequent indication for revision surgery. In this situation, the extent of decompression required must be considered when selecting whether to perform a fusion and use instrumentation (and, if so, what instrumentation option would be appropriate). For example, persistent neuroforaminal stenosis, or the need to address a foraminal disc herniation amidst extensive scar tissue, may require a partial or complete facetectomy to address safely. In this situation, consideration should be given to fusion with rigid fixation given the iatrogenic destabilization that will be created to address the pathology.
In addition, certain instrumentation options may address decompression through indirect means, and may be the only options to adequately address neural element compression. In the setting of a collapsed disc, with stenosis primarily in the cephalad to caudal direction, distraction may be the most reliable way to increase neuroforaminal dimensions. Chen and colleagues conducted a cadaveric study examining the effects of interbody distraction on neuroforaminal size and showed that distraction of the disc space can increase neuroforaminal area up to 30% to 35%. Their findings have been confirmed by multiple studies.
Need for Solid Arthrodesis
Some authors suggest that obtaining a solid arthrodesis is vital to achieving good clinical outcomes. Although achieving a solid arthrodesis is an important goal in every lumbar fusion, in the setting of revision surgery it may be particularly challenging. Many factors influence fusion rates, including patient factors, the type of graft, and the surgical material used for the technique. Instrumentation options in the revision lumbar spine setting allow the surgeon to control the local biomechanical environment in which arthrodesis will occur. In particular, creating a rigid construct for arthrodesis can result in higher rates of fusion. Increased rigidity may be achieved by using larger or longer screws, by altering screw trajectory, by manipulating rod size or material, by adding bone cement, or by pelvic fixation. Also, the addition of an interbody spacer to the fusion construct likely improves the chance of obtaining a successful arthrodesis.
Need for Deformity Correction
The treating surgeon must closely evaluate alignment parameters to successfully choose the instrumentation to be used in the revision setting. In particular, overall sagittal and coronal alignment, as well as lumbar lordosis and pelvic incidence, must be examined on full-length standing films. Often, deformity present in the revision setting is as a result of iatrogenic causes from creating a flat arthrodesis or from a failure to correct a previously present deformity in an effective manner. The location and severity of the deformity may dictate the surgical approach and whether instrumentation is used. For example, significant kyphosis in the lower lumbar spine may call for interbody distraction, with or without an anterior column release, to achieve alignment goals. Alternatively, if a significant correction is required through a posterior osteotomy, larger or multiple rods may be required to decrease the chance of instrumentation failure.
Patient Bone Quality
The patient’s bone quality, particularly the presence of osteoporosis, should play an important role in instrumentation selection in the revision setting. Osteoporosis can lead to poor fixation, early implant loosening, and subsidence with both posterior- and anterior-based instrumentation. Osteoporosis has been shown to decrease the pullout strength of pedicle screws in the lumbar spine, as well as in sacral fixation. Although the gold standard to measure patient bone mineral density is dual energy x-ray absorptiometry (DEXA) scan, the treating surgeon may be able to obtain some information about the patient’s bone quality via routine imaging studies done for surgical planning. Hounsfield units (HU), measured on computed tomography (CT) scans, can estimate patient bone quality. In terms of practical application, decreasing HU is associated with pedicle screw loosening in the lumbar spine. An L1 HU of 110 and 135 is 90% specific for osteoporosis and osteopenia, respectively, and may provide a quick measure of osteoporosis in patients who have or will get a CT scan.
Posterior Instrumentation
A large number of posterior-based instrumentation options exist, including wires, hooks, various screw fixation techniques (e.g., pedicle, cortical, facet, pars, etc.), and interspinous and dynamic fixation devices. Although many of these may be reasonable options in the primary setting, the options in the revision setting are more limited. For example, although hooks may provide easy and reliable fixation during deformity correction, their use in a revision setting may not provide adequate fixation to accomplish surgical goals.
Pedicle Screws
Pedicle-based fixation was introduced in the 1980s and quickly became popular for its ability to create rigid constructs via fixation of all three columns of the thoracolumbar spine. In the revision setting, pedicle screw fixation allows the surgeon to create a rigid environment to allow for deformity correction, prevent iatrogenic destabilization, and encourage arthrodesis. As discussed in a previous section, obtaining rigidity via pedicle screw fixation is of utmost importance, and a number of nuances must be considered by the treating surgeon.
First, location of fixation is important to consider. Depending on trajectory, pedicle screws extend from either the superior articular process or the mamillary body, through the pedicle, into the vertebral body. The majority of the fixation is obtained in the pedicle, with 80% of the cephalocaudal strength and 60% of the pullout strength based on the quality of bone in the pedicle. Thus, to achieve stronger fixation, larger-diameter screws that engage the cortical bone of the pedicle should be the primary focus, with increasing length of the screw to engage the cancellous bone of the vertebral body being a secondary goal. In the revision setting, when a previous screw has already been placed in the tract of the pedicle, the surgeon may question whether to reinsert a previously placed pedicle screw. Some authors have suggested that a previously strong pedicle screw may be replaced without compromising fixation. However, others have disputed this and have found that reinserting pedicle screws may compromise the principles of pedicle screw fixation, particularly in the revision setting. Increasing the diameter with a new pedicle screw to reengage cortical bone in the pedicle may be prudent. Consideration of a longer pedicle screw is another option.
Apart from the decision to use pedicle screws as well as selecting the size of the pedicle screw that is to be placed, the treating surgeon can optimize screw placement to maximize the purchase obtained. Barber and colleagues showed that screws placed at 30 degrees of convergence had significantly higher pullout strength than parallel screws. Convergence also allows longer screws to be placed compared with parallel screws in the axial plane. Alternative screw paths, such as placement of screws into the endplate, transdiscally, or into the anterior cortex with S1 pedicle screws, may also allow more rigid fixation in patients with inferior bone quality.
In patients with osteoporosis, placement of screws augmented with bone cement is an option that may allow for stronger pullout strength. Hydroxyapatite cement may be placed before screw placement or through fenestrated screws to increase fixation strength. In addition, hydroxyapatite cement can be used to rescue failed screw trajectories or a previously pulled out screw ( Fig. 9.1 ). Polymethylmethacrylate (PMMA) is another polymer that may also be used to augment pedicle screws in patients with poor bone quality. When planning screw augmentation, bone cement is most effective closer to the pedicle. However, cement augmentation with pedicle screws is not without complications. Although the rate of cement leakage into the canal or the rate of cortical defects is quite low, vascular leakage is extremely common, with rates as high as 75% into the segmental veins and 25% into the inferior vena cava. The clinical consequence of this vascular extravasation is unclear, as pulmonary consequences secondary to cement extravasation are uncommon.
In general, insertional torque is related to the pullout strength of the screw. The surgeon should be able to tell whether a screw has a satisfactory “bite” based on the amount of force required to insert the screw. If there is not a healthy amount of resistance while placing the screw, alternative options, such as cement augmentation or screw redirection to obtain more convergence, may be considered. In some newer screw designs, there are altered thread patterns which lower the insertional torque of the screw while allowing higher pullout strength. The surgeon should still, however, expect some feedback during screw insertion in these cases.
Cortical Screws
Cortical screws provide an alternative screw trajectory that begins on the lateral pars and is directed in a medial to lateral and caudal to cephalad direction. Although these screws are shorter in length, they are reported to have similar pullout strength to traditional pedicle screw trajectory because they engage more robust cortical bone, and may even have greater pullout strength in osteoporotic bone. In addition, cortical screws obviate the need to dissect laterally to the transverse process to find the starting point for pedicle screws, as the starting point is on the pars. Early biomechanical analysis of cortical screw and rod fixation showed similar strength to pedicle screw and rod fixation techniques, regardless of whether or not an interbody fusion was performed. Additionally, early clinical reports in prospective studies showed no significant difference in fusion rates at 1 year, with significantly less operative time, blood loss, and shorter surgical incisions. Larger meta-analyses have confirmed the early findings in terms of equivalent clinical outcomes between cortical and pedicle screw fixation, with faster operative times, lower intraoperative blood loss, and decreased length of stay.
However, these findings apply mostly in the primary setting. The outcomes of cortical screws in the revision setting are less clear. As discussed earlier, a revision setting may require more bony decompression, resulting in iatrogenic destabilization that is not typically the case in the primary setting. Sin and Heo found that cortical screws had inferior mechanical strength compared with pedicle screws, particularly when the midline ligaments were taken down. Nomoto et al. agreed with these findings, and found the lumbar cortical screws were inferior to pedicle screws in terms of mechanical strength in a destabilized setting, such as with a wide laminectomy. These findings suggest that cortical screws may have limited application in the revision situation.
Some authors have advocated for the use of cortical screws in the setting of failed pedicle screw fixation, which may also be encountered in a revision setting. In a biomechanical cadaveric study by Calvert et al. cortical screws placed at the same side and level after pedicle screw failure provided similar levels of stiffness with flexion, extension, and axial rotation compared with primary pedicle screw fixation. Other authors have found similar results, and have credited the alternative trajectory that cortical screws use as a reason for robust fixation in the revision setting.
Pelvic Fixation
Pelvic fixation may be an important consideration in revision lumbar spine surgery. In particular, patients who are undergoing long fusion constructs to the sacrum, have high-grade spondylolisthesis, have unstable sacral fractures, have poor bone quality in the sacrum, or require a destabilization osteotomy close to the lumbosacral junction may require pelvic fixation to add stability to the construct to allow fusion at the lumbosacral junction. In revision settings that require lumbosacral fusion, pelvic fixation should be strongly considered given that fusion constructs to the sacrum have a high failure rate without pelvic fixation. There are a multitude of options for sacropelvic fixation, including S1 tricortical screws, S2 and sacral alar anchors, iliosacral anchors, Jackson intrasacral rods, and the Galveston technique. The most commonly used techniques, however, are iliac screws through the posterior superior iliac spine (PSIS) and S2 alar iliac (AI) screws.
Iliac screws are typically placed through the PSIS and extended between the inner and outer tables of the ilium. These screws provide rigid fixation at the end of a long construct, can enhance fusion rates, and can protect the S1 pedicle screws from stress and cutout. Although these screws can be difficult to line up with the rest of the pedicle screw construct given its lateral starting point, side-to-side connectors can be used to include it in the construct. In addition, in thin patients, screws at the PSIS can be very prominent and bothersome. This can be addressed by recessing the starting point on the PSIS with a high-speed burr or Leksell rongeur.
S2 AI screws also provide sacropelvic fixation with the added benefit of lining up better with the rest of the construct and being much less prominent than PSIS screws. In addition, they provide fixation of the sacroiliac joint which can sometimes see altered or increased stress after a long lumbosacral fusion construct. The starting point for an S2 AI screw is usually just distal and lateral to the S1 neuroforamen. After this, the screw path is created approximately 40 degrees to the horizontal plane, and 20 to 30 degrees caudally. Once the sacroiliac joint is crossed, the screw should follow a path between the inner and outer tables of the ilium. This technique has been shown to have excellent outcomes with low revision rates when pelvic fixation is required.
Rod Selection
Rod selection may have important implications in the rigidity of the construct created and should be thought about carefully in the revision setting. Rods can be selected based on size, number, material, and precontoured versus noncontoured. In general, larger-diameter rods afford a higher level of rigidity and may give the overall construct more stability. In the setting of revision surgery for any type of pseudarthrosis or bony resection osteotomy, use of larger rods may be required.
With regard to size and number of rods, some authors have seen acceptably low rates of instrumentation failure when using smaller rods (4.75 mm). However, this has been documented in the setting of degenerative pathologies primarily, without demanding large corrections from the rods used. The deformity literature has seen much lower rates of instrumentation failure, with better maintenance of correction with use of larger-diameter rods (i.e., 6.35 mm compared with 5.5 mm rods). However, the use of larger rods is not without complications/consequences. Asher and colleagues found that although larger 6.35 mm rods had increased stiffness compared with 4.75 mm rods, much higher stress shielding was present, with no difference in fusion rates.
Although increased rigidity may allow for a better mechanical environment to allow for arthrodesis, too much rigidity may stifle the micromotion necessary for bone formation. In the setting of pedicle subtraction osteotomy, for example, the use of extra rods decreases the rate of instrumentation failure, but lengthens the amount of time necessary to heal the osteotomy site. In addition, some authors report that increased rigidity of the construct may accelerate breakdown at an adjacent segment. Han and colleagues found that although increasing the stiffness of rods decreases the rate of rod breakage, it increases the rate of proximal junctional kyphosis after deformity correction. Semirigid rods may create less stress adjacent to a previous instrumentation construct, thereby potentially decreasing the rate of adjacent segment disease. To this end, some authors have proposed the use of polyetheretherketone (PEEK) rods to allow for semirigid fixation and have found similar levels of stability to titanium constructs in biomechanical studies. However, it is unclear whether the use of semirigid rods would provide the stability needed in a revision setting.
Rod contouring can bring another point of potential failure into the fixation construct in a revision setting. The notches created by manual contouring by a rod bender in the operating room slightly weakens the rod and is usually the point of failure in titanium rods. In contrast, cobalt chromium rods that are precontoured typically fail at the head of the titanium screw, rather than at the rod itself, suggesting that they are not weakened by rod contouring to the extent that titanium rods are. In a biomechanical study by Serhan et al., whereas contoured cobalt chromium rods were able to produce higher corrective forces than titanium rods, they still underwent deformation when asked to correct very rigid curves. Precontoured rods were the best of all materials tested at maintaining their shape in response to high corrective forces.
Crosslinks
Crosslinks are rod-to-rod links that may be used to improve the rigidity of the overall construct, particularly in multilevel constructs. A single crosslink connector can improve the strength of nonsegmental posterior pedicle screw constructs to the level of segmental constructs. Additionally, diagonal placement of a crosslink may help resist torsional forces on the construct, and may be a consideration in patients who may require more rigidity from their fixation. Although crosslinks may improve the torsional strength of posterior instrumentation, they likely do not improve resistance to axial, flexion/extension, or lateral bending forces.
Some authors have questioned the usefulness of crosslinks in long constructs. Dhawale et al. compared maintenance of correction, clinical outcomes, and complications in patients who underwent instrumented fusion for adolescent idiopathic scoliosis, with or without the use of a crosslink, and found no difference between the two groups. The authors suggest that in a construct with multiple levels of fixation, the use of crosslinks may not add a significant amount of rigidity and may not improve outcomes.
Anterior Instrumentation
Anterior instrumentation options in revision lumbar surgery most commonly consist of interbody or corpectomy devices/cages placed in the anterior column. In the primary setting, some authors have questioned the need for interbody fusion for degenerative spinal conditions, asserting its association with longer operative times, higher blood loss, and higher complication rates without improving clinical outcomes. However, in the revision setting, certain goals such as deformity correction, indirect neural element decompression, and obtaining solid arthrodesis may require anterior-based intervention. In particular, patients with collapsed discs or other anterior column pathology may be more likely to benefit from anterior instrumentation. Interbody instrumentation options will first be discussed based on approach to the interbody space, as approach dictates the use of certain instrumentation options, followed by cage profile. The specific surgical approaches will be discussed in more detail in other chapters in this textbook.
Posterior Approach
The posterior approaches to the interbody space were popularized by Cloward, and have subsequently become commonly used to access the posterior, middle, and anterior columns of the lumbar spine via a posterior approach. Although described by Cloward originally as the posterior lumbar interbody fusion (PLIF), subsequently, the transforaminal lumbar interbody fusion (TLIF) allowed access to the disc space with minimal root retraction, decreasing complications.
Although posterior-based interbody devices allow the surgeon to achieve multiple goals through a single approach, the working corridor of a TLIF is much smaller than lateral and anterior-based approaches. Thus the profile of the cages used, as well as the ability to prepare the disc spaces well, is limited compared with lateral and anterior-based approaches. In a retrospective review comparing surgical approaches with lumbar interbody fusion, Watkins and colleagues found that both anterior and lateral lumbar interbody fusions were successful at restoring lumbar lordosis and sagittal alignment at almost 2 years, whereas TLIF was not. Other authors have documented similar findings when comparing lumbar interbody fusion approaches and it is likely caused by the type of device that can be used in each setting. TLIFs seem to require more aggressive posterior releases to achieve greater lordosis when attempting to achieve alignment goals in the revision setting.
Lateral Approach
The lateral approach to the interbody space, or the lateral transpsoas approach, was developed with the intention of allowing access to the interbody space proximal to L5–S1without the need for an approach surgeon. The lateral lumbar interbody fusion (LLIF) is relatively readily accomplished from L1 to L3 or L4, with the L4–L5 disc sometimes in close proximity to the lumbar plexus. However, safe access to the L4–L5 disc with the use of real-time neural monitoring has been reported. Thigh pain and weakness after this approach has been described, presumably caused by psoas muscle dissection. However, these effects are thought to be temporary, with more than half resolving by 3 months, and greater than 90% resolving at 1 year. The lateral approach allows for a larger cage that can easily span the entire medial/lateral profile of the interbody space, affording powerful sagittal correction. Some authors have asserted that LLIF cages can cause overall favorable increases in lumbar lordosis and sacral slope.
Although LLIF cages may allow for large corrections anteriorly, they may require rigid fixation to allow for arthrodesis and maintenance of sagittal alignment, particularly in the revision setting. In a large meta-analysis of 1492 patients, Alvi et al. found that standalone LLIF without posterior fixation was associated with a higher rate of reoperation and cage subsidence. Their results suggest that LLIF may require supplementation posteriorly, particular in patients with poor bone quality or in the revision setting.
Placement of the lateral cage in a particular position in the disc space also allows the treating surgeon control over changes in sagittal alignment depending on surgical goals in the revision setting. Anand and colleagues conducted a retrospective radiographic analysis on LLIF cage placement, and found that lower lumbar lateral cage placement, as well as placement of cages in the middle and anterior one-third of the disc space, had a more powerful impact on lumbar lordosis compared with high lumbar and posterior one-third disc-space cage placement.
In settings where even more sagittal correction is desired, anterior column release (ACR) can be considered. ACR uses the release of the anterior longitudinal ligament to allow for more distraction of the disc space. This may allow for corrections up to 30 degrees of lordosis in certain clinical scenerios. Some authors have asserted similar levels of sagittal correction as with pedicle subtraction osteotomy, although it is unclear whether or not these authors have taken into account the flexibility of the deformity undergoing correction.
As discussed earlier, use of lateral cages in revision surgery may provide indirect decompression required to adequately address persistent neural element compromise. Interbody distraction is a powerful tool to increase neuroforaminal size, and thus relieve any nerve root compression caused by cephalocaudal compression ( Fig. 9.2 ).
