Transforaminal Lumbar Interbody Fusion

Published on 17/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1477 times

CHAPTER 50 Transforaminal Lumbar Interbody Fusion

Alan S. Hilibrand, MD, Harvey E. Smith, MD

The intervertebral disc space is an environment that is conducive to fusion because of the advantageous biomechanics of compressive force and the blood supply provided via the endplate after curettage of the cartilage. This is in contrast to the posterolateral space, in which the fusion mass is under tensile forces, must bridge a greater distance between the transverse processes, and is surrounded by muscle. Involvement of the intervertebral disc space is ubiquitous with degenerative disc disease because the disc space undergoes progressive loss of height as part of the degenerative process. This loss of height results in progressive micromotion and is hypothesized to contribute to instability and degeneration of the posterior elements. Restoration of the intervertebral disc space height provides for an indirect decompression of the neural foramen, while allowing one potentially to address issues of sagittal imbalance by restoring lumbar lordosis. Consequently, achieving an interbody fusion with restoration of the disc space height is one of the component goals of surgery to address degenerative disease and deformity of the lumbar spine.

Anterior lumbar fusion was first described by Muller in 1906.¹ Evolutions of the technique have included modifications of the abdominal approach to a less invasive, mini-open technique and the use of interbody cages or structural allografts augmented with autogenous iliac crest graft, local bone graft, or, more recently, recombinant human bone morphogenetic protein (rhBMP-2).²^–⁵ Without the augmentation of posterior support, it has been well recognized that there is an increased rate of graft subsidence.^1,⁶ The combination of anterior lumbar interbody fusion with a posterolateral instrumented fusion (360-degree fusion) has been shown to yield fusion rates of greater than 95%.⁷ Anterior lumbar interbody fusion has significant potential morbidities, however, including potential injury to the great vessels, abdominal hernia, injury to the sympathetic chain with subsequent sexual dysfunction, and thromboembolus secondary to retraction of the artery.⁸^–¹⁰

Recognition of the biomechanical advantages of an anterior lumbar interbody fusion augmented with rigid posterior instrumentation led to the development of posterior-only approaches to the disc space to eliminate the approach-related morbidity of anterior interbody fusions. Posterior lumbar interbody fusion was developed to provide access to the disc space via a bilateral posterior approach with retraction of the thecal sac. Posterior lumbar interbody fusion is limited in its use to below the level of the conus, however, owing to the degree of thecal sac retraction necessary. Because of the retraction of the neural elements with posterior lumbar interbody fusion, there is concern for injury to the nerve roots, pain syndromes secondary to injury of the dorsal root ganglion, and cerebrospinal fluid leaks.¹¹

Anterior lumbar interbody fusion via a transforaminal posterior approach (transforaminal lumbar interbody fusion [TLIF]) was first described by Harms and Rolinger in 1982.¹² Using a transforaminal approach via osteotomy of the pars interarticularis and inferior articular facet allows for facile access to the disc space with minimal retraction of the neural elements, avoids the morbidity of an anterior approach,^13,¹⁴ and has a lower complication rate than direct posterior lumbar interbody fusion.¹¹ Accessing the disc space via a posterior annulotomy, approximately 56% of the endplate can be prepared for fusion¹⁵ with placement of bone graft and interbody implant. Combined with standard posterolateral instrumentation, decortication, and bone grafting, radiographic fusion rates greater than 90% can be achieved.¹⁶ Published outcomes of single-level TLIF and multilevel TLIF are equivalent or better than anterior and 360-degree fusions, with restoration of the disc space height.^16,¹⁷ TLIF has been shown to result in significant savings relative to anteroposterior interbody fusion, owing in part to the decreased operating room time, less blood loss, and shorter hospital stays.¹⁸

Although some degree of restoration of disc space height with TLIF has been documented, the magnitude of restoration has been shown to be less than that achieved via anterior lumbar interbody fusion.^19,²⁰ Similarly, longitudinal studies have shown a progressive loss of restored lordosis, as the fusion construct over time tends to drift back to the preoperative sagittal balance. The clinical significance of radiographic differences between anterior lumbar interbody fusion and TLIF is unknown because these radiographic differences have not been shown to correlate to differences in outcome measures.²⁰

Radiculopathy secondary to a foraminal disc herniation is difficult to address adequately from a midline approach and decompression owing to limited access to the lateral foramen. A far-lateral or foraminal disc herniation may be managed with a facetectomy, facilitating a transforaminal approach to the nerve root.²¹ TLIF may be considered as treatment for a foraminal disc herniation, particularly in the setting of significant loss of disc space height, because one may directly decompress the nerve root with the facetectomy, achieve an interbody and posterolateral fusion, and achieve indirect decompression with increase in the disc space height.

Advances in less invasive spine surgery, mini-open approaches, tubular retractors, and percutaneous instrumentation systems have been applied to TLIF, and minimally invasive TLIF procedures have been proposed to address the morbidity of the posterior midline approach. Preliminary studies have suggested that minimally invasive approaches may result in less blood loss and shorter hospital stays.²²^–²⁴ These preliminary reports are case series, however, so the results must be extrapolated to general practice with caution. Prospective randomized studies are needed comparing less invasive or minimally invasive TLIF procedures with the standard open midline approach to compare clinical outcomes and radiologic indices (disc space height, graft placement, lordosis, and fusion).

Surgical Procedure

The patient should be placed prone in standard fashion on a Jackson table, with care taken to pad all bony prominences. An Andrews table or frame may also be used, but care should be taken to consider if proper lordosis can be achieved in that position. Alternatively, the leg portions of hinged tables such as an Andrews or Wilson frame may be elevated before locking down posterior instrumentation to induce lordosis across the instrumented segments.

A standard midline incision is made through the skin and subcutaneous tissues, and subperiosteal dissection is carried down to the spine in standard fashion. The transverse process and pars interarticularis at the cephalad and caudal levels should be exposed in their entirety, with care taken not to violate the cephalad facet joint capsule (i.e., for L4-5 TLIF, the pars of L4 and L5 and transverse process of L4 and L5 must be entirely exposed, with care taken not to violate the L3-4 facet while exposing the landmarks for the L4 pedicle screws). An intraoperative marker film or fluoroscopy should be obtained to confirm the level of exposure. The interspinous ligament between the operative levels may be resected, and the interspinous space may be skeletonized. In the event of a tight interspace, a lamina spreader may be placed between the spinous processes to provide temporary distraction. It is recommended that vigorous distraction not be done on the pedicle screws because this weakens their biomechanical fixation. A consideration before taking down the interspinous ligament is the biomechanical support it contributes to the stability of the motion segment.

For transforaminal access to the disc space, it is necessary to remove the entire facet joint effectively on one side. This is accomplished by removing the inferior articular process of the cephalad level with an osteotomy through the pars interarticularis. This osteotomy should be performed on the side of greatest neural pathology; the osteotomy through the pars should be done as cephalad as possible to maximize the exposure but with care not to violate the pedicle itself (Fig. 50–1). Before the osteotomy, the ligamentum flavum should be completely freed from the lamina with a curet; particularly in the setting of a revision procedure, removal of the ligamentum and associated adhesions after the osteotomy may prove difficult and increase the risk of incidental durotomy. The transforaminal approach may be done either in concert with a midline laminectomy or with preservation of the spinous processes and midline ligamentous structures. To preserve the midline structures, one first makes the osteotomy up the medial aspect of the ipsilateral lamina, with the second osteotomy across the pars interarticularis as close as possible to the inferior margin of the cephalad pedicle. The inferior articular facet may be grasped with a Leksell rongeur or pituitary and rotated away from the underlying dura, taking care to remove any ligamentous adhesions with a curet.

FIGURE 50–1 Surgical exposure after completion of osteotomy with removal of facet joint. A, Thecal sac. B, Exiting nerve root. C, Intervertebral disc.

Using Leksell and Kerrison rongeurs, the superior facet of the caudal level should be resected as flush with the pedicle as possible, removing the overhanging portions so that the remaining bone is flush with the superior and medial aspects of the pedicle. The lateral recess may be completely decompressed, and the nerve root should be confirmed to be free of compression. At this point, careful identification of the exiting nerve root and medial border of the thecal sac is essential. Vessels that are in the field of approach to the disc space should be cauterized and divided with bipolar electrocautery. Although the exiting nerve root should be confirmed to be mobilized, excessive retraction should be avoided because injury to the ganglion can result in debilitating postoperative pain.

After completion of the surgical exposure of the transforaminal zone as described, a nerve root retractor should be placed medially to protect the thecal sac. With appropriate surgical exposure, only minimal retraction should be necessary as protection against incidental durotomy during the annulotomy. The posterior anulus should be incised widely with a box cut, with care taken to visualize the exiting and traversing nerve roots (Fig. 50–2). After completion of the annulotomy, the discectomy may be started with a pituitary rongeur and curets. A forward angled pituitary is necessary to enter the disc space on the contralateral side of the patient. When removing the disc and cartilage from the endplates, care must be taken not to violate the subchondral bone. Angled curets and chondrotomes should be used to remove the cartilaginous endplate from the far lateral side. A thorough discectomy and endplate preparation are essential, but absolute care must be taken not to violate the anterior anulus or subchondral bone (Fig. 50–3).

FIGURE 50–2 Incision of posterior anulus. Box cut should be as large as possible in posterolateral margin of disc, with nerve root and thecal sac as lateral and medial boundaries.

FIGURE 50–3 Angled curets (A) and chondrotomes (B) are necessary for thorough discectomy and endplate preparation, particularly far contralateral space. When working curets anteriorly and posteriorly, care must be taken not to violate anterior anulus.

As the discectomy progresses, sequential dilation of the disc space should be accomplished with serial dilators. When impacting the dilators into the disc space, care must be taken to follow the sagittal plane of the interspace to avoid driving the dilator through the endplate, particularly in an osteoporotic patient (Fig. 50–4). If necessary, exposure may be aided by distracting between the spinous processes with a lamina spreader. After distraction, the assistant must hold the lamina spreader to stabilize it because inadvertent dislodgment may place the contents of the spinal canal at risk. The posterior lips of the vertebral bodies may be removed with osteotomes so that they are flush with the concavity of the endplate; if not removed, the interbody implant may be undersized (Fig. 50–5). After dilation to an appropriate size, the trial implant should be placed, with care taken to tamp it appropriately medial and anterior.

FIGURE 50–4 Sequential dilators should be used to distract intervertebral disc space. Care must be taken in insertion of dilator in collapsed disc space to parallel endplate; particularly in an osteoporotic patient, inadvertent angulation of dilator may fracture endplate.

FIGURE 50–5 Posterior lips of vertebral bodies, if significantly different than concavity of disc space, may result in undersizing of implant. To address this, posterior lips may be removed flush with concavity with osteotome.

After implant trialing, the interspace should be bone grafted with the clinically indicated graft material. Bone graft should be placed into the anterior interspace and the implant. After anterior grafting of the interspace, the implant with graft is placed, and tamps are used to direct the implant anterior and medially (Fig. 50–6).

Based on surgeon preference, the pedicle screws may be placed before or after the discectomy and interbody fusion. It is recommended that the screws not be used for vigorous distraction of the interspace because this is deleterious to their fixation. The surgeon should place the screws in the standard fashion with which he or she is most comfortable. A complete decortication of the fusion bed should be completed, with bone graft placed in the posterolateral gutters and in the facet joint and dorsal elements, if preserved, on the contralateral side of the osteotomy.

Rods should be measured, cut, and bent into the appropriate lordosis and then captured in the pedicle screws. An intraoperative image must be obtained at this point to confirm pedicle screw and interbody cage placement, with care taken to note the anterior position of the cage. The wound should be irrigated and closed in standard fashion. Before wound closure, it should be confirmed with a probe that no bone graft has fallen into the canal or foramen.

A modified TLIF procedure may be used to address isthmic spondylolisthesis with reduction of the slip, if clinically indicated. A bilateral transforaminal approach should be done via the standard midline exposure, and the foramen on both sides should be decompressed. Using posted screws, a reduction maneuver may be done while using electrophysiology to monitor the neural elements. In the setting of an isthmic spondylolisthesis, the posterior elements are not attached to the vertebral bodies, and the TLIF procedure is a biomechanically sound approach. It provides a stable anterior interbody fusion, restores posterior stability through the placement of pedicle screw instrumentation to increase likelihood of obtaining a successful fusion, and includes a wide decompression of the foramen to facilitate direct visualization of the neural elements.

rhBMP-2 is currently approved by the U.S. Food and Drug Administration (FDA) for use in an anterior interbody cage but is not FDA approved for use in a posterior transforaminal interbody fusion. Several studies have investigated the off-label use of rhBMP-2 in TLIF, with published fusion rates similar to rates achieved with autograft bone in the cage. There have been reports of an increased incidence of radiculopathy postoperatively on the side of the transforaminal approach when rhBMP-2 is used, and an increased incidence of heterotopic bone formation has been noted.²⁵ When rhBMP-2 is used posterolaterally, there have been reports of sterile seroma formation and local tissue swelling.²⁶ Some investigators have hypothesized that rhBMP-2 may incite an inflammatory response along the nerve root if it leaks out of the intervertebral space, but this has not been confirmed. An increased incidence of transient vertebral body osteolysis has been noted^27,²⁸; it is hypothesized that with perforation of the endplate, the rhBMP-2 may incite a transient remodeling of the cancellous bone. The surgeon should take into consideration these reported complications when contemplating the use of rhBMP-2 in an “off-label” setting.

The TLIF procedure provides for an effective interbody and posterolateral fusion with fusion rates greater than 90%. TLIF also facilitates a wide direct and indirect decompression of the neural elements that cannot be accomplished through an anterior approach and anterior lumbar interbody fusion. There are, however, several technical aspects that should be considered.

The biomechanical load-sharing properties of the interbody graft are predicated on the bony stability of the endplate. Fracture or violation of the endplate may result in subsidence of the graft into the vertebral body. If increased bleeding is appreciated during the procedure, it may be an indication that the endplate has been violated. Similarly, visualization of cancellous bone during the discectomy indicates violation of the endplate. Endplate violation is also more likely in the setting of a collapsed disc space in an osteoporotic patient. To minimize this risk, particular care must be taken to appreciate the sagittal orientation of the endplates with intraoperative imaging and always to maintain the trajectory of the instruments parallel to this orientation. The authors have found it helpful to use gradual “blunt, bulleted distraction” into the interbody space and always avoid the use of sharp cutting tools such as endplate shavers in these patients. If there is a significant violation of the endplate, the surgeon should consider aborting the interbody portion of the fusion.

Although the transforaminal approach is advocated over a direct posterior approach to the intervertebral space in part because it necessitates comparatively less retraction of the thecal sac, there have been reports of postoperative radiculopathy on the ipsilateral side of the transforaminal approach. Some instances of radiculopathy may result from overly vigorous retraction of the exiting nerve root. There have been reports more recently of an increased incidence of radiculopathy when rhBMP-2 is used in the intervertebral space and incidences of heterotopic bone formation.^25,²⁹ This is an “off-label” use of rhBMP-2. This issue is significant, however, owing to the increasing use of biologics in spine surgery. It is hypothesized that biologics that upregulate the transforming growth factor pathway may promote an inflammatory response if they are introduced to the neural space. If the surgeon thinks it is clinically indicated to use a biologic osteoinductive promoter in the disc space, it may be advisable to consider strategies to minimize the possibility of migration from the disc space to the spinal canal and neural foramen.

There is some debate regarding the optimal position of the interbody graft. Harms and Rolinger ¹² originally described the procedure with a middle to posterior third graft position, to take advantage of the stronger posterolateral endplates. Subsequently, other authors have advocated a more anterior position of the interbody graft to improve its load-sharing properties and to facilitate restoration of lumbar lordotic angle.³⁰ In disc space preparation and implant insertion, care must be taken to preserve the anterior anulus. Violation of the anterior anulus may result in inadvertent entry into the retroperitoneum with either an instrument or an implant, which may be catastrophic because of the nearby location of the large vessels. During preparation of the disc space, dilation, trialing, and implant insertion, the surgeon must maintain direct visualization of the working space. Intraoperative x-ray or fluoroscopic imaging should be used if there is any ambiguity regarding the position of instruments, implants, or the angle of the disc space. If there is a decline in hemodynamic status at any point after beginning disc space preparation, the potential of a vascular catastrophe must be considered.

TLIF does restore disc space height, but it does not allow an anterior release, and multilevel TLIF may be associated with prolonged operative time and blood loss and less restoration of lordosis compared with a direct anterior approach with subsequent posterolateral fusion. Consequently, in the setting of significant kyphotic deformity, a direct anterior decompression with release may be a more appropriate consideration. In the setting of a rigid coronal plane deformity, an anterior interbody approach also may be more appropriate because multiple anterior releases may be needed for correction of coronal deformity, adequate distraction to achieve indirect decompression, and restoration of sagittal balance.

Case Presentation

A 38-year-old woman who had a prior right-sided L4-5 microdiscectomy presented with recurrent right L5 radiculopathy and significant back pain (Fig. 50–6). She was managed initially with selective L5 root blocks and physical therapy without improvement. Plain films and magnetic resonance imaging (MRI) showed a recurrent right-sided L4-5 disc herniation and significant loss of disc space height at L4-5 with a new degenerative spondylolisthesis. Given the recurrent disc herniation and degeneration with spondylolisthesis of the L4-5 interspace, L4-5 TLIF was offered to stabilize the L4-5 interspace, restore the disc space height, and decompress the nerve root.

FIGURE 50–6 A-C, Radiograph (A) and MRI (B and C) of recurrent L4-5 disc herniation with significant degenerative changes and loss of height of L4-5 interspace. D and E, Postoperative lateral (D) and anteroposterior (E) images show L4-5 transforaminal lumbar interbody fusion with significant restoration of interbody height.

Complications

TLIF has a complication rate that is comparatively low relative to that of anteroposterior lumbar interbody fusion.¹⁴ For single-level TLIF, fusion rates are reported to be greater than 90%, but they are less than 90% for multilevel procedures. As with any procedure with pedicle screws, there is the possibility of screw misplacement; the incidence of a misplaced pedicle screw with TLIF procedure is approximately 5%.¹³ Transient postoperative neurologic deficit has been reported to range from 2% to 7%. Neurologic deficits lasting longer than 3 months were reported to occur in 4% of patients undergoing minimally invasive TLIF approaches in one series of 73 minimally invasive cases.¹⁴ In cases reported to have a postoperative radiculopathy, the most commonly affected nerve root is L5.¹⁶

Minor complications have been reported to occur in 16% of cases, with an approximately 5%/level risk of transient radiculopathy and 4%/level risk of incidental durotomy.^13,¹⁶ One large series of 124 consecutive TLIF procedures (51 open, 73 minimally invasive) reported a 20% incidence of cerebrospinal fluid leak in the open cases.¹⁴ Postoperative infections have an incidence of approximately 5%.^13,¹⁶ Postoperative hematomas have been reported to have an incidence of approximately 4%.¹⁴ In addition to postoperative radiculopathy on the ipsilateral side of transforaminal approach, there has also been a report of postoperative radiculopathy on the contralateral side,³¹ hypothesized to occur in the setting of asymptomatic contralateral stenosis that is exacerbated by the increased lordosis resulting from the procedure. Reflex sympathetic dystrophy has been reported postoperatively,¹⁶ presumably from irritation of the dorsal root ganglion. Postoperative gastrointestinal complications, predominantly ileus, were reported in 4% of patients in the series reported by Potter and colleagues¹⁶ of 100 TLIF procedures; one patient developed pseudomembranous colitis.