Chapter 172 Lateral Lumbar Interbody Fusion

Indications and Techniques

Stabilization and arthrodesis of the lumbar spine may be achieved by various surgical approaches. Traditionally, direct anterior or posterior surgical approaches have been employed that take advantage of natural tissue planes and midline orientation. A lateral approach to the lumbar spine provides a unique surgical corridor to the spinal column but is often avoided due to the risk of nerve injury of exiting lumbar nerves as they traverse the psoas muscle. Though lateral approaches for vertebral corpectomy are sometimes necessary for trauma or tumor, the traditional indications for lateral interbody discectomy and arthrodesis are rare. Recent advances in neuromonitoring and surgical access have enabled lateral lumbar interbody fusion (LLIF) to be performed with increasing safety and thus have been reinvigorated over the past decade. Today, LLIF can be performed in both openly and with minimal invasion and can be done at single or multiple levels. As with all surgery, patient selection is of utmost importance for LLIF to be successful, and potential complications and limitations must be recognized.

The traditional lateral approach to the thoracic and lumbar spines was originally developed by Capener for the management of tuberculous spondylitis (Pott’s disease) in the 1950s.¹ Through thoracotomy and retroperitoneal approaches, wide exposure of the lateral spinal column could be achieved for debridement and stabilization of the spine. In the 1970s, Larson et al. modified and popularized the lateral extracavitary (LEC) approach to the thoracic and lumbar spines to treat a variety of pathologies, effectively allowing for an entirely posterior approach to reach the anterior and lateral spine.² However, aside from major trauma, infection, or tumor cases, which typically required removal of vertebral bodies, these approaches were not popular for interbody discectomy and stabilization. The risk of nerve injury and the extensive disruption of the psoas attachments were perceived as too great for benign conditions such as degenerative disc disease and back pain. Therefore, LLIF was relegated to rare use, namely, for anterior release in deformity/scoliosis cases.

The LLIF approach has gained increasing popularity with the development of minimally invasive surgery (MIS) techniques. Advantages of MIS techniques include smaller incisions, less tissue dissection, improved cosmesis, decreased blood loss, less postoperative pain, and thus, shorter recovery time and hospital stay. The use of an MIS technique in lateral approach allows access to the lumbar spine through the retroperitoneal fat and psoas major muscle via a small incision using a muscle splitting technique. Advances in electromyography (EMG) neuromonitoring have been incorporated into these techniques, which in turn have made them safer. Pioneering work in minimally invasive LLIF is credited to Bergey et al. for describing an endoscopic transpsoas discectomy in 2004 ³ and to Ozgur et al. for describing the minimally invasive lateral interbody fusion technique in 2006.⁴ The minimally invasive LLIF technique has been trademarked under the proprietary name Extreme Lateral Interbody Fusion (XLIF, NuVasive, San Diego, CA) or Direct Lateral Interbody Fusion (DLIF, Medtronic, Memphis TN), but the procedure can be performed through a variety of tubular retractors outside of these proprietary formats.

In the minimally invasive technique, access to the disc space is achieved through a minimally disruptive lateral, retroperitoneal, transpsoas approach to the spine using neuromonitoring electrodes to identify a safe corridor and expandable tubular retractors for exposure. This chapter reviews the indications, surgical anatomy, techniques, benefits, and potential complications of open and minimally invasive LLIF techniques. A comparison to anterior lumbar interbody fusion (ALIF) and posterior lumbar interbody fusion (PLIF)/transforaminal lumbar interbody fusion (TLIF) techniques is also provided.

Indications and Contraindications

The goals of LLIF are to access the lumbar disc space safely, release the lateral annulus attachments, remove disc material, and place a structural graft. The results should be increased interbody height, restoration of collapse or deformity, and stabilization of interbody motion. LLIF is most suitable for interbody access from L2 to L4 for degenerative disc disease with or without instability, adjacent segmental disease, degenerative spondylolisthesis (grade I or II), and complex degenerative scoliotic deformity. LLIF can be performed at L1-2, but requires either removal of or maneuvering around the descending 12th rib (Fig. 172-1); it also can be performed at L4-5 but with a higher chance of nerve root injury. Also, the positions of the iliac crests determine whether L4-5 can be accessed. LLIF at L5-S1 is generally contraindicated due to obstruction by the iliac wing (Fig. 172-1). Other relative contraindications include grade III or greater degenerative spondylolisthesis, greater than 30-degree lumbar deformities, and bilateral retroperitoneal scarring. In addition, LLIF is generally not used alone when direct posterior decompression is necessary, such as with lumbar stenosis or disc rupture. It can, however, be combined with staged posterior decompression and posterior–lateral fusion if necessary. Patients with radicular symptoms and neuroforaminal stenosis can be considered for indirect decompression by restoring disc height and increasing foraminal diameter via LLIF, which is an active area of investigation.

FIGURE 172-1 Lateral lumbar x-ray with an outline of the overlying 12th rib and iliac crest, showing a clear view of the L2-3 and L3-4 disc spaces. Access to the L1-2 and L4-5 spaces would require maneuvering around the overlying bony anatomy.

Clinical Results

Unfortunately, there are limited clinical data concerning outcomes using LLIF techniques (including XLIF, DLIF, or other lateral lumbar interbody approaches). In a prospective series of 100 patients with adjacent segment degeneration after prior lumbar fusion, Rodgers et al. reported an average improvement of the visual analogue score (VAS) for pain from 8.6 to 2.8 within 6 months using the XLIF technique.⁵ Though this was not a comparison study, the mere improvement in pain scores in a patient who underwent prior lumbar fusion surgery with a minimally invasive XLIF is clinically meaningful.

LLIF approaches may be especially beneficial in treating certain complex scoliotic deformities, as they provide excellent coronal and sagittal corrective ability (Fig. 172-2). Pimenta et al. reported 23 symptomatic adult scoliosis patients at levels between L2 and L5 using LLIF approaches and achieved significant changes in coronal and sagittal alignment, as well as improved pain score.⁶ Benglis et al. also presented favorable short-term outcomes with mid- to high-lumbar coronal deformities treated with LLIF techniques.⁷ All patients showed improvement in preoperative pain and solid arthrodesis at 6 months. Similarly, Diaz et al. reported a 3-year follow-up for 39 patients treated with LLIF for symptomatic degenerative scoliosis and showed consistent improvement of the VAS and scoliotic deformity improvement in 3-year follow-up.⁸ LLIF can also be combined with other minimally invasive techniques, such as trans–axial lumbar interbody fusion.⁹

FIGURE 172-2 Degenerative thoracolumbar scoliosis corrected with staged anterior release at L1-2 and L2-3 with lateral lumbar interbody grafting and posterior thoracolumbar–pelvic fixation. A, Preoperative x-ray. B, Postoperative x-ray. Arrows indicate L1-2 and L2-3 interbody PEEK grafts.

Minimally invasive LLIF procedures might be also used for lumbar total disc replacement (TDR). Pimenta et al. described a series of 25 patients who has TDR placed using minimally invasive LLIF for degenerative disc disease with positive discography.¹⁰ The authors reported an improvement of the VAS from 7.5 to 2.6 and Oswestry Disability Index from 60 to 30. They also found this approach to be quick, to have minimal morbidity, and to avoid the need for anterior longitudinal ligament (ALL) removal; therefore, it has at least a theoretical advantage in segmental stability over the anterior approaches. Artificial discs placed in the lateral position have not yet been evaluated or approved by the Food and Drug Administration in the United States.

Surgical Anatomy

As in all surgical disciplines, a thorough understanding of the surgical anatomy involved in LLIF is crucial for maximal patient benefit and complication avoidance. For LLIF, the most critical anatomy is the distribution of the lumbar plexus within the psoas muscle, because the approach inevitably requires the use of a dilator or retractors to traverse the psoas muscle, which places the lumbar plexus at risk of injury.

The paired psoas major muscles form a major part of the posterior abdominal wall. They are long, thick, and fusiform shaped, and they lie lateral to the lumbar vertebrae. They arise from the roots of the lumbar transverse processes; pass inferolaterally, deep to the inguinal ligament to reach; and insert into the lesser trochanter of the femur. Another major component of the posterior abdominal wall is the paired quadratus lumborum muscles, which form a thick muscular sheet in the posterior abdominal wall alongside the lumbar vertebral column. They lie posterior and lateral to the origin of the psoas muscle.

The lumbar plexus is embedded mainly in the posterior portion of the psoas, anterior to the lumbar transverse process (Fig. 172-3). It is composed of the ventral rami of the L1 through L4 roots. Major cutaneous branches include (1) the ilioinguinal and iliohypogastric nerves (L1), which supply the skin of the suprapubic and inguinal regions; (2) the genitofemoral nerve (L1 and L2), which supplies the cremaster muscle and the skin over femoral triangle; and (3) the lateral femoral cutaneous nerve (L2 and L3), which supplies the skin on the anterolateral surface of the thigh. The genitofemoral nerve pierces the anterior surface of the psoas muscle and runs inferiorly, deep to the psoas fascia. The two major motor branches of lumbar plexus are the obturator nerve (L2-L4), which emerges from the lower part of the medial border of psoas muscle and supplies the adductor muscles, and the femoral nerve (L2-L4), which emerges from the lower part of the lateral border of psoas muscle and supplies the hip flexors and knee extensors.

FIGURE 172-3 Branches of the descending lumbar root plexus as they course through the psoas muscle lateral to the vertebral bodies and disc spaces. Lumbar disc levels are labeled to the left. The middle and anterior portions of the disc spaces at L2-3 and L3-4 are generally clear of traversing nerves, while the L5-S1 space is frequently impeded by nerves in the lateral trajectory. The L4-5 space has a variable degree of clearance with respect to traversing nerves and therefore warrants more attention.

(Courtesy of Medtronic, Memphis, TN.)

In 2003, Moro et al. defined the relationship between the psoas major muscle and the lumbar plexus using cadaveric dissection.¹¹ Excluding the genitofemoral nerve, the roots and critical branches of the lumbar plexus were found to be overlapping with the dorsal half of the vertebral column above L4-5 in a lateral projection view. The genitofemoral nerve, however, traverses through the psoas to emerge on the ventral surface between the rostral third of the L3 and the L4 vertebral bodies. When genitofemoral nerve is taken into account, only the ventral half of the vertebral column above L2-3 is free of lumbar plexus. The safest corridor, then, for the minimally invasive LLIF approach is the ventral half of the vertebral body above L2-3. Damage to the genitofemoral nerve usually causes only a transient sensory disturbance to the ipsilateral scrotum and medial thigh, which rarely becomes a serious problem. The ventral half of the vertebral column above L4-5 is therefore considered safe if transient genitofemoral nerve dysfunction is acceptable.

In comparison, accessing the lateral vertebral body at or below L5 carries significant risk of damaging critical structures such as the L4 and L5 nerve roots, femoral nerve, and/or obturator nerve. Thus, although there is considerably more space between the psoas major muscle and the quadrates lumborum muscle at L5-S1 compared to at L4-5 and above, L5 and below may not be used for lateral approaches to the lumbar spine. Benglis et al. also found an obvious dorsal to ventral migration of the lumbar contribution to the lumbosacral plexus within the psoas muscle from L2 to L5 in their cadaver studies.¹²

Using the lateral transpsoas approach in cadaveric dissection to identify the structures at risk with transpsoas K-wire and dilator placement, Banagan et al. found a more serious potential anatomic problem ¹³: The nerve roots and the genitofemoral nerve could be at risk in all their dissections in which the transpsoas approach is re-created. K-wire placement caused damage in 25% of cases at L3-4 and L4-5, including one direct L4 nerve root piercing. It was also found that the lumbar plexus was under tension after sequential dilator placement even when no direct injury happened during insertion. In addition to the lumbar plexus, the sympathetic chain was identified in the anterior third of the psoas over the disc spaces of L1 to L4, putting it at risk of potential damage with the transpsoas approach. Using a similar approach, Davis et al. found the femoral nerve is consistently at risk as it crosses the L4-5 interspace and can be compressed against the L5 transverse process when retractors are opened during the XLIF/DLIF procedures.¹⁴

Using a morphometric analysis of the ventral lumbar nerve roots and large vessels with the vertebral end plate, from hundreds of MRI studies, Regev et al. found the overlap of either roots or large vessels with the end plates gradually increased from L1-2 to L4-5. At the L4-5 level, the overlap can reach up to 87%, resulting in a very narrow corridor for potential LLIF procedures.¹⁵ Scoliosis was found to further decrease the potential safe corridor for LLIF. The preceding anatomic studies indicated the importance for neuromonitoring when establishing safe passage through the psoas muscle during LLIF procedures.

Surgical Technique

Open LLIF

Access to the lateral lumbar spine using an open approach allows several potential surgical dissection planes. In the LEC approach popularized by Larson et al., the lumbar spine is accessed with the plane of dissection posterior to the quadrates lumborum and anterolateral to the erector spinae muscles.² The erector spinae muscle group is elevated and retracted medially to expose the lateral elements of the spine. In 2002, Wolfla et al. described a retroperitoneal L2 to L5 lumbar interbody fusion using a true lateral trajectory to treat symptomatic nonunion.¹⁶ The access to the lumbar spine was provided by retraction of the psoas and quadrates lumborum muscles posteriorly. In their series of 15 patients with painful pseudarthrosis from 1 or more (average 2.1) previous posterior lumbar operations, 87% had significant improvement after open LLIF, and a 90% radiographic fusion rate was reported.

In open LLIF, the patient is placed in the lateral decubitus position in a plane perpendicular to the floor to facilitate obtaining lateral radiographs (Fig. 172-4). Generally, a left-sided approach is preferred for preferential retraction of the descending aorta as opposed to the inferior vena cava. In addition, for upper lumbar levels, the liver may prevent right-sided exposure. The ipsilateral lower extremity is preferably flexed at the hip to reduce tension on the psoas muscle.