History and Description

Dr. Ralph Cloward, who spent the majority of his professional life in Hawaii as one of the only practicing neurosurgeons during the period encompassing World War II, is credited as one of the original proponents of the posterior lumbar interbody fusion (PLIF). Although Mercer had theorized an interbody fusion as the “ideal” operation for spinal stabilization in 1936 and reportedly Jaslow performed the first PLIF in 1946, Cloward popularized the technique the following decade through presentations of his own cases at various meetings of organized neurosurgery.^1–4 In his 1952 description of 321 patients treated with PLIF for ruptured intervertebral discs, Cloward mentions the unacceptably poor durability of symptomatic relief afforded by lumbar discectomy at the time as an impetus for reexamining the more conservative treatment of disc herniations. He implied a general lack of appreciation in the spine community of the primary etiology, namely instability resulting from a broken or damaged vertebral joint as the causative factor in the generation of lumbar disc disease. Cloward argued that effective treatment for the radicular symptoms as well as the mechanical symptoms of low back pain present in many of his patients required not only decompression but immobilization of the damaged joint. He advocated for intervertebral body fusion as opposed to fusion of the dorsal elements alone because this allowed for restoration of the intervertebral space (and thus indirectly the neural foramen and central canal) while incorporating the chief load-bearing elements into the fusion.¹ As Cloward wrote in 1953, “The purpose of this procedure was to maintain the normal width of the intervertebral space and the intervertebra foramen. At the same time al1 false movement between the vertebra resulting from injury and collapse of the intervertebra disc was arrested.”²

Approximately 30 years after Cloward’s initial presentations regarding PLIF, Lin modified the original concept to include four central principles.⁵ Preservation of the posterior elements of the motion segment (by maintaining the supraspinous and interspinous ligaments and limiting bony decompression to a medial facetectomy) and total (80%) discectomy (thought to improve fusion rates) represented the first two principles. Partial decortication, but not complete removal of the bony end plates, and the concept of the “unigraft” represented the remaining two principles. The “unigraft” concept was analogous to the model of the interbody graft used in anterior cervical discectomy and fusion surgery, namely the packing of all remaining interbody space with autologous bone to achieve a single solid fusion mass.⁵

Lin also described the notion of “dynamic decompression,” which refers to the combining of two motion segments into one in a state of relative decompression. Widening of the neural foramen is achieved through both direct decompression and indirectly through the use of an interbody graft.⁵ This restoration of the normal anatomic relationship between the motion segment and neural structures achieved in part through maintenance of disc height and restriction of motion was thought to protect the nerve roots. Immobilization of the unstable degenerative area arrested further degeneration; some authorities thought that this helped to protect the nerve roots. Investigators have cited reestablishment of weight bearing to ventral structures, prevention of recurrent disc herniations (at that level), and placement of the anulus under tension as advantages of PLIF over more conservative decompressive procedures such as discectomy, laminectomy, and foraminotomy.^6,7

The PLIF has seen many changes and procedural variations since its first description nearly 80 years ago.^1,2,8 Modifications to the surgical technique have mainly focused on the expansion of available methods for achieving decompression, access to the disc space, and fusion. Unilateral versus bilateral exposures with unique methods of introducing grafts into the disc space and decompression strategies ranging from midline to far lateral, incorporating laminotomy, laminectomy, various extents of facetectomy, or foraminotomy have been devised. Additionally, the application of minimally invasive options for achieving interbody fusion has likely increased its attractiveness. An assortment of interbody grafts has been proposed, including various autologous sources such as morselized or structural elements of the posterior neural arch, bicortical or tricortical grafts, cancellous sources, or iliac crest.⁹ Allograft sources ranging from morselized cancellous bone chips to numerous shapes of structural cadaveric bone graft such as bone pegs, trapezoids, crescents, or other configurations designed in part to prevent retropulsion or graft migration have also been described. Titanium mesh cages, threaded cages (BAK cage, Ray cage), polymeric rectangular cages (Brantigan carbon cage, polyetheretherketone [PEEK] cage, bioabsorbable cages (Hydrosorb cage), and ceramic cages (hydroxyapatite blocks) have been developed, all with variable biomechanical and bioabsorptive properties and similar clinical outcome in most reviews.^10–13 Facilitation of bony fusion and sound construction, applying concepts such as load sharing to complement the native spine successfully in its endurance of biomechanical forces, are among the fundamental properties used by these various grafts. Studies characterizing available interbody grafts generally describe variations such as sagittal contour, posterior disc height, fusion rates, clinical outcome, and biomechanical advantages in load sharing and the modulus of elasticity.^9,14–19 One investigation found cylindrical cages to be associated with a higher rate of nerve root damage than wedge-shaped bone allografts.²⁰ Some class III evidence supports general trends such as improvement in radiographic and functional outcome using artificial cages over bone chips. However, incorporating bilateral pedicle screw fixation may attenuate some of the differences described for the various interbody cages, and no consensus has been reached regarding efficacy among interbody grafts.^21–25

Although some early pioneers of PLIF believed that a posterior fusion was redundant and perhaps unnecessary given the construction of a good anterior fusion, a description of supplementation with posterolateral bony fusion appeared early in the history of the PLIF.^1,26 More recently, evidence has accumulated for the use of posterior instrumentation, which, among other advantages, provides immediate internal fixation and facilitates fusion, as well as helps prevent loss of disc space height, progressive kyphosis, and graft migration.^13,27–30 The widespread use of transpedicular screw fixation for posterior spinal fusion began in the late 1980s and has generated necessary adaptations to the original concepts of PLIF as described by Mercer, Cloward, Lin, Jaslow, and others.^{1–6,31–33} This has spawned a plethora of outcomes studies describing new techniques.^6,34,35

The use of PLIF was not widespread until the 1990s, when advances in instrumentation and surgical technique revitalized its popularity.^36,37 Although the procedure is biomechanically sound conceptually and most reviews demonstrate significant improvement in various radiographic and functional outcomes when applied to spinal disorders commonly treated with decompression and fusion, comparison studies with other fusion techniques show mixed results.^13,37–46 Spondylolisthesis, which is thought by many surgeons to require rigid internal fixation to better withstand inherent instability from factors such as shear forces, has varying results regarding need for interbody fusion.^42,47–50 A recent international study comparing open pedicle screw instrumented fusion with and without posterior interbody fusion for spondylolisthesis found the PLIF group to have a better fusion rate and the other group to have a higher rate of hardware complications related to biomechanics; however, there were no significant differences in clinical outcome.⁷ Studies of degenerative spondylolisthesis suggest evaluation of preoperative segmental instability may be an important factor in determining whether to perform additional interbody fusion.⁴⁷

More recently, the transforaminal approach to the intervertebral disc, known as transforaminal lumbar interbody fusion (TLIF), has gained popularity.^13,19,51,52 Originally described by Harms in the late 1990s, the TLIF has arguably developed into the most commonly performed and efficacious posterior interbody fusion method in modern spine surgery.^51,53,54 Although there are many variations on specific surgical techniques, the underlying concept is access to the intervertebral disc space from a more lateral trajectory; this is generally accomplished through unilateral exposure of the neural foramen and exiting nerve root using a greater degree of facetectomy.⁵⁵ It should be noted, however, that although the “traditional” TLIF uses a less invasive unilateral approach to the disc space, whereas the “traditional” PLIF uses a more extensive bilateral exposure, both can be performed in a unilateral or bilateral fashion.⁵⁵ The impetus for the development of the TLIF grew out of concern regarding damage to the cauda equina and lumbosacral nerve roots because the more midline exposure performed for the PLIF requires more retraction of the dura to achieve adequate operative exposure for the interbody fusion. Reflective of this concern is the limitation of PLIF to the levels of L3-S1 to avoid damage to the conus medullaris, whereas the TLIF may be performed at higher levels. Therefore, a major advantage to the TLIF is the potential for less damage to the dura or nerves.^36,56 A recent prospective study evaluating TLIF for degenerative and isthmic spondylolisthesis found a median decrease in the Oswestry Disability Index (ODI) score of 10 points (23.5 to 13.5), with a fusion rate of 95%. There was a 7.6% serious postoperative complication rate requiring operative revision.⁵⁷

Although both PLIF and TLIF are regarded as technically demanding, repeated studies have shown them to be safe and effective means of establishing circumferential fusion with similar clinical outcomes.⁵⁸ Numerous outcomes studies evaluating the efficacy of PLIF and the plethora of technical modifications are varied, however.⁴⁵ With regard to fusion, it is now widely accepted that interbody grafts result in higher fusion rates, often exceeding 90% to 94%.⁵⁹ Multilevel PLIF cases appear safe and effective, and although mention is made regarding increased invasiveness, good clinical outcomes and lumbar lordosis have been described.⁶⁰ However, although both techniques include positioning of an interbody graft under compression, maintenance of the posterior tension band, and to an extent correction of deformity through restoration of lumbar lordosis, recent reports have found additional benefits of TLIF over PLIF.^13,52,61 These include better improvement in lumbar lordosis given placement of interbody graft within the anterior column, greater enlargement of the neural foramen, and the option for using an effective unilateral approach; all these preserve other aspects of the posterior column integrity such as the contralateral lamina, facet, and pars, which may also provide a greater surface area for bony arthrodesis.^*13,48,57

Compared with TLIF, it would appear that the larger dural exposure required for PLIF may carry with it added risk of durotomy during dissection of scar tissue in revision surgery. Fusion rates, however, are similar between the two and range between 89% and 95% in most studies.^13,48,57 Humphreys et al. performed a comparison of operative characteristics and complications between TLIF and PLIF.⁵⁷Although for one-level surgeries, there was no significant difference in blood loss, operative time, or length of hospitalization, there was a statistically significant decrease in overall complications as well as blood loss for two-level surgeries in the TLIF group.⁵⁶ Preservation of the interspinous ligaments and preservation of the contralateral laminar surface were cited as additional advantages of TLIF in this study.⁵⁶ Biomechanical studies focusing on the destabilizing effects of surgical approach for TLIF suggest no significantly increased spinal flexibility with the exception of axial rotation at L4-5. However, this was corrected with bilateral (but not unilateral) pedicle screws.⁶² No consensus can be made on the effect of exposure on destabilization or the ability of pedicle screws to correct this effect between TLIF and PLIF, however, because other studies have demonstrated good results with unilateral PLIF surgery.⁶³

Many additional variations to both TLIF and PLIF have been described. TLIFs incorporating more aggressive decompressions through total bilateral facetectomies and resection of the pars (in combination with instrumented fusion), enabling placement of bilateral interbody graft placement as in PLIF, have been reported.⁵⁵ Alternatively, less invasive procedures (e.g., LI-PLIF), which advocate the use of percutaneous pedicle screws if needed and “preservation of posterior elements and avoidance of far lateral dissection over the transverse processes,” have been described.³⁵

Although outcomes studies often cite a steep “learning curve” for the necessary acquisition of new skills and familiarity of equipment as limitations, minimally invasive techniques (described elsewhere) have several unique advantages that make them attractive alternatives to open surgery.³⁵ Quicker and less painful postoperative recovery and less destruction of adjacent tissue are commonly mentioned advantages.^64–66 Recent studies have also demonstrated reduced hospital charges and lower transfer rates to inpatient rehabilitation facilities.⁶⁷ Investigators have described similar clinical results for spondylolisthesis, including similar reductions in listhesis when original slip is less than 50%, between open and minimally invasive PLIF.⁶⁶ Researchers recently reported a minimally invasive method of interbody fusion for isthmic spondylolisthesis.⁶⁸ This technique, called the extraforaminal lumbar interbody fusion (ELIF), uses a prone position and several small parasagittal incisions and fluoroscopic guidance to achieve decompression and percutaneous pedicle screw fixation. This approach appears to be limited to decompression of the exiting nerve root, however. Another recent series describes minimally invasive TLIFs using unilateral percutaneous pedicle screw fixation and demonstrates results similar to those of previously published open TLIF/PLIF procedures.⁶⁴ Reports of other approaches such as minimally invasive presacral approaches for intervertebral discectomy and fusion at L5-S1 exist in the literature.⁶⁹ Although exposure to radiation and operative times may be longer for minimally invasive PLIF or TLIF, reported differences in long-term clinical outcomes with open surgery are not consistent.^70,71

Although new reports refining and validating old techniques and innovative descriptions of novel concepts continue to represent a majority of the available literature, one appreciates some commonly debated concepts. These include trends in the preservation of various posterior elements, trajectories for accessing the disc space, strategies for achieving fusion, less invasive techniques, and advancements in materials research.^72–74

Indications for Surgery

Controversy regarding the indications for PLIF has existed since its inception.^1,2,26,45,75 Cloward designated broad indications for PLIF that included essentially all symptomatic lumbar disc disease (low back pain with or without radiculopathy resultant from a pathologic disc). Current treatment guidelines consist mainly of class III evidence, because class I and II data are lacking. Outcomes studies have generated a more extensive and specific list of indications; however, in light of the current trend of cost-benefit analysis and increasing costs of health care, controversy remains.

Degenerative disc disease (generally associated with Modic changes), lumbar segmental instability (iatrogenic, degenerative, or other causes), spondylolisthesis, degenerative scoliosis, pseudarthrosis after previous fusion surgery, spinal stenosis, deformity, and recurrent disc herniations are common indications for PLIF or TLIF surgery.^{13,59,76–78} Cloward, in his review of 100 patients with 30 to 40 years of follow-up after treatment for spondylolisthesis, has described PLIF in combination with laminectomy of the entire separate neural arch as a “superior operation.”³² Citing poor durability of posterior decompressive surgery for symptomatic spinal stenosis (15–20% short-term failure and 50% long-term failure), Hutter combined the benefits of anterior fusion and posterior decompression afforded by the PLIF and applied them to the treatment of this disorder. In 142 patients with a minimum of 3 years of follow-up, he described good or excellent results in 78% and a fusion rate of 91%.⁷⁹ In a relatively large, long-term (12-year), single-surgeon comparison study of PLIF versus standard laminotomy and discectomy for lumbar disc disease, Hackenberg described similar results in clinical outcome; however, he found PLIF to reduce the rate of revision surgery, and overall he thought that it represented an improvement in the management of lumbar disc disease.⁴⁹ Although one-level disc disease (at L5-S1) was well treated with either technique, additional instrumentation and stabilization were recommended for spondylolisthesis and severe multilevel lumbar disc disease.⁴⁹ Lumbar instability has been debated for some time, and clear evidence of instability is not always found. Some authors state that a difference of 10 degrees of angulation on lumbar flexion and extension films or spondylolisthesis of 4 mm or more may support this diagnosis, whereas other surgeons rely more heavily on intraoperative findings of instability such as “rocking of the vertebral bodies one on another with two Kocher clamps on adjacent spinous processes” (as described by Brown).^80,81 Segmental instability as defined by Frymoyer et al. refers to a “loss of motion stiffness such that force application to that motion segment produces greater displacement than would be seen in a normal structure, resulting in a painful condition and the potential for progressive deformity.”⁸² In a prospective study comparing discectomy to PLIF for massive disc herniations and/or segmental instability, some superiority of PLIF over discectomy was demonstrated within 5 years of follow-up.⁷⁵ PLIF and TLIF have been shown to be more effective at deformity correction than posterolateral fusion and more cost-effective than anterior-posterior lumbar interbody fusion with similar outcomes.^39,41,44 Most proponents of PLIF cite protection from pain resultant from recurrent disc herniations (attributed to the complete discectomy performed in preparation for the interbody fusion) as a primary indication for fusion surgery over more conservative decompressive procedures.^49,75,83 Data regarding the durability of lumbar disc surgery for controlling radiculopathy describe recurrence rates between 10% and 29%, and reduction in the recurrence of radicular symptoms subsequent to disc herniations has been demonstrated with PLIF.⁴⁹ Controversy remains regarding how many conservative lumbar discectomies should be performed (if any) prior to PLIF.

Attempts have been made to clarify the more subjective indications for PLIF. Evidence positively correlating preoperative level of disability (in the setting of degenerative disc disease) to functional recovery has been described.⁷² Acute disc herniations with significant and protracted pain, chronic mechanical axial low back pain (“discogenic” pain), central disc prolapse at the L4-5 level, high demand activity, the “failed back syndrome,” and chronic back pain following chemonucleolysis are all supported to some degree as indications for decompression and/or interbody fusion.^26,36,81 Chemonucleosis, although currently not as commonly performed due to concerns regarding side effects such as allergic reactions and neurologic deficits, is proposed to lead to loss of intervertebral disc height and progression of degenerative disc disease similar to the standard model of degenerative lumbar disc disease.⁸¹

Contraindications to PLIF may include arachnoiditis, active infections, short life expectancy (<3 months), severe osteoporosis, severe subchondral sclerosis with failure to demonstrate viable bone marrow with MRI, and potentially significant epidural fibrosis.^6,84

Techniques

The PLIF as proposed by Cloward in 1953 was described as a 2- to 3-hour surgery for an experienced surgeon. Recent studies have demonstrated similar operative times and have associated longer operative times with increased perioperative complications.⁸⁵ Cloward described three separate surgical phases: harvest of the iliac crest bone graft, laminectomy and discectomy, and spinal fusion.¹ Modern technical advances have generated some necessary adaptations to the original descriptions.

In contrast to most current positioning techniques, which maintain a more neutral position through slight hip extension and elevation in an attempt to preserve lumbar lordosis (i.e., Jackson spine frame which in addition facilitates use of 360-degree fluoroscopy), some surgeons may prefer the original position described by Cloward, which used a prone flexed position (knee-chest position) favoring an opened posterior lumbar spine. Although flexion of the hip reduces lumbar lordosis, it facilitates access to the spinal canal. Interestingly, however, a recent study failed to find a difference in the maintenance of lumbar lordosis after PLIF whether a Wilson frame or a Jackson table was used.⁸⁶