Posterior Lumbar Interbody Fusion

History

Hibbs and Albee were the first to introduce stabilization of the spine in 1911. They reported the use of an interlaininar fusion technique for the treatment of the unstable spine secondary to Pott's disease. This technique became a popular treatment for lumbar instability secondary to infectious diseases of the spine but ultimately was extended to other pathological lesions and regions of the spine. In 1936, Mercer theorized that the ideal operation for stabilization of the spine was an interbody fusion. He lamented that the procedure was impossible to perform at that time given the existing equipment and sophistication of spinal surgery. The first posterior lumbar interbody fusion (PLIF) was reported by Jaslow in 1946 when he utilized a bone peg, that was placed within the lumbar interspace after discectomy. He augmented this with autogenous bone chips harvested from the posterior elements placed posteriorly. Variations on this procedure were sporadically reported over the years including the anterior lumbar interbody approach introduced by Hodgson and Stock in 1956.

The visionary report of Cloward in 1945 brought PLIF into mainstream spinal surgery. Cloward, in his landmark presentation at the American Association of Neurological Surgeons Meeting, reported his results in treating one hundred lumbar inter–vertebral disc herniations with PLIF. It was his belief that interbody fusion should be utilized as a first line strategy in treating ruptured lumbar disc disease. While it is now known that simple lumbar disc herniations refractory to conservative therapy have excellent long term outcomes with limited discectomy alone, the treatment for recurrent disc herniations is still less clear.

Indications

The indications for PLIF in recent years have ranged in spectrum from established, universally accepted indications such as gross instability to the borders of pain medicine for the treatment of low back pain. We are all familiar with the passionate arguments for and against the surgical treatment of low back pain secondary to segmental instability. Our indications for this procedure are based on an understanding of the pathophysiology of the degenerative cascade and the magnitude of disability and dysfunction the patient experiences. This is coupled with the correlation of objective diagnostic (and potentially therapeutic) management interventions, and an overall philosophy that has developed by our experience in treating patients with lumbar degenerative disease.

These include patients with degenerative disc disease characterized by segmental disc degeneration with Modic changes with or without foramina stenosis demonstrated on NIRI. These patients must have a clinical history consistent with mechanical low back pain with or without leg pain. Other candidates include patients with degenerative lumbar instability, spondylolisthesis (less than grade III), postsurgical spinal instability, and
pseudoarthrosis after attempted posterior fusions.

Plain films should demonstrate disc space collapse and parasagittal CT reconstructions usually show impingement of the neural foramen by the superior–facet of the inferior vertebrae.

Provocative discography is occasionally utilized in patients with multi–level degenerative changes to identify the possible pain generator. Whereas discography traditionally focused on the disc morphology, its current application is based on reproduction of a patient's particular pain (i.e. concordant pain). Specifically, needles are placed into the disc of interest as well as adjacent discs. Sterile saline solution or contrast material is injected.

A positive discograin recreates the patient's pain at the appropriate level of interest. Injection at adjacent levels should produce no pain or "discordant" pain (dissimilar from the patient's described pain in terms of location and intensity).

Patients considered to be poor candidates for interbody fusion have inappropriate responses to discography. The usefulness of discography is controversial and patient selection can be complicated.

Front, Back, or Neither?

We would consider absolute contraindications to interbody fusion (either anterior or posterior) to be: multilevel disc disease with a non diagnostic discogram, patients with single level disease and radicular pain without symptoms of instability (little or no mechanical back pain), and patients with severe osteoporosis. Relative contraindications for interbody fusion include patients who are actively smoking, patients with multilevel disease with more than one level positive on discogram, and patients with significant co–morbid illnesses.

When attempting to determine the best approach for an interbody fusion (i.e. anterior vs. posterior) it is important to consider evidence of posterior pathology such as a disc fragment in the canal, facet hypertrophy, lateral recess or central spinal stenosis. These findings would clearly cause us to favor a posterior approach. Contraindications specific to anterior approaches may include multiple prior abdominal procedures, which have led to adhesions, and severe atherosclerosis which may preclude mobilization of the aorta and iliac vessels in fusions above the L5–S1 level. Relative contraindications to posterior procedures include patients with multiple prior posterior surgeries suggesting significant scar formation. With advances in technology it is anticipated that some posterior pathology such as free fragments may eventually be addressed from an anterior approach similar to anterior cervical disectomy.

Modern Age PLIF

Over the last ten years it has been fascinating to see the development of the modern age PLIF. There are few spine procedures that have undergone such revolution in technology and frequency.

The introduction of pedicle screw fixation by Roy–Camille for posterior stabilization stimulated a resurgence of posterolateral fusions. This was followed by the development of better instrumentation systems for the delivery and stability, of the intervertebral disc space. Placement of threaded cylindrical titanium cases began to overcome standard posterolateral fusion techniques with pedicle screw fixation because they provided better biomechanical, biological, and functional advantages.

The improved instrumentation systems stimulated an explosion of interbody procedures because the original procedure was refined to the point that interbody fusions were easier to perform with less risk. After the initial reports of good success with threaded cylindrical titanium cages various modifications of the prosthesis have been created.

Our group has experienced an evolution of philosophy regarding instrumentation for patients undergoing PLIF. Several years ago we began utilizing stand alone threaded titanium cages packed with morselized autograft as our front line PLIF procedure. Utilizing this technique, we began to find similar stability rates and clinical outcomes compared to our previous standard posterolateral fusion techniques. Although the early stability rates for our "stand alone" PLIFs were the same (approximately 91% in our series) compared to previous posterolateral fusion procedures, we were concerned about the extensive disruption of the posterior tension band required to place these devices. Because of this and to increase our stability percentage, we began to supplement the stand alone PLIF patients with pedicle screw fixation. This made more sense to us from a theoretical and biomechanical standpoint because we were restoring the posterior tension band. With the addition of pedicle screw fixation we were able to achieve a
100% stability rate. Despite this, we still struggled with the difficulty of determining fusion in these patients due to metallic artifact. We also wondered what the long term implications were of implanting a device with such a different modulus of elasticity compared to native bone.

Because of this we switched to threaded cortical bone dowels that were packed with morselized autograft hoping to find a better way to determine fusion and provide patients with a more physiologic implant. We found that it was easier to determine fusion status. Long term follow up demonstrated osseous integration of the cortical bone with identical stability rates to our metallic implants with tension band constructs.

Unfortunately, during the period of utilizing cylindrical cages (both metallic and allograft bone) we also experienced our highest number of nerve root injuries and CSF leaks. The problem with cylindrical devices is that they are as wide as they are tall. To achieve optimal distraction and increase joint stiffness, a prerequisite for successful fusion, large cylinders had to be placed. This required extensive dural and nerve root retraction and caused our subsequent morbidities.

The introduction of rectangular cages added a significant advantage to PLIF. Rectangular implants are considerably easier to navigate through the window of the superior and inferior nerve roots because of the more slender width in the sagittal plane.
Although width is decreased, height is not sacrificed. Therefore the same distractive force achieved with large cylindrical devices is replicated with rectangular cages without the additional nerve root retraction. We were finally able to achieve the desired disc space distraction without the associated morbidity.

The most commonly used rectangular devices include carbon fiber, titanium, and allograft bone. The vast majority of metallic rectangular devices are utilized in Europe. We prefer using allograft bone implants because they offer structural stability under compression with a similar modulus of elasticity to the vertebral bodies. They also do not cause artifact allowing for better assessment of fusion, and they can undergo osseous integration. Moreover, the allograft implants are lordotically shaped in the axial plane, allowing the surgeon to restore lumbar lordosis at the diseased segment. In addition to the placement of the allograft implant we always supplement this construct with morselized autologous iliac crest graft within the interspace between the implants and pedicle screw fixation.

Technique

Despite the technology and instrumentation that has made this procedure considerably easier to perform than years ago, PLIF is still a difficult and technically demanding procedure. Only expeditious completion of each step of the procedure coupled with meticulous hemostasis will result in technical success with minimal morbidity. We have come to order the sequence of this operation in such a way that formalizes the teaching of the procedure and allows for a logical streamlined approach to completing the task at hand. Rarely do we deviate from this sequence.

The Procedure

It has our belief that a wide exposure with a generous skin incision and deep muscle relaxing incisions offers the best visualization for the surgeon and least postoperative pain for the patients. We routinely utilize "one twitch' paralysis to improve muscle relaxation. Adequate muscle relaxation coupled with wide exposure decreases the paraspinous muscular injury incurred with vigorous retraction. Prior to skin incision the proper size of the implant is determined by templating an adjacent normal level to determine optimal distraction. Pedicle screw length is also determined by measuring the length of the pedicles on axial CT scans. This information is conveyed to the scrub nurse before beginning to minimize intraoperative delays.

Step 1: Harvest autologous bone

At the beginning of the procedure after the midline incision is made we dissect a suprafacial plane through the same lumbar incision to identify the posterior superior iliac spine (PSIS). The fascia is then opened over the PSIS and the area is dissected in a subperiosteal fashion to expose the entire roof of the PSTS. The gluteus is not dissected off of the ileum when exposing the PSTS, as this results in significantly increased post–operative graft site pain. Using an osteotome the cortical bone of the PSIS is chipped off to expose the cancellous undersurface. A large bone gouge is utilized to harvest the cancellous bone from the ileum. Using the gouge in a rotating fashion, the cancellous bone can be removed without violating the cortex, and the bone can be undermined within the intact pelvic wing. Usually in an adult
on. The bone approximately 20–30 ccs of autologous graft can be harvested from this region. This then morselized and stored for use later in the procedure, We have found that this technique, first described by Cloward as a PUCA style graft, yields high quality bone graft with less graft site pain compared to other grafting procedures.

The graft site is then irrigated with antibiotic irrigation and packed with gelfoam. The facial opening is then closed in a watertight fashion and the dead space created by the suprafascial approach is also closed. Although the graft can be taken at any time in the procedure, we have found that harvesting graft first, before any epidural dissection, will minimize blood loss later in the procedure.

Step 2: Obtain Wide Bony Exposure

After the bone graft has been harvested the midline incision is deepened and the spinous processes are identified. In a subperiosteal fashion the muscle is dissected laterally to the base of the transverse processes. The pars intra–articularis must be identified and clearly dissected from the soft tissues. The intersection of the pars intra–articularis and the base of the transverse process will mark the pedicle entry site. The pars intra–articularis will also mark the region of the exitino, nerve root. Because we supplement all PLIFs with pedicle screw tension band constructs we prefer to mark our preliminary pedicle entry sites before the bony decompression. These preliminary sites will later be reconfirmed by fluoroscopy and by direct palpation of the pedicle from within the canal.

Before proceeding to the next step the base of the transverse process and the pars intra–articularis of the superior and inferior vertebrae to be fused should be well visualized. The surgeon should have a "mind's eye" view of where the exiting and traversing nerve roots are under the bone, before the decompression.

Step 3: Perform the Decompression

At this point the spinous processes and lamina are removed to decompress the spinal canal and to give the surgeon an intra–canal
view of the pedicies. Although many stylized sketches of a minimal laminectomy defect have been propagated by various instrumentation companies, we have found that a complete laminectomy of the superior and inferior vertebrae is the best way to ensure adequate visualization of the nerve foots. We prefer to use a high–speed drill to thin the posterior elements and the lateral bony overhang. This tends to minimize blood loss during this portion of the procedure, and eases the bony removal with the Kerrison
rongeur. After the lamina have been removed we utilize a bone chisel to remove the medial portion of the facet joint. We have found chisels to be especially effective at removing the medial portion of the superior facet of the inferior vertebrae. Another approach we have used is to create an iatrogenic pars defect through the superior vertebrae. This disarticulates the inferior facet that can then be removed completely exposing the superior facet of the inferior vertebrae. During the removal of the lateral bony wall we try to rongeur only bone and leave the ligamentum flavum in the lateral for a later point in the dissection. This removal of bone without the ligament again minimizes blood loss from the epidural plexus during this step of the procedure.

Step 4: Follow the Nerve Roots

After a wide decompression and facetectomy is performed the pedicles of the superior and inferior vertebrae can be visualized and palpated. Using a Woodson or dental dissector the course of the nerve should be followed out laterally into the foramen. Then a generous foraminotomy is performed to approximately one centimeter lateral to the origin of the nerve root at the superior and inferior level. This bony removal allows for adequate exposure of the true axilla of the superior nerve root to be performed in the next step. Moreover, this generous foraminotomy ensures adequate decompression of any symptomatic nerve roots.

Step 5: Find the True Axilla of the Superior Nerve Root

After completing the foraminotomy it is important to remove the lateral bony overhang comprised by the inferior facet of the superior vertebrae that sometimes interferes with placement interbody devices. After this is removed the epidural veins within the lateral gutter are coagulated and cut and the remaining ligamentum flavum is removed. The coagulation of epidural veins and removal of ligamentum flavum should only be performed after both the superior and inferior nerve roots are well visualized. Utilizing meticulous bipolar electrocautery to the veins located in the axillia of the superior nerve root, the surgeon can then divide the tethering venous plexus to expose the true axilla of the nerve root. This dissection affords the surgeon a slightly larger portal of entry for the interbody prothesis, gives more laxity to the superior nerve root during dural retraction, and minimizes the chances for dural tears or nerve root injury.

Step 6: Place the Pedicle Screws

At this point in the operation the surgeon can easily construct a three dimensional mind's eye view of the size, diameter, and can trajectory of the pedicles. The preliminary pedicle entry sites, are checked and verified with c–arm fluoroscopy. At this point necessary adjustments can be made and the pedicles are then probed, tapped, and screwed under fluoroscopic guidance. We prefer to use the TSRH instrumentation system. When tapping the pedicle it is important to remember that if the entry site has been chosen correctly then the tap will almost guide itself down the pedicle. No downward pressure needs to be at)t)lied to the tap for it to enter the pedicle. After tapping, the screw is placed and should easily follow the tapped channel. The pedicle screw depth can be estimated by lateral fluoroscopy and should incorporate approximately 80% of the vertebral body. It has been our preference to place the pedicle screws prior to performing the PLIF. The reason for this is that if a dural tear is encountered during the PLIF portion of the procedure the epidural bleeding can be difficult to control due to the loss of the tamponade of the thecal sac turgor. This epidural bleeding can be significant enough to prohibit placement of the interbody device. If the dural tear is encountered prior to placement of pedicle screws then the epidural bleeding will continue during the fifteen minutes of pedicle screw placement time.

Step 7. Perform a generous disectomy

After placement of the pedicle screws and generous debulking and coagulation of the lateral recess epidural plexus and ligamentum flavum a generous discectomy is performed on both sides of the thecal sac. We generally perform the discectomy utilizing a standard annulotomv, curretage and pituitary rongeur removal of the neucleus pulposis. Although the PLIF instruments perform a significant discectomy with the reamers and scrapers, we have found that these instruments' effectiveness are enhanced if the disc has been debulked. A standard discectomy does not add significant time to the procedure, and augments the surface area for fusion. Moreover, particularly with the newer Tangent system, inadequate discectomy can occasionally lead to the distracter plugs kicking out after initial placement.

Step 8. Perform the PLIF

At this point we proceed with the posterior lumbar interbody technique. If is our preference to utilize the Tangent instrumentation system and place rectangular lordotically shaped cortical allograft bone implants. It has been our experience that this system offers the best distraction and structural support with the least amount of nine root retraction. There are numerous other instrumentation systems available and the details of their implementation can be obtained from the manufacturer's literature. For the purposes of this communication, we will address only the Tangent system method.

After a generous standard discectomy is performed distractors are hammered into place under direct fluoroscopic guidance. These distracters are then sequentially dialed up to larger sizes until the desired distraction is achieved. After the desired distraction is obtained, one distracter is left in place on one side of the thecal sac. On the contralateral side of the thecal sac a rotating cutter is then hammered into position under fluoroscopic guidance. The guide is typically hammered to a depth of 30 – 40 mm depending on the size of the vertebral body. Because the rotator cutter is flanged in the middle it is important to rotate the instrument several times as it is hammered down. This prevents the rotator cutter from being too difficult to rotate at its peak depth. Next the remaining loose disc material is removed from the interspace. An endplate scraper is then placed within the disc space and the endplates are denuded of their cartilage. We utilize short stroking movements with the endplate scraper to minimize pullout of the instrument.

After the endplate has been prepared a channel cutter is then placed at the dorsal aspect of the interspace. Utilizing fluoroscopic guidance, the ideal entry and trajectory of the channel cutter is approximated to cut a channel of bone approximately 1–2mm into the endplate of both the superior and inferior vertebral bodies. The trajectory of this instrument is checked several times as it is hammered into the interspace. If the trajectory veers off the ideal pathway the channel cutter is backed out and a new trajectory is taken. This is important because if a channel is not cut in the ideal position the graft can be placed partially into the vertebral body. This is undesirable because the distractive force of the graft will be lsot and it will not be aligned with the contralateral graft. After the channel has been cut, a pituitary rongeir again is used to remove any loose debris within the interspace. At this point the pre–chosen allograft implant is then tamped into place. Usually we countersink the graft approximately 2mms.

Once the first graft is placed attention is then turned to the contralateral side where the, rotator cutter, endplate scrapper, channel cutter sequence is repeated. After the channel is cut and the remaining interspace debris is removed we then pack the disc space with approximately 10 ccs of morselized autologous bone graft that was harvested earilier in the operation. Using the distracters and a bone spatula this bone is packed within the interspace and medially towards the contralateral allograft implant. Then, the second allograft implant is tamped and countersunk into position.

The wound is thoroughly irrigated and the remaining bone surfaces are decorticated. The rod connector construct is then placed onto the pedicle screws and tightened under some mild compression. Because these grafts are only countersunk 1–2mms the implants sit within the joint's instantaneous axis of rotation. We feel that gentle compression of the graft increases the chances for osseous union because the graft is placed under stress, a necessary component for the bone growth (Wolff's Law). Also because a posterior tension band construct is created with the "locking down" of the posterior elements we have found our patients to have less postoperative pain that patients undergoing stand alone devices. Because the interbody fusion is performed in conjunction with pedicle screw fixation we do not perform a traditional posterolateral fusion with muscle dissection out to the tips of the transverse processes. Any remaining bone is important to check all exposed nerve roots to verify adequate decompression and make sure that loose autograft has not accidentally fallen into the foramen. After this step, a large drain is placed and the wound is closed in three layers with a watertight fascial closure.

Summary

The treatment of low back pain has undergone, a revolution in the last twenty years. Newer techniques and better instrumentation systems have allowed spine, surgeons to do today what previously was impossible. PLIF, is an excellent example in the medical literature of an idea that came before its time. Some of the earliest descriptions of PLIF, once abandoned by surgeons as dangerous, high morbidity operations, are now well accepted techniques that are strikingly similar to their original descriptions. It would behoove us to think more about the ideas of our predecessors.