Screws, Cages or Both?
Threaded titanium interbody fusion cages were developed as a standalone device to augment arthrodesis through an anterior or posterior lumbar interbody approach. Their introduction in 1996 caused resurgence in lumbar interbody fusion for degenerative disorders with over 33 thousand procedures performed in 1997, the first full year after release. This increased revenues by $111 million to the spine implant industry, propelling them as the fastest growing segment of the $9 billion worldwide orthopaedic market. This fastpaced growth continued with 58 thousand cage procedures performed in 1998 generating $182 million for the $950 million worldwide spine implant market. (Merrill Lynch: personal communication, August, 1999)
These threaded titanium cages trace their roots to veterinary medicine. During the mid 1970s and early 1980s, Bagby and colleagues' began treating "Wobbler Syndrome", a chronic cervical instability causing myelopathy in thoroughbred horses, by means of a smooth, stainless steel fenestrated cylinder (Bagby Basket) placed through an anterior approach. The standard Cloward technique had resulted in unacceptable morbidity due to the necessity of autogenous iliac bone graft harvest (2). Bagby eliminated the need for autograft harvest by packing his cage with cancellous bone chips obtained from the reaming of the cervical decompression. This novel device was designed with perforations in its walls to allow bone ingrowth and enhance arthrodesis. They coined the term "distractioncompression stabilization", referring to their technique of distraction of the cervical interspace with this implant, achieving early stability while improving arthrodesis. Animal studies demonstrated excellent clinical results, particularly in comparison to previous techniques utilizing interbody allografts or xenografts (2,3,4), with up to 88% fusion success'. This standalone interbody fusion technique continued to evolve with material changes and the design of threaded cages to increase stability and decrease displacement rates.(56) Similar to the method of Wiltberger(7), bilateral, parallel implants were designed for use in the lumbar spine. This ultimately resulted in the current Bagby and Kuslich design (BAK, SpineTech, Minneapolis, MN), with the first human implantation occurring in 19928. This cylindrical titanium cage has threads to screw into the endplates, thereby stabilizing the device and allowing for increased fusion rate with a standalone anterior device. Ray (6) developed a similar titanium interbody fusion device (Ray TFC, Surgical Dynamics, Norwalk, CT) which was initially used in posterior lumbar interbody fusions (PLIF), but expanded to include anterior lumbar interbody fusion (ALIF) procedures. In 1985, OteroVich (5) reported using threaded bone dowels for anterior cervical arthrodesis, and femoral ring allograft bone has subsequently been fashioned into cylindrical threaded dowels for lumbar application.
Currently, there are a wide number of available interbody fusion devices of varying design and material, not all of which have gained Food and Drug Administration (FDA) approval in the setting of a standalone device.
1) Cylindrical threaded titanium interbody cages (BAK, SpineTech, Minneapolis, MN), (RTFC, Surgical Dynamics, Norwalk, CT), and (inter Fix, Sofamor Danek Group, Memphis, TN)
2) Cylindrical threaded cortical bone dowels (MD II, MD III, MD IV) (Sofamor Danek Group, Memphis, TN)
Figure 1: Inter Fix threaded interbody fusion device. (Sofamor Danek Group, Memphis, TN) 2) Cylindrical threaded cortical bone dowels (MD II, MD III, MD IV) (Sofamor Danek Group, Memphis, TN)
Figure 2: MD 11 (left) and MD IV (right) threaded bone dowels. (Sofamor Danek Group, Memphis, TN)
3) Vertical interbody rings or boxes (Harms titaniummesh cage, DePuyAcromed, Cleveland, OH), (Brantigan carbon fiber cages, DePuyAcromed, Cleveland, OH), and (Femoral Ring Allograft FRA Spacer, Synthes, Paoli, PA).
Steffee (9) popularized posterior pedicle screw internal fixation in North America for augmentation of a posterolateral lumbar fusion. Zdeblick (10) demonstrated a high fusion rate with a standalone rigid posterior pedicle screw and rod device for degenerative lumbar disorders. Spinal fusion has become a widely used option in the treatment of degenerative conditions of the lumbar spine. Posterior, posterolateral, and interbody fusions, both anterior and posterior, have been used successfully alone or in combination. Although interbody fusion cages and pedicle screw devices have enjoyed some success as standalone devices, universal acceptance has not occurred for either strategy in the setting of degenerative lumbar disorders. This chapter will review the important issues and controversies regarding the appropriate use of standalone interbody cages (implanted through an ALIF or PLIF approach), standalone posterior screw constructs, and combined interbody cage and posterior screw techniques.
Anterior Lumbar Interbody Fusion
The earliest reports of anterior interbody arthrodesis were in association with the treatment of tuberculosis and lumbar spondylolisthesis (11, 12, 13). Initially they were transperitoneal approaches (14), and later, retroperitoneal approaches were developed (11,12,15). The first description of an anterior transperitoneal approach occurred in 1906 by Mueller, (14) with lwahara (15) reporting the first lumbar arthrodesis performed through a retroperitoneal approach. In 1948 Lane and Moore (16) in a classic description, were the first to report anterior lumbar interbody fusion (ALIF) for the treatment of lumbar degenerative disc disease. In 1950, Harmon (17) described a retroperitoneal transabdominal approach for cases of acute intervertebral disc prolapse caused by disc degeneration. Capener (18) considered fusion of the lumbar spine by an anterior approach biomechanically ideal but technically impossible in 1932, however, over the ensuing decades surgical technical advances allowed anterior lumbar interbody fusion to become a common procedure. The anterior approach to the lumbar spine was increasingly utilized in the management of a variety of spinal pathologies, using a number of different grafting materials, including corticocancellous blocks (19,20), corticocancellous dowels (21,22), and femoral ring allografts (23). Hodgson (19,20) pioneered the anterior approach for spinal tuberculosis using corticocancellous blocks. Cylindrically shaped corticocancellous dowels were first used for an anterior lumbar fusion in 1963 by Harmon (21) and 1965 by Sacks (22). Ralph Cloward (24,25,26) pioneered the dowel technique. While he utilized a posterior approach, his methods for disc removal, endplate preparation and grafting came to be used extensively. Later, Henry Crock (27) adapted Cloward's dowel technique for use with an anterior approach to the lumbar spine using cylindrical allograft. O'Brien (23) devised a hybrid interbody graft using a biological fusion cage (femoral cortical allograft ring) packed with autogenous cancellous bone graft. The concept of this hybrid is that the femoral allograft ring provides the acute stability of the construct, while the autogenous iliac crest graft provides for longterm stability. Although the technical feat of exposing the anterior lumbar spine safely was reliable in the 1970s 1980s, standalone anterior lumbar interbody fusion fell out of favor due to low fusion rates. Despite initial reports encompassing a heterogeneous group of patients and surgical techniques indicating fusion rates of 95% by Harmon (21), 70% by Hoover (28), 90% by Crock (27), and 96% by Fujimaki (29), other reports demonstrated significantly poorer fusion rates. Calandruccio (30), Nisbet (3l), Raney (32), and Flynn (33) respectively cited fusion rates of 19%, 40%, 45%, and 56%, but the 1972 study from the Mayo Clinic authored by Stauffer and Coventry (34) drove the final nail in the coffin of standalone ALIF. They reported on 83 patients who had an anterior lumbar interbody arthrodesis without instrumentation between 1959 and 1967. They found an alarmingly low success rate with pseudarthrosis occurring in a discouraging 44%, and concluded that the only justification for this procedure was as salvage for failed posterolateral fusions. These outcomes resulted in a reassessment of ALIFs as a standalone procedure and a gradual decline in its popularity, particularly for the indication of lumbar degenerative disc disease and lumbar axial back pain.
Posterior Lumbar lnterbody Fusion
Cloward (35) popularized the PLIF procedure and advocated its use after excision of a ruptured intervertebral disc. Lin (36) modified Cloward's technique and in 1977, reported on 75 cases with a fusion rate of 94%. Lin acknowledged the lack of acceptance of this technique despite over 3 decades after Cloward's popularization. Lin believed the fear of technical difficulties prohibited general use of this technique including epidural bleeding, nerve root trauma, and cerebrospinal fluid leak. Instability due to destruction of the facet joints with bone graft extrusion and subsequent neurologic deficits occurred. Nonunion with collapse of the bone graft and recurrent segmental stenosis were reported. Lin advocated meticulous control of the epidural vessels, preserving the integrity of the facet joints through a more limited interiaminar approach, and perforation of the cortical endplate to allow punctate bleeding to the fusion bed. Steffee (37) described profound instability after a PLIF procedure resulting from the extensive posterior element removal and subsequent graft collapse, displacement and resorption, nonunion, and nonrelief of pain.
Interbody Fusion Combined with Pedicle Screws
Standalone anterior and posterior interbody fusion fell out of favor due to low fusion rates. In the late 1980's, one technique (pedicle screw plating) revived interest in both types of interbody techniques. The combination of anterior interbody fusion with a posterior fusion technique was developed with the aim of obtaining higher rates of fusion and improved outcome. Because of the significant drop in the fusion rate especially over multiple levels, combined anterior interbody fusion with posterior fusion and internal fixation 38 became common. Steffee (37) repopularized the PLIF procedure due to failures encountered with posterior standalone pedicle screw constructs especially over multiple levels. This "PLIF and Plates" technique significantly increased the success rate over either procedure performed in isolation. Although conceptually the fixation of the spine with a posterior pedicle screw device is in a neutralization mode which neutralizes torsional, shear, and bending forces, in reality, the posterior plate fixation acts similar to the mechanics of a long bone fracture with an "uncompleted tension band". The spine clearly has a compression side and a tension side in the erect loaded spinal column. The compression side of the spine is a composite structure that has an elastic disc component. Implants placed across the intact disc space from posterior behave similarly to a long bone fracture plated on the tension side without continuity of the compression side. This loads the platescrew construct with a flexion moment. This is less stable than a construct loaded in tension with a stable compression side. This concept is based on an engineering principal first applied surgically by Pauwels (39). Theoretically, any posterior spinal plating with intact discs will always have a bending moment at the screwplate interface. The "spinaltensionband" can be completed only by the use of an anterior or posterior interbody fusion, removing the elastic disc and replacing it with bone. This significantly increases the rigidity of the construct and allows a greater surface area for fusion to occur. The advantage of a very high fusion rate with these circumferential (360 degrees) procedures, however, must be balanced against the increased risk of morbidity related to the increased magnitude of the procedure.
Posterior Pedicle Screws
Standalone pedicular instrumentation has increased the fusion rate in degenerative lumbar conditions (40,41,42). Zdeblick (10) showed, in a wellconceived prospective, randomized study, that a rigid pedicle screw/rod construct statistically significantly increased the fusion rate (95%) of degenerative lumbar conditions compared with noninstrumented (65%) or a semirigid pedicle screw/plate construct (77%). Unfortunately, simply exposing the posterior lumbar spine can result in profound paraspinal muscle damage with postoperative muscle fibrosis, as well as muscle and facet joint denervation. This posterior fusion disease" causes severe damage to the posterior spinal musculature, not only by the direct dissection but also by the denervation that must inevitably occur as the result of the destruction of its nerve supply during the exposure.
Posterior lumbar muscles are injured after posterior lumbar spine surgery, as demonstrated by findings on histology, computed tomography, and magnetic resonance imaging. Mayer (43) found weakness in paraspinal muscle strength with atrophy detected by measuring crosssectional area and density on postoperative CT scans 3 months after posterior lumbar surgery. These pathologic changes likely contribute to poor clinical outcome. Alterations in electromyographic activity have been documented up to 4 years after surgery (44). Macnab (45) reported that denervation of paravertebral muscles occurred in 96% of 113 patients who underwent posterior lumbar surgery based on results of an electromyographic study. Denervation potentials were demonstrated within 1 year after surgery. Degeneration of the back muscle occurs just after surgery and the muscle in most reoperated patients shows severe histologic damage, including denervation, reinnervation, and early aging. Sihvonen (46) demonstrated CT and EMG abnormalities and correlated these with postoperative failed back syndrome.
External compression by a retractor increases the intramuscular pressure and decreases local muscle blood flow. The pathologic condition of the back muscle beneath the retractor blade is similar to that of skeletal muscle beneath a tourniquet. Metabolic changes and microvascular abnormalities occur. A pathogenic mechanism for the muscle injury is based on compression and ischemia of the affected muscle. Two hours of continuous retraction caused significant histologic changes and neurogenic damage including degeneration of the neuromuscular junction and atrophy of the muscle (47). In an animal model, muscle injury after surgery was related to the retraction time and the pressure load generated by the retractor4a. Posterior surgical intervention to the lumbar spine always produces a risk of back muscle injury. Degeneration of the multifidus muscle (49) was found after surgery and human back muscle in patients who underwent repeat surgery showed severe neurogenic damage (50). This muscle injury after posterior surgery might cause postoperative low back pain and compromise the functional integrity of the muscle (51). Rantanen et al (51) also found selective type 2 muscle fiber atrophy and pathologic structural changes in the back muscles of the patients who had severe handicap after posterior lumbar surgery. The medial branch of the dorsal primary ramus, which courses around the superior articular process, innervates the muitifidus. The medial branch sits in a groove between the mammary process and the accessory process, and retraction of the multifidus lateral to the midpoint of the facet joint stretches the nerve. This dorsal (posterior) ramus is damaged by posterior lumbar procedures (52). The nerve root has no perineurium and is only covered by a thin root sheath (53). Moreover, the nerve root has a poorly developed vascular network54 compared to peripheral nerves; thus nerve root compression induces structural change more readily than occurs with peripheral nerve.
Furthermore, posterior lumbar fusions have been associated with an increased incidence of adjacent level degeneration (transitional syndrome). Lehmann55 reported a 30% rate of stenosis above a posterior fusion with an average followup of 21 years. Aota (56) found a 25% incidence of postfusion instability after posterior pedicle screw fusion with CotrelDubousset instrumentation in 65 patients with lumbar degenerative disorders. Conversely, an ALIF procedure does not significantly alter the rate of development of adjacent level degenerative changes over that of natural history (57). Fraser (57) found better outcomes are obtained after anterior interbody fusion than after posterolateral fusion with internal fixation, despite a higher fusion rate in the latter group. Late spinal stenosis adjacent to a fusion is more likely to occur with posterior fusion procedures than with anterior fusion alone. A posterolateral fusion carries the distinct disadvantage of causing damage to important stabilizing muscles and damage to the nerve supply of these muscles, in itself a possible mechanism for continuing pain and loss of function. In one study with 16year followup after ALIF, the rate of adjacent level degenerative changes was similar to an agematched control population (58). Luk et al. (59) found no increased compensatory motion in the transition zone immediately above an ALIF. Penta et al. (60) concluded that the rate of degenerative changes adjacent to an ALIF at 10 years, as assessed by MRI, was not significantly increased.
The advantages of anterior lumbar fusion in comparison to posterior lumbar interbody fusion are many, including ease of dissection, reduced operative time and blood loss, noninterference with the potentially painful posterior elements of the lumbar spine, and avoidance of scarring within the spinal canal. In addition, the disc can be resected in its entirety, advantageous from a structural and biochemical perspective. Pain from a degenerative disc can remain despite a solid posterolateral fusion, which is resolved with an anterior discectomy and fusions (61).
In an independent review (62) of a prospective comparative series of anterior interbody fusions and posterolateral fusions with pedicle screw and plate fixation, ALIFs did better despite a lower fusion rate. Although the fusion rate for anterior interbody fusion was less than that for posterolateral fusion with internal fixation, there was no difference in the subjective opinion of fusion between the two groups. Patients treated with ALIF were statistically significantly better in regards to functional outcome as assessed by the Low Back Outcome Score. One surgeon performed all procedures, and there was a minimum followup period of 2 years. Posterolateral lumbar fusion with pedicle screw instrumentation (135 patients) was compared to a group of 151 patients who underwent anterior lumbar interbody fusion. The improved outcome in the anterior fusion group, despite the higher pseudarthrosis rate, supports the concept that part of the benefit with anterior fusion is removal of the pain source itself. Another possible explanation is that some of the patients' continuing pain and disability is related to the effects of posterior surgery on the spinal musculature and the presence of a rigid pedicle screw fixation system.
lnterbody fusion cages are classified by their structure (geometry) and by material. Horizontal cylinders, vertical rings, and open boxes are standard designs. Cages can be made of metal, carbon fiber, or allograft bone. The seductive expectation of titanium threaded lumbar interbody fusion cages when released in 1996 was the ability to reliably and safely perform an interbody fusion without the need of pedicle screw augmentation. Previous attempts of standalone ALIF and PLIF were not universally successful, required pedicle screw stabilization for reliability, and suffered the complications of posterior pedicle screw fusion as detailed in the previous section. Some clinical studies have verified this expectation while others are less optimistic.
Many authors have reported excellent clinical results with the use of threaded cylindrical devices for ALIF. In a prospective, multicenter trial of the BAK device, Kuslich et al (8) reviewed 947 patients. An anterior approach was used in 591 operations with 93% obtaining fusion at 24 months postoperatively. Pain was eliminated or reduced in 84%. Function was improved in 91 %. Major complications occurred in 2%. Implant migration occurred in 1.2% with all requiring reoperation. Vessel damage or iliac vein tears (1.2%) were all repaired without apparent longterm problems. The overall rate of device related reoperation was 4.4% with most requiring additional posterior instrumentation to relieve ongoing pain. There were no instances of implant fracture or other forms of structural failure. There were no deaths, major paralyses, or deep infections. Fusion rate at 12 months after ALIF was 88.3%. At 24 months, the fusion rate increased to 93% of ALIF procedures and at 3 years after surgery (118 patients), 98.3% of patients had fused operative segments.
Blumenthal et al (63) also found a low revision surgery rate (3.3%) among their series of 130 consecutive standalone open and laparoscopic threaded interbody cage patients. However, in a prospective nonrandomized study of 51 standalone open and laparoscopic BAK patients, O'Dowd et al (64) recently reported an overall failure rate requiring revision of 31% due to clinical failures at a mean of 15 months. Furthermore, 75% of their patients had residual symptoms at 2 years postoperatively and 47% had the same or poor selfassessment. The authors believe the unacceptable failure rate and poor clinical results were due to use of the cage as a standalone device. Based on these findings, in order to avoid such poor outcomes, the authors recommend supplemental posterior stabilization for all threaded interbody ALIF patients. (64,65)
In the only direct comparison of threaded bone dowels and titanium cages to date, 100 anterior interbody fusion patients were randomized in a prospective study comparing titanium interbody cages (BAK) with threaded cortical bone dowels (MD II). (66) At 12 months, there were no significant differences in clinical outcome or radiographic evaluations.
Ray (6) reported a large series (236) treated with his threaded cages, however, through a posterior approach. Of 208 followed for a minimum of 24 months, 203 (96%) had radiographic evidence of fusion. Clinical outcome as described by Prolo was excellent for 84(40%), good for 53(25%), fair for 44(21 %), and poor for 30(14%). Ray (67) also reported decreased cost with a standalone PLIF cage compared to a circumferential fusion. In a prospective, nonrandomized study, 25 patients were treated with a Ray threaded fusion cage and 25 had combined anterior and posterior arthrodesis (360degree technique) with pediclescrew instrumentation. The average combined cost (surgeon, hospital, and anesthesiologist) of onelevel procedures performed with the Ray cage was $25,171 and that for the 360degree procedures was $41,813, a difference of $16,642 (40 percent). In addition, ten patients who had the 360degree procedure later had removal of the pediclescrew instrumentation, which added $8635 to the cost for each patient. The final average cost of the 360degree procedures was $22,889 higher than that of the corresponding procedures performed with use of the Ray cage.
Complications, however, are higher with PLIF cage cases compared to ALIF standalone threaded titanium cages. Scaduto (68) found perioperative complications were 3.6 times higher in the PLIF group and major postoperative complications were 7.1 times higher. The ALIF group had less blood loss, shorter operative time and hospitalization.
The greatest strength of the vertebral body is present in the subchondral bone of the cortical endplate. The maximal endplate strength is peripheral near the ring apophysis. Two techniques of endplate preparation during interbody fusion are practiced. One involves purposeful endplate cavitation to provide an optimal bleeding bed of cancellous bone. Two cylindrical grafts are screwed into adjacent circular holes oriented parasagitally across a disc space prepared by a reamer which partially removes the subchondral bone and at the apex of the cavity, exposes weak but very vascular cancellous bone. The outer portion of the perforated, cylindrical cage has a continuous threadform that engages the adjacent vertebral bodies and endplates. This threaded design permits insertion by screwing the device into the disc space, which provides resistance to device migration and stabilization to the vertebral bodies, which facilitates spinal fusion. The intervertebral disc space is predrilled such that a hole is created which spans the entire height of the disc and includes semicircular concavities in the vertebral bodies above and below the disc space. The cages, packed with autogenous bone graft, screw into the predrilled holes.
The second technique of endplate preparation involves preservation of the subchondral bone. The advantage of preserving much of the endplate and filling the disc space with a greater quantity of bone graft should reduce the risk of graft collapse and increase the fusion rate. This leaves the strongest bone adjacent to the implanted graft and requires a precisely cut graft to exactly match the interspace. The disadvantage, however, is that the endplate is minimally vascularized and the recipient bed is less vascular. Technically, the dowel technique is easier to perform and consistently allows accurate fitting of the cage to the prepared graft bed. By reaming a cavity, the recipient bed is reliably created for the cylinders. The disadvantage with this technique is that the strong trabeculae adjacent to the endplates are breached, increasing the risk of graft settling. The second technique is much more difficult because a perfect fit between host and graft is mandatory and the graft must be perfectly cut to match the subchondral bone surfaces. The accurate insertion of individual blocks is less reproducible than with the dowel technique.
During daily activity, the lumbar spine is exposed to significant biomechanical forces. Studies indicate that a motion segment may experience axial compressive loads ranging from 400 N during quiet standing to more than 7000 N during heavy lifting (69, 70). The ultimate compressive strength of a nonosteoporotic vertebral body has been reported to be slightly over 10,000 N. (71) Corticocancellous autografts demonstrate inadequate initial mechanical strength for lumbar interbody loading, often leading to collapse or extrusion. (5,6,72) The compressive forces across the grafted interspace should be less than that required for failure of the graft construct. The graft should be able to transmit force without significant motion so that immediate mechanical load transfer is achieved, and the technique should induce arthrodesis as quickly as possible with minimal to no morbidity associated with its use. Threaded titanium alloy (Ti6A14V) interbody fusion cages have undergone extensive in vitro and in vivo biomechanical testing, demonstrating rigidity sufficient to withstand lumbar spinal loading forces without fracture or deformation (8,73,74,75). Longterm clinical studies have reported no cases of structural cage failures (6,8) and cages have been shown to impart increased stiffness as compared to the intact spine (75,76,77,78).
One study comparing biomechanical stability performance among three interbody devices (Threaded Bone Dowels, BAK, and RTFC) found no significant difference under physiological loading conditions. (79) Bone dowels performed as well as titanium cages. In flexion, bone dowels increased stiffness by 970%, Ray increased by 253% and BAK increased by 96%. Under extension, Bone dowels increased local stiffness by 166%, BAK 71 %, and Ray 56%. In torsion, bone dowels increased intact stiffness by 20%; BAK also increased global intact stiffness by 20%, while Ray decreased intact stiffness by 5%. Bone dowels increased intact stiffness by 91 % under lateral bending, while Ray cages and BAK cages maintained the intact stiffness under lateral bending. None of the implants fractured during failure tests. The vertebral endplate and the sacroiliac joint were found to be the most common failure sites. All devices withstood load to failure, with the vertebral body end plate failing before the implant (79). Threaded cortical bone dowels (Sofamor Danek MD II and MD III) provide an increase in construct stiffness of 68334% over the intact motion segment. (80) Additionally, the threaded cortical bone dowels demonstrated static compressive strengths of over 24,000 N, well above maximal physiologic loads.
Several biomechanical studies have shown that these threaded anterior interbody devices improve overall stiffness, but are least rigid in extension and axial rotation. (65,75,81,82) Oxland et al (65) compared the stability of a traditionally paired anterior implantation with that of a lateral implantation technique (preserving the ALL and anterior annulus). The purpose was to test whether this decreased extension rigidity was due to resection of the anterior longitudinal ligament (ALL) and anterior annulus during cage insertion. They found no significant improvement in extension stability with lateral insertion, leading them to conclude this lack of rigidity was associated with distraction of the facet joints after interbody cage placement. Additional posterior instrumentation can provide the added stability required in extension and axial rotation. Supplemental transiaminar facet screws that can be placed in a minimal invasive fashion significantly reduce the motion of a BAK biomechanical model in extension and axial rotation. Rathonyi and associates (83) found that using transiaminar screw fixation can substantially stabilize the problematic loading directions of extension and axial rotation. Volkman84 also demonstrated in a cadaveric biomechanical model that motion segment stiffness of an anteriorly placed threaded spine cage was increased, especially in extension, with transfacet screws.
The biomechanics of posterior lumbar interbody fusion was well described by Brodke, et al (76) in a calf spine model. They found the PLIF approach is a destabilizing procedure. The PLIF with bone graft construct was less stiff than the intact spine and also the destabilized spine (which had removal of the facet joints.) In two of the eight specimens, the bone graft dislodged posteriorly into the canal during torsional testing. The PLIF threaded titanium cage model was similar in flexionextension and torsional stiffness to the PLIF bone graft with pedicle screw instrumentation group. These two groups, however, were less stiff than the destabilized model with posterior pedicle screw instrumentation in flexionextension and torsional testing. This demonstrates the significant additional instability caused by removing the posterior annulus and intervertebral disc after the facet joints are destroyed.
The ideal interbody graft combines a strong mechanical construct to withstand compressive loads across the disc space while providing an osteogenic, osteoinductive, and osteoconductive matrix. The gold standard for this matrix is autogenous cancellous bone. The compressive strength of this bone, however, is very poor. Combining this with a strong titanium or cortical allograft shell (cage) is sensible. The cancellous autograft iliac crest bone is packed into the cylindrical screwin cages with the goal that the cage provides mechanical strength to prevent collapse, subsidence, shear, and torsional forces. This produces an optimum stable environment while the autogenous graft grows through the cages into the vertebral bodies above and below. The mechanical strength of the cage is combined with the biologic strength of the autograft. This graft, however, is not biomechanically loaded while it is inside the cage, and the surface area available for the graft to grow through the cage is not large and varies between cage types. Optimally, graft should be packed around and between the cages to maximize the surface area of bone available for fusion and to allow bone graft to undergo physiologic loading. Maximally packing the interspace with bone graft also ensures removal of all disc material and cartilaginous endplate that is avascular and inhibits fusion. With this concept in mind, performing a subtotal "channel discectomy" (only removing a cylindrical channel of disk material using a drill) that occurs in the laparoscopic technique, is not optimal. (85) This partial "reamed channel" discectomy results in a limited fusion confined to a small crosssectional area (the fenestrations in the cage). In a prospective, randomized study (85) of BAK cages packed with autogenous iliac bone graft, a complete discectomy vs. partial reamed channel discectomy was performed in 100 patients. All 50 patients in the complete discectomy group achieved a solid arthrodesis at a mean follow up 25 months with no revision surgical procedures. In contrast, 7 patients in the partial reamed channel discectomy group had a pseudarthrosis with 8 patients required revision surgery. The difference between the groups was significant (p=0.019).
One conceptual problem associated with cylindrical interbody fusion devices, titanium cages and threaded cortical bone dowels, is their geometric shape. The volume available for bone graft in cylinders is less than that in vertical ring devices, such as the femoral ring allograft. (64) A tapered device as opposed to a cylindrical shape (75) (which has identical anterior and posterior height), better restores lordosis and sagittal balance. The segmental lordosis and wedge shaped anatomy present in the human intervertebral disc space results in nonuniform implant contact, anterior to posterior.(87) Additionally, it has been calculated that the BAK cage allowed a maximum interface with only 10% of the total surface area of the end plate (86). Some authors have concluded that the interbody bone graft area should be significantly greater than 30% of the total end plate area to prevent failure.(88) However, to increase interbody graft contact with high quality bony bed, greater amounts of subchondral bone need to be removed, increasing the risk of subsidence.(88) In a sheep invivo model, a threaded titanium interbody fusion device was compared to an anterior fusion using autogenous iliac crest dowel graft. After surgery, interbody distraction successfully occurred in cage and autograft sites. Loss of interbody height ensued in both groups during the first 2 months. Percentage loss of height was lowest in the cage sites. Both techniques effectively distracted the intervertebral spaces beyond their baseline measures. The cylindrical cages nearly doubled the normal vertical span of the disc spaces. All, however, experienced subsidence of disc height during the first 2 months. Although the absolute reduction in intervertebral height was similar between the groups, the cage sites lost a smaller fraction of their initial distraction. At final measure, the cageimplanted sites had lost 19.6% of their postoperative height but remained well above the normal disc height (82).
Unlike titanium interbody cages, threaded cortical bone dowels are subject to supply shortages and processing problems. Presently, the majority of threaded cortical allograft bone dowels are obtained through aseptic harvest techniques with subsequent processing steps occurring in "class 1O certified" clean rooms. After appropriate donor screening tests and chemical processing with hydrogen peroxide and 70% ethanol, the bone dowels are freezedried or frozen. This often avoids the necessity of terminal sterilization by highdose gamma irradiation or ethylene oxide methods that can impair the mechanical and physiologic properties of the allograft.(39) The biomechanical properties of allograft bone can be altered by the methods chosen for its preservation and storage. These effects are minimal with deepfreezing or lowlevel radiation. Freezedrying, however, markedly diminishes the torsional and bending strength of bone allografts but does not deleteriously affect the compressive or tensile strength. Irradiation of bone with more than 3.0 megarad or irradiation combined with freezedrying appears to cause a significant reduction in breaking strength.(90)
There have been no documented cases of HIV transmission from musculoskeletal allografts since 1985, although there have been over 7 million bone and soft tissue transplants performed since that time. Utilizing currentgeneration PCR screening tools, the risk of HIV transmission is estimated to be approximately one in eight million. (91)
Cages and Screws
Although originally designed as a standalone device, threaded cylindrical lumbar interbody cages may not be appropriate by themselves without additional posterior stabilization in various circumstances. In many ways, we are relearning the lesson spine surgeons of three to six decades ago realized about interbody constructs only now we are using cages rather than iliac crest bone graft. Some lumbar segments are too unstable for a standalone interbody graft, whether it is simple bone graft or a cage. The strongest argument for routine posterior screw augmentation can be made for cages inserted through a PLIF approach. The geometry or material of the cage did not matter in a comparative biomechanical study of posterior lumbar interbody fusion implants by Tsantrizos, et al.(92) They tested the Ray cage (titanium cylinder), Contact Fusion Cage (titanium box), and PLIFS (allograft trapezoid). The data clearly indicate the need for posterior instrumentation in all three of these models to achieve adequate initial segmental stability. Dimar, et al (93) found that posterior lumbar interbody cages do not augment segmental biomechanical stability in a human cadaveric model. They concluded that the use of posterior threaded interbody cages as an isolated procedure should be avoided unless supplemented with posterior instrumentation.
If a PLIF approach is used, a box or trapezoidal geometry is more reasonable than a cylindrical cage. This is due to the fact that a cylinder is as tall as it is wide and the minimal width of the construct is two times the diameter of the cage. In order to maximize distraction, grafthost contact, and ligamentotaxis, a large diameter cage is required. The size of the cage has major neurologic and facet joint (stability) implications. This intrinsic problem of the height determining the width of the cage is resolved by a trapezoidal geometry where significant distraction can be achieved with a narrow cage.
Biomechanically, the most compelling indication for posterior pedicle screw augmentation of an interbody cage placed through an ALIF approach is with spondylolisthesis. Cagli, et al.(94) evaluated the biomechanics of lumbar cages and pedicle screws for treating spondylolisthesis in a human cadaveric model. They found that cylindrical cages add only a small amount of stability to pedicle screws, but pedicle screws add a large amount of stability to cylindrical cages.
|Figure 3: Failed anterior lumbar interbody fusion with titanium threaded cages in Grade I isthmic spondylolisthesis. (Figs. 3A, 3B, 3C, 3D)|
|Fig 3A: Grade I isthmic spondylolisthesis: Preoperative lateral xray.|
|Fig 3B. lntraoperative lateral xray: Excellent placement of interbody cages with proper distraction of the interspace and reduction of the spondylolisthesis. L5 radiculopathy resolved immediately postoperatively.|
|Fig 3C: 6 month postoperative lateral xray: Collapse of interspace with cavitation of cages through subchondral endplate of L5 into cancellous vertebral body. L5 radiculopathy recurred bilaterally and back pain became severe.|
|Fig 3D: Lateral xray after salvage posterior Gill laminectomy and posterior pedicle screw instrumentation and fusion.|
The disadvantage, however, of posterior pedicle screw instrumentation has been detailed earlier. Due to the wide exposure and disruption of the juxtalevel facet joint capsule required for placement of the pedicle screws, significant problems may occur from the approach alone. A less invasive alternative for posterior stabilization is transiaminar facet screw fixation.
Devised by Magerl,(95) this technique requires a small incision with dissection only out to the facet joints. The transverse processes and cephalad juxtalevel facet joints are not exposed. Clinical studies have reported a high success rate with minimal complications (96,97,98,99). Magerl's technique is a modification of Boucher, which is a modification of King's description of facet joint screws. King'(100) in 1948 reported his operation whereby short screws are placed horizontally directly across the facet joint. The screw enters the inferior articular process just medial to the joint and crosses the joint into the ipsilateral superior articular process. In 1959, Boucherlo' described his method that uses the same starting point as King, but the screw is directed more vertical into the pedicle thereby increasing the length of the screw in the caudal vertebrae.
Magerl's screw is significantly longer because the entry point is at the base of the contralateral spinous process. This increases the effective working length of the screw on both sides of the facet joint thus increasing strength of the fixation. The anatomic angle of screw insertion and screw length 102 at the various levels in the lumbar spine has been studied for this technique and transiaminar facet screw stabilization has been successfully used after selective decompression for spinal stenosis and disc protrusion (103). Biomechanical studies have demonstrated significant stability in flexion, extension, and rotation (104).
Translaminar facet screws significantly increase the stiffness of spinal motion segments (105). When coupled with threaded cylindrical interbody fusion devices, translaminar facet screws provide substantial stability in the weakest loading directions, extension and axial rotation (83,84). lnterbody cages separate the facet surfaces with distraction, which reduces the role of the facets in extension and axial rotation (81). Translaminar facet screws stabilize this facet uncoupling caused by the interbody distraction. Translaminar facet screw technique has also been evaluated in a biomechanical model of PLIF. Zhao(106) compared the segmental stiffness of three different PLIF constructs: two posterior cages, a single long diagonally placed threaded cylindrical cage from a posterolateral position, and the single long posterolateral cage with simultaneous facet joint fixation. The two, standard PLIF cages construct was the weakest due to the need for bilateral facetectomy and posterior element destruction, which is detrimental to segmental stiffness. The single posterolateral cage technique requires only a unilateral facetectomy and conserves more of the posterior elements. As expected, this model was more stable than the twocage construct. The addition of translaminar joint fixation to the remaining facet provided significantly more stability in compression, extension, flexion, bending, and torsion. This study clearly proves the advantage of even unilateral facet stabilization, and the disadvantage of the standard PLIF approach, which results in a profound decrease in biomechanical stiffness. Extensive removal of the posterior elements is required to insert the cylindrical cages of appropriate size and kyphosis may occur when larger cages are used. Also, cauda equina retraction is necessary during insertion of these cages and may be severe with potential neurologic damage when appropriate, larger cages are employed.
In conclusion, technical and biomechanical advantages support the combination of interbody cages and least invasive posterior translaminar facet screw fixation. An ALIF approach is less damaging to the soft tissues and supporting structures of the spine than a PLIF technique for interbody fusion. Clinically, Vamvanij (107) found simultaneous ALIF with BAK cages and posterior facet fusion offered the highest fusion rate, pain relief, and clinical success compared to three other lumbar fusion techniques. Limited, posterior soft tissue dissection only to the facet joints appears to be important. lnterbody fusion cages are least able to resist extension due to distraction and restoration of disc height, which uncouples the posterior facet joints. Insertion of transfacet screws significantly increases the stiffness in an interbody cage model, especially in extension.(83, 84) Extension moments on a standalone interbody cage without posterior stabilization tends to separate the vertebral endplates from the interbody cage, potentially resulting in nonunion, loosening, or migration of the cage. Stiffness of a cage model loaded in compression is also significantly greater with the addition of facet screws (84). Thus, transiaminar facet screws should help resist collapse and subsidence of the cage as well as loss of lordosis and foraminal narrowing. In the future, this concept may be developed even further with the minimally invasive percutaneous delivery of transiaminar facet screws under realtime image guided control.