For destructive lesions, particularly tumors, the degree of vertebral body and/or pedicle involvement and the degree of deformity enter into in the determination of instability.⁵^–⁷ At present, no method of systematically assessing impending instability exists for infectious processes. Other forms of instability, such as inflammatory, congenital, degenerative, and postoperative instability, are not well assessed with grading systems such as those described by Gaines and coworkers.

Once instability is determined, surgical treatment generally requires adequate correction of deformity, reconstruction of bony defects, and fusion of the affected segments. The choice of method by which surgical correction is performed remains a challenging task.

TECHNIQUE

The method by which surgery is performed depends on the cause and character of the instability. The available methods of treatment include anterior, posterior, or combined anterior–posterior surgical approaches. General guidelines have some degree of support either in the literature or by consensus (Fig. 102.1).

Fig. 102.1 Algorithm for the surgical treatment of instability via fusion. Instability due to trauma, tumor, infection, as well as iatrogenic induced instability are discussed. Instability due to degenerative conditions such as spondylolisthesis, scoliosis, and disc disease are discussed in previous chapters.

Anterior lesions

When instability is due to a lesion that is primarily anterior, surgical stabilization can be performed from an anterior approach. This anterior approach generally involves discectomy, perhaps corpectomy, structural grafting and anterior instrumentation. Anterior surgery is indicated if there is canal compromise due to the anterior lesion, if anterior column stability is deficient, or a previous posterior procedure has failed to stabilize the segment(s). The load-sharing classification system of Gaines and coworkers is one method of determining the integrity of the anterior column.^5,⁶ The load-sharing classification system describes the integrity of the anterior column in the setting of trauma. However, this concept can be applied to other lesions of the anterior column. If there is no significant osteoporosis or deformity that requires correction, then anterior strut grafting with anterior instrumentation will provide sufficient stabilization (Fig. 102.2).

Fig. 102.2 L2 pathologic fracture due to metastatic adenocarcinoma. (A) T2-weighted sagittal MRI showing pathologic fracture of L2. (B) T1 axial image with contrast showing canal stenosis. (C) Postoperative radiographs after L2 corpectomy, anterior cage strut, and anterior instrumentation.

However, certain conditions warrant posterior surgery, either together with the anterior surgery or posterior surgery alone. Posterior surgery alone can be used to treat lesions that are otherwise anterior if the anterior column has sufficient stability and there is no need for decompression. In some cases, such as when there is no neurologic deficit, posterior-only surgery can be used to indirectly reduce retropulsed fragments of bone. Posterior surgery may be further entertained if there are anteriorly exposure risks, such as previous abdominal surgery or severe obesity. Posterior fusion may also be necessary if there is significant osteoporosis. In the setting of osteoporosis, anterior surgery alone may be inadequate in achieving spinal stability (Fig. 102.3).

Fig. 102.3 Construct failure after anterior-only treatment of a osteoporotic burst fracture. (A) Lateral radiograph immediately after vertebral corpectomy, allograft strut fusion, and anterior instrumentation. Black dotted lines highlight endplates above and below the fusion site. (B) Lateral radiograph 4 weeks after surgery. Black dotted lines highlighting the endplates show interim kyphosis above the construct. The vertebral body above has collapsed into the superior screws.

When instability stems from a posterior lesion, posterior surgery via instrumented fusion is usually the best option (Fig. 102.4). Unfortunately, instability due to a purely posterior lesion is relatively uncommon. The most common such lesion is the Chance fracture involving a flexion–distraction injury. If there is severe osteoporosis, additional levels may be added to the construct. The length of posterior instrumentation can be extended as necessary to achieve stability without much added difficulty. If necessary, pedicle screws can also be reinforced with bone cement to achieve better fixation in osteoporotic bone. Interbody fusion can be performed posteriorly via transforaminal lumbar interbody fusion (TLIF) or posterior lumbar interbody fusion (PLIF) to improve fusion rates, especially in long fusions or in fusions to the sacrum.

Fig. 102.4 Metastatic erosion of posterior elements. (A) Lateral radiograph showing posterior element bony destruction. (B) Sagittal MRI showing canal compromise. (C) Axial CT image showing bone erosion of facets, right pedicles, lamina, and spinous process. (D) Lateral radiograph after resection, decompression, and posterior instrumented fusion.

Combined anterior–posterior lesions

In combined lesions of both the anterior and posterior columns, the surgical approach is dictated by several factors. The first issue is whether anterior surgery is necessary. Anterior surgery is necessary if there is isolated or predominantly anterior canal compromise requiring decompression or if anterior column stability is compromised. Anterior surgery is usually in the form of strut or structural grafting. If there is good bone stock and no significant deformity, then anterior instrumentation may be all that is needed. If there is no canal compromise and the anterior column is stable, posterior instrumented fusion alone is sufficient to obtain stability. If there is osteoporosis, the levels of posterior fusion can be increased for improved fixation. When the anterior column is deficient and it is combined with severe deformity, osteoporosis, or infection, then anterior strut graft fusion combined with posterior instrumentation is necessary (Fig. 102.5).

Fig. 102.5 Spinal osteomyelitis and epidural abscess. (A) Contrast-enhanced sagittal MRI showing osteomyelitis of L5 and S1 along with anterior and posterior epidural abscesses. (B) Contrast-enhanced axial MRI showing circumferential infection. There are also abscesses in the right paraspinal muscles. (C) Lateral radiograph after surgical treatment. Treatment involved L5 and partial S1 laminectomies and right L5–S1 facetectomy. (D) Second stage partial L5–S1 corpectomies with cage strut graft were followed by third stage posterior instrumented fusion.

FACTORS AFFECTING THE OUTCOME OF FUSION SURGERY

Achieving arthrodesis and determining when this occurs are still problems that confront the spine surgeon. In general, spinal fusion is a nonphysiological operation that attempts to achieve bone formation in a site where this does not normally occur. In addition to the local and systemic factors that influence typical fracture healing, mechanical factors also influence the healing process during spinal fusion. These mechanical factors include the distance to be bridged by the fusion mass, the magnitude and planes of motion, and the structural properties of the graft. For convenience, the factors that affect successful spinal arthrodesis can be grouped into local, biologic, and surgical factors.

Local factors

Location

The physiologic and biomechanical forces acting on the healing of anterior interbody and posterolateral fusions are quite different. Anterior interbody fusions are revascularized through the vertebral bodies themselves and the graft material used in the interspace is under compressive loading. Bone healing is favored under conditions of axial loading and stability. As such, interbody fusion tends to occur more readily than posterolateral spinal fusion. In a posterolateral fusion, revascularization is primarily derived from the bony surface of the transverse process and surrounding muscle tissue. In this region, there are few or no compressive forces acting on the graft material. Consequently, the posterolateral spinal fusion environment generally does not tolerate the use of purely osteoconductive materials as stand-alone substitutes. At this site, osteoinductive substitutes are more likely to be successful as extenders or enhancers for spinal fusion. In the anterior spine, purely osteoconductive substitutes may be suitable when they are rigidly immobilized.

Fusion site

Arthrodesis depends on ingress of osteoprogenitor and inflammatory cells from the fusion bed and from the surviving bone cells located in the bone graft. Bone healing is greatly affected by the local blood supply along with the availability of osteoprogenitor cells. The fusion bed vascular supply is a source of supportive nutrients, a vehicle for endocrine signals, and a pathway for the recruitment of osteoprogenitor and inflammatory cells. The preparation of the bony surfaces where the graft material is to be placed also plays an important role in determining the outcome of fusion. Decortication allows vascularization of the fusion bed. In addition, it is important to remove avascular tissue such as scar tissue during surgery since a fusion bed with excessive scar tissue is less likely to achieve successful fusion.

Biologic factors

Nutrition

Malnutrition is associated with reduced cognitive function, poor wound healing, impaired muscle function, decreased bone mass, immune dysfunction, anemia, delayed recovery from surgery, and ultimately increased morbidity and mortality.⁷ Nutritional deficiencies can be confirmed using various studies such as total white blood cell count, serum albumin and total protein, transferrin levels, nitrogen balance, and anthropometric measurements.

Hormones

Thyroid hormones have a direct stimulatory effect on cartilage growth and maturation, and promote bone healing.⁸ These hormones are required for the synthesis of somatomedins by the liver.⁹ Growth hormone promotes bone healing by increasing calcium absorption from the intestine, and promoting bone formation and mineralization.⁸ Androgens and estrogens are considered important for skeletal development and preventing age-related bone loss. Estrogens may increase bone mineralization through their effect on increasing serum parathyroid hormone (PTH).¹⁰ Excess corticosteroids can negatively affect bone healing by decreasing synthesis of the major components of the bone matrix. These steroids have also been shown to inhibit the differentiation of osteoblasts from mesenchymal cells.¹¹

Smoking

The pathophysiological effects of smoking are multidimensional and include arteriolar vasoconstriction, cellular hypoxia, demineralization of bone, and delayed revascularization. Cigarette smoking also inhibits osteoblastic activity resulting in decreased rate of successful unions, with slowed bone healing and prolonged treatment.^12,¹³ It appears that cigarette smoke has a greater impact upon cancellous bone than cortical bone, and causes decreased bone healing around implants. The constituents of cigarette smoke, and not only nicotine, seem to be more important.^14,¹⁵

Complication rates after surgery and the rate of nonunion in smokers has been shown to be consistently higher than in nonsmokers.^16,¹⁷ In a recent prospective clinical study designed to determine the effects of smoking on the healing of tibial deformities, half of the smokers and only one-fifth of nonsmokers developed complications.¹⁸ Smokers required an average of 16 days more treatment. Delayed healing and pseudoarthroses were more common in smokers than nonsmokers. The risk ratio for smokers to develop complications as compared to nonsmokers was 2.5.¹⁸ For spinal fusion, an increased rate of pseudoarthrosis has been observed following posterolateral lumbar spine grafting or anterior cervical interbody fusion in smokers.¹⁹ Smoking had a significant negative impact on healing and clinical recovery after these procedures.

Drugs

Many drugs such as methotrexate and Adriamycin can inhibit bone formation if administered in the early postoperative period. Nonsteroidal antiinflammatory drugs (NSAIDs) act by inhibiting action of cyclooxygenase (COX) enzymes and decreasing production of prostaglandins (PGs). This can result in diminished bone formation, healing, and remodeling. These effects seem to be greater with nonselective COX inhibitors, such as ketorolac, than with selective COX-2 inhibitors at therapeutic levels.²⁰ Ketorolac also decreases bone mineralization and endochondral ossification in a dose-dependent manner.²¹ Other NSAIDs, such as diclofenac, can also significantly delay fracture healing.²² The timing of administration of NSAIDs also appears important. In animal studies, the earlier indometacin was resumed postoperatively, the greater was its negative effect on intertransverse process arthrodesis.²³ Recombinant bone morphogenetic protein-2 (BMP-2), however, seems to show promise for combating the inhibitory effect of ketorolac on bone formation during spinal arthrodesis.²⁴

It has been observed that longer duration of administration of selective COX inhibitors decreases the rate of successful fusion.²⁵ In a recent study, however, it was found that the perioperative (short-term) administration of celecoxib had no apparent effect on the rate of nonunuion at 1 year following surgery. In this study, patients had elective decompressive lumbar laminectomy and instrumented spinal fusion. In addition, the use of the selective COX-2-specific NSAID resulted in a significant reduction in postoperative pain and opioid use following surgery.²⁶ As such, short-term administration of selective COX inhibitors for perioperative pain, unlike long-term use, appears to be a reasonable therapeutic option.

AGE The effects of aging on spinal fusion are multifactorial. With the elderly there is often a decrease in food intake secondary to reduction in appetite. If prolonged, this can lead to a decline in nutritional status. Bone healing, including spinal fusion is affected by the nutritional status of the individual.⁷ Common medical conditions, such as renal and cardiac insufficiency, may exist in the elderly and can affect bone healing. The clinical observation that incisional wounds also heal less well in the aged may be secondary to these coexisting pathologies.²⁷

The number of osteogenic stem cells may be more deficient in the elderly than in younger individuals. In animal studies, the numbers of osteoprogenitor cells present in the bone marrow from aged rats were 65% lower than the amount found in young adults. These cells were also 10 times less likely to differentiate in vitro and to form bone in vivo.²⁸ However, bone-stimulating growth factors may have a role in increasing the bone-forming capacity of aged bone in the future.^29,³⁰

During aging, bone mass loss, structural continuity, and strength gradually diminish. Osteoporotic bone is weak and difficult to stabilize with spinal instrumentation. Although gradual bone loss starts at 30 years of age, an accelerated rate of loss occurs after menopause in women. Men also experience bone loss, but in a more insidious fashion. The rate of successful fusion in elderly individuals of both sexes with osteoporotic bone tends to be lower.³¹

Surgical factors

Approaches

The surgical approach can affect the outcome of lumbar fusion. Fritzell et al. studied the outcomes of fusion surgery using three different, well-accepted techniques of lumbar fusion.³² In their randomized prospective study, they show that posterolateral fusion without instrumentation achieves arthrodesis in 72% of patients. With the addition of posterior instrumentation, the fusion rate increased to 87%. For patients treated with combined posterior instrumented fusion and interbody fusion (i.e. circumferential fusion) the fusion rate was 91%. Overall, the fusion rates for grafts used in circumferential, posterior interbody, anterior interbody, and posterolateral fusions methods are about 91%, 89%, 86%, and 85%, respectively.¹

Instrumentation

It is well established that spinal instrumentation increases the likelihood of successful fusion. Fischgrund et al. demonstrated that the fusion rate with instrumentation was twice that obtained without instrumentation for patients with degenerative spondylolisthesis.³³ Zdeblick observed statistically greater fusion rates in patients with rigid instrumentation than those without. In his study, the fusion success with semirigid instrumentation (plate/screw system) was not different from that obtained without instrumentation.³⁴ In contrast, Thomsen et al. found that the fusion rates were not significantly different between instrumented (supplementary pedicle screw fixation) and noninstrumented groups for patients undergoing posterolateral lumbar spinal fusion.³⁵

Perhaps the best perspective regarding the impact of instrumentation on the fusion rate was revealed in a meta-analysis of the outcomes of lumbar fusion surgeries that were reported in the literature from 1970 to 2000.¹ In this study, the average instrumented fusion rate (89%) was significantly higher than the noninstrumented fusion rate (84%). The fusion rates with semirigid instrumentation and rigid instrumentation were not significantly different (91% and 88%, respectively). In addition, both semi-rigid and rigid stabilization resulted in significantly higher fusion rates than no instrumentation (p=0.019, p=0.022, respectively).

Number of levels fused

The number of levels fused is an important factor that can dramatically affect the outcome of spinal fusion. In a retrospective study, the length of hospital stay, operative time, amount of intraoperative blood loss, and transfusion requirements were closely related to the number of levels fused for patients having revision posterior lumbar spine decompression, fusion, and segmental instrumentation. The arthrodesis rate also decreased with number of levels fused.³⁶ Exact figures of fusion rates with number of levels fused vary according to the disease or condition in question, the approach, and whether the fusion was augmented with bone growth enhancers. In a meta-analysis study which assessed data obtained from fusions done over the past 20 years, the mean fusion rates were 89% in one-level operations, 69% in two-level operations, and 71% when three or more levels were fused.¹

Graft source

The choice of graft material can influence the outcome of a spinal fusion. During spinal arthrodesis, graft material may not only provide structural support but may also possess osteogenic, osteoconductive, and/or osteoinductive properties. The osteogenic potential of a graft is determined by the number of viable cells that can form bone or differentiate into bone-forming cells. Osteogenic grafts are usually obtained from fresh autologous bone and bone marrow cells. Osteoconduction is the physical property of a graft that allows vascular ingrowth and infiltration of osteogenic precursor cells. Osteoconductive grafts act as nonviable scaffolds that support the healing response. Examples are autologous and allograft bone, bone matrix, collagen, and calcium phosphate ceramics. Osteoinduction is the ability of a substance or factor to stimulate an undetermined osteoprogenitor cell to differentiate into an osteogenic precursor.

The most common source of graft material is autogenous bone from the iliac crest. This is still considered the gold standard. However, due to the morbidity of autograft harvesting along with its limited supply, allograft bone and synthetic bone substitutes are being used with increasing frequency. Autograft cancellous bone has all three ideal transplant properties: osteogenicity, osteoconductivity, and osteoinductivity. It possesses a large trabecular surface area for new bone formation and contains viable bone cells. Cortical autologous bone grafts have less capacity for new bone formation but are ideal when structural support is needed.

Allograft bone can be procured in greater quantities than autografts. Allografts are incorporated more slowly and to a lesser degree than autografts. Fresh-frozen grafts are stronger, more immunogenic, and more completely incorporated than freeze-dried grafts. In most studies, fusion using autograft bone leads to higher fusion rates than fusion, using allograft or synthetic bone substitutes alone. In a selected comparison, uninstrumented interbody fusion using autograft was compared with uninstrumented interbody fusion using autograft/allograft mixtures. The fusion rates were 88% and 82%, respectively, and were not statistically different.³⁷ If used anteriorly, allografts are well suited for reconstructive procedures and have good fusion rates, especially if combined with posterior fusions. When used with BMP-2, the effectiveness of an allograft can be increased. In one study, patients undergoing anterior lumbar fusion surgery with structural threaded cortical allograft bone dowels showed higher rates of fusion, improved neurologic status, and lower back and leg pain when compared with the autologous control group.³⁸ Considering the possible complications associated with harvesting iliac crest bone, the use of allogenic bone appears justified.^39,⁴⁰

Xenografts have been used in orthopedic surgery and have included cow horn and bovine bone. Due to the immune response evoked by the host to these materials, xenografts are not commonly used in spine surgery.

Graft morphology

The size of a structural bone graft, its positioning, and elasticity should be considered when used for lumbar fusions. A larger graft with low stiffness should be favored from a mechanical point of view. Stiff bone grafts increase stresses on adjacent endplates.⁴¹ Placement of a bone graft in a more anterior location can better resist flexion moments and decreases the tensile forces acting on the posterior ligaments. Bone grafts placed at a posterior site are better in resisting torsional moments and decreasing contact force of the facet joint in the fused segment.⁴² The degree to which these aforementioned observations may influence the long-term clinical outcome of spinal fusion is unclear.

Bone morphogenic proteins

The osteoinductive potentials of bone morphogenic proteins have been well described. Recombinant human BMP-2 has been approved by the FDA for single-level interbody fusions of the lumbar spine. In the first prospective, multicenter study involving 279 patients with degenerative lumbar disc, the fusion rate at 24 months was higher for the BMP-treated group compared with those who received autogenous iliac-crest bone graft (94.5% versus 88.7%).⁴³ Clinical trials have shown that rhBMP-2-filled fusion cages can eliminate the need for harvesting iliac crest graft.^44,⁴⁵ The benefits of BMPs in other applications of spinal fusion are currently under study.

Interbody cages

Intervertebral cages or spacers were developed to prevent the collapse and pseudoarthrosis seen with bone-only interbody fusions. Initially, the early types consisted of vertically placed titanium mesh cylinders and rectangular carbon fiber cages. Later, this was followed by the development of stand-alone threaded Bagby and Kuslich (BAK) intervertebral cages. Threaded intervertebral cages usually stabilize a segment through distraction and tensioning of the annular and ligamentous structures. By partially reaming the endplates, one may expose cancellous bone for arthrodesis, increasing the likelihood for successful fusion.⁴⁶ Anteriorly placed threaded cages significantly stabilize the motion segment in all directions except extension. With posteriorly placed cages, there is less stability as a result of the facetectomy required for placement of the device. Interbody cages may be implanted via a variety of surgical approaches to the disc space, with or without supplemental posterior fixation.

For single-level degenerative disc disease, anteriorly placed intervertebral threaded cages have shown a high degree of clinical success.⁴⁷ Both patient selection and technical expertise seem equally important in determining patient outcome.⁴⁷ Patients with multilevel disease, normal (tall) looking disc on radiographs, or with osteoporotic bone, have a higher incidence of complication and failure. Posteriorly or transforaminally placed cages which are supplemented by pedicle screws seem successful for the treatment of spondylolisthesis.⁴⁷ For the management of pure discogenic pain without collapse, multilevel disc disease, or disc degeneration in elderly patients, the use of cages is less certain.⁴⁷

Electrical stimulation

During the last 20 years, electrical stimulation has shown promise for improving the success rate of spinal fusion. The maximum benefit, in terms of an increased successful fusion rate, is estimated to be approximately 10%. At present, the types of electrical stimulation used as adjuncts to spinal fusion includes direct current electrical stimulation (DCES), pulsed electromagnetic fields (PEMF), combined magnetic fields (CMF), and capacitive coupling. These devises are usually implanted close to the fusion mass intraoperatively or worn externally.

Internal devices utilize DCES. Their mechanism of action is thought to involve the attraction of charged proteins, growth factors, and bone-forming cells to the fusion site. They may also alter the function of voltage-gated channels with activation of cyclic AMP and triggering of a second messenger cascade. Other proposed mechanisms include Faradic reactions at the cathode–bone graft interface with formation of OH and H₂O₂. This reduces the local oxygen tension (PO₂) and slightly increases the pH. The end result of both the electric field effects and the Faradic products is a stimulation of calcium uptake.⁴⁸^–⁵⁰ Increased pH stimulates osteoblastic bone formation and mineralization but inhibits osteoclastic bone resorption. The end result is an increased rate of new bone formation.⁵¹

In contrast to DCES, PEMF devices generate an electromagnetic field across the spinal fusion area and are worn externally, usually 3–8 hours per day for 3–6 months. As these braces can be removed by patients, noncompliance can potentially hinder treatment.⁵² Successful arthrodesis can also be obtained with shorter application times. In a prospective study by Linovitz and coworkers, successful outcomes for one-level or two-level fusions (between L3 and S1) without instrumentation, either with autograft alone or in combination with allograft, were obtained when the combined magnetic field device was worn for 30 minutes per day for 9 months.⁵³ Such devices with shorter wearing times may have better patient acceptance.

The mechanism of action of PEMF is not well understood, but is thought to involve alterations in cell membrane potentials and parahormone receptors of bone cells. An increase in calcium influx into bone cells, promotion of calcification, and vascularization also occurs with PEMF. PEMF may also be capable of altering the production of cytokines, and increasing the expression of bone morphogenic protein (BMP)-2 and BMP-4 mRNA by cells. In ovariectomized rats, PEMF stimulation can decrease local production of tumor necrosis factor-α (TNF-α), interleukin 1β (IL-1β), and interleukin 6 (IL-6) and inhibit osteoclastogenesis.⁵⁴ A possible effect on prostaglandin E₂ (PGE₂) levels is also suggested.⁵⁵ PEMF seems to enhance cellular differentiation and does increase the number of bone-forming cells. This stimulatory effect is greatest during the early stages of bone healing.⁵⁶ However, there may be some place for PEMF stimulation in the later stages of treatment in cases of nonunion. For example, in patients with symptomatic pseudoarthrosis after lumbar spinal fusion, pulsed electromagnetic field stimulation has been shown to be an effective nonoperative salvage approach to achieving fusion.⁵⁷ In one study, 67% of patients who were treated with a pulsed electromagnetic field device worn consistently 2 hours a day for at least 90 days achieved fusion. Treatment was equally effective for posterolateral fusions (66%) as with interbody fusions (69%).⁵⁷

Since 1985 when the first report of the clinical efficacy of PEMF on spinal fusion was published, numerous studies have demonstrated the benefits of electrical stimulation for improving the success rate of spinal fusion surgery.^58,⁵⁹ Not all modes of electrical stimulation are equally effective in promoting successful spinal arthrodesis. Clinical data reveal superiority of DCES to PEMF, particularly when used to enhance posterior spinal fusions. Capacitive coupling seems superior to PEMF, and perhaps DCES as an adjunct to posterior spinal fusion.⁶⁰