Chapter 57 Dorsal and Lateral Thoracic and Lumbar Fusion
When surgical treatment of degenerative disorders is required, following neurologic decompression, fusion may be indicated in cases associated with spondylolisthesis or significant scoliosis. One well-known study,1 for example, demonstrated better clinical and radiographic outcomes following decompression and noninstrumented fusion, compared to decompression alone in the treatment of degenerative spondylolisthesis. Spine fusion may also be useful in cases in which significant iatrogenic instability has been created by wide decompression or by resection of significant lesions. Lumbar spine fusion as a treatment of degenerative disc disease alone, however, is generally not indicated.
The addition of segmental instrumentation has been shown to increase fusion rates and improve clinical outcomes.2,3 While internal fixation may be associated with complications, particularly in the setting of osteoporosis, it should also be considered for patients who are thought to be candidates for spine fusion.
Historical Background
Albee4 and Hibbs5 were the first surgeons to independently describe the technique of arthrodesis. They created greenstick fractures of spinous processes in patients with Pott disease. Albee used tibial autograft to augment the fusion. Mackenzie-Forbes6 and Hibbs7 introduced the technique of laminar decortications extending the fusion further laterally. Hibbs also described the concept of facet joint fusion in a series of his scoliosis patients.7 The technique was further modified by Howorth,8 McBride,9 and Moe.10
The subsequent introduction of intertransverse process fusion by Mathieu and Demirleau provided an alternative technique to dorsal fusion that proved beneficial for patients in whom laminectomy was contemplated.11 Adkins12 implemented this technique in spondylolisthesis patients. Watkins13 described a paramedian approach to gain access to transverse processes lateral to paraspinous musculature, thus sparing the spinous processes and interspinous ligaments. Wiltse et al.14 devised a muscle-splitting approach between the longissimus and multifidus muscles to gain access to the intertransverse process region. The introduction of spinal instrumentation further improved the fusion results.
Biology of Bone Grafting and Spine Fusion
Bone healing following fracture involves a regenerative process, usually leading to complete restoration of the structural integrity of the affected bone. The physiology of spine fusion is similar to the physiology involved in fracture healing. The process of fusion begins by hematoma formation at the surgical site. The rich vascular arcade around the dorsal elements delivers oxygen to the fusion site and also leads to an influx of various inflammatory cells. The healing process begins within hours after surgery. The macrophages infiltrating the hematoma remove the necrotic debris at the fusion site. During this process, the macrophages secrete a number of inflammatory cytokines that stimulate migration of other inflammatory cells into the fusion area.15 Platelets within the hematoma secrete platelet-derived growth factor, which is chemotactic to fibroblasts and other mesenchymal cells in the area.16 The inflammatory markers secreted by platelets and other inflammatory cells promote proliferation and differentiation of primitive mesenchymal cells into cells of osteoid or chondroid lineage.17 Bone morphogenetic protein plays an active role in bone synthesis.18–20
The second phase of fusion involves a repair process.21 Inflammatory markers secreted during the repair process promote ingrowth of new blood vessels. Formation of cartilaginous matrix by chondroblasts as well as osteoid formation by osteoblasts leads to new bone formation. Mineralization of the matrix leads to primary or woven bone formation. Compression and distraction forces22 at the fusion site may provide a stimulus for remodeling21 and consolidation of the fusion mass. Incomplete ossification may lead to fibrous nonunion or pseudarthrosis.
Graft Selection for Dorsal and Lateral Thoracolumbar Fusions
The ideal graft contains an ample number of viable osteoprogenitor cells. Cortical and cancellous autografts and allografts are the most commonly used graft23–26 materials for thoracic and lumbar fusions. Autogenous cancellous grafts are considered more likely to promote fusion than are cortical bone grafts because they are more rapidly vascularized and are more osteoinductive.19,20,27,28 However, cortical autografts are used for their structural integrity in ventral fusion procedures of the thoracic and lumbar spine. Autograft is limited in quantity, however, and is associated with significant morbidity from the autologous bone harvest.27,29–32
Allografts may provide an osteoconductive matrix necessary for fusion. They have limited osteoinductive properties, however, and are inferior to autogenous grafts. Allografts may evoke significant host immune response.33,34 Allografts are considered useful in patients with a limited bone supply or when a large quantity of graft material is needed.35 Adequate fusion rates can be achieved with autograft or allograft when supplemented with segmental instrumentation in young patients with thoracolumbar scoliosis.23–26 Thoracolumbar spine fusion rates are much higher, however, when autograft is used in comparison to allograft alone. A higher pseudarthrosis rate has been demonstrated with the use of allograft as compared to autograft for intertransverse and interlaminar fusion in the lumbar spine.36–39
Surgical Technique
Paraspinal (Wiltse) Approach to the Lumbar Spine
In the paraspinal (Wiltse) approach,14 two paramedian incisions are made approximately 5 cm from the midline just medial to the posterior superior iliac spine. The incision can be planned from the preoperative MRI images by measuring the distance of the plane of separation between the multifidus and longissimus muscles and the midline (Fig. 57-1). The dissection is advanced to the deep dorsal fascia. The fascial incisions are curved medially at their caudal ends to provide adequate muscular retraction. The dissection is further advanced through muscle fascia. The interval between the longissimus and the multifidus muscles can be separated by blunt finger dissection. The dissection should be directed medially to reach the transverse processes and lateral surface of the facet joints. Self-retaining retractors are placed deep to the fascia, and the transverse processes are exposed subperiosteally.
Preparation of Fusion Bed and Bone Grafting
Preparation of Interlaminar and Intertransverse Region
The dorsal surface of the lamina, lateral surface of the facet joint, mamillary body, and lateral pars should be cleared of soft tissues. The exposed area is decorticated by using either a rongeur or a high-speed bur. Graft material is then placed over the decorticated bony elements (Fig. 57-2A).
Facet Joint Fusion
In the thoracic spine, the inferior articular process can be excised by using an osteotome or a Capner gouge to expose the facet joint and cartilaginous surface of the superior articular process.40 The cartilage is then removed to expose bleeding cancellous surface. The bone graft can then be placed on this exposed surface (Fig. 57-2B).
In the lumbar spine, the facet joints can be exposed by excising the facet capsule. Once exposed, the cartilage from the joint can be excised by using a narrow rongeur or a high-speed bur. The exposed bleeding joint surface can then be packed with cancellous bone chips to facilitate fusion across the joint (see Fig. 57-2A).
Complications and Avoidance
Hemorrhage
Perforation of terminal branches of lumbar and thoracic segmental arteries during the exposure of the facet joints, lateral pars interarticularis, and transverse processes can lead to significant blood loss. These arteries are found deep to the paraspinal muscles at the superior and inferior aspects of the facet joints (rostral and caudal articular arteries) and on the dorsal surface of the transverse process (communicating artery) (Fig. 57-3). Bleeding from these vessels can be reduced by identification and cauterization prior to dissection. Venous bleeding in the lumbar spine can be avoided by careful positioning of the patient to avoid increased intra-abdominal pressure. Autologous blood donation and the use of cell savers should be considered in patients in whom excessive intraoperative blood loss is anticipated. Intraoperative venous bleeding can be greatly decreased by reducing pressure on the abdomen. This decreases pressure on the vena cava, thus reducing the pressure in the epidural and paravertebral venous plexus.41
Pressure and Traction Injuries
Pseudarthrosis
Successful fusion is defined as the presence of continuous bridging trabeculae of bone between spinal segments. A successful fusion inherently requires the absence of motion between the segments and may provide relief of symptoms caused by mechanical instability. Failure of fusion at the surgical site at or after 1 year from index surgery indicates a pseudarthrosis and needs further investigation into etiology and treatment. Pseudarthrosis is one of the leading causes for revision lumbar spine surgery.43
A variety of factors are thought to be responsible for pseudarthrosis, including metabolic abnormalities, smoking, infection, and persistent motion at the fusion site. Smoking has been shown to be associated with increase in nonunion rates from 8% in nonsmokers to 40% in smokers.44 Persistent motion across the fusion segments is thought to be associated with pseudarthrosis. The incidence of pseudarthrosis increases with increase in the number of levels fused owing to the presence of a longer area that needs to be bridged by the fusion process as well as the presence of more motion across the fusion area.45,46 Spinal instrumentation has been shown to increase the fusion rate by limiting the motion across the fusion segments.3,46–48 A prospective randomized study by Fischgrund et al. showed significantly improved fusion rates at the end of 2 years when lumbar decompression and dorsolateral fusion were combined with instrumentation as compared to noninstrumented decompression and dorsolateral fusion.49
Use of flexion/extension radiographs for diagnosis of pseudarthrosis is controversial and is affected by high interobserver variability.50 Thin-cut CT scans are considered more accurate than are plain radiographs in determining the integrity of the fusion. The presence of metallic artifacts in instrumented fusion, however, decreases the sensitivity of CT scan as the diagnostic modality of choice. Thin-slice CT scan combined with a high index of clinical suspicion may be considered a most reliable diagnostic option. However, exploration of the fusion mass is considered the most specific and sensitive test for diagnosis of pseudarthrosis.
Failure of fusion may present with persistence of back pain, progression of deformity in scoliosis surgery, or recurrence of symptoms in spondylolisthesis. Pseudarthrosis is associated with worse clinical outcomes.51 The treatment of pseudarthrosis should begin with identification of biologic and mechanical factors that contribute to poor bone healing after spine fusion. Correction of endocrine and nutritional factors may assist in achieving solid fusions. Addition of nonrigid mechanical fixation has also been shown to lead to higher fusion rates. The incidence of pseudarthrosis is higher at the thoracolumbar junction as well as the lumbosacral junction. Surgical treatment includes repair of pseudarthrosis by exposure of the fusion area, removal of instrumentations, thorough decortication, and bone grafting with large quantity of autogenous iliac crest bone graft. Instrumentation is then reapplied with compressive forces across the fusion area. RhBMP-252 and osteogenic protein-1 (OP1) have been found to be useful in increasing the fusion rates in lumbar spine fusion surgeries.
Infection
The rate of infections following spine surgery procedures has been shown to correlate with the duration of the procedure, associated patient comorbidities, and the use of instrumentation. The rate of infection is 2% to 4% in noninstrumented spine fusion and 6% to 11% in instrumented spine fusion.53 Staphylococcus aureus is the most common organism responsible for postoperative spinal infections. Postoperative infections may result from direct inoculation or through hematogenous spread of infection from a remote source of infection. Statistically significant preoperative risk factors that are associated with increased risk of infection include age more than 60 years, smoking, diabetes, previous surgical infection, increased body mass index, and alcohol abuse. Intraoperative factors that are associated with increased risk of infection include staged procedure, operative time more than 5 hours, and fusions involving 7 to 13 levels.54
