Phẫu thuật điều trị cột sống pptx

Thoracolumbar Fracture Management: Anterior Approach John S. Kirkpatrick, MD Abstract The goals of thoracolumbar fracture management are to preserve or restore the neurologic and biomechanical functions of the spine. Both nonsurgical and surgical management have arole. Optimal surgical management of thoracolumbar fractures requires understanding the patient’s clinical situation, the fracture classification, and the strengths and weak- nesses of a variety of approaches and stabilization techniques. The anterior approach can be used for both management of the neurologic deficit and restoration of stability to the spine. In most patients, neurologic deficit is caused by impact and/or compression tothe ventral sur- face of the spinal cord. The anterior approach provides optimal direct exposure for visualization of the ventral aspect of the dura mater during surgical decompression. 1 Additional- ly, for fracture patterns involving marked comminution with loss of support of the anterior and middle columns of the spine, the anterior approach provides excellent exposure for reconstruction with structural grafts or implants. This allows restoration of height and correction of kyphosis while limiting the number of motion segments fused. This is especially use- ful in patients whose general condition prevents posterior reduction in the first 7 to 10 days after injury. The anterior approach also avoids additional injury to the paraspinal muscles and disruption of their innerva- tion. However, comparative studies between anterior and posterior approaches are limited. Thus, it is difficult to present objective evidence that one approach is better than the other, especially in the first 7 to 10 days after injury.Additionally, nonsurgical management as well as anterior and posterior approaches each have their unique role in the treatment of patients with thoracolumbar spine injury. Patient Evaluation Appropriate treatment of patients with thoracolumbar injuries is guided by a thorough history and physical examination, with particular attention paid to the neurologic assessment. The history should include the mechanism of injury; the presence of pain, weak- ness, and loss of sensation; and a re- view of comorbidities. The physical examination should include log roll- ing to allow visual inspection and pal- pation of the back and spine. Local tenderness, swelling, gaps between spinous processes, gibbous deformity, and ecchymosis should be noted. Neurologic examination of the lower extremities and perineum is critical and should include evaluation of sensation, motor function, and reflex- es. Radiographs, computed tomogra- phy (CT), and, on occasion, magnetic resonance imaging help delineate the nature of the fracture and extent Dr. Kirkpatrick is Associate Professor, Division of Orthopaedic Surgery, University of Alabama at Birmingham, Birmingham, AL. Neither Dr. Kirkpatrick nor the department with which he is affiliated has received anything of value from or owns stock in a commercial company or institution related directly or indirectly to the subject of this article. Reprint requests: Dr. Kirkpatrick, 940 Faculty Office Tower, 510 20th Street South, Birmingham, AL 35294-3409. Copyright 2003 by the American Academy of Orthopaedic Surgeons. The surgeon who treats patients with spine trauma must be able to apply a variety of management techniques to achieve optimal care of the patient. The anterior surgical approach is appropriate for some thoracolumbar burst fractures in patients with neurologic deficit and without posterior ligamentous injury. Surgery is most often indicated for patients with incomplete deficit, especially those with a large retropulsed fragment, marked canal compromise, severe anterior comminution, or kyphosis >30°. This approach provides excellent visualization of the anterior aspect of the dura mater for decompression. Reconstruction of the anterior body defect can be done with autograft, allograft, or a cage. Supplementation of the graft with anterior internal fixation helps prevent kyphosis. Clinical results demonstrate improved neurologic function in most patients as well as low pseudarthrosis rates. In patients with incomplete deficit, improvement in neurologic function usually can be expected with few complications. J Am Acad Orthop Surg 2003;11:355-363 Vol 11, No 5, September/October 2003 355 of injury. 2 Magnetic resonance imaging is indicated in the presence of an unexplained neurologic deficit, progressive deterioration, or notable soft- tissue injury. Patients with fracture in the presence of multisystem trauma require special consideration. Blood loss can be significant (up to 1,500 mL) with the anteriorapproach, which can lead to coagulopathy and hypoperfusion in the presence of other injuries. Both the transdiaphragmatic and the thoracic approaches can temporarily impair pulmonary function, leading to marked pulmonary compromise, especially in the presence of closed chest injury. Intra-abdominal injury can lead to notable peritoneal disten- sion, making retroperitoneal exposure difficult. Patients with multiple extremity and/or pelvic fractures may be better served by reduction and stabilization of unstable thoracolumbar fractures through a posterior approach, thus allowing for more efficient care of the other injuries. Assessment of these factors, in addition to fracture pattern and neurologic status, are critical to making the appropriate treatment selection for the patient with thoracolumbar trauma. Classification of Injury The mechanistic fracture classification, as modified by McAfee et al, 2 is the preferred thoracolumbar fracture classification system because it can provide important insight into reduction mechanisms and stabilization needs as well as guide the surgeon to the appropriate surgical procedure. 3 This classificationscheme com- prises compression, stable burst, and unstable burst fractures; flexion- distraction injury; and fracture-dis- location. McAfee et al 2 emphasized the importance of Denis’ middle column and classified fractures as failure of the middlecolumn in compression, distraction, or translation. Indications The anterior approach is most commonly indicated for an unstable burst fracture from T10 to L3 (although it can be used up to T5) associated with an incomplete neurologic deficit and radiologically demonstrated neural compression. 1,4 Recent studies 5,6 have refined the radiologic and clinical features of patients with incomplete deficits in whom the anterior approach is indicated. These features may include a large retropulsed fragment with marked (>67%) canal compromise, anterior comminution and kyphosis >30°, and time of more than 4 days from injury. 5 Patients with incomplete fracture reduction after a posterior approach may be candidates for anterior decompression if neurologic recovery is incomplete and re- sidual compression persists. Parker et al 6 reviewed their institution’s experience with the load- sharing classification, which is based on the extent of comminution and displacement. Patients who had fractures with a high degree of comminution, displacement, and kyphosis had a better result with anterior stabilization because it provided greater restoration of anterior column support than did short-segment pedicle screw fixation. Others have found that longer posterior fusion constructs provide adequate stability. Controversy exists about whether patients with complete neurologic injuries warrant decompression in addition to stabilization. Patients with complete thoracic level paraplegia (above T10)have a poor prognosis for recovery and generally are best treated with posterior stabilization. Pa- tients with thoracolumbar junction or lumbar injuries who initially appear to be neurologically complete but are still in spinal shock can be treated more aggressively with prompt surgery. No data indicate a difference in recovery between anterior or posterior approach when adequate decompression is obtained. Because of alack of definitive studies demonstrating a clear advantage of one approach over the other, both are considered for management in appropriate circumstances. A summary of relative indications for the anterior approach is listed in Table 1. The anterior approach is more difficult for low lumbar fractures (ie, L4 and L5) because of anatomic con- straints, especially for restoring alignment and attaining satisfactory fixation. However, these injuries rarely require surgery and generally should be managed posteriorly in patients in whom stabilization is required. Be- cause of the large ratio of canal area to neural element, low lumbar fractures behave differently from upper lumbar fractures and generally do well with nonsurgical management, except in cases of instability. For pa- Table 1 Indications for Anterior Approach Incomplete paraplegia with stable or unstable burst fracture Fractures with poor reduction potential with posterior approach Large retropulsed fragment with >67% compromise 5 Anterior comminution with kyphosis >30° 5 >4 days since injury 5 Fractures with inadequate canal reduction and incomplete neurologic recovery after posterior stabilization and suboptimal neural recovery Reconstruction of anterior column after short-segment posterior stabilization Traumatic disk herniation with flexion-distraction injury Thoracolumbar Fracture Management: Anterior Approach 356 Journal of the American Academy of Orthopaedic Surgeons tients with instability, the posterior approach is preferable. Anterior decompression is rarely used alone for injuries other than burst fractures. Most other fracture types are either well treated nonsur- gically (ie, compression fractures) or managed with a posterior approach to restore the integrity of the posterior elements and prevent kyphosis (ie, flexion-distractioninjuries). In cer- tain limited circumstances, these injuries may require an adjunctive anterior decompression because of herniated disk, marked comminution of the middle column, or concern about additional displacement of fragments into the canal. In such cases, a combined approach should be considered. Most fracture-dislocations, because of their extreme instability, are best managed with a posterior approach. Spinal stability of fracture-dislocations after anteriordecompression, even with internal fixation, is insufficient; thus, the anteri- or approach should be avoided. Contraindications Preexisting medical conditions and concurrent traumatic injuries to the abdomen and chest may represent relative contraindications to the anterior approach. Patients with severe pulmonary disease may have limited reserve for pulmonary function and may not tolerate thoracic or thoracoabdominal approaches. Severe chest or abdominal injuries also may limit pulmonaryreserve or impair exposure. Marked osteoporosis in ad- jacent vertebrae may result in impaction of the strut graft and failure of screw purchase, leading to nonunion and/or kyphosis. Morbid obesity may impair exposure and lead to inadequate visualization for safe decompression. When these conditions are present, the surgeon must balance the relative merits of anterior and posterior approaches for the specific fracture. Timing The timingof decompression remains controversial, with a divergence between the results of animal studies and clinical reports. Some basic sci- ence data indicate that early decompression results in improved neurologic recovery. In surgically created acute cord compression in a canine model, neurologic recovery was better when release was done within 1 hour rather than after longer time pe- riods. 7 Carlson et al 8 studied regional blood flow, the interface pressure between the spinal cord and a com- pressing piston, and somatosensory evoked potentials in a canine model with decompression at 5 minutes and with no decompression. Regional blood flow returned to normal within 3 hours after onset of compression. The viscoelasticityof the cordallowed the interface pressure to decrease to <20% of the maximum in the first hour ofcompression and was approximately 10%by 3 hours. Despite these changes, somatosensory evoked potentials did not show improvement, indicating the multifactorial nature of spinal cord injury. The authors suggested that sustained displacement initiated a secondary phase of phys- iologic events.Acorrelation of evoked potential recovery with regional blood flow during compression was later reported, supporting the concept of an ischemic mechanism of secondary injury. 9 These factorsled tothe im- pression that there is a limited window of opportunity for obtaining optimal neurologic recovery in such injuries. Unfortunately, this window appears to be too brief to allow clinical rescue, resuscitation, diagnosis, and inductionof anesthesiafor urgent decompression. Recovery of cauda equina or root injuries does not appear to be as time-dependent. Clinical data about the timing of cord decompression are limited and mostly relate to cervical injuries. Ear- ly closedreduction of cervical subluxation has been shown to be possible as early as 2 hours after injury, with no neurologic deterioration directly attributed to traction and reduction. 10 Comparisons of early and late decompression are few. Vaccaro et al 11 did a prospective, randomized study of surgery done early (<72 hours) versus late (>5 days) after spinal cord injury. They found no significant benefit to neurologic function, length of stay, or length of rehabilitation when the surgery was done early. Authors of a more recent ret- rospective study comparing experience at two different institutions found that early surgery (within 72 hours of injury) was not associated with a higher complication rate. 12 They also suggested that the early surgery group may have had improved neurologic recovery in spite of early closed reduction in both groups. This difference could be ex- plained by variations in surgeons, sites, methods of neurologic evaluation, and preoperative function, but it warrants further study. Although these studies involve cervical injuries, they do provide some insight into the issues related to the timing of surgery for cord injuries. To date, there is no clear difference in outcomes based on timing of surgery. The timing and use of the anterior approach in multiply injured patients requires coordination between trauma surgeons and spine surgeons. Acute life-threatening injuries, such as unstable pelvis fractures, head, chest, and abdominal injuries, and limb-threatening injuries, such as open long bone fractures, should be handled first, followed by spine care. It is important to consider early decompression of incomplete spinal cord injuries within 24 to 48 hours as an emergent or urgent procedure, be- fore the typical onset of pulmonary complications resulting from the patient’s injuries. Early decompression has been shown to be safe and effec- tive, with the major difference between urgent and early treatment being the extent of blood loss. 13 John S. Kirkpatrick, MD Vol 11, No 5, September/October 2003 357 Management ofinjuries involvingthe cauda equina or root deficits are planned as soon as the patient’s overall condition is satisfactory, usually within the same time period, but oc- casionally up to 7 to 10 days in patients with life-threatening chest and abdominal injuries. Patients initially treated nonsurgi- cally or with a posterior approach who have persistent cord compression may be candidates for late decompression and may obtain clinical improvement. The anterior approach can be used months or years after initial injury. 1 Bohlman et al 14 studied patients treated for late pain and/or paralysis a mean of 4.5 years after injury (range, 3 months to 21 years). They noted improvement in pain for 41 of 45 patients (91%) and improvement in neurologic function in 21 of 25 patients (84%). Reconstruction Patients withburst fractures have failure of the anterior column in compression, producing akyphotic deformity andinability ofthe spineto resist axial load. When associated with loss of the posterior column tension band, the result is an extremely unstable spinal injury.Anteriordecompression of the neural elements further desta- bilizes the spinal column by removing whatever anterior support remains. Thus, the primary principle for reconstruction after anterior decompression is restoration of the anterior column so that it can resist axial compression. If the posterior tension band also has failed, posterior stabilization may be required, as well. Reconstruction generally involves two components: immediate stability and restoration of normal alignment. Immediate stability can be obtained with a variety of devices, such as cages, rod-and-screw or plate-and- screw constructs, and external braces. Iliac crest autograft was initially used for the strut graft, but more re- cently, tibial or humeral allograft has been used. 15 Ventral cages containing cancellous autografthave beendevel- oped, but there have been few studies of their efficacy. Long-term stability can be achieved with fusion of the strut graft. Implant types commonly used anteriorly for fracture indications include both rigid and nonrigid plate- and-screw and rod constructs. Most studies report the use of either rigid screw and rod constructs 16,17 or rigid plate constructs. 18 Semirigid or dynamized constructs using screws and rods have been reported with satisfactory outcomes. 19 Implants should be placed laterally to avoid contact with the aorta because such contact has been reported to cause late vascular disruption and death. 20 In addition, to avoid problems with the iliac vessels, anterior plate-and-screw or rod-and-screw constructs should not be used below L4. Biomechanics of Anterior Reconstruction The biomechanics of anterior reconstruction have been studied in a variety of models, including animal (both in vivo and in vitro), cadaveric, and biomechanical synthetic. Devic- es used anteriorly can be divided into two categories: interpositional and splinting. Interpositional devices, which substitute for the anterior and middle columns, are usually biologic (eg, iliac crest strut). Splinting devices are used to stabilize the construct during biologic incorporation of theinterpositional device.The multifactorial nature of the biomechanical properties of these devices makes comparison of different studies difficult. Construct strength and stiffness are affected by biologic variability in patient size and bone density, purchase strength of the anchors (usually screws), load sharing with the graft, and the mechanical properties of the implants. Biologicvariability between individual patients must be considered when planning reconstruction after corpectomy. Although little has been published about the purchase strength of anchors, bicortical purchase has been shown to increase screw pullout strength over unicor- tical purchase; however, the effect is less pronounced when bone mineral density is low. 21 Synthetic models standardized by the American Soci- ety for Testing and Materials provide a methodfor comparing implant stiffness and fatigue strength without the confounding variables introduced by biologic variability, specimen avail- ability, and anchor purchase failure. 22 The load sharing of the graft may contribute to stability, depending on the construct chosen. Lee et al 23 reported differences in stability testing depending onthe anteriorreconstruc- tion method used. They compared a polymethylmethacrylate block, iliac crest bone graft, two small cages, and one large cage.The large cagewas superior in axial rotation and sagittal motion, and the two small cages and iliac crest bone graft were superior in lateral bending. The PMMA block was approximately the same as iliac crest bone graft in all modes tested except lateral bending, where it was inferior. Many surgeons use internal fixation inaddition to anterior column reconstruction. The properties of the interpositional and internal fixation devices may be combined for improved overall stability. In an older in vitro study, the Kaneda device (DePuy Acromed, Raynham, MA, formerly Acromed) was found to restore torsional stiffness as well as a posterior Cotrel-Dubousset construct 24 and better than Harrington rods or the AO fixateur interne (neither ofwhich isused anymore). 25 Oth- er authors emphasized the importance ofconnecting parallel rods used in anterior or posterior constructs. 26 In a canine study, the fusion rate was Thoracolumbar Fracture Management: Anterior Approach 358 Journal of the American Academy of Orthopaedic Surgeons higher with the Kaneda device and graft (86%) than with graft alone (29%) at 24 weeks (P = 0.028). 27 Ax- ial, flexural, and torsional stiffness were tested; only torsional stiffness was found to be significantly (P < 0.05) stronger in the instrumented fu- sions. Other studies have included comparisons of different devices in animal or cadaveric in vitro models, with varying results. The models using stability testing favor the Kaneda device over plate-and-screw constructs in general, especially in torsion. 28-30 Results of stability testing using a synthetic model have favored plate over rod constructs, with the ex- ception of the Z-plate (ZPLATE-ATL; Medtronic Sofamor-Danke, Mem- phis, TN), which tends to be the least stiff of all constructs tested. 31 The use of synthetic models tends to favor screw anchors over bolt-type anchors, which may have had some effect on the results. The material used in manufactur- ing theimplant affectsits stiffness and strength. There are few direct comparison data. Kotani et al 31 compared Kaneda devicesmade oftitanium and of stainlesssteel. Thesedevices are as- sumed to have similar dimensional tolerances. The titanium implant was found to have both greater bending strength and higher stiffness than the stainless steel implant. The importanceof subtledifferenc- es instiffness remains tobe seen.Clin- ical studiesof all devices demonstrate high fusion rates and good perfor- mance, probably making the subtle differences in biomechanical properties unimportant. In addition, the stiffness needed in each planeappears to bedifferent. Oneclinical studywith an implantconstruct thatallowed dy- namic axial compression showed that no patients had pseudarthrosis, and kyphosis worsened by only 4° at long- term follow-up (mean, 42 months; range, 24 to 84 months). 19 Although the required degree of stiffness remains unclear, it appears that any reconstruction technique that provides rotational stiffness equal to or greater than that of the intact spine will provide a stable construct that leads to fusion. In youngpatients with qualitative- ly good bone quality, a biologic strut (eg, iliac crest bone graft) is optimal for anterior column support. Alterna- tively, an appropriately sized allograft can be used. When preparing the end plates, a curette should be used to re- move the cartilaginous end plate to bleeding bone. This removes less of the subchondralbone andthus reduc- es the chance of graft settling. If the bone quality is poor, or if the patient prefers to avoid the discomfort of graft harvest, a tibial allograft provides a broad base of contact between the graft and vertebrae. Either type of graft can be supplemented with cancellous bone from the vertebrec- tomy. Alow-profile plate construct to augment the reconstruction is placed laterally, away from the aorta, avoid- ing thepotential for late erosion ofthe aorta (Fig. 1). Clinical Results Because of thehighly variable nature of thoracolumbar burst fractures and the lack of randomized prospective studies, clinical results after direct anterior decompression are difficult to Figure 1 A, Lateral radiograph of a 48-year-old man who sustained an L2 burst fracture associated with paraparesis. The patient had grade 4 quadriceps strength and dysesthetic pain in his thighs. B, Axial CT scan demonstrates >67% canal compromise and comminution anteriorly. Anteroposterior (C) and lateral (D) radiographs after anterior decompression and application of tibial allograft with lumbar plate. The patient had resolution of leg pain and normal strength after the decompression and returned to full function and work as a contractor. John S. Kirkpatrick, MD Vol 11, No 5, September/October 2003 359 compare with those of other techniques. However, McAfee et al 1 reported that 37 of 42 patients treated with anterior decompression at a mean of 60 days after initial injury had some degree of neurologic improvement. Of the 37 patients, 30 pre- operatively had motor strength of grade 3 or less. Fourteen of these 30 patients became community ambula- tors; 9 others regained function adequate for household ambulation, although some required short leg braces and/or crutches. Radiograph- ic results indicated that 12 of the 42 patients developed kyphosis >20° postoperatively. Kanedaet al 16 reported a series of 150 patients with burst fractures and showed comparable results in78 withneurologic deficit who underwent anterior decompression supplemented with anterior fixation. The time from injury to decompression variedfrom <48 hours to>1 year. Seventy-two percent of the neurologically compromised patients (56/78) recovered completely. Eighty-six percent of all patients who had been em- ployed preinjury (112/130) returned to their jobs. 16 No patient in either study 1,16 was made neurologically worse, and those with the most incomplete deficits recovered one Frankel grade or more. In both studies, CT done after decompression demonstrated adequate decompression in all but two cases. Comparative studies between anterior and posterior approaches are few, use multiple surgical techniques, and have relatively small numbers of patients in each group. Esses et al 32 compared 18 patients treated withan- terior decompression, iliac strut, and Kostuik-Harrington anterior fixation with 22 patients who had posterior distraction instrumentation with the fixateur interne and posterolateral fusion. Postoperative CT demonstrated better canal decompression with the anterior approach, but this did not correlate with neurologic recovery, which was no different between the groups. Schnee and Ansell 33 compared 14 patients treated with anterior decompression, allograft strut, and platefixation with 9 patients who underwent combined anterior decompression and posterior fixation and 2 patients who had transpedic- ular decompression, fusion, and fixation. Choice of technique apparent- ly was related to the severity of injury to the spinal column. The authors concluded that anterior decompression is critical to success in manag- ing fractures with significant vertebral destruction. Been and Bouma 34 studied 27 patients treated with anterior decompression and iliac strut combined with posterior fixation and compared them with 19 patientstreat- ed with the AO fixateur interne for posterior distraction and stabilization. There was no statistical difference between the two groups; neurologic recovery of more than one Frankel grade occurred in10 of 10 patients in the anterior group and 7 of 8 in the posterior group. Bladder recovery was obtained in 3 of 7 in the anterior group and in 1of 3in thepos- terior group. Pain relief occurred in 85% of the anterior group and 79% of the posterior group. Complications occurred in 15% of the anterior group and 26% in the posterior group. Surgical Technique The right-sided approach is used for upper thoracic burst fractures to avoid the aortic arch, common carot- id artery, esophagus, and trachea. Thoracic fractures (T6 through T11) can be approached from the right, although a left-sided approach is possible because the aorta is easily mo- bilized. The left side is preferred for thoracolumbar fractures requiring a thoracoabdominal (T12-L1) or retroperitoneal (L2-3) approach; this avoids the approach being obscured by the liver or having to mobilize the vena cava.However,a right-sidedap- proach can be used for thoracolumbar fractures, if necessary. Lower lumbar fractures (L4 and L5) usually are approached posteriorly, although rarely they may require reconstruction of the anterior column through a retroperitoneal (L4) or transperito- neal (L5) approach. The patientshould beplaced in the true lateral position to help the surgeon maintain orientation of the vertebral body. The region of the fracture is positioned over the break in the table because flexing the table will improve exposure. For a fracture at the thoracolumbar junction (T12-L2), an oblique incision is made either along the 12th rib (T12-L1) or just inferior to it (L1-L2) extending toward the umbilicus for a retroperitoneal approach. For T11 and some T12 fractures, especially when internal fixation is planned, a 10th or 11th rib approach is used. Special attention is required in these exposures to remain extrapleural, and meticulous repair is needed when the diaphragm is in- cised. When such an incision is needed, it is important to leave approximately 1 to 2 cm of the periphery attached to the chest wall for later repair. Sutures may be placed to mark normal anatomic locations around the periphery to aid anatomic closure after reconstruction. For a 12th rib exposure, the rib is excised and the ret- roperitoneum identified where the transversalis fascia, pleura, and diaphragm meet near the tip of the 12th rib. Incision of the abdominal muscles provides access to the retroperi- toneum. The peritoneum, retroperitoneal fat, and kidneys are reflected anteriorly using blunt dissection, thereby exposing the quadratus lum- borum and psoas major muscles. The level of the fracture is identified and the psoasgently elevatedfrom the anterior portion of the vertebral body. The segmental vessels then are identified between the disk spaces on the vertebral bodyand are ligatedand divided. Subperiosteal dissection of these structures allows exposure nearly to the opposite pedicle. A Finochietto rib retractor or other self- Thoracolumbar Fracture Management: Anterior Approach 360 Journal of the American Academy of Orthopaedic Surgeons retaining retractor is placed between the 11th rib and the iliac crest. Decompression is accomplished with the aid of loupe magnification and a fiberoptic headlight. Diskecto- my is done above and below the fractured vertebra, followed by corpectomy for the decompression. 1 Bone removed with rongeurs is saved for later use to supplement the strut graft with cancellous bone. Direct visualization of the dura through the pedicle resection and the diskectomy sites helps avoid premature penetration of the posterior wall of the vertebra and injury to the dura. The transverse width of the vertebral body, as noted through the diskectomy sites, serves as aguide for the extent of the decompression. The medial wall of the contralateral pedicle is the landmark for the adequacy of the decompression. A common mistake is not to decom- press across the vertebral body to the contralateral pedicle. The dura will resume its normal contour after decompression. Some modifications of this technique are needed for thoracic fractures. A double-lumen endotracheal tube often is used because some patients do not tolerate packing of the lung for exposure. The rib above the fracture can be used as the landmark for the approach. Some surgeons re- move the rib; others use the costal in- terspace for the exposure. Removal of the rib has the disadvantage of in- creased pain, but it provides some local autograft and a much wider exposure. The rib head attachment to the fractured vertebra is removed to expose the pedicle and foramen. If the spine is not already in adequate sagittal alignment after decompression, pressure on the skin posteriorly at the level of injury and/or distraction within the corpectomy defect, using a large laminar spreader or implant instrumentation, can help restore normal alignment. The vertebral end plates should be prepared by removing the cartilaginous end plate with a curette or bur. A compromise must be made between obtaining adequate bleeding bone for vascularsup- ply to the graft and removing so much of the subchondral bone that the mechanical support for the graft is de- creased. Some surgeons prefer to make indentations or seating holes into the vertebral bodies to accommodate the ends of the graft, but generally this should be combined with additional support (eg, internal fixation) to prevent graft impaction and kyphosis. Measurements are made for the height and width of the graft, which then are applied to shaping the iliac crest autograft, allograft tibia or femur, or cage device. The graft is impacted into position with direct visualization of the dura to avoid impingement. Placing the tricortical portion of the graft on the contralateral side can help prevent settling and coronal deformity when internal fixation is used. The break in the operating table is then removed, thereby eliminating the lateral bending induced in the spine by position- ing the patient. This tends to lock the graft in place and prevents the spine from being left with a coronal plane deformity. If implants are being used, they are applied according to man- ufacturer recommendations. After instrumentation, a check should be done to make sure the hardware does not impinge on vascular or visceral structures. Anatomic closure of the wound is done after insertion of an appropriate drain or thoracostomy tube. If the pleural cavity was involved in the exposure, clinical and radiographic monitoring forpneumothorax should be done postoperatively. Thoracosto- my tubes generally are removed when there is no pneumothorax on radiograph and no air leak and when drainage has subsided. In patients who develop a postoperative ileus, diet should be restricted and/or na- sogastric suctionused until the return of bowel sounds. External support with a total-contact or thoracolum- bosacral orthosis often is used, and ambulation is begun once brace fit- ting is accomplished (usually 48 hours postoperatively). Progressive exercises for rehabilitation of lumbar and abdominalmuscles are begun approximately 3 months postoperatively and, when necessary, followed by work hardening programs. Complications Causes of complications can be grouped into three generalcategories: surgical approach, decompression, and structural (reconstruction). 35 Complications may occur as a result of using the retroperitoneal or thoracoabdominal approach to the spine. Pneumothorax, recognized intraoper- atively and on postoperative radiographs, is managed with insertion of a chest tube. Atelectasis, and occa- sionally postoperative pneumonia, can occur and appears in the contralateral or dependent lung in 3% to 5% of patients. 35 Superficial and deep wound infections are rare; most re- spond to antibiotic therapy. Infections unresponsive to initial antibiotic therapy may require surgical débride- ment, usuallywith retention ofthe reconstruction. Genitofemoral nerve injury, nerve root (eg, lumbar plexus) traction injury, and injury to the sym- pathetic plexusoccur inapproximate- ly 2% to 4% of patients. 1,16,35 Intraop- erative laceration of the inferior vena cava has been reported. Blood loss is variable, but the need for transfusion should be anticipated. Ileus is common with retroperitoneal approaches but generally resolves within 24 hours. Reported, but rare, complications include peritoneal entry, dam- age to the ureter, interruption of lymphatic channels with resulting chylothorax or chylous leak, and splenic rupture. Late complications from the retroperitoneal approach also may include incisional hernia and permanent abdominal swelling on the side of the approach. Complications related to decompression are relatively uncommon. John S. Kirkpatrick, MD Vol 11, No 5, September/October 2003 361 Iatrogenic neurologic injury is not reported in major series, likely because of the safety resulting from direct anterior visualization of the thecal sac. Iatrogenic dural lacerations should be isolated and closed, if possible. Sub- arachnoid drainageshould beconsid- ered for persistent cerebrospinal flu- id leak. Kyphosis and pseudarthrosis are the main structural complications of anterior arthrodesis. Without internal fixation, kyphosis occurs in approximately 25% of patients. 4 This has not been felt to be detrimental to the recovery ofneurologic function because the kyphosis results from settling of the graft without compression of the neural elements. Instrumentation can reduce this rate of kyphosis to between 5% and 10%. 29 Rates of pseudarthrosis are generally 5% to 10% and appear to be at the lower end of this range when fixation is used. 16,17,32,35 Pain at the iliac crest donor site is reported in approximately 5% of cases, although this may be an artificially low estimate. 15,17,35 Implant complications are device- or technique-related. 18 Device-related complications include screw or bolt breakage and rod or interconnection failure, both of which are associated with progressive kyphosis. In one series using the first-generation Kane- da device in 20 patients, three screw failures and one pseudarthrosis occurred. 17 Implant technique complications are rarely reported, but concern exists regarding canal penetration and vascular injury to the vessels on the contralateral side. In- adequate exposure of the superior end vertebra may prevent proper placement of the screw or bolt flush against the vertebral body, complicat- ing proper placement of the plate. Al- though the prominence of implants is a potential problem in the thoracic spine, the anterior position ofthe aorta and the overlying psoas major muscle eliminates this concern at the thoracolumbar junction. Summary The anterior approach for thoracolumbar fractures may be preferred in patientswith incompleteneurolog- ic deficit from burst fractures without substantial posterior element injury. Excellent visualization of the anterior dura mater allows safe decompression and leads to some degree of neurologic recovery in most patients. Reconstruction generally in- cludes the use of iliac crest strut graft, cages, or allograft. Supplementation with internal fixation can improve biomechanical stability and may lead to improved fusion rates and reduction in ultimate kyphosis. Complica- tions are rare and do not generally af- fect long-term outcome. References 1. McAfee PC, Bohlman HH, Yuan HA: Anterior decompression of traumatic thoracolumbar fractures with incomplete neurological deficit using a retroperitoneal approach. J Bone Joint Surg Am 1985;67:89-104. 2. McAfee PC, Yuan HA, Fredrickson BE, Lubicky JP: The value of computed to- mography in thoracolumbar fractures: An analysis of one hundred consecutive cases and a new classification. J Bone Joint Surg Am 1983;65:461-473. 3. Mirza SK, Mirza AJ, Chapman JR, Anderson PA: Classifications of thoracic and lumbar fractures: Rationale and supporting data. J Am Acad Orthop Surg 2002;10:364-377. 4. Bohlman HH: Treatment of fractures and dislocations of the thoracic and lumbar spine. J Bone Joint Surg Am 1985; 67:165-169. 5. McCullen G, Vaccaro AR, Garfin SR: Thoracic and lumbar trauma: Rationale for selecting the appropriate fusion technique. Orthop Clin North Am 1998; 29:813-828. 6. Parker JW, Lane JR, Karaikovic EE, Gaines RW: Successful short-segment instrumentation and fusion for thoracolumbar spine fractures: A consecutive 4½-year series. Spine 2000;25:1157-1170. 7. Delamarter RB, Sherman J, Carr JB: Pathophysiology of spinal cord injury: Recovery after immediate and delayed decompression. J Bone Joint Surg Am 1995;77:1042-1049. 8. Carlson GD, Warden KE, Barbeau JM, et al: Viscoelastic relaxation and regional blood flow response to spinal cord compression and decompression. Spine 1997;22:1285-1291. 9. Carlson GD, Gorden CD, Nakazowa S, Wada E, Warden K, LaManna JC: Perfusion-limited recovery of evoked potential function after spinal cord injury. Spine 2000;25:1218-1226. 10. Grant GA, Mirza SK, Chapman JR, et al: Risk of early closed reduction in cervical spine subluxation injuries. J Neuro- surg 1999;90(1 suppl):13-18. 11. VaccaroAR, Daugherty RJ, Sheehan TP, et al: Neurologic outcome of early versus late surgery for cervical spinal cord injury. Spine 1997;22:2609-2613. 12. Mirza SK, Krengel WF III, Chapman JR, et al: Early versus delayed surgery for acute cervical spinal cord injury. Clin Orthop 1999;359:104-114. 13. McLain RF, Benson DR: Urgent surgical stabilization of spinal fractures in polytrauma patients. Spine 1999;24: 1646-1654. 14. Bohlman HH, Kirkpatrick JS, Delamar- ter RB, Leventhal M: Anterior decompression for late pain and paralysis after fractures of the thoracolumbar spine. Clin Orthop 1994;300:24-29. 15. Finkelstein JA, Chapman JR, Mirza S: Anterior cortical allograft in thoracolumbar fractures. J Spinal Disord 1999; 12:424-429. 16. Kaneda K, Taneichi H, Abumi K, Hash- imoto T, Satoh S, Fujiya M:Anterior decompression and stabilization with the Kaneda device for thoracolumbar burst fractures associated with neurological deficits. J Bone Joint Surg Am 1997;79: 69-83. 17. Kirkpatrick JS, Wilber RG, Likavec M, Emery SE, Ghanayem A: Anterior stabilization of thoracolumbar burst fractures using the Kaneda device: Aprelim- inary report. Orthopedics 1995;18:673-678. 18. Ghanayem AJ, Zdeblick TA: Anterior instrumentation in the management of thoracolumbar burst fractures. Clin Or- thop 1997;335:89-100. Thoracolumbar Fracture Management: Anterior Approach 362 Journal of the American Academy of Orthopaedic Surgeons 19. Carl AL, Tranmer BI, Sachs BL: Antero- lateral dynamized instrumentation and fusion for unstable thoracolumbar and lumbar burst fractures. Spine 1997;22: 686-690. 20. Jendrisak MD: Spontaneous abdominal aortic rupture from erosion by a lumbar spine fixation device: A case report. Surgery 1986;99:631-633. 21. Breeze SW, Doherty BJ, Noble PS, LeBlanc A, Heggeness MH: A biomechanical study of anterior thoracolumbar screwfixation.Spine1998;23:1829-1831. 22. Standard test methods for static and fatigue for spinal implant constructs in a corpectomy model. Annual Book of ASTM Standards 1996;13.01:1097-1112. 23. Lee SW, Lim TH, You JW, An HS: Bio- mechanical effect of anterior grafting devices on the rotational stability of spinal constructs. J Spinal Disord 2000;13: 150-155. 24. Gurr KR, McAfee PC, Shih CM: Biome- chanical analysis of anterior and posterior instrumentation systems after corpectomy: A calf-spine model. J Bone Joint Surg Am 1988;70:1182-1191. 25. Shono Y, McAfee PC, Cunningham BW: Experimental study of thoracolumbar burst fractures: A radiographic and biomechanical analysis of anterior and posterior instrumentation systems. Spine 1994;19:1711-1722. 26. Gaines RW Jr, Carson WL, Satterlee CC, Groh GI: Experimentalevaluationofseven different spinal fracture internal fixation devices using nonfailure stability testing: The load-sharing and unstable-mechanism concepts. Spine 1991;16:902-909. 27. Zdeblick TA, Shirado O, McAfee PC, deGroot H, Warden KE:Anterior spinal fixation after lumbar corpectomy: A study in dogs.J Bone Joint Surg Am 1991; 73:527-534. 28. Zdeblick TA, Warden KE, Zou D, McAfee PC, Abitbol JJ: Anterior spinal fixators: Abiomechanical in vitro study. Spine 1993;18:513-517. 29. An HS, Lim TH, You JW, Hong JH, Eck J, McGrady L: Biomechanical evaluation of anterior thoracolumbar spinal instrumentation. Spine 1995;20:1979-1983. 30. Hitchon PW, Goel VK, Rogge T, Gros- land NM, TornerJ: Biomechanical studies on two anterior thoracolumbar implants in cadaveric spines. Spine 1999; 24:213-218. 31. Kotani Y, Cunningham BW,Parker LM, Kanayama M, McAfee PC: Static and fatigue biomechanical properties of anterior thoracolumbar instrumentation systems: A synthetic testing model. Spine 1999;24:1406-1413. 32. Esses SI, Botsford DJ, Kostuik JP: Eval- uation of surgical treatment for burst fractures. Spine 1990;15:667-673. 33. Schnee CL, Ansell LV: Selection criteria and outcome ofoperative approaches for thoracolumbar burst fractures with and without neurological deficit.J Neurosurg 1997;86:48-55. 34. Been HD, Bouma GJ: Comparison of two types of surgery for thoraco-lumbar burst fractures: Combined anterior and posterior stabilisation vs. posterior instrumentation only. Acta Neurochir (Wien) 1999;141:349-357. 35. McAfee PC: Complications of anterior approaches to the thoracolumbar spine: Emphasis on Kaneda instrumentation. Clin Orthop 1994;306:110-119. John S. Kirkpatrick, MD Vol 11, No 5, September/October 2003 363