Rockwood & Green’s Fractures in Adults
6th Edition

Chapter 40
Fractures and Dislocations of the Thoracolumbar Spine
Robert K. Eastlack
Christopher M. Bono
P.1544

Reports of trauma to the thoracolumbar spine with associated neurologic injury were described as early as 3000 BC in the Edwin Smith Papyrus. With the introduction of motorized vehicles and greater exposure to high-energy blunt trauma, the[jyoccurrence of thoracolumbar fractures and dislocations has[jyincreased substantially. Recent data have indicated that motorcycle accidents are associated with a greater chance of severe and multiple-level spinal column injuries than other types of vehicular trauma (1). As with other spine trauma, thoracolumbar injuries occur most frequently in male patients between 15 and 29 years of age (2).
Approximately 90%of all spinal fractures occur in the thoracic and lumbar spines (3). In fact, the majority of thoracic and lumbar injuries occur within the region between T11 and L1, commonly referred to as the thoracolumbar junction (2). A variety of factors can explain this increased susceptibility. The thoracolumbar junction is a transition zone between the relatively stiff thoracic spine, stabilized by the costovertebral articulations, and more mobile lumbar spine. This area also exhibits significant alterations in flexion-extension and rotational degrees of freedom, as well as morphological and biomechanical changes in intervertebral disc architecture. Neurologic deficit reportedly occurs in approximately 15% to 20% of thoracolumbar fractures and dislocations; this is an injury combination that affects about 1 in every 20,000 people in the US (4,5). Although blunt trauma remains responsible for the majority of thoracolumbar fractures, other mechanisms such as gunshot wounds and osteoporosis have become increasingly more common (6,7).
BIOMECHANICS AND INJURY PATHOMECHANICS
Normal Mechanics
The thoracolumbar junction represents a biomechanical transition zone between the rigid thoracic spine and its more flexible lumbar counterpart. The rib cage, thoracic musculature, and facet joint configuration contribute to the relatively stiff thoracic spine. Additionally, the thoracolumbar junction is the area of change in sagittal alignment between the kyphotic thoracic segment and a lordotic lumbar region.
Between T1 and T6, there is approximately 4 degrees of flexion-extension permitted at each intervertebral segment, whereas approximately 12 degrees is permitted at T12–L1. In contrast, lateral bending between vertebrae decreases from proximal to distal, with 8 degrees allowed between thoracic levels, but only 2 degrees in the upper lumbar segments (8). This pattern is partially due to the more sagittal orientation of the facet joints in the lumbar spine.
Structural Injury and Pathomechanics
Determining the extent and nature of damage relies on a complex evaluation of the injury mechanism, the clinical examination, and x-ray studies. This analysis facilitates the choice of the most appropriate treatment plan.
Panjabi et al (9) demonstrated that only 1 mm of shear normally occurs between thoracic vertebrae. Furthermore, they showed that that failure of the intervertebral disc occurs with approximately 2.5 mm of displacement. Disruption of the costovertebral joints results in substantial increases in intervertebral motion within the thoracic spine (9).
Various forces act on the spine in traumatic situations, including axial loading, flexion, extension, shear, and axial rotation (10). Most often, damage occurs from a combination of these forces. Pure axial loads, or compressive forces, have been shown to result in end plate fractures, wedge (or compression) fractures, and burst fractures (11,12,13). The extent of the fracture depends largely on the energy and speed with which the force is delivered.
Flexion forces applied to the spine can result in a variety of injury patterns dictated by the axis of rotation. If the center of rotation occurs near the posterior longitudinal ligament, there will be a compressive load applied to the anterior vertebral body and a corresponding distraction force within the posterior elements. When the sagittal rotation centers about a point anterior to the spine, as may occur in a so-called seat-belt injury, primary distraction forces act on both the anterior and posterior elements. This leads to substantial midsubstance ligament failure and/or avulsion fractures, as the spinal column is poorly equipped to withstand such tensile forces.
Extension-type mechanisms, also known as lumberjack injuries, are not as common as flexion injuries (14). Such injuries produce tensile forces in the anterior spine with compressive or tensile forces applied to the posterior elements. Although pure axial load has long been thought to be the primary mechanism leading to burst fractures, recent biomechanical evidence suggests that adding some extension load is necessary to produce the characteristic burst fracture pattern with pedicle widening and fragment retropulsion (15).
INITIAL EVALUATION AND EMERGENCY CARE
Prehospital Care
The high incidence of concomitant adjacent spinal injuries warrants strict precautions for the entire spine when managed acutely in the field. A cervical collar and a flat spine board should be standard protocol during in-field evaluation and transport. Advanced trauma life support protocols should also be strictly adhered to during transport and resuscitation, with an emergent focus on airway, breathing, and circulation.
In-Hospital Resuscitation
Hypotension and bradycardia may be a sign of spinal shock in the unconscious trauma victim. Early supraphysiologic oxygen delivery, fluid resuscitation, and vasopressors can be helpful in their clinical management. It is critical to distinguish neurogenic
P.1545

shock from hypovolemic shock, as overly aggressive fluid administration in the former could lead to intravascular overload and pulmonary edema.
Strict spinal precautions should be maintained until a full and thorough evaluation can be performed on the alert patient. Moving the patient to and from various hospital beds or tables may require the use of a rigid transfer board in order to adequately protect an unstable spinal column from deleterious motion. A high index of suspicion should be maintained for spinal injury in polytraumatized patients. One study revealed that thoracolumbar injuries were missed in 24% of multitrauma patients (16).
History and Physical Examination
Accurate assessment of the spinal column in a trauma victim requires thorough initial and secondary evaluation. Distraction injuries in polytrauma patients often make the initial physical examination unreliable, which underscores the importance of the secondary survey. The initial evaluation should consist of a history and physical examination, focusing on a detailed neurologic assessment and direct palpation of the spinal column.
Neurologic evaluation is an essential step in the care of a trauma victim and should be performed before log-rolling a patient for posterior examination. Ideally, in a patient who is awake, these tests consist of a careful motor, sensory, and reflex examination, but a complete assessment is not always possible in an unconscious, obtunded, sedated, or polytraumatized patient. Grading motor strength should be performed on a scale from 0 to 5 (see Table 39-2 in Chapter 39). Sensory testing should include pain/temperature and vibratory/position/deep touch examination in all dermatomes. When spinal cord injury is present, neurologic injury can be scored according to the American Spinal Injury Association scale.
Perianal sensation, rectal tone, and the bulbocavernosus reflex must be assessed. The bulbocavernosus reflex can be elicited by stimulating the glans penis or clitoris or by a gentle tug of the Foley catheter. With these maneuvers, a corresponding involuntary anal sphincter contraction should occur. An absent bulbocavernosus reflex in a trauma victim can indicate spinal shock, provided the patient is not pharmacologically paralyzed. The prognosis for the potential of neurologic recovery cannot be determined until spinal shock has resolved. Spinal shock is generally considered to have ended either after 48 hours from injury or once the bulbocavernosus reflex returns, making this reflex an important clinical sign in the evaluation of spinal cord injury.
Anterior inspection can reveal abdominal and/or chest ecchymoses from a seat belt, which may be a clue to an underlying flexion-distraction injury (chance or seatbelt injury). The presence of a seat-belt injury should alert the clinician to the probability of an intra-abdominal injury, such as a splenic or liver laceration, which can be present in up to 45% of cases (17). The posterior soft tissues should be inspected, and the spine should be palpated by log-rolling the patient, making note of irregularities such as ecchymoses, bogginess, crepitus, open wounds, focal sites of tenderness, malalignment, or areas of palpable step-off.
Initial Imaging Protocol
In the alert and cooperative patient, the thoracic and lumbar spine can be “cleared” with the absence of pain or tenderness or distraction injuries, and a normal neurologic examination. In the presence of distraction injuries or after high-energy trauma, however, plain x-rays are usually obtained. Anteroposterior (AP) and lateral x-rays of the thoracic and lumbar spine should be obtained in such circumstances. With the advent of helical scanners, computed tomography (CT) with sagittal and coronal reconstructions is becoming an increasingly popular method of initial spine screening. If a spinal injury is detected, a dedicated CT scan and a series of plain x-rays of the region should be obtained. AP and lateral views centered at the thoracolumbar junction are useful in assessing suspected injuries in this region. They may not be optimally visualized at the extremes of standard thoracic and lumbar x-rays. Magnetic resonance imaging (MRI) should be considered acutely in the presence of a neurologic deficit despite normal x-rays and CT imaging.
IMAGING AND INJURY DESCRIPTION
Plain X-Rays
Injury Detection
Plain x-ray evaluation should begin with AP and lateral views. The AP view best demonstrates changes in coronal alignment, as well as changes in interpedicular distance and the space between spinous processes. Coronal translational deformity suggests a high-energy injury and mechanical instability. Interpedicular distance should generally increase as one moves from cranial to caudal along the spinal column, but comparison with adjacent levels provides a more reliable means of assessment. Abnormal widening of the interpedicular distance signifies lateral displacement of vertebral body fragments, typical of burst fractures. Abnormally increased distances between adjacent spinous processes suggest posterior ligamentous complex disruption (PLC) (18,19). Relative coronal or sagittal plane translation greater than 2.5 mm strongly suggests gross discoligamentous failure and instability (9).
The lateral view is, in general, more useful in characterizing and detecting injuries. Sagittal plane malalignment should be studied carefully. Quantification of sagittal plane alignment can be performed using the Cobb method (Fig. 40-1), which involves measuring the angle created by the intact superior and inferior endplates of the adjacent uninjured segments (20). Vertebral body height loss can be measured by comparing the height of the injured level with adjacent uninjured vertebrae. It is expressed as a percentage of the expected normal height (Fig. 40-1). Separate comparisons between anterior and posterior vertebral body height measurements can provide greater accuracy in assessing height loss in the presence of a uniformly
P.1546

compressed vertebra. Breaks in the posterior vertebral body line can indicate sagittal translational deformity (Fig. 40-2). The anterior vertebral body line can also be used, but this measurement is often limited by the presence of degenerative change.
FIGURE 40-1 Vertebral body height loss and kyphosis are thought to be important indicators of PLC disruption. Vertebral body height loss can be measured anteriorly and posteriorly. Measurements of the injured vertebra (A,B) are compared to those of the suprajacent (a′ and p′) and infrajacent (a″ and p″) uninjured vertebra. The Cobb kyphosis is measured by the angle subtended between a line drawn along the superior endplate of the suprajacent uninjured vertebra (T12, dashed line) and the inferior endplate of the infrajacent uninjured vertebrae (L2, dashed line).
Before the availability of MRI, the determination of posterior ligamentous disruption was based solely on the physical and plain x-ray findings. This led to the development of various threshold criteria above which PLC disruption would be strongly suspected (21,22,23,24,25,26). Segmental kyphosis greater than 30 degrees has been suggested by several investigators as a critical point beyond which PLC disruption is likely (21,22,23,24,25,26). This finding is particularly relevant at the thoracolumbar junction, where absolute measurements of kyphosis represent both true and relative kyphosis, as the thoracolumbar junction is normally straight in the sagittal and coronal planes. Based on in vitro biomechanical data (27), a loss of more than 50% of vertebral body height is another strong indicator of posterior instability. Despite the reproduction of these criteria in various textbooks and articles throughout the years, there is little to no supportive clinical evidence that they are reliable indicators of PLC injury. The presence or absence of PLC disruption can occur with injuries that do not fulfill or exceed these criteria, respectively.
FIGURE 40-2 A line drawn along the posterior vertebral bodies (PVB) should normally be continuous. A break in this line indicates translational deformity, as is seen in the lateral x-ray of a lumbar dislocation.
Computed Tomography
CT scans provide finer detail of the bony involvement in thoracolumbar injuries. Plain films can underestimate the extent of canal compromise by more than 20% when compared to CT scans (28). Recent data have suggested that rapid chest, abdomen, and pelvis screening with CT scans is equally sensitive to (but more efficient than) standard plain x-ray protocols in assessing thoracolumbar spine trauma (29,30). In areas of concern or where injury has been detected, fine cuts (less than 2 mm) should be obtained, along with sagittal and coronal reconstructions (31). Axial images readily demonstrate the degree of canal compromise from retropulsed fragments, but canal compromise from translational deformities, such as dislocations, can be easily missed with axial images alone. This fact highlights the importance of sagittal and coronal reconstructions if CT is used for injury detection alone. CT images also allow accurate assessment of the degree of comminution within the affected vertebral body, which may influence treatment (32,33).
Posterior element fractures are easily missed on plain x-rays.
P.1547

Fractures of the pedicles, laminae, facets, and transverse processes are best detected and assessed with CT. Lamina fractures in conjunction with a burst fracture can be associated with dural tears and nerve root entrapment (34,35). The so-called empty or naked facet sign is a characteristic sign of facet dislocation and may present concurrently with a so-called double-body sign (Fig. 40-3).
Magnetic Resonance Imaging
MRI provides superior visualization of the spinal cord and soft tissues. An MRI is useful in evaluating a patient with a neurologic deficit that either does not correspond to the level of bony or ligamentous injury or is present in the absence of a bony or ligamentous injury. In such settings, MRI can help visualize disc herniations, epidural hematomas, or spinal cord edema that would not be detectable by other imaging modalities. Such findings can have a significant impact on treatment and prognosis. For example, spinal cord edema, seen as fusiform increased intensity on T2-weighted images (Fig. 40-4), may be associated with a poorer prognosis for motor recovery (36).
With a growing trend toward nonoperative treatment of stable burst fractures, MRI has become an important tool in assessing the integrity of the PLC (37,38,39,40). T2-weighted imaging with fat suppression (also known as a Short Tau Inversion Recovery, or STIR, sequence[mh55]) [fn3]highlights edematous signals within soft tissues, such as ligaments, and has proven useful for detecting PLC injury (40). Determining the severity of ligamentous injury can be difficult, as the amount of edema may under- or overestimate the degree of disruption. Some investigators have established grading systems to provide assistance in interpreting MRI findings (38,39), however, such systems have yet to be prospectively evaluated. In the authors’ experience, inspecting the anatomical continuity of the ligamentum flavum and interspinous process ligaments can be a useful method of determining PLC disruption (Fig. 40-4).
FIGURE 40-3 The empty or “naked” facet sign is an indication of a high-energy dislocation of the thoracolumbar spine. It represents an unmatched articular process.
FIGURE 40-4 Fusiform swelling and increased signal intensity on T2-weighted images suggest substantial injury to the spinal cord.
The use of MRI to evaluate gunshot wounds of the thoracolumbar spine is controversial (6,41). Bullet fragment migration from the magnet’s pull can potentially lead to further neurologic or soft tissue damage (41), although published reports of its use in this setting have not supported this concern (42). While it can sometimes allow better visualization of the neural elements, artifact scatter from the ferromagnetic fragments often compromises image quality (41,43,44). In the authors’ experience, the most common complaint from patients undergoing MRI scans after gunshot wounds is a complaint of a heat sensation in the area of the bullet (particularly jacketed types), which can lead to discomfort and premature termination of the study. Ultimately, the use of MRI should be judged on an individual case-by-case basis.
P.1548

Additional Studies
There are no formal guidelines for the use of flexion-extension x-rays in assessing traumatic injury of the thoracolumbar spine; however, such imaging is reportedly used in various institutions. In the authors’ practice, standing x-rays in a brace are routinely obtained to assess increasing deformity associated with nonoperative treatment of burst fractures. A recent study found that information from upright weight-bearing x-rays after thoracolumbar trauma changed the treatment from nonsurgical to surgical in 25% of cases (45).
CLASSIFICATION
Various classification schemes have been proposed to describe thoracolumbar fractures and dislocations. Advancements in imaging technology have increased our understanding of their pathomechanics, and have therefore improved their classification. The ideal classification scheme should allow reliable clinical application, provide a consistent prognosis and, ultimately, optimization of treatment decisions. As with other fractures, this has not, as yet, occurred.[iu5]
In 1943, Watson-Jones identified comminuted wedge fractures as a subset of thoracolumbar injuries (46); these are now commonly referred to as burst fractures. Later that decade, Nicoll (47) created the first well-known classification system for thoracolumbar trauma, which was stimulated by his review of injuries sustained by coal miners. The system focused on morphological differences between various injury patterns. About a decade later, Holdsworth (19) proposed that burst fractures were the result of extreme flexion and further divided these injuries into stable or unstable patterns. His assessment of stability depended on the integrity of the PLC. Based on Holdsworth’s work, Kelly and Whitesides (48) developed a two-column concept of spinal stability in which they delineated an anterior column consisting of the vertebral body, disc, anterior longitudinal ligament (ALL), and posterior longitudinal ligament (PLL), and a posterior column, which was formed by the posterior neural arch, facet joints, and PLC (Fig. 40-5).
FIGURE 40-5 Holdsworth suggested that the spine should be divided into two columns. The anterior column is composed of the vertebral body, disc, and associated ligaments The posterior column is composed of the pedicles and neural arch, facet joints, and associated ligaments. Denis suggested that the spine should be divided into three columns. Whereas the posterior column remained the same, the anterior column consisted of the anterior portions of the vertebral body, disc, and ALL, and the middle column consisted of the posterior portion of the vertebral body, disc, and PLL.
The next step was Denis’ (5) development of the three-column theory of spinal instability upon which he later based a classification system of thoracolumbar injuries (Fig. 40-5). The system contained four main categories of injury: compression fractures, seat-belt injuries, burst fractures, and fracture-dislocations (Figs. 40-6, 40-7, 40-8, 40-9). These were further divided into 16 subgroups. Aimed mainly at providing information on the morphology and injury mechanism, this classification scheme has been less useful for grading stability or assigning treatment. The distinguishing feature of the Denis classification is the importance placed on the middle column, made up of the posterior vertebral body, posterior disc, and PLL. Using Denis’ definition, injury to this middle column indicates mechanical instability, and by nature of their fracture morphology, all burst fractures would therefore be considered unstable. Although Denis’ classification scheme is an attractive system that gained popularity throughout the years, its basic premise has not enjoyed clinical support, as the literature contains a number of reports of burst fractures that have been successfully treated by nonoperative methods, provided the PLC is intact. In addition to these shortcomings, recent analysis has found the inter- and intraobserver reproducibility of the Denis system to be relatively low (38).
Recognizing this problem with the Denis system, McAfee et al (32) proposed another classification scheme. Using x-rays and CT analysis of 100 thoracolumbar injuries, they proposed that the mechanism of failure of the middle column, by axial compression, axial distraction, or translation in the transverse plane, could be determined, which would influence stability. The McAfee system divided injuries into six categories: wedge-compression fractures, stable burst fractures, unstable burst fractures, Chance fractures, flexion-distraction injuries, and translational injuries. This system was devised before pedicle screw instrumentation was available at a time when hook-based systems were the most commonly used instrumentation systems. The McAfee system indicated which injuries would be best treated by distraction or compressive constructs—an important consideration at the time. Since its publication, its inter- and intraobserver reliability has not been tested, nor has it been validated prospectively.
The most comprehensive system in current use is the AO classification system (49), which was developed after reviewing the x-rays and CT scans of 1,400 injuries in a number of trauma centers. It divides injuries into three basic groups based on the primary mechanism of failure. These groups are compression (group A), distraction (group B), and torsional or rotational forces (group C) (Fig. 40-10). Further subgroups were developed to characterize the fracture location and morphology, as well as to distinguish between osseous or ligamentous disruption and the direction of displacement. Group A injuries result
P.1549

primarily from axial or compressive loads, with or without flexion, with limited involvement of the posterior elements. With greater forces, there may be significant height loss, and vertebral body fragments can be retropulsed into the spinal canal in neurologic compromise. Group B injuries are the result of distraction forces and are by definition inherently unstable. Group C, or rotational injuries, also imply gross ligamentous injury. These injuries are often associated with transverse process fractures, costovertebral dislocations, translational malalignment between vertebral bodies, and frequently, neurologic deficit. Like the Denis classification, however, the inter- and intraobserver reliability of the AO classification system is limited (38).
FIGURE 40-6 Denis classification of compression fractures. Type A involves both endplates, type B involves the superior endplate, and type C involves the inferior endplate. In type D fractures, there is a compression fracture of the anteriovertebral body.
After the introduction of pedicle screws, posterior stabilization of thoracolumbar burst fractures became popular, as did so-called short-segment stabilization, which involved posterior instrumentation of only the levels above and below the injury. Reports of high failure rates with this construct prompted McCormack et al (33) to devise a classification scheme specifically intended to predict when a short-based construct would fail. Named the “Load-Sharing Classification,” it assigned a point value to the degree of vertebral body comminution, fracture fragment apposition, and kyphosis (Fig. 40-11). Based on their primary outcome of hardware failure, McCormack et al concluded that injuries with scores greater than 6 points would be better treated with the addition of anterior column reconstruction. A recent study demonstrated very high inter- and intraobserver reliability of this classification system (50).
Descriptive Classification of Thoracolumbar Injuries
Despite disagreement about a unified classification system, a number of thoracolumbar injury patterns are consistently reported in the literature. It must be kept in mind, however, that these injuries often represent different stages of a continuum; many
P.1550

of these injuries share similar characteristics and many are considered to be “unclassifiable” by most systems. In general terms, however, thoracolumbar fractures can usually be described as one of the following:
FIGURE 40-7 Denis classification of burst fractures. Type A involves fractures of both endplates, type B involves fractures of the superior endplate, and type C involves fractures of the inferior endplate. Type D is a combination of a type A fracture with rotation. Type E fractures exhibit lateral translation.
Compression Fractures
Compression fractures are isolated wedge-type fractures of the anterior and middle aspects of the vertebral body. Unlike burst fractures, there is no involvement of the posterior vertebral body. Whereas most compression fractures are thought to be stable, more severe injuries may be associated with PLC disruption.
Burst Fractures
Burst fractures show comminution of the vertebral body with involvement of the posterior cortex. Most of these fractures (but not all) will show some degree of fracture fragment retropulsion. Burst fractures can occur with or without PLC disruption. Unlike fracture-dislocations, burst fractures do not demonstrate translational deformity. This is an important distinction, as many fracture-dislocations can be associated with comminuted vertebral body fracture patterns. Keep in mind that translational malalignment implies circumferential ligamentous disruption.
Flexion-Distraction (Seat Belt or Chance Type) Injuries
Despite their name, most of these injures probably do not occur from a pure flexion-distraction mechanism. In many cases, a component of vertebral body compression or lateral translation may be noted (108), suggesting that the center of rotation was probably not anterior to the spine. The characteristic feature of a flexion-distraction injury is primarily tensile failure of the PLC, facet capsules, and intervertebral disc or bone itself. Injuries rarely occur through bone alone and are most commonly the result of osseous and ligamentous failure. As in the subaxial cervical spine, flexion-distraction may result in facet dislocation or subluxation in conjunction with PLC failure. These injuries are often overlooked and are easily missed during initial evaluation (51).
Fracture-Dislocations
Fracture-dislocations of the thoracolumbar spine are highly unstable injuries. Although they may display various combinations
P.1551

of damage to both the anterior and posterior elements, the unique injury feature is translational deformity, which can occur in the sagittal and/or coronal planes.
FIGURE 40-8 Denis classification of flexion-distraction injuries. Types A and B occur at one level, either through bone (A) or ligament (B). Type C and D occur at two levels (motion segments). Type C denotes that the middle column failed through bone. Type D denotes that the middle column failed through ligament and disc.
APPLIED ANATOMY OF THE THORACOLUMBAR SPINE
The thoracolumbar junction is subject to unique stresses within the spinal column. This exposes the underlying structures to a greater risk of injury. Understanding the anatomy of the osseo-[jyligamentous and neurologic structures is important in optimizing initial evaluation and treatment of injuries in this region.[fn2]
Anterior Elements
The vertebral bodies of the thoracic and lumbar spine are relatively similar in cylindrical configuration, although their dimensions increase from cranial to caudad. The intervertebral discs lie between the endplates of adjacent vertebral bodies and comprise the outer annulus fibrosis and the nucleus pulposus (Fig. 40-12). The annulus fibrosis is composed of an organized collagen framework that intimately integrates with both the anterior and posterior longitudinal ligaments. In the central portion of the disc lies the soft, gel-like nucleus pulposus. In general, intervertebral disc height progressively increases from cranial to caudad.
The ALL is composed of superficial fibers that span several vertebral levels. Deeper fibers connect adjacent vertebrae. The main function of the ALL is to restrict extension of the spinal column. The PLL has a similar layered configuration, but it functions to limit overdistraction of the disc space during flexion.
Posterior Elements
Pedicles originate from the posterosuperior aspects of the vertebral bodies and serve to connect the anterior elements with the
P.1552

P.1553

P.1554

posterior neural arch. A relatively thick cortex and short tubular morphology makes the pedicles the strongest portion of the vertebrae and an uncommon site of fracture (52). From each pedicle, laminae extend posteromedially to fuse in the midline to give rise to a spinous process.
FIGURE 40-9 Denis classification of fracture-dislocations. Type A are bony one-level injuries. Type B are one-level ligamentous injuries. Type C injuries are two-level injuries that occur through bone and/or ligament.
FIGURE 40-10 The three major injury groups of the AO classification. Type A injuries are compressive to the anterior column only. Type B injuries involve the anterior and posterior columns by translational forces. Type C injuries involve the anterior and posterior columns by rotational forces.
FIGURE 40-11 The Load-Sharing classification of thoracolumbar burst fractures assigns points according to the degree of comminution, apposition of the fragments, and the degree of kyphosis.
FIGURE 40-12 The intervertebral disc is composed of a laminate, layered fibrous ring known as the annulus fibrosis. It surrounds a soft, gel-like center known as the nucleus pulposus.
The ligamentum flavum is an elastic, yellow, collagenous tissue that spans the interlaminar spaces. The stronger and more fibrous interspinous and supraspinous ligaments connect the spinous processes (Fig. 40-13). This collection of posterior ligaments is often termed the PLC. Its integrity is critical to stability after trauma. Radiographic widening of the space between the bispinous processes or discontinuity detected on a magnetic resonance image may represent disruption of this complex.
Transverse processes [fn3]extend laterally from the junction of the pedicle, the superior articular process, and the pars interarticularis (Fig. 40-13). In the thoracic spine, they are angled posterolaterally to accommodate the medial aspect of the rib, which articulates with the anterior aspect of the transverse process and spans the posterolateral disc space. The transverse processes are important sites for ligamentous and paraspinal muscle attachments. Each costovertebral junction is stabilized by ligaments that attach to two adjacent vertebral bodies, as well as the transverse processes. This interaction is responsible for much of the increased stability of the thoracic spine compared to the cervical and lumbar spines (27).
FIGURE 40-13 Anatomical diagram of the thoracolumbar spine.
Facets, or zygapophyseal joints, are complex synovial articulations between the posterior elements of adjacent vertebrae. The superior articular process for each of these facets arises from the cephalad portion of the laminae, while the inferior articular process arises from the caudal border of the laminae. Facet joint orientation varies markedly between and within the thoracic and lumbar spines. Coronal plane orientation in the upper and middle thoracic spine provides greater resistance to anteroposterior translation, while the more sagittally oriented joints of the lower thoracic and upper lumbar spine allow greater flexion-extension flexibility with more resistance to medial-lateral translation (52).
Spinal Canal and Canal Compromise
The borders of the spinal canal in the thoracic and lumbar spine are similar to those in the cervical spine:
  • Anterior borders: vertebral body, intervertebral disc, posterior longitudinal ligament
  • Lateral borders: pedicles, medial aspect of facet joints
  • Posterior borders: ligamentum flavum, laminae
P.1555

Dimensional compromise of the spinal canal can occur after injury, and the most common cause is posterior bony retropulsion from a burst fracture of the vertebral body. Canal space can also be diminished in dislocations and fracture-dislocations where there is translation between adjacent vertebrae. Less frequently, anteriorly displaced fractures of the laminae can result in canal compromise (Fig. 40-14). Hematomas and disc herniations are potential nonbony sources of neural compression.
Spinal Cord
The most distal aspect of the spinal cord, known as the conus medullaris, is variably located between T12 and L3. The conus medullaris contains the distal sacral nerve cell bodies. The nerve rootlets distal to the conus medullaris are bathed with cerebrospinal fluid within the dural sac of the cauda equina. They descend to form nerve roots.
The ratio of the axial [fn2]spinal canal dimensions to the spinal cord dimensions are smallest in the T2–T10 region, which makes this area prone to neurologic injury after trauma. In fact, complete spinal cord injury is six times more common than incomplete injury with high-energy trauma to the midthoracic spine (36). The vascular supply to the spinal cord also plays an important role in the susceptibility for thoracic spinal cord injury. The region between T2 and T10 is a circulatory watershed area, deriving its proximal blood supply from antegrade vessels in the upper thoracic spine and distally from retrograde flow from the artery of Adamkiewicz, which can be variably located between T9 to L2.
FIGURE 40-14 Canal compromise rarely occurs from displaced fractures of the posterior elements. In the axial CT scan, a large piece of fractured lamina intruded on the spinal canal of this patient who had sustained a high-energy injury to the thoracic spine.
Spinal cord anatomy involves a complex spatial arrangement of gray (neurons) and white (axons) matter, which can be assessed indirectly using physical examination to detect functional losses. Upper motor neuron axons cross from the contralateral cerebral cortex and travel predominately in the lateral corticospinal tracts. Motor axons destined for cervical synapses and upper-extremity supply are most central within the cord, whereas those for the lower extremities are peripherally located. These motor axons ultimately terminate in the anterior horn of the gray matter, which synapse with corresponding lower motor neurons or interneurons near their root exit point.
Sensory input occurs through peripheral afferent neurons, which terminate in the dorsal root ganglia and pass information onto the dorsal horn gray matter of corresponding cord levels. Sensory input then ascends ipsilaterally along various pathways to the brain. The dorsal columns are composed of axons transmitting pressure, touch, vibratory, and proprioceptive sensation.
The lateral spinothalamic tracts transmit pain and temperature sensation, but the impulse crosses to and ascends within the contralateral side of the cord within a few levels of entry. Consequently, injury to one side of the spinal cord, as with Brown-Sequard syndrome, results in ipsilateral light touch, proprioception, and motor loss with contralateral pain and temperature loss.
TREATMENT OF SPECIFIC INJURIES
Compression Fractures
Most compression fractures in the thoracolumbar spine are considered stable, and can therefore be treated nonoperatively. In the setting of normal neurologic function, vertebral body height loss and kyphosis have been used as radiologic parameters to assess the degree of instability rendered by the injury. Patients with fractures exhibiting minimal height loss (less than 10%) can usually be safely mobilized without the use of a brace or an external support. Most practitioners consider a compression or wedge fracture with less than 30% or 40% of vertebral body height loss and less than 20 degrees to 25 degrees of kyphosis to be inherently stable. It is the authors’ preference, however, to prescribe a brace in these cases. A Jewett hyperextension brace is applied for 6 to 8 weeks. Standing x-rays in the brace should be obtained both initially and at regular follow-up visits to monitor fracture healing and alignment.
A height loss of more than 50% or more than 30 degrees of kyphosis in nonosteoporotic bone strongly suggests the possibility of PLC disruption, which places the patient at risk of
P.1556

increasing kyphotic deformity or neurologic deficit (53). Interspinous widening on lateral x-rays may strengthen this suspicion, and give cause for concern if nonoperative management is used. In such cases, posterior stabilization is usually recommended (21,22,23,24,25,26,27,54). In the authors’ practice, an MRI is obtained in equivocal cases to examine the integrity of the PLC.
Overview of Braces
External immobilization (bracing or casting) is the most common form of nonoperative treatment used today. Hyperextension devices, such as a Jewett brace, can counteract sagittal flexion forces although they offer minimal resistance to rotation or lateral flexion. Custom fit, clamshell devices, such as a thoracolumbar sacral orthosis (TLSO) provide better multiplanar support, which may be more appropriate for burst fractures. A number of studies have evaluated the ability of external orthoses to immobilize the thoracic and thoracolumbar spine (55,56,57). Most braces have shown increased motion across the lumbosacral junction (55,57). One study compared a corset, a rigid brace, and a plaster jacket with and without a thigh cuff. It demonstrated that plaster jackets provided the best immobilization of the L1–L3 segments. Prefabricated, adjustable thoracolumbar or thoracolumbosacral orthoses have recently been developed to accommodate body habitus changes such as weight loss. They have demonstrated similar efficacy in sagittal and lateral bending resistance, however, their efficacy in rotational immobilization has not been well studied (58). For fractures at or above the T5 level, a cervical component should be added to the brace, as under-arm braces are ineffective in immobilizing these levels.
Folman and Gepstein (59) retrospectively reviewed their results of the nonoperative treatment of 85 patients with traumatic wedge-compression fractures of the thoracolumbar spine. With a minimum of 3 years follow-up, 69% of patients continued to have chronic low back pain and pain intensity correlated with the degree of focal kyphotic deformity. The use of physical therapy or bracing did not influence outcomes. Gertzbein’s (2) findings from a prospective, multicenter study were similar. He found that the presence of a kyphotic deformity correlated with chronic back pain, and more severe pain was related to focal kyphosis greater than 30 degrees.
Pearls
  • Fractures with <10% vertebral height loss do not need external support.
  • Fractures with <30% to 40% height loss and <20 degrees to 25 degrees kyphosis can be treated with a Jewett brace for 6 to 8 weeks.
  • In fractures below T5, a plaster jacket or TLSO can be used.
  • In higher fractures, a cervical component should be added to the brace.
Pitfalls
  • Not all compression fractures should be considered stable.
  • 50% height loss or >30 degrees kyphosis suggests PCL disruption, and posterior stabilization is recommended.
  • An MRI scan should be used to examine the integrity of the PCL.
Burst Fractures
Nonoperative Treatment
Burst fractures usually occur as a result of higher-energy trauma. As with compression fractures, the use of nonoperative management should be decided by the presence of an intact PLC. The x-ray guidelines for burst fractures are similar to those described for compression fractures. Some surgeons have used the amount of canal compromise as a parameter for surgical decision making (60,61). Trafton et al (60) found early neurologic deterioration in patients with greater than 50% canal compromise if there were retropulsed fragments, but it is unclear if these patients also had PLC disruption. Despite such reports, there is little published evidence to suggest a critical threshold for canal compromise that would indicate the need for surgical treatment in stable thoracolumbar burst fractures, because of the potential for neurologic decline (62,63,64).
Bracing can be used to immobilize stable burst fractures. A form-fitting TLSO or extension cast is usually prescribed for 3 months. Once the brace is fitted, the patient should have standing x-rays in the brace to ensure stability. Look for substantial loss of height or increase in kyphosis, as they indicate underlying posterior ligamentous insufficiency. If the alignment is minimally changed, log-roll precautions can be discontinued, and the patient should be mobilized with physical therapy. Frequent x-ray follow-up is recommended.
A nonoperative approach may rarely be warranted in patients with a neurologic deficit or an unstable burst fracture but who are medically unfit for surgery. Burke and Murray (65) supported this approach and found that surgery provided no additional neurologic benefit in patients with thoracolumbar injuries, although their report was not limited to burst fractures. Dendrinos et al (66) demonstrated that neurologic outcomes were similar with or without surgery after thoracolumbar burst fractures associated with a complete neurologic injury. Despite a lack of substantive clinical evidence, most surgeons would
P.1557

agree that the presence of a neurologic deficit with a burst fracture is usually an indication for operative management.
FIGURE 40-15 AP (A) and lateral (B) x-rays of a stable compression fracture of L3. The patient was successfully treated in a brace (C).
In a review of 42 patients with stable thoracolumbar burst fractures, Weinstein et al (67) found that nonoperative treatment, consisting of bed rest or body casts, offered promising results. All patients were neurologically iflntact, and the average kyphosis was 26 degrees. In contrast to the outcomes reported by Folman and Gepstein (59) and Gertzbein (2) for compression fractures, there was no association between the degree of deformity and persistent back pain. No neurologic deterioration was noted, and 88% of the patients returned to work.
Shen and Shen (68) recently reviewed the results of nonsurgical treatment of 38 patients who sustained thoracolumbar burst fractures without neurologic deficit. Most patients were treated with early ambulation without a brace. The average follow-up was 4 years. Mean focal kyphosis increased from 20 degrees to 24 degrees, although persistent pain was present in only 10% to 15% of cases. No correlation was shown between kyphotic deformity or canal compromise and clinical outcome.
Aligizakis et al (69) studied the results of nonoperative treatment using an orthosis in 60 patients with thoracolumbar burst fractures without neurologic deficit. The average follow-up was 42 months. After treatment, 91% of patients had a satisfactory functional outcome, and 83% had little or no pain. The average initial kyphosis was only 6 degrees, which worsened to 8 degrees at final follow-up.
Chow et al (70) presented the results of brace treatment in 24 neurologically intact patients with so-called unstable thoracolumbar burst fractures. The fractures were deemed unstable according the Denis classification system. Only 10 fractures demonstrated interspinous process widening that would suggest PLC injury and true instability, and nine patients returned to work after treatment was completed.
Nonoperative Versus Operative Treatment
The results of surgical and nonsurgical treatment for thoracolumbar burst fractures have been compared. One study prospectively compared outcomes of patients treated with either posterior surgery or nonoperative management for thoracolumbar burst fractures without neurologic deficit (71). Injuries that involved the posterior arch, such as facet dislocations, were excluded. A hyperextension brace and early ambulation was the treatment in 47 nonoperatively managed patients. Short-segment pedicle screw fixation was used in 33 patients in the surgical group. Although outcome scores were better in the operative
P.1558

group at 3 months, the 6-month and 2-year scores were not statistically different. Kyphosis correction was also initially better in the surgical group, but this benefit was not demonstrable at final follow-up. Pain scores followed a similar pattern.
FIGURE 40-16 Lateral x-ray (A) of a seemingly benign compression fracture of T11. Because a horizontal fracture through the pedicle was suggested, an MRI was obtained, which demonstrated a suspicion of a PLC injury on the STIR images (B) noted by increased signal intensity in the interspinous process space and on the T2-weighted images (C) by discontinuity of the ligamentum flavum. Intraoperative examination confirmed the presence of posterior ligamentous injury. The fracture was stabilized with a posterior pedicle screw construct (D,E).
In a well-designed, randomized prospective study, Wood et al (72) compared operative and nonoperative treatment of stable thoracolumbar burst fractures in patients without neurologic deficit. Importantly, they excluded injuries with suspected or confirmed posterior ligamentous disruption. Nonoperative care included a brace or cast followed by early mobilization. Operative treatment was either anterior or posterior surgery. There were no statistical differences in kyphosis, functional outcome, or pain scores between the two groups. Operatively treated patients, however, tended to have higher pain scores and more complications.
P.1559

Surgical Treatment
Anterior and posterior approaches have been advocated for decompression and stabilization. Both have advantages and disadvantages. Although there is considerable controversy regarding the optimal surgical treatment of thoracolumbar burst fractures, there is a general consensus that a progressive neurologic deficit is a strong indication for urgent decompression of the spinal canal and instrumented stabilization. Otherwise, the relative indications for surgery after a thoracolumbar burst fracture are the presence of a PLC injury, the presence of a neurologic deficit with or without an unstable fracture, or a rapidly increasing kyphotic deformity. In addition, fractures in multiply injured patients or those in whom bracing cannot be effectively employed, due to other injuries or body habitus, may benefit from internal stabilization.
Posterior Surgery
Advocates of posterior surgery cite various advantages compared to anterior surgery for thoracolumbar burst fractures. First, it avoids the morbidity of anterior exposure in patients who potentially have concomitant pulmonary or abdominal injuries. Second, several studies have shown that the posterior approach involves shorter operative times and decreased blood loss, and functional outcomes are similar to those following anterior surgery (73,74,75). Posterior instrumentation alone cannot reconstitute anterior column support, however, and is therefore somewhat weaker in compression than anterior instrumentation (76). This has led to a higher incidence of progressive kyphosis and instrumentation failure when treating highly comminuted fractures (33).
The role of posterior [fn2]surgery for burst fractures is primarily for realignment and stabilization (Fig. 40-17). Although hook-based and sublaminar wire constructs have been used in the past, pedicle screw instrumentation is now the most frequently used method of stabilizing thoracolumbar fractures (77,78,79,80,81,82). In contrast to hooks or wires, pedicle screw constructs provide greater three-point fixation strength that allows the use of shorter constructs that fuse fewer levels. Many surgeons have opted to use short-segment fixation, which involves pedicle screw placement one level above and one level below the injured segment. Short-segment fixation has been shown to have higher rates of pedicle screw failure, however, which is thought to result from cyclic loading combined with poor anterior column support (Fig. 40-18) (79,83). Transpedicular bone grafting has been suggested as a means of improving the anterior column support in this setting, but the results of its use in several studies have been disappointing (79,84). Some surgeons have suggested using fixation two levels above and below the injured segment in severely comminuted vertebrae, particularly if the bone quality is poor, or in areas of higher stress concentration, such as the thoracolumbar junction (85). Increasing the length of a posterior fusion segment should be considered carefully, especially in the more mobile segments of the lumbar spine, as short-segment fixation can also be supplemented by an adjunctive anterior procedure.
FIGURE 40-17 Lateral x-ray of a short-segment construct used to stabilize an L3 burst fracture.
P.1560

Although the authors prefer to perform an anterior decompression if there is a neurologic deficit, the use of laminectomy decompression with posterior stabilization has been reported (85). Laminectomy should only be performed as an adjunct to instrumented stabilization, as simple decompressive laminectomy will lead to additional posterior instability and progressive kyphotic deformity. If a laminectomy is performed, the inferior half of the lamina from the vertebra above the injury should be removed to allow adequate decompression, as the retropulsed fragments usually originate from the superior aspect of the vertebral body.
FIGURE 40-18 Lateral x-ray of a failed posterior pedicle screw construct used to stabilize a burst fracture with substantial height loss of the vertebral body.
Indirect canal decompression relies on ligamentotaxis generated by longitudinal distraction. Theoretically, it requires continuity of the posterior longitudinal ligament and/or posterior annulus to produce fragment reduction. Clinical experience with indirect decompression shows varied results (86,87). Katonis et al (87) found only 19% improvement in canal clearance after longitudinal distraction, and speculated that failures were probably due to PLL disruption. In cases with suspected PLL injury, the investigators recommended an anterior approach for decompression. Timing is another important factor with indirect decompression. Delay in surgical intervention beyond 1 week allows considerable consolidation of the fractures, and therefore diminishes the potential for successful indirect decompression (88). Some surgeons have suggested even shorter time-constraints, with results being compromised if surgery is delayed more than 24 to 72 hours (89).
A posterolateral technique of clearing the anterior spinal canal has been described (4,90,91). Removing the facet joint and/or the pedicle provides improved access to the posterior vertebral body. The retropulsed fragments are then reduced by pushing them anteriorly into the residual vertebral body. Despite improved anatomical canal clearance with this technique, studies have demonstrated no additional neurologic benefit over posterior instrumentation alone (92).
The presence of a vertical lamina fracture in conjunction with a thoracolumbar burst fracture is reportedly a risk factor for traumatic dural lacerations and nerve root entrapment (34,35). In this setting, some surgeons have recommended posterior surgery with decompression of the nerve roots and repair of the dura before any planned anterior procedures. It is unclear, however, if this treatment results in any improvement in clinical or neurologic outcomes.
Anterior Surgery
Anterior surgery for thoracolumbar burst fractures is primarily indicated for decompression of the neural elements. It provides direct visualization of the anterior thecal sac and is the most reliable method of spinal canal decompression. Although the literature seems to suggest that anterior decompression results in greater neurologic improvement than posterior decompression (21), no prospective, randomized comparison has been undertaken to demonstrate this difference. Bradford and McBride (21) found that return of normal bowel and bladder function occurred more frequently in patients decompressed anteriorly than in those treated posteriorly (69% vs. 33%, respectively). Another benefit of an anterior approach is restoration of anterior column support. Severely comminuted or unstable burst fracture patterns are more prone to hardware failure or late collapse with posterior fixation alone. Reconstruction of the anterior column can be effected with or without instrumentation, which provides greater mechanical stability and helps prevent late collapse in more unstable comminuted burst fractures than posterior constructs alone (93,94).
If an anterior corporectomy is undertaken to decompress the spine, it must be understood that the spine is always destabilized. To remedy this condition, the vertebral body defect must be replaced with a supportive strut that will also result in bony fusion. Structural autograft can be harvested from rib, fibula, or the iliac crest. Allograft bone, typically femoral or humeral shafts, can also be used. Recently, titanium mesh cages have become popular (95). Such devices can be filled with harvested morcelized autograft. Anterior instrumentation is usually recommended to add further stability.
Numerous constructs can be used for anterior stabilization. A number of in vitro biomechanical studies have compared various anterior devices (83,93,96,97). Among the more rigid devices are plates with fixed-angle vertebral body screws and
P.1561

cross-linked rod-screw-staple designs, such as the Kaneda system (Fig. 40-19).
Kostuik performed anterior surgery for thoracolumbar burst fractures in 35 patients with neurologic injury (98). Iliac crest autograft struts were used to reconstruct the anterior column, which was stabilized with a modified Harrington rod-screw system extending between the vertebral bodies above and below the fractured segment. High rates of neurologic recovery were documented in incomplete injuries, but patients with complete neural damage showed no improvement. Nineteen of 21 patients with initial incomplete loss of bowel or bladder function showed some improvement at final follow-up. Twelve of 13 patients that were treated within 10 days of injury recovered useful bladder function, whereas only five of eight patients treated more than 10 days from injury gained useful bladder function without incontinence.
Kaneda et al (99) have reported the largest series of surgical stabilization for burst fractures. This series included 150 consecutive patients with thoracolumbar burst fractures with neurologic deficits treated with a single-stage anterior decompression, strut grafting, and instrumentation using a rod-sleeve-staple device. The average canal clearance was nearly 100%, whereas the fusion rate was 93%. Anterior pseudarthroses were successfully treated by posterior instrumentation and fusion. One hundred forty-two patients (95%) improved at least one Frankel grade.
The use of plate fixation has also been reported. Haas et al (100) documented excellent rates of neural recovery and maintenance of alignment using anterior decompression, strut grafting, and an anterior plate for a variety of thoracolumbar fractures. Ghanayem et al (101) reported their results using an anterior plate in 12 patients with thoracolumbar burst fractures. In 10 cases, postoperative kyphosis correction was maintained at 1-year follow-up. The other two patients had 10 degrees to 20 degrees of correction loss. Both of these patients had a preoperative kyphosis measuring greater than 50 degrees, suggesting that anterior surgery alone may not be sufficient to stabilize injuries associated with significant posterior ligamentous disruption. Construct failure may also have been related to the choice of Z-plate implant.
FIGURE 40-19 The Kaneda device is a cross-linked rod-staple-screw construct that can be used to stabilize thoracolumbar fractures after anterior corpectomy. Note that a titanium mesh cage (filled with salvaged fractured bone and supplemented by allograft) was used to reconstruct the anterior column.
Combined Anterior and Posterior Surgery
Combined anterior and posterior surgery can also be performed to treat thoracolumbar burst fractures. The specific indications and benefits of this fixation method remain unclear, although many surgeons feel that evidence of posterior ligamentous disruption is an important factor to be taken into account. Poor bone quality from osteopenia or osteoporosis is another relative indication for combined anterior and posterior reconstruction (Fig. 40-20). The theoretical advantages of a combined approach include maximization of canal clearance, immediate circumferential stability, and optimized fusion rates. The added morbidity of two separate procedures is a potential disadvantage.
The decision to proceed with a second stage (either anterior or posterior) operation is often delayed and is based on the results of the first stage of surgery. Postoperative CT scans can be helpful in identifying patients who may benefit from second-stage surgery. For example, patients with inadequate decompression of the canal after posterior surgery, in the presence of neurologic deficit, may be considered for a second-stage anterior surgery. Conversely, after performing an anterior approach, patients with poor bone quality or evidence of PLC disruption might be considered for a second-stage posterior stabilization.
There are few series that have examined the results of combined surgery for thoracolumbar burst fractures. In one study, Been et al (102) compared the results of combined anterior-posterior surgery to those of posterior surgery for the treatment of thoracolumbar burst fractures. The method of treatment had no effect on neurological recovery, but combined surgery resulted in better maintenance of kyphosis correction. Note that as with many of the series dealing with nonoperative treatment, increased loss of correction was not associated with a higher rate of back pain.
Methods of less-invasive surgery for thoracolumbar fractures have recently been developed (103). Although still highly controversial and not universally accepted, thoracoscopic anterior surgery and percutaneous posterior instrumentation may help to expand the role of combined anterior-posterior fracture treatment. Early results have been both encouraging and educational.
P.1562

P.1563

In one study, Khoo et al (103) reported on a series of 371 patients treated with thoracoscopic anterior surgery for fractures. Despite reasonable alignment and functional outcomes, they reported a high rate of vascular and neurologic complications. Faced with a difficult procedure associated with a slow learning curve, the potential benefit of less-invasive treatment of thoracolumbar fractures remains to be determined.
FIGURE 40-20 Lateral x-ray (A) and axial CT scan (B) of a 71-year-old woman with an incomplete neurologic injury who had sustained a burst fracture after a motor vehicle collision. Because of the relative osteopenia of the bone, posterior stabilization was undertaken after anterior corpectomy for decompression and anterior stabilization (C).
P.1564

P.1565

P.1566

P.1567

P.1568

Flexion-Distraction Injuries
Chance first described flexion-distraction injuries in 1948, as he observed the mechanism of posterior element distraction from a flexion moment centered anterior to the spine. Because of its association with the mechanism, they are also commonly referred to as seat-belt injuries. This term requires some clarification, however. Before the introduction of three-point motor vehicle restraints (lap-shoulder belts), most cars were equipped only with lap seat belts in [fn2]both the front and back seats. With sudden deceleration, such as may occur with a head-on collision, the torso and lower extremities are flexed and thrust forward about an axis of rotation within the lap belt. Because the lap belt acts as the fulcrum of rotation, as it lies anterior to the spine, it effectively produces tensile forces in both the anterior and posterior elements of the spine (Fig. 40-25). Thus, the associated lesion has been mechanistically described as a flexion-distraction injury.
FIGURE 40-21 Axial (A) and sagittal (B) CT scan of a 47-year-old woman who had sustained a complete spinal cord injury at the T5 level. The T2 and STIR MRI images (C,D) showed gross disruption of the PLC, although there was minimal, if any, residual spinal canal compromise. The injury was stabilized by a posterior pedicle screw construct without a formal decompression (E).
FIGURE 40-22 Open reduction maneuver for kyphotic deformities using pedicle screws. After pedicle screws have been placed above and below the injury (top), a rod is precontoured to the desired shape at final correction. (Note: at the thoracolumbar junction, this is usually straight.) The rod is then locked to the screw heads above the injury and gently reduced into the screw heads below the injury (bottom).
FIGURE 40-23 Preoperative AP (A) and lateral (B) x-rays of a 19-year-old man who had sustained an L3 burst fracture and a cauda equina injury with a large retropulsed bone fragment noted on axial CT scan (C). An anterior corpectomy was performed, followed by titanium mesh (filled with salvaged autograft and allograft) reconstruction. As the PLC appeared to be intact, anterior stabilization alone (D,E) was effected using a Kaneda device.
FIGURE 40-24 Illustrations of the surgical steps for anterior corpectomy and reconstruction for a burst fracture (A). First, the comminuted bone fragments of the vertebral body are removed (B), paying special attention to take out the retropulsed bone fragment(s) from the spinal canal (black arrow). After decompression, the vertebral body staples and screws can be placed (C). A distractor can be placed directly onto the anterior screws to reduce the kyphotic deformity and allow easier placement of the interbody device (D,E). With the interbody space slightly overdistracted, a cage or strut graft can be more easily placed (F). Rods are then fixed to the screw heads (G). Gentle compression of the construct locks the cage or strut in place. Two cross-links are then applied to provide better torsional stability to the construct (H).
Fortunately, most car manufacturers today provide lap-shoulder belts as standard equipment. More recent studies have recognized a trend toward a lower incidence of flexion-distraction injuries in patients restrained with a lap-shoulder belt compared to a lap belt alone (106,107). Notwithstanding the life-saving benefits of seat belts, unrestrained passengers have a lower rate of this injury than those restrained with a lap belt (107,108).
In the presence of a flexion-distraction injury, there should be a high index of suspicion of possible intra-abdominal injury. Rates of intra-abdominal hollow viscus injury have approached 50% in some studies (17,108). Conversely, the presence of the hallmark transabdominal or anterior chest wall band of ecchymosis in conjunction with an intra-abdominal injury should highlight the strong possibility of an underlying spinal injury. Even with careful evaluation, the diagnosis is often missed or delayed (51,109).
Abdominal evaluation should include CT scanning, ultrasound and/or deep peritoneal lavage. Laparotomy with intestinal repair and/or staged diversion takes precedence over spinal injuries, as it is a life-saving procedure. Strict spinal precautions should be maintained at all times, however, to avoid further displacement and potentially neurologic decline.
Although conveniently grouped into one category, flexion-distraction injuries present with considerable variability in their osseous and ligamentous disruption. Gertzbein and Court-Brown (110) developed a classification scheme that helps distinguish between these characteristics (Fig. 40-26). Disruptions typically occur within the transverse or axial plane. They may be purely ligamentous, purely osseous, or a combination of osseous and ligamentous injury. One important feature of this categorization is that purely bony injuries can potentially be treated with hyperextension casting or bracing due to the possibility of direct bone-to-bone healing. Flexion-distraction injuries should be differentiated from shear fracture-dislocations in that there is typically minimal translational deformity.
Nonoperative Treatment
There are few indications for nonoperative management of flexion-distraction injuries of the thoracolumbar spine. By definition, they involve both PLC and intervertebral disc disruption, making them an unstable injury. Some injuries occur primarily through bone. In this setting, provided that the fracture can be reduced and maintained in a brace or cast and the patient is neurologically intact, nonoperative treatment can be effective.
The argument for surgical treatment based on the presence of a neurologic deficit is less compelling. Spinal cord injury uniquely occurs from overdistraction, in distinction to the direct bony compression that occurs with burst fractures or dislocations. With the exception of direct neural compression from a herniated disc or hematoma, the primary goal of surgery for flexion-distraction injuries is restoration of alignment and stability
P.1569

P.1570

to enable early patient mobilization and prevent secondary displacement.
FIGURE 40-25 Diagram depicting the mechanism of injury of a flexion-distraction injury (also known as a Chance or seat-belt injury). Note that the fulcrum of forces lies anterior to the spine. With high energy (top), distraction occurs in both the posterior and anterior elements of the spine. With lower energy (bottom), anterior compression forces occur in concert with posterior distraction forces.
FIGURE 40-26 Classification system of flexion-distraction injuries as described by Gertzbein and Court-Brown (110).
Successful employment of nonoperative management depends on the ability to obtain a reasonable reduction of spinal alignment, as well as the likelihood of maintaining that reduction through healing processes. Because of the general reliability of bony healing, purely osseous injuries can usually be treated nonsurgically, provided an adequate reduction has been achieved. Reduction can be performed by applying various hyperextension forces to the spinal column, with subsequent casting or bracing in hyperextension. Anderson et al (106) recommended that nonsurgical treatment should be attempted in neurologically intact patients with less 15 degrees of kyphosis provided the injury is primarily bony. X-ray follow-up should be frequent to ensure that the reduction is maintained. External immobilization is recommended for a minimum of 3 months. Lateral flexion-extension can be useful in confirming stability. With a mixture of bone and ligamentous injury, effective healing and long-term stability is not reliably obtained.
Surgical Treatment
In most cases, flexion-distraction injuries involve disruption of the posterior intervertebral disc and the posterior ligaments, but leave the ALL and anterior annulus intact. Unless a herniated disc noted on a preoperative MRI warrants anterior discectomy, posterior reduction and compressive stabilization is usually adequate. In fact, anterior surgery would lead to greater destabilization of the injured segment. Compression across an injured or herniated disc can lead to further displacement into the canal and carries the potential risk of neurologic decline. Generally, a single-segment posterior segment fusion can stabilize the spinal column after flexion-distraction mechanisms. Preoperative CT or MRI images should be studied to ensure that the pedicles at the levels to be instrumented are intact and in continuity with their vertebral bodies. There is usually minimal vertebral body involvement, which permits the application of compressive forces through posterior instrumentation.
In 15% of cases, a burst-type fracture configuration occurs in conjunction with a flexion-distraction mechanism (17). It is an exercise in semantics as to whether this is labeled an unstable burst fracture or a flexion-distraction injury with a burst component (Fig. 40-27). In some cases, anterior decompression may be indicated for neural decompression, and anterior column reconstruction may be used to restore stability. Intervertebral disc herniations can occur in approximately 5% of cases, which may also warrant anterior discectomy for decompression (17). Alternatively, a posterolateral or transpedicular approach may be used to extract fragments anterior to the conus or spinal cord. A laminotomy or laminectomy can be used to remove disc fragments below the level of the conus medullaris. Laminectomy may be required for evacuation of an injury-related epidural hematoma or for repair of dural tears, but it should not be performed without spinal stabilization.
There are relatively few studies of flexion-distraction injuries in the thoracolumbar spine, but most suggest that patients typically do well regarding pain resolution. LeGay et al evaluated 17 patients with flexion-distraction injuries of the thoracolumbar spine of which 82% were treated nonoperatively (111). Nearly
P.1571

P.1572

80% of these patients had mild or no pain at follow-up, but only 65% were able to return to their preinjury activities. The patients who had significant alterations of their activity choices sustained multiple concomitant injuries in addition to spinal trauma.
FIGURE 40-27 Lateral x-ray (A) and axial CT scan (B) of a 48- year-old man who had sustained a flexion-distraction injury. Although the fracture could be described as a so-called burst fracture due to the involvement of the posterior vertebral body, there was not a substantial amount of height loss. A T2-weighted (C) and STIR (D) MRI clearly demonstrate the failure of the PLC. AP (E) and lateral (F) x-rays show the fracture well aligned using a posterior pedicle screw construct.
Anderson et al (106) promoted the successful use of nonoperative treatment if specific x-ray criteria could be met. In their series, 85% of patients with less than 15 degrees of initial kyphosis and no neurologic deficit did well after extension casting. Patients with greater than 15 degrees of kyphosis, neurologic deficits, or purely ligamentous injuries underwent surgery. Twelve of 13 patients had good results at 1 year.
Glassman et al (112) presented a group of 12 children with flexion-distraction injuries treated either surgically or nonsurgically. More than 80% of the patients had little or no pain at follow-up, and the investigators found no difference in outcome between treatment groups. Triantafyllou and Gertzbein (113) reviewed a series of thoracolumbar flexion-distraction injuries treated surgically with short-segment (one-level) fusions. Fourteen of 16 patients (87%) had little or no pain at follow-up.
Most recently, Liu et al (114) reviewed their surgical treatment of 23 patients who sustained flexion-distraction injuries of the thoracolumbar spine. Short-segment posterior pedicle screw instrumentation was employed, and the patients used a TLSO for 3 months postoperatively. Final average kyphosis improved 9.5 degrees to 5.4 degrees, and pain was absent or mild for all but two patients.
Pearls
  • Up to 50% of flexion/distraction injuries may be associated with intra-abdominal pathology.
  • They may be ligamentous, osseoligamentous, or osseous in nature.
  • Osseous injuries can be treated nonoperatively.
  • Single-segment posterior fusion is usually adequate.
Pitfalls
  • If the injury is ligamentous or osseoligamentous, surgical stabilization is indicated.
  • Surgeons should check that the pedicles at adjacent levels are intact prior to surgery.
  • In about 15% of cases, there is a burst fracture configuration.
  • In about 5% of cases, there is an associated herniated disc.
FRACTURE-DISLOCATIONS
Thoracolumbar fracture-dislocations are high-energy injuries that result in a markedly unstable spine. They are associated with a very high rate of complete neurologic deficit (14,115). The mechanism of injury usually involves a combination of forces, including flexion, extension, shear, torsion, and compression. Lateral views usually demonstrate a translational deformity or listhesis at the site of injury (Fig. 40-30). More frequently with dislocations of the thoracolumbar spine, however, the facets are fractured due to substantial translational or rotational forces. Under these circumstances, displacement may be appreciated on both AP and lateral x-rays. CT scanning and reconstruction sequences have been invaluable in characterizing these injuries more accurately.
As previously discussed, the concomitant presence of posterior element fractures can be associated with dural tears. In a review performed by Denis and Burkus (14), 50% of patients with thoracolumbar fracture-dislocations were found to have dural tears intraoperatively.
Nonoperative Treatment
Generally, fracture-dislocations are highly unstable injuries that require surgical stabilization. In the event a patient is incapable or unwilling to proceed with surgery, reduction should be attempted through postural manipulation. Such circumstances are rare. Nonoperative treatment requires an extended period of bracing and bed rest.
P.1573

P.1574

Surgical Treatment
Fracture-dislocations are almost always treated with surgical stabilization. The energy imparted in this type of disruption creates marked instability that may compromise residual neurologic function or hamper progress with rehabilitation.
FIGURE 40-28 Lateral x-ray (A) and MRI (B) of a young man who had sustained a flexion-distraction fracture of L4. Instead of performing a fusion, the fracture was reduced and stabilized using a compressive pedicle screw and hook construct (C,D). As the construct did cross the L3–L4 motion segment, removal of the hardware was planned as a secondary procedure after fracture healing (E). (Courtesy of Richard Rooney, MD, and Behrooz Akbarnia, MD.)
FIGURE 40-29 In some flexion-distraction cases, the facet joints can be dislocated without fracture, which can make postural reduction difficult. A direct open reduction of the dislocated articular surfaces can be effected by distracting between the lamina and/or spinous processes to “unlock” the joints.
Posterior surgery is usually most useful for achieving reduction and stability after these injuries. A spinal frame, such as a Jackson table, that encourages normal sagittal plane alignment is usually adequate for operative positioning, but prone positioning on the operative table largely depends on whether the facets are intact. Once the midline incision has been made, a step-off or interspinous gap may be easily appreciated. Additionally, marked soft tissue stripping and disruption can be seen while exposing the posterior elements.
The characteristic deformity of fracture-dislocations is translational malalignment of the involved vertebrae. Realigning the spine is often difficult and is best performed by direct manipulation of the vertebra with bone clamps or elevators. In the authors’ experience, however, realignment is more easily achieved by delivering the appropriate forces through pedicle screw posterior instrumentation.
As has already been detailed for flexion-distraction injuries, facet dislocations can also occur with fracture dislocations. Reduction can be difficult if the articular processes are not fractured and thus must be achieved through gradual application of longitudinal distraction forces directly at the site of dislocation (116). Despite the similarity in mechanism to cervical dislocations, indirect reductions (as seen in the cervical spine) are not possible for these injuries, mainly due to the substantially greater forces required.[iu30] In some cases, resection of the superior articular processes may be necessary to reduce the fracture.
Hyperextension is a rare cause of fracture-dislocation injuries that must be recognized because of its association with various comorbidities. This extension-distraction mechanism is more commonly seen in individuals with ankylosing spondylitis, diffuse idiopathic skeletal hyperostosis, and advanced degenerative spondylosis (117).
Because of the severe ligamentous disruption and gross instability that occurs, long fusions have been advocated (118), although some reports suggest that short-segment posterior fixation may provide adequate stability in most cases (115,119). Due to the rarity of these lesions, definitive comparative studies are difficult to perform, but technologic advancements with instrumentation may allow more frequent use of these shorter constructs. Earlier instrumentation included nonsegmental hook fixation and sublaminar wiring, but segmental hooks and pedicle screws offer substantially greater stability. In long constructs, cross-linking can provide additional torsional stability (118,120).
P.1575

In the middle and upper thoracic spine, some surgeons are reluctant to place pedicle screws because of the dimensional constraints (121,122), and lamina and pedicle hooks remain a viable option for proximal fixation.
FIGURE 40-30 AP x-ray of a high-energy fracture dislocation with lateral translation. This degree of translation implies nearly complete discoligamentous disruption with instability.
Provided that spinal stability is restored, the patient can be mobilized as required to improve pulmonary function and general medical care. External immobilization or bracing is usually not necessary, although these types of treatments should be judged on a case-by-case basis.
There is a relative paucity of outcome studies detailing the results of the treatment of thoracolumbar fracture-dislocations. This is due largely to their rarity in comparison to compression and burst fractures. In addition, many of the reports in the literature involve surgical treatment with outdated instrumentation systems that were less reliable than contemporary pedicle screw constructs. These series also contain a variety of thoracolumbar injuries, making analysis of results specifically of fracture-dislocations difficult (123,124,125).
Kinoshita et al (126) used a spinous process wiring technique after open reduction in 30 patients. They reported some success with control of sagittal displacement, although lateral angulation and displacement was poorly controlled with this method. Later, Denis and Burkus (14) studied the use of Harrington distraction rods for hyperextension fracture-dislocation injuries and found that the rods alone could not maintain stability. Importantly, the addition of a compression rod or interspinous wiring resulted in better anatomic restoration.
Short-segment fusion constructs using pedicle screws for fracture-dislocations have been evaluated in two different studies, but the results are somewhat contradictory (115,119). Yu et al (119) found a higher rate of failure with short-segment fusion when used in low lumbar injuries compared with thoracolumbar junction injuries. On the other hand, Razak et al (115) reported that 15 patients with thoracic or lumbar fracture-dislocations treated with short-segment fixation showed no loss of reduction at 1-year follow-up.
Pearls
  • Stabilization and spinal realignment are the main goals of surgery.
  • Lateral x-rays usually show translational deformity or a lithesis at the fracture site.
  • Long posterior pedicle screw constructs are best for thoracolumbar fracture-dislocations.
Pitfalls
  • There is a high incidence of complete neurologic deficit.
  • Up to 50% dural tears have been noted.
  • Short-segment spinal fixation may not provide adequate stabilization.
TIMING OF SURGERY
Surgical timing is an important consideration. This is particularly true when treating polytraumatized patients. In a retrospective review of patients with spinal fractures, including cervical, thoracic, and lumbar injuries, patients with injuries that were fixed within 3 days of injury had a lower incidence of pneumonia and a shorter hospital stay than those fixed more than 3 days of injury (129). The two groups were not randomized and the late-fixation group could have included patients that were more severely injured. Although the ISS scores were not significantly different between the early- and late-fixation groups, the average age, admission Glasgow Coma Scale, and chest injury scores were significantly higher in the late-fixation group. These variables could certainly have accounted for the poorer results found with delayed spinal stabilization.
In patients without neurologic deficit, urgent surgery is rarely indicated, and careful planning should take place. Surgery can be performed when the patient has been adequately stabilized medically. A similar approach should be employed in patients that have complete neurologic injuries when there is little chance for significant recovery. These patients often require surgical stabilization in order to optimize their rehabilitation course, but intervention can usually be delayed until medical stability has been achieved. Greater controversy surrounds surgical timing in the setting of incomplete spinal cord injury (130,131). Animal data suggest that early decompression is beneficial, but there is a lack of clinical evidence to support this belief. In a retrospective review, Chipman et al (132), although finding that early surgery within 72 hours was associated with fewer complications, shorter hospital stays, and shorter stays in intensive care, did not demonstrate any neurologic benefit.
P.1576

It important to note that there is little evidence that emergency decompression is indicated in patients with documented neurologic decline (133).
When emergency surgery is not required for neurologic decompression, the patient’s physiological state should be carefully assessed. Hemodynamic instability could have deleterious effects on spinal cord perfusion and increase the risk of further neurologic compromise during surgery. Additionally, the substantial stress of spine surgery may exacerbate pulmonary problems.
GUNSHOT WOUNDS
Gunshot wounds are a distinct group of thoracolumbar injuries (6). The mechanism of injury is that of a missile passing through the spine rather than the spine being forcibly accelerated or decelerated until structural failure. In general, fractures associated with low-velocity gunshot wounds are stable fractures. This is the case with most handgun injuries, as they are associated with a low infection rate and can be prophylactically treated with 48 hours of a broad-spectrum antibiotic. Transintestinal gunshot wounds require special attention. In these cases, the bullet passes through the colon, intestine, or stomach before passing through the spine. This type of wound carries a significantly higher rate of infection (134). Whereas open debridement offers no advantages, broad-spectrum antibiotics should be continued for 7 to 14 days (134,135). High-energy wounds, as caused by a rifle or military assault weapon, require open debridement and stabilization.
The extent and type of neurologic deficits from gunshot wounds to the thoracolumbar spine can vary. Unlike the situation in blunt trauma, complete neurologic deficits are more common than incomplete deficits after gunshot wounds (136,137). Neural injury is secondary to a blast effect in which the bullet’s energy is absorbed and transmitted to the soft tissues. Because of this unique mechanism, decompression is rarely indicated. One exception is when a bullet fragment is found in the spinal canal between the level of T12 and L5 in the presence of a neurologic deficit (138). Delayed bullet extraction is rarely indicated for lead toxicity or late neurologic deficits due to migration of a bullet fragment. Steroids after gunshot wounds to the spine are not recommended, as they have shown no neurologic benefit and appear to be associated with a higher rate of nonspinal complications (139,140,141).
Instability after thoracolumbar gunshot wounds is most commonly iatrogenic and is the result of laminectomy without fusion or instrumentation. In rare cases, destruction of bilateral facets and posterior elements from the gunshot wounds can result in instability. Initially, patients are best treated in an orthosis, followed by x-ray evaluation to ensure maintenance of spinal column alignment. Decisions to intervene surgically are based on progression of deformity or signs of instability.
OSTEOPOROTIC FRACTURES
Osteoporosis leaves the spine highly susceptible to fracture, even if there is little or no trauma. Individuals with osteoporotic compression fractures tend to have progressive collapse of one or more of the vertebral bodies, which is occasionally accompanied by radicular pain and rarely by neurologic deficit. Bracing is not well tolerated in elderly patients who sustain these injuries, and if they develop significant neurologic deficits, disabling deformity, or marked pain, surgery can be a reasonable option. Vertebroplasty or kyphoplasty are techniques involving the injection of polymethylmethacrylate cement into the involved vertebral bodies, thereby providing mechanical support in areas of insufficient bony strength. Although no clear benefit has been established for one technique over the other, kyphoplasty has the theoretical advantage of enabling vertebral body height restoration through the inflation of balloon tamps (142,143) (Fig. 40-31). Both procedures have shown high success rates, with greater than 90% of patients reporting pain relief (142,143,144).
Open surgery for osteoporotic compression fractures is difficult because of the poor bone quality and there are often significant complications. In most cases, open surgery is reserved for patients with neurologic deficits, usually following burst fractures, or intractable pain associated with a severe spinal deformity. An anterior surgical approach allows for greater deformity correction, but fixation is limited by poor bone quality. Posterior segmental instrumentation, usually with pedicle screws, is often performed to supplement anterior surgery in osteoporotic patients with more significant thoracolumbar trauma.
FIGURE 40-31 Kyphoplasty, which utilizes an inflatable balloon tamp, can be used to restore vertebral body height and create a cavity within osteoporotic compression fractures (top). The balloon is removed, and polymethylmethacrylate cement is inserted into the void to provide stabilization (bottom).
P.1577

COMPLICATIONS
Thoracolumbar spine injuries can lead to a variety of potential complications. It is incumbent on the surgeon to examine the unique characteristics of each injury so that the risk of adverse events can be minimized. Understanding the natural history of thoracolumbar injuries aids in this process, but one must also consider the complications from surgical intervention. The controversy between operative and nonoperative treatment is derived as much from concerns about complications as anything else. In a recent study by Rechtine et al (145), the complication rates of operative and nonoperative treatment were similar, although the infection rates were considerably higher in the operative group (but hospital stays were longer in the nonoperative group).
Nonsurgical Complications
Neurologic deficits remain one of the major problems associated with thoracolumbar injuries. The complexity of their pathophysiology is poorly understood and this has led to disagreement about the optimal treatment for these patients. Progressive neurologic decline is generally treated by urgent surgery, whereas more static neurologic deficits have been successfully treated both surgically and nonsurgically. Delayed neurologic changes can also occur due to syrinx formation after thoracolumbar trauma. Careful physical examination should be performed to identify this possibility, and an MRI will usually corroborate its presence.
[dc9]Acute or progressive deformity may occur following thoracolumbar trauma, and nonsurgical management can lead to increased kyphosis or malalignment. Despite the increased risk of deformity progression without surgery, most patients do not benefit from surgery when outcome measures such as pain and function are assessed (67). Initial kyphotic deformities greater than 30 degrees may be associated with increased pain and should be considered for surgical correction and stabilization (2).[dc0]
Dural tears may occur in thoracolumbar trauma and are particularly associated with vertical lamina and facet fractures (34,35). They may also suggest the possibility of nerve root entrapment. Neurologic deficit in the setting of a burst fracture with laminar involvement should lead the surgeon to consider nerve root exploration when performing posterior surgery for these injuries. Dural repairs can often be primarily repaired, and a subarachnoid drain should be employed in the case of persistent leaks.
Ileus may follow many thoracolumbar injuries, regardless of the treatment that is chosen. Careful attention must be paid to postinjury bowel function, and a suspicion of ileus warrants aggressive workup and treatment. An increased incidence of peptic ulceration has also been found in the trauma population and can be worsened by the administration of corticosteroids for spinal cord injury. Prophylactic use of histamine-2 blockers should be considered.
Constipation often occurs as a result of fluid imbalance, stress-response, as well as narcotic use and should be managed with appropriate dietary and pharmacologic interventions. In the setting of spinal cord injury, a specific program, often called a neurobowel protocol, is necessary. Urinary dysfunction must be formally evaluated, and assistance with regular voiding should be provided in order to prevent urinary tract infections.
Deep venous thrombosis and venous thromboembolism are not unusual occurrences after spinal trauma. In a recent meta-analysis, approximately 12% of patients suffering spinal trauma developed deep venous thrombosis, whereas only 1.5% incurred a thromboembolic event (146,147). It is also clear that these incidences substantially increase if there is a spinal cord injury. Mechanical compression devices should be used, but chemical prophylaxis has been controversial. In general, the authors recommend chemical prophylaxis for spinal cord injury and spinal trauma, except in the first 4 to 5 days or when a compressive hematoma is suspected. If chemical prophylaxis is contraindicated, the use of a vena cava filter must be strongly considered.
Pain can be persistent after thoracolumbar trauma, and it may be recalcitrant to pharmacologic treatment. When associated with progressive or severe deformity, surgery may be indicated as a means for pain reduction, but complete resolution may not occur. Despite a favorable overall functional outcome for many patients with thoracolumbar fractures, a number continue to have at least a mild level of pain.
The trauma patient frequently requires long-term hospitalization, and malnutrition may complicate their care during this period. Many trips to the operating room, intubation, or other reasons for poor dietary intake can lead to impaired physiology and overall health as well as a general decline of the patient’s health. Appropriate investigation and treatment should be undertaken.
Surgical Complications
Despite our rapidly evolving techniques for managing spinal trauma surgically, there are still considerable risks inherent in spinal surgery. There are few prospective studies comparing nonoperative and operative treatment in thoracolumbar fractures, which makes it difficult to assess the relative risk of complications when choosing an appropriate treatment method.
Although surgery is generally favored, if there is neurologic compromise, there are risks of further neurologic injury during the operative process. Transferring the patient, retracting the spinal cord or nerve roots, placing instrumentation, or postoperative epidural hematoma can all damage the nervous system. Intraoperative monitoring should be strongly considered in a patient with intact or incomplete neurologic status. preoperatively.
Bleeding can be substantial during spine surgery and may be poorly tolerated by patients already under severe physiologic stress from trauma. Freeing the abdomen of pressure may help reduce the rate of blood loss, but coagulopathy can be a major problem in the trauma patient, and transfusion may be required. A cell-saver device can be helpful in reducing the overall impact of blood loss.
P.1578

Intraoperative iatrogenic dural tears undoubtedly occur during the surgical care of these patients, but their true frequency is unknown. These tears should be treated by repair, if possible, with the use of a subarachnoid drain and postoperative recumbency.
Infection is a major concern in the patient undergoing surgery for spinal trauma. Rates of postoperative infection are about 10%, and infection should be dealt with aggressively (145). In addition to intravenous antibiotics, debridement and irrigation must be performed early. Once fusion has occurred, persistent or new infections can be treated with hardware removal and at least 6 weeks of parenteral antibiotics. Various surgical techniques have been described for managing severe infections, such as the use of irrigation with drains, vacuum devices, and antibiotic-laden cement beads.
Hardware failure can occur after instrumented spinal fusion. Metal breakage, screw pullout, or subsidence may occur during the postoperative period, leading to at least a partial loss of correction. Fortunately, poor outcomes have not necessarily resulted from this loss of correction (148,149).
Technical errors in difficult surgical approaches and corrections may cause neurologic, pulmonary, and vascular complications. Pneumothorax from pleural penetration during insertion of instrumentation and injury or erosion into the great vessels has been documented (150). Thoracotomy can have a negative effect on pulmonary function. In addition, it places the major vascular structures at risk. Anterior abdominal approaches risk injury to the iliac vessels, ureter, and sympathetic plexus.
Pseudarthrosis occurs in approximately 4% of cases following surgery for thoracolumbar trauma (151), can lead to chronic pain, and may result in worsening deformity. X-ray workup with CT scans and/or MRI should be performed, and other causes, such as infection, should be excluded.
FUTURE DIRECTIONS
It is likely that in the future, the treatment of thoracolumbar trauma will include minimally invasive fracture stabilization and decompression. Methods such as thoracoscopic surgery have shown encouraging early results, but it is difficult to know if such techniques will be universally adopted, as they require considerable specialist training and are resource dependent. Percutaneous techniques of osteoporotic fracture treatment, such as vertebroplasty and kyphoplasty, however, have already become accepted by both community as well as academic surgeons, and given the rapid expansion of the elderly population, it seems probable that the use of these techniques will increase further.
The other recent innovation has been computer-guided surgery, and it is possible that guided instrument insertion, controlled by computer, and facilitated by intraoperative imaging will become a popular technique for the treatment of more complicated injuries such as burst fractures and fracture dislocations.
REFERENCES
An asterisk denotes a Level 1 evidence-based reference.
1. Robertson A, Branfoot T, Barlow IF, et al. Spinal injury patterns resulting from car and motorcycle accidents. Spine 2002;27:2825–2830.
2. Gertzbein S. Scoliosis Research Society. Multicenter spine fracture study. Spine 1992;17:528–40.
3. DeWald RL. Burst fractures of the thoracic and lumbar spine. Clin Orthop 1984;189:150–161.
4. Benson DR. Unstable thoracolumbar fractures, with emphasis on the burst fracture. Clin Orthop 1988;230:14–29.
5. Denis F. The three columns of the spine and its significance in the classification of acute thoracolumbar spine injuries. Spine 1983;8:817–831.
6. Bono CM, Heary RF. Gunshot wounds to the spine. Spine J 2004;4:230–240.
7. Boukhris R, Becker KL. The inter-relationship between vertebral fractures and osteoporosis. Clin Orthop 1973;90:209–216.
8. Panjabi MM, Brand RA, White AA. Mechanical properties of the human thoracic spine as shown by three-dimensional load-displacement curves. J Bone Joint Surg (Am) 1976;58:642–652.
9. Panjabi MM, Hausefeld JN, White A. A biomechanical study of the ligamentous stability of the thoracic spine in man. Orthop Scand 1981;52:315–326.
10. Hipp JA. Biomechanics of thoracolumbar fractures. In: Reitman CA, ed. Management of Thoracolumbar Fractures. Rosemont: American Academy of Orthopaedic Surgeons, 2004:9–18.
11. Shirado O, Kaneda K, Tadano S, et al. Influence of disc degeneration on mechanism of thoracolumbar burst fractures. Spine 1992;17:286–292.
12. Fredrickson BE, Edwards WT, Rauschning W, et al. Vertebral burst fractures: an experimental, morphologic, and radiographic study. Spine 1992;17:1012–1021.
13. Panjabi MM, Oxland TR, Lin RM, et al. Thoracolumbar burst fracture: A biomechanical investigation of its multidirectional flexibility. Spine 1994;19:578–585.
14. Denis F, Burkus J. Shear fracture-dislocation of the thoracic and lumbar spine associated with forceful hyperextension (lumberjack paraplegia). Spine 1992;17:156–161.
15. Langrana NA, Harten RD, Lin DC, et al. Acute thoracolumbar burst fractures: a new view of loading mechanisms. Spine 2002;27:498–508.
16. Anderson S, Biros MH, Reardon RR. Delayed diagnosis of thoracolumbar fractures in multiple-trauma patients. Acad Emerg Med 1996;3:832–839.
17. Eismont FJ. Flexion-distraction injuries of the thoracic and lumbar spine. In: Levine AM, Eismont FJ, Garfin SR, et al, eds. Spine Trauma. Philadelphia: WB Saunders, 1998:402–413.
18. Gehweiler JA, Daffner RH, Osborne RL. Relevant signs of stable and unstable thoracolumbar vertebral column trauma. Skeletal Radiol 1981;7:179–183.
19. Holdsworth FW. Fractures, dislocations, and fracture-dislocations of the spine. J Bone Joint Surg (Br) 1963;45:6–20.
20. Cobb JR. Outline for the study of scoliosis. Instr Course Lect 1948;5:261–275.
21. Bradford D, McBride G. Surgical management of thoracolumbar spine fractures with incomplete neurologic deficits. Clin Orthop 1987;218:201–215.
22. Denis F. Thoracolumbar injuries. Instr Course Lect 1988;230.
23. Esses SI. The placement and treatment of thoracolumbar spine fractures: an algorithmic approach. Orthop Rev 1988;17:571–584.
24. Jacobs RR, Asher MA, Snider RK. Thoracolumbar spinal injuries: a comparative study of recumbent and operative treatment in 100 patients. Spine 1980;5:463–477.
25. Jacobs RR, Casey MP. Surgical management of thoracolumbar spinal injuries. Clin Orthop 1984;189:22–35.
26. Weitzman G. Treatment of stable thoracolumbar spine compression fractures by early ambulation. Clin Orthop 1971;176:116–122.
27. White A, Panjabi M. Clinical biomechanics of the spine. 2nd ed. Philadephia: Lippincott-Raven, 1990.
28. Keene J, Fischer S, Vanderby R, et al. Significance of acute posttraumatic bony encroachment of the neural canal. Spine 1989;14:799–802.
29. Hauser CJ, Visvikis G, Hinrichs C, et al. Prospective validation of computed tomographic screening of the thoracolumbar spine in trauma. J Trauma 2003;55:228–234.
30. WIntermark M, Mouhsine E, Theumann N, et al. Thoracolumbar spine fractures in patients who have sustained severe trauma: depiction with multi-detector row CT. Radiology 2003;227:681–689.
31. Harris MB, Reitman CA. Spinal trauma-initial evaluation and imaging. In: Reitman CA, ed. Management of Thoracolumbar Fractures. Rosemont: American Academy of Orthopaedic Surgeons, 2004:27–34.
32. McAfee P, Yuan H, Fredrickson BE, et al. The value of computed tomography in thoracolumbar fractures. An analysis of one hundred consecutive cases and a new classification. J Bone Joint Surg (Am) 1983;65:461–473.
33. McCormack T, Karaikovic E, Gaines RW. Load sharing classification of spine fractures. Spine 1994;19:1741–1744.
34. Cammisa F, Eismont F, Green B. Dural laceration occurring with burst fractures and associated laminar fractures. J Bone Joint Surg 1989;71-A:1044–1052.
35. Denis F, Burkus J. Diagnosis and treatment of cauda equina entrapment in the vertical lamina fractures of lumbar burst fractures. Spine 1991;16:S433–S439.
36. Singh K, Vaccaro AR. Thoracic and lumbar trauma. In: Bono CM, Garfin SR, eds. Orthopaedic Surgery Essentials: Spine. Philadelphia: Lippincott Williams & Wilkins, 2004:45–57.
P.1579

37. Brightman R, Miller C, [fn3]Rea G, et al. Magnetic resonance imaging of trauma to the thoracic and lumbar spine. The importance of the posterior longitudinal ligament. Spine 1992;17:541–550.
38. Oner FC, Ramos LM, Simmermacher RK, et al. Classification of thoracic and lumbar spine fractures: problems of reproducibility. A study of 53 patients using CT and MRI. Eur Spine J 2002;11:235–245.
39. Oner FC, van Gils APG, Faber JAJ, et al. Some complications of common treatment schemes of thoracolumbar spine fractures can be predicted with magnetic resonance imaging. Spine 2002;27:629–636.
40. Lee HM, Kim HS, Kim DJ, et al. Reliability of magnetic resonance imaging in detecting posterior ligament complex injury in thoracolumbar spinal fractures. Spine 2000;25:2079–2084.
41. Teitelbaum GP. Metallic ballistic fragments: MR imaging safety and artifacts. Radiology 1990;175:855–859.
42. Finitsis S, Falcone S, Green B. MR of the spine in the presence of metallic bullet fragments: is the benefit worth the risk? Am J Neuroradiol 1999;20:354–356.
43. Ebraheim N, Savolain E, Jackson W, et al. Magnetic resonance imaging in the evaluation of a gunshot wound to the cervical spine. J Orthop Trauma 1989;3:19–22.
44. Bashir E, Cybulski G, Chaudhri K, et al. Magnetic resonance imaging and computed tomography in the evaluation of penetrating gunshot injury. Spine 1993;18:772–773.
45. Mehta JS, Reed MR, McVie JL, et al. Weight-bearing radiographs in thoracolumbar fractures: do they influence management? Spine 2004;29:564–567.
46. Watson-Jones R. Fractures and Joint Injuries. 3rd ed. Edinburgh: ES Livingston, 1943.
47. Nicoll EA. Fractures of the dorso-lumbar spine. J Bone Joint Surg (Br) 1949;31:376.
48. Kelly RP, Whitesides TE. Treatment of lumbodorsal fracture-dislocations. Ann Surg 1968;167:705.
49. Magerl F, Aebi M, Gertzbein S, et al. A comprehensive classification of thoracic and lumbar injuries. Eur Spine J 1994;3:184–201.
50. Dai LY, Jin WJ. Interobserver and intraobserver reliability in the load sharing classification of the assessment of thoracolumbar burst fractures. Spine 2005;30:354–358.
51. Burdi M, Bono CM, Kauffman CP, et al. Delayed diagnosis of a flexion-distraction (Seatbelt-type) injury in a patient with multiple abdominal injuries: a case report. Am J Orthop. In press.
52. Ebraheim NA, Xu R. Surgical anatomy of the thoracolumbar spine. In: Reitman CA, ed. Management of Thoracolumbar Fractures. Rosemont: American Academy of Orthopaedic Surgeons, 2004:1–7.
53. Willen J, Anderson J, Toomoka K, et al. The natural history of burst fractures at the thoracolumbar junction. J Spinal Disord 1990;3:39–46.
54. Ferguson RL, Allen BJ. A mechanistic classification of thoracolumbar spine fractures. Clin Orthop 1984;189:77–88.
55. Norton PL, Brown T. The immobilizing efficacy of back braces. J Bone Joint Surg Am 1957;39:377–386.
56. Lumdsden RM, Morris JM. An in vivo study of axial rotation and immobilization at the lumbosacral joint. J Bone Joint Surg (Am) 1968;50:1591–1602.
57. Fidler MW, Plasmans CMT. The effect of four types of support on the segmental mobility of the lumbosacral spine. J Bone Joint Surg (Am) 1983;65:943–947.
58. Cholewicki J, Alvi K, Silfies SP, et al. Comparison of motion restriction and trunk stiffness provided by three thoracolumbosacral orthoses (TLSOs). Spinal Disord Tech 2003;16:461–468.
59. Folman Y, Gepstein R. Late outcome of nonoperative management of thoracolumbar vertebral wedge fractures. J Orthop Trauma 2003;17:190–192.
60. Trafton PG, Boyd CA. Computed tomography of thoracic and lumbar spine injuries. J Trauma 1984;24:506–515.
61. Hashimoto T, Kaneda K, Abumi K. Relationship between traumatic spinal canal stenosis and neurological deficits in thoracolumbar burst fractures. Spine 1988;13:1268–1272.
62. Dai LY. Remodeling of the spinal canal after thoracolumbar burst fractures. Clin Orthop 2001;382:119–123.
63. Mohanty SP, Venkatram N. Does neurological recovery in the thoracolumbar and lumbar burst fractures depend on the extent of canal compromise? Spinal Cord 2002;40:295–299.
64. Limb D, Shaw DL, Dickson RA. Neurologic injury in thoracolumbar burst fractures.[jyJ Bone Joint Surg (Br) 1995;77:774–777.
65. Burke DC, Murray DD. The management of thoracic and thoracolumbar injuries of the spine with neurological involvement. J Bone Joint Surg (Br) 1976;58B:72–78.
66. Dendrinos GK, Halikias JG, Krallis PN, et al. Factors influencing neurological recovery in burst thoracolumbar fractures. Acta Orthop Belg 1995;61:226–234.
67. Weinstein JN, Collalto P, Lehmann TR. Thoracolumbar “burst” fractures treated conservatively: a long-term follow-up. Spine 1988;13:33–38.
68. Shen WJ, Shen YS. Nonsurgical treatment of three-column thoracolumbar junction burst fractures without neurologic deficit. Spine 1999;24:412–415.
69. Aligizakis A, Katonis P, Stergiopoulos K, et al. Functional outcome of burst fractures of the thoracolumbar spine managed non-operatively with early ambulation, evaluated using the load sharing classification. Acta Orthop Belg 2002;68:279–287.
70. Chow GH, Nelson BJ, Beghard JS, et al. Functional outcome of thoracolumbar burst fractures managed with hyperextension casting or bracing and early mobilization. Spine 1996;21:2170–2175.
71. Shen WJ, Liu TJ, Shen YS. Nonoperative treatment versus posterior fixation for thoracolumbar junction burst fractures without neurologic deficit. Spine 2001;26:1038–1045.
72. Wood K, Butterman G, Mehbod A, et al. Operative compared with nonoperative[jy[om0p3]treatment of a thoracolumbar burst fracture without neurological deficit. J Bone Joint Surg 2003;85-A:773–781.
73. Danisa OA, Shaffrey CI, Jane JA, et al. Surgical approaches for the correction of unstable thoracolumbar burst fractures: a retrospective analysis of treatment outcomes. J Neurosurg 1995;83:977–983.
74. Stancic MF, Gregorovic E, Nozica E, et al. Anterior decompression and fixation versus posterior reposition and semirigid fixation in the treatment of unstable burst thoracolumbar fracture: prospective clinical trial. Croat Med J 2001;42:49–53.
75. Wood KB, Bohn D, Mehbod A. Anterior versus posterior treatment of stable thoraco-[jy[om0p3]lumbar burst fractures without neurologic deficit. A prospective, randomized, study. J Spinal Disord Tech 2005;18:S15–S23.
76. Slosar PJ, Patwardhan AG, Lorenz M, et al. Instability of the lumbar burst fracture and limitations of transpedicular instrumentation. Spine 1995;20:1452–1461.
77. Annicchiarico N. Internal A-O Fixator for the reduction and stabilization of thoracolumbar fractures. Chir Organi Mov 2000;85:167–175.
78. Alanay A, Acaroglu E, [fn3]Yazici M, et al. The effect of transpedicular intracorporeal grafting in the treatment of thoracolumbar burst fractures on canal remodeling. Eur Spine J 2001;10:512–516.
79. Alanay A, Acaroglu E, Yazici M, et al. Short-segment pedicle instrumentation of thoracolumbar burst fractures: does transpedicular intracorporeal grafting prevent early failure. Spine 2001;26:213–217.
80. Arino VL. Our experience in 33 cases of thoracolumbar fracture treated by transpedicle instrument. Chir Organi Mov 2000;85:161–165.
81. Oda T, Panjabi MM. Pedicle screw adjustments affect stability of thoracolumbar burst fracture. Spine 2001;26:2328–2333.
82. Oda T, Panjabi MM, Kato Y. The effects of pedicle screw adjustments on the anatomical reduction of thoracolumbar burst fractures. Eur Spine J 2001;10:505–511.
83. Brodke DS, Gollogly S, Bachus KN, et al. Anterior thoracolumbar instrumentation: stiffness and load sharing characteristics of plate and rod systems. Spine 2003;28:1794–1801.
84. Knop C, Fabian HF, Bastian L, et al. Fate of the transpedicular intervertebral bone graft after posterior stabilisation of thoracolumbar fractures. Eur Spine J 2002;11:251–257.
85. Bellbarba C, Mirza SK, Chapman JR. Surgical treatment of thoracolumbar fractures-posterior approach. Management of Thoracolumbar Fractures. Rosemont: American Academy of Orthopaedic Surgeons, 2004:65–78.
86. Sjostrom L, Karlstrom G, Pech P, et al. Indirect spinal canal decompression in burst fractures treated with pedicle screw instrumentation. Spine 1996;21:113–123.
87. Katonis P, Kontakis G, Loupasis G, et al. Treatment of unstable thoracolumbar and lumbar spine injuries using Cotrel-Dubousset instrumentation. Spine 1999;24:2352–2357.
88. Scheffer MM, Currier BL. Thoracolumbar burst fractures. In: Levine AM, Eismont FJ, Garfin SR, et al, eds. Spine Trauma. Philadelphia: WB Saunders, 1998:428–451.
89. Yazici M, Gulman B, Sen S, et al. Sagittal contour restoration and canal clearance in burst fractures of the thoracolumbar junction (T12–L1): the efficacy of timing of the surgery. J Orthop Trauma 1995;9:491–498.
90. Garfin SR, Mowery CA, Guerra J, et al. Confirmation of the posterolateral technique to decompress and fuse thoracolumbar spine burst fractures. Spine 1985;10:218–223.
91. McAfee P, Yuan HA, Lasada NA. The unstable burst fracture. Spine 1982;7:365–373.
92. Lemons VR, Wagner FC, Montesano PX. Management of thoracolumbar fractures with accompanying neurological injury. 1992;30:667–671.
93. Shono Y, McAfee P, Cunningham BW. Experimental study of thoracolumbar burst fractures. A radiographic and biomechanical analysis of anterior and posterior instrumentation. Spine 1994;19:1711–1722.
94. Gurwitz GS, Dawson JM, McNamara MJ, et al. Biomechanical analysis of three surgical approaches for lumbar burst fractures using short-segment instrumentation. Spine 1993;18:977–982.
95. Dvorak MF, Kwon BK, Fisher CG, et al. Effectiveness of titanium mesh cylindrical cages in anterior column reconstruction after thoracic and lumbar vertebral body resection. Spine 2003;28:902–908.
96. Oda I, Cunningham BW, Abumi K, et al. The stability of reconstruction methods after thoracolumbar total spondylectomy. An in vitro investigation. Spine 1999;24:1634–1638.
97. Hitchon PW, From AM, Brenton MD, et al. Fractures of the thoracolumbar spine complicating ankylosing spondylitis. J Neurosurg 2002;97:218–222.
98. Kostuik J. Anterior fixation for fractures of the thoracic and lumbar spine with or without neurologic involvement. Clin Orthop 1984;189:103–115.
99. Kaneda K, Taneichi H, Abumi K, et al. Anterior decompression and stabilization with the Kaneda device for thoracolumbar burst fractures associated with neurological deficits. J Bone Joint Surg (Am) 1997;79:69–73.
100. Haas N, Blauth M, Tscherne H. Anterior plating in thoracolumbar spine injuries. Indication, technique, and results. Spine 1991;16:S100–S111.
101. Ghanayem AJ, Zdeblick TA. Anterior instrumentation in the management of thoracolumbar burst fractures. Clin Orthop 1997;335:89–100.
102. 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 1999;141:349–357.
103. Khoo LT, Beisse R, Potulski M. Thoracoscopic-assisted treatment of thoracic and lumbar fractures: a series of 371 consecutive cases. Neurosurgery 2002;51:52–117.
104. Silvestro C, Francaviglia N, Bragazzi R, et al. Near-anatomical reduction and stabilization of burst fractures of the lower thoracic or lumbar spine. Acta Neurochir (Wien) 1992;116:53–59.
105. Bono CM, Garfin SR. Spinal Osteoporosis [American Academy of Orthopaedic Surgeons], 2003, 2002.[su2]
106. Anderson PA, Henley MB, Rivara FP, et al. Flexion-distraction and chance injuries to the thoracolumbar spine. J Orthop Trauma 1991;5:153–160.
107. Ball ST, Vaccaro AR, Albert TJ, et al. Injuries of the thoracolumbar spine associated with restraint use in head-on motor vehicle accidents. J Spinal Disord 2000;13:297–304.
P.1580

108. Gertzbein SD, Court-Brown CM. [fn3]Rationale for the management of flexion-distraction injuries of the thoracolumbar spine based on a new classification. J Spinal Disord 1989;2:176–183.
109. Mangiadi J, Moser F, Spitzer D, et al. Spinal injuries. In: Schartz G, Cayten C, Mangelsen M, et al., eds. Principles and Practice of Emergency Medicine. 3rd ed. Philadelphia: Lea and Febiger, 1992:955–993.
110. Gertzbein SD, Court-Brown CM. Flexion-distraction injuries of the lumbar spine: mechanisms of injury and classification. Clin Orthop 1988;227:52–60.
111. LeGay D, Petrie D, Alexander D. Flexion-distraction injuries of the lumbar spine and associated abdominal trauma. J Trauma 1990;30:436–444.
112. Glassman SD, Johnson JR, Holt RT. Seatbelt injuries in children. J Trauma 1992;33:882–886.
113. Triantafyllou S, Gertzbein S. Flexion-distraction injuries of the thoracolumbar spine. A review. Orthopedics 1992;15:357–364.
114. Liu YJ, Chang MC, Wang ST, et al. Flexion-distraction injury of the thoracolumbar spine. Injury 2003;34:920–923.
115. Razak M, Mahmud MM, [fn3]Hyzan MY, et al. Short segment posterior instrumentation, reduction and fusion of unstable thoracolumbar burst fractures–a review of 26 cases. Med J Malaysia 2000;55:9–13.
116. Levine AM. Facet fractures and dislocations of the thoracolumbar spine. In: Levine AM, Eismont FJ, Garfin SR, et al, eds. Spine Trauma. Philadelphia: WB Saunders, 1998:415–427.
117. Hendrix RW, Melany M, Miller F, et al. Fracture of the spine in patients with ankylosis due to diffuse skeletal hyperostosis: clinical and imaging findings. Am J Roentgenol 1994;162:899–904.
118. Mikles MR, Stchur RP, Graziano GP. Posterior instrumentation for thoracolumbar fractures. J Am Acad Orthop Surg 2004;12:424–435.
119. Yu SW, Fang KF, Tseng IC, et al. Surgical outcomes of short-segment fixation for thoracolumbar fracture dislocation. Chang Gung Med J 2002;25:253–259.
120. Lynn G, Mukherjee DP, Kruse RN, et al. Mechanical stability of thoracolumbar pedicle screw fixation. The effect of crosslinks. Spine 1997;22:1568–1572.
121. Vaccaro AR, Rizzolo SJ, Allardyce TJ, et al. Placement of pedicle screws in the thoracic spine. Part I: Morphometric analysis of the thoracic vertebrae. J Bone Joint Surg (Am) 1995;77:1193–1199.
122. Vaccaro AR, Rizzolo SJ, Balderston RA, et al. Placement of pedicle screws in the thoracic spine. Part II: An anatomical and radiographic assessment. J Bone Joint Surg (Am) 1995;77:1200–1206.
123. Bedbrook GM. Treatment of thoracolumbar dislocation and fractures with paraplegia. Clin Orthop 1975;112:27–43.
124. Rosenthal RE, Lowery ER. Unstable fracture-dislocations of the thoracolumbar spine: results of surgical treatment. J Trauma 1980;20:485–490.
125. Devilee R, Sanders R, de Lange S. Treatment of fractures and dislocations of the thoracic and lumbar spine by fusion and Harrington instrumentation. Arch Orthop Trauma Surg 1995;114:100–102.
126. Kinoshita H, Nagata Y, Hirakawa H. Thoracolumbar fracture dislocation: a study of 30 patients. Paraplegia 1989;27:289–298.
127. Heary RF, Bono CM, Black M. Thoracic pedicle screws: postoperative computerized tomography scanning assessment. J Neurosurg Spine 2004;100:325–331.
128. O’Brien M, Lenke L, Mardjedtko S, et al. Pedicle morphology in thoracic adolescent idiopathic scoliosis. Spine 2000;178:2285–2293.
129. Croce MA, Bee TK, Pritchard E, et al. Does optimal timing for spine fracture fixation exist? Ann Surg 2001;233:851–858.
130. Bohlman HH, Bahniuk E, Raskulinecz G, et al. Mechanical factors affecting recovery from incomplete cervical spinal cord injury: a preliminary report. Johns Hopkins Med J 1979;145:115–125.
131. Delamarter RB, Sherman J, Carr JB. Pathophysiology of spinal cord injury. Recovery after immediate and delayed compression. J Bone Joint Surg (Am) 1995;77:1042–1049.
132. Chipman JG, Deuser WE, Beilman GJ. Early surgery for thoracolumbar spine injuries decreases complications. J Trauma 2004;56:52–57.
133. Silber JS, Vaccaro AR. Summary statement: the role and timing of decompression in acute spinal cord injury: evidence-based guidelines. Spine 2001;26:S110.
134. Roffi R, Waters R, Adkins R. Gunshot wounds to the spine associated with a perforated viscus. Spine 1989;14:808–811.
135. Kumar A, Wood G, Whittle A. Low-velocity gunshot injuries of the spine with abdominal viscus trauma. J Orthop Trauma 1998;12:514–517.
136. Waters RL, Adkins RH, Yakura J, et al. Profiles of spinal cord injury and recovery after gunshot injury. Clin Orthop 1991;267:14–21.
137. Kane T, Capen DA, Waters R, et al. Spinal cord injury from civilian gunshot wounds: the Rancho experience 1980–88. J Spinal Disord 1991;4:306–311.
138. Waters R, Adkins R. The effects of removal of bullet fragment retained in the spinal canal. A collaborative study by the National Spinal Cord Injury Model Systems. Spine 1991;16:934–939.
139. Prendergast MR, Saxe JM, Ledgerwood AM, et al. Massive steroids do not reduce the zone of injury after penetrating spinal cord injury. J Trauma 1994;37:576–580.
140. Heary RF, Vaccaro AR, Mesa JJ, et al. Steroids and gunshot wounds to the spine. Neurosurgery 1997;41:576–583.
141. Levy M, Gans W, Wijesinghe H, et al. Use of methylprednisolone as an adjunct in the management of patients with penetrating spinal cord injury: outcome analysis. Neurosurgery 1996;39:1141–1148.
142. Garfin SR, Yuan HA, Reiley MA. New techniques in spine: kyphoplasty and vertebroplasty for the treatment of painful osteoporotic compression fractures. Spine 2001;26:1511–1515.
143. Lieberman IH, Dudeney S, Reinhardt MK, et al. Initial outcome and efficacy of “kyphoplasty” in the treatment of painful osteoporotic vertebral compression fractures. Spine 2001;26:1631–1638.
144. Barr JD, Barr MS, Lernley TJ, et al. Percutaneous vertebroplasty for pain relief and spinal stabilization. Spine 2000;25:923–928.
145. Rechtine GR, Cahill D, Chrin AM. Treatment of thoracolumbar trauma: comparison of complications of operative versus nonoperative treatment. J Spinal Disord 1999;12:406–409.
146. Velmahos GC, Kern J, Chan LS, et al. Prevention of venous thromboembolism after injury: an evidence-based report–Part I: analysis of risk factors and evaluation of the role of vena caval filters. J Trauma 2000;49:132–138.
147. Velmahos GC, Kern J, Chan LS, et al. [fn3]Prevention of venous thromboembolism after injury: an evidence-based report–Part II: analysis of risk factors and evaluation of the role of vena caval filters. J Trauma 2000;49:140–144.
148. Gaebler C, Maier R, Kukla C, et al. Long-term results of pedicle stabilized thoracolumbar fractures in relation to the neurological deficit. Injury 1997;28:661–666.
149. McLain RF, Benson DR. Urgent surgical stabilization of spinal fractures in polytrauma patients. Spine 1999;24:1646–1654.
150. Minor ME, Morrissey NJ, Peress R, et al. Endovascular treatment of an iatrogenic thoracic aortic injury after spinal instrumentation: case report. J Vasc Surg 2004;39:893–896.
151. Edwards CC, Levine AM. Complications associated with posterior instrumentation in the treatment of thoracic and lumbar injuries. In: Garfin SR, ed. Complications of Spine Surgery. Baltimore: Williams and Wilkins, 1989:164–199.