Thoracolumbar Trauma

General Assessment

Spinal trauma patients, particularly those with high-energy–induced fracture-dislocations of the spine, often present with acute polytrauma and therefore require a thorough multispecialty assessment. Because these patients often present in shock, either hemorrhagic or neurogenic, they require extensive efforts to resuscitate and stabilize them, including intubation, intravenous (IV) fluids or blood products, chest tube placement, intra-abdominal assessment (for liver, spleen, pancreas, and bowel injury), and long bone splinting or external fixation. The trauma evaluation is critically important because more than half of patients with flexion-distraction injuries of the thoracolumbar spine have intra-abdominal injuries, including hollow viscus perforations. This number may be even higher in the pediatric population. An acute abdomen secondary to undetected intra-abdominal injury is associated with significant morbidity and mortality. It is recommended that a general assessment of the patient's overall mental functioning (Glasgow Coma Scale), polytrauma status (Injury Severity Score), neurologic status (American Spine Injury Association [ASIA] ), International Standards for Neurological Classification of Spinal Cord Injury, and comorbidities (Charlson Comorbidity Index ) be conducted. Although plain radiographs of the chest, abdomen, pelvis, spine, and long bones remain useful, most trauma centers now use more advanced imaging, such as computed tomography (CT) scanning and magnetic resonance imaging (MRI) of targeted areas. This is particularly true for obtunded patients, where a rapid CT scan of the chest, abdomen, pelvis, and total spine is required to rule out any unknown injuries.

Physical Examination

The physical examination in the conscious patient should include the following: evaluation of the entire spine from the occiput to the sacrum for any points of tenderness, painful gaps, or step-offs between the spinous processes that may be indicative of a ligamentous injury; subcutaneous bruises or contusions; subcutaneous air; and the general alignment of the spine. In flexion-distraction injuries, soft tissue injuries can be detected by noting hematoma, significant paraspinous tenderness, and palpation of gaps between spinous processes. In fracture-dislocations, a gibbous deformity and step-off between spinous processes are palpable.

The neurologic examination includes a complete assessment of motor strength, sensory examination, proprioception, and rectal tone. If the patient is in spinal shock, the anal wink and bulbocavernosus reflex are less relevant for the prognosis of the injury because most of these patients should have had emergent reduction and stabilization of the spine within 72 hours of the injury. ASIA standards for the neurologic examination are used to guide physicians through a complete neurologic examination.

Imaging in Thoracolumbar Fractures

Imaging investigations are mainstays in diagnosis, decision making based on classification, and treatment of thoracolumbar fractures. Patients are managed with spine precautions until the imaging investigations and spine clearance are completed. In the past, anteroposterior (AP) and lateral radiographs were used for clearance of the TL spine. However, the use of radiographic images for the diagnosis of TL injuries had several limitations. These radiographs were usually completed in an area away from the emergency room (ER). Obtaining optimal visualization of the fracture site can be challenging in the scenario of trauma. The presence of multiorgan trauma also poses potential difficulties in obtaining necessary radiographs. Plain radiographs only have a specificity of 58% and a positive predictive value of 64%. Due to these limitations, helical CT scans of the chest, abdomen, and pelvis obtained for the diagnosis of visceral injuries have become the investigative tool of choice for the diagnosis of spinal injuries. CT scans have a sensitivity for thoracic and lumbar spine fractures of 97% and 99%, respectively.

CT scanning not only offers higher sensitivity and specificity, but also the time required for obtaining the CT scans has been found to be significantly less than that for the appropriate plain radiographs. The use of higher-quality CT of the chest, abdomen, and pelvis in the polytrauma setting is recommended. Dedicated CT scans are required for better definition if surgery is planned. Critical assessment of the sagittal and coronal CT imaging sequences is necessary to diagnose the fracture pattern and mechanism of injury. Axial sequences are used to diagnose coronal or sagittal plane injuries to the vertebral bodies, assess the severity of canal encroachment, and determine the presence of laminar fractures.

MRI can be used in a stabilized patient when there is a neurologic compromise. Other indications for MRI are the presence of unexplained neurologic deficits or neurologic deterioration and to determine the status of the posterior ligamentous complex. Further, MRI can aid in determining the acuity of osteoporotic-related fractures. The modern classification systems discussed herein are based on CT and, when indicated, MRI.

Evolution of Thoracolumbar Classification Systems

In 1938, Watson-Jones described a thoracolumbar classification system based on three distinct fracture patterns: the wedge fracture, the comminuted fracture, and the fracture-dislocation. Others identified unique fracture types that have become eponyms of the authors and have persisted to this day. One example is the Chance fracture, a flexion-distraction injury characterized by a wedge compression of the vertebral body and tension failure of the posterior elements, which was described by Chance in 1948. In 1970, Holdsworth first used the term burst fracture to describe a vertebral body compression fracture with involvement of the vertebral wall that he identified in a review of 1000 spinal injuries.

In 1968, Kelly and Whitesides first proposed the column theory to describe spinal stability. They described the spine as consisting of two columns, an anterior and posterior column. The anterior column includes the vertebral body and intervertebral disc and both longitudinal ligaments. The posterior column represents the entire neural arch and ligamentous attachments. They hypothesized that an injury to either one of the columns would be stable, but an injury to both columns would result in an unstable injury. This classification was modified by Denis, who divided the spine into three columns: the anterior, middle, and posterior columns ( Table 33.1 ). In this classification, the anterior column is composed of the anterior longitudinal ligament and the anterior two-thirds of the vertebral body and disc. The middle column represents the posterior third of the vertebral body/disc and the posterior longitudinal ligament. Lastly, the posterior column is described as including all structures from the pedicles on posteriorly ( Fig. 33.1 ). From this three-column theory, there are four major injury types (compression fractures, burst fractures, seat-belt injuries, and fracture-dislocations). The major injury types are further subdivided into 16 distinct subtypes. In the Denis classification, it is postulated that the middle column is the critical element of spinal stability. Therefore injury to either the anterior or posterior column in isolation would result in a stable injury, whereas injury of either of these columns and the middle column would create an unstable injury. This system has shown moderate interrater reliability with regard to injury type; however, it has shown poor reliability with regard to injury subtype. Although attracting some criticism because it is not comprehensive, this classification still has gained widespread acceptance. Further, it is often cited in providing justification for the operative treatment of burst fractures even in neurologically intact patients. Subsequent classification systems, described later in this chapter, have challenged this management principle.

McAfee et al. reviewed 100 thoracolumbar fractures utilizing the recently developed CT scan. This classification scheme combines the fracture morphology with the injury mechanism and groups the injuries into six main types: wedge compression, stable burst, unstable burst, flexion-distraction, Chance fractures, and translational injuries. This classification system does not challenge the need for surgical intervention being based on the integrity of the middle column; rather, this classification attempts to help determine the type of surgery needed. If the middle column injury is secondary to either a compressive force, such as a burst fracture, or a distractive force, such as a flexion-distraction injury, the authors recommended a posterior-based approach with distraction or compression instrumentation. However, if the injury is a translational injury, posterior distraction is not recommended, and instead, segmental instrumentation of each level is recommended.

Ferguson and Allen published a more comprehensive mechanistic classification, built on Denis's three-column theory. In the Ferguson and Allen system, there are seven main injury types with an additional five subtypes. These main types include vertical compression, compression-flexion, distraction-flexion, lateral flexion, translation, torsional flexion, and distractive extension ( Box 33.1 ). Although the classification nomenclature has become widely accepted, the validity of the classification and its usefulness for guiding treatment are unclear; it has never been independently validated. This is likely due to its reliance on an inferred injury mechanism as opposed to objective and verifiable imaging characteristics.

In 1990, White and Panjabi proposed an important new concept of spinal stability that has altered the subsequent interpretation of spinal injury. They defined stability as the ability of the supporting elements of the spine to resist physiologic loads so as to prevent neurologic injury, deformity, or pain. This was an important step in refuting the notion that all fractures that involve the middle column are unstable. However, in spite of the work by White and Panjabi, the next major classification system proposed by Magerl and colleagues failed to challenge the assertion that all fractures involving the middle column are unstable. The Magerl system is a hierarchical system that first divides the injury into one of three major types based on mechanism and then subdivides each injury based on morphology. The major types, identified as type A, B, and C, refer to compression, distraction, or rotational injuries, respectively. However, although this classification increased the granularity of differing fracture morphologies, with 53 unique injuries described, the complexity has resulted in only poor to fair interobserver reliability. Furthermore, because of its reliance in the middle column to determine stability, it does not alter the treatment algorithm of thoracolumbar injuries. This, along with its failure to appreciate the importance of the neurologic status of the patient or the integrity of the posterior ligamentous complex, has limited the clinical utility of the Magerl classification.

The Thoracolumbar Injury Classification System

The Thoracolumbar Injury Classification System (TLICS) was developed by Vaccaro and colleagues in 2005 with the goal of serving as a compressive classification system that would guide clinical decision making. Classification of an injury with this system is first evaluated by three parameters: morphology, integrity of the posterior ligamentous complex (PLC), and neurologic status of the patient. The morphology is described as one of three major categories: compression injuries, translational/rotational injuries, and distraction injuries. The PLC is qualified as either intact, disrupted, or indeterminate. Lastly, the patient's neurologic status is described as intact, nerve root deficit, complete spinal cord injury, or incomplete spinal cord injury/cauda equina syndrome. An associated injury severity score can be applied to the TLICS to assign a score to each injury. That score (between 1 and 7) can be used to help guide clinical decision making in that a score of less than 4 would suggest nonoperative management, a score greater than 4 suggests operative treatment, and a score of 4 is an equivocal recommendation that makes no specific recommendation on treatment ( Table 33.2 ).

The TLICS is distinguished from its predecessors by several unique characteristics. First, with regard to its classification, it is the first system to include either the integrity of the PLC or the patient's neurologic status in its injury categorization. Second, implicit in its injury severity score and treatment recommendation is the idea that not all injuries to the middle column are unstable. Lastly, the TLICS represents the first thoracolumbar injury classification system that has been externally validated. The TLICS utilizes modern imaging such as multidetector CT and MRI, which historical systems did not have available. Furthermore, the TLICS has been validated in the pediatric population.

Most likely the most difficult fracture type to determine treatment for is the burst fracture. Despite the usefulness of the TLICS, controversy persists regarding the best treatment for burst fractures in neurologically intact patients. In this scenario, if the PLC is intact, then nonoperative treatment is recommended (injury score of 2). If the PLC is disrupted, then operative management is recommended (injury score of 5). However, if the integrity of the PLC is unknown, then no formal recommendation is given (injury score of 4). However, using the PLC as a key determinant of spinal stability is challenging because, despite the high-quality imaging that is available to most surgeons today, multiple studies have shown that there is poor reliability in assessing the integrity of the PLC. Schroeder et al. found poor reliability (κ = 0.11) when more than 500 spine surgeons from around the world evaluated 10 compression-type injuries for the integrity of the PLC. Similarly, Harrop et al. found only marginally better reliability (κ = 0.34) when 48 highly trained academic spine surgeons reviewed 56 injuries to the thoracolumbar spine.

Another critical flaw in the TLICS system is that there are dramatic differences across the globe in the treatment for a neurologically intact patient with a burst fracture. So although the TLICS makes a firm recommendation against operative fixation for neurologically intact patients with a burst fracture, surgeons in some parts of the world still consider the middle column the cornerstone to spinal stability and therefore recommend surgical intervention for all of these fractures. This has caused these regions of the world to reject the TLICS classification.

The Aospine Thoracolumbar Spine Injury Classification System

In an effort to create a globally accepted thoracolumbar classification system, Vaccaro and colleagues combined the Magerl classification and TLICS to create a new system in 2013: the AOSpine Thoracolumbar Spine Injury Classification System. The Magerl classification was modified to create three main patterns with a total of nine subtypes. The three main patterns are as follows: type A—compression injuries, type B—tension band injuries, and type C—translation injuries. Type A injuries are from compression forces and are subclassified into five subtypes: A0—a trivial injury that does not affect the stability of the spine, such as spinous or transverse process fracture; A1—a fracture of one endplate without posterior wall involvement; A2—a fracture involving the cephalad and caudal endplates without posterior wall involvement; A3—a fracture of one endplate with posterior wall involvement (incomplete burst); A4—a fracture involving the cephalad and caudal endplates with posterior wall involvement (complete burst). Type B injuries are distractive injuries with three subtypes: B1—an osseous injury to the tension band injury (a bony Chance fracture), B2—an injury that disrupts the posterior ligamentous tension band, and B3—a distractive injury of the anterior tension band. Type C includes injuries that result in intervertebral translation and may be along any major axis, including anteroposterior, axial, and vertical.

Kepler and colleagues conducted a reliability analysis with data from 100 surgeons around the world, and for the three main types, they found substantial (κ = 0.74) interobserver reliability and excellent (κ = 0.81) intraobserver reliability. When the subtypes were evaluated, the results were not as impressive, but they did find moderate interobserver (κ = 0.56) and intraobserver (κ = 0.43 to 0.57) reliability of the fracture subtypes. Furthermore, these results have since been validated by multiple unaffiliated authors, with Urrutia et al. reporting substantial interobserver reliability of the main types (κ = 0.62) and moderate reliability of the subtypes (κ = 0.55); similarly, Azimi and colleagues found excellent interobserver reliability of the three main types of injuries (κ = 0.83 to 0.89).

Schroeder and colleagues verified the hierarchical nature of the classification, and Kepler and colleagues used these results to develop the Thoracolumbar AOSpine Injury Score (TL AOSIS). The TL AOSIS is similar to the TLICS injury severity score such that it assigns a value to each injury pattern that can be used to guide clinical decision making ( Table 33.3 ).

Lastly, Vaccaro et al. reported on a worldwide survey of over 500 surgeons to determine the treatment algorithm for the AOSpine Classification. The algorithm itself was developed with input from over 500 surgeons from around the world, with the hope that with this worldwide input, the global community would implement the treatment algorithm. A vignette of all controversial thoracolumbar injury patterns was described to the participating surgeons, and they were asked to recommend surgical or nonsurgical care. A recommendation of nonoperative care was given to injuries in which less than 30% of the participating surgeons suggested initial surgical management. Similarly, a recommendation of surgical management was established when greater than 70% of surgeons recommended operative intervention for the injury. Using these results and the TL AOSIS, a surgical algorithm that is simple and easy to use was proposed. Patients with a total TL AOSIS of less than 4 should undergo a trial of nonoperative treatment, and early operative intervention is appropriate for patients with a TL AOSIS of more than 5. Treatment of patients with a TL AOSIS of 4 or 5 should be individualized based on surgeon and patient variables because operative or nonoperative care may be appropriate ( Table 33.4 ).

Physical Examination

The posterior spine is examined to note areas of abrasion/contusion and hematoma. The posterior midline spine can be palpated to assess for step-offs, gaps between spinous processes, or hematoma formation. Posterior midline edema or significant pain with palpation may increase the clinician's suspicion for PLC injury.

The neurologic examination, including perineal examination, is performed as outlined later in this chapter. Patients with thoracolumbar burst fractures may present with upper- or lower-motor-neuron signs due to the transition from cord to nerve roots depending on the level of injury. The severity of neurologic injury may be related to the initial displacement and amount of canal encroachment of the retropulsed fragments at the time of injury. Careful evaluation of perineal sensation and sphincter tone should be performed to rule out conus medullaris or cauda equina syndrome.

Imaging

Plain radiographs may have been obtained in some cases in patients with a low-energy mechanism but are less sensitive than CT. Characteristic radiographic findings of burst fractures include loss of posterior vertebral body height, narrowing of the spinal canal on lateral views, and increased interpedicular width on AP views.

The modern practice for trauma patients widely uses CT, which can easily be formatted to assess the thoracolumbar spine. On the sagittal CT scan, findings include a decrease in vertebral body height; a lack of smooth, defined cortices; retropulsion of the posterior cortex into the canal; and an increase in local kyphosis. Clinicians should have a high level of suspicion for instability if there is evidence of interspinous widening. If there is an indeterminate PLC injury or there is suspicion of worsening neurologic compromise, MRI should be obtained. Other indications for MRI include unexplained neurologic deficits, progressive neurologic deterioration, the timing of the acuity of fracture, and preoperative planning in some cases.

Nonoperative Treatment

It remains the standard to treat stable burst fractures without significant PLC disruption and absence of neurologic injury nonoperatively. Wood et al., in a multicenter, randomized clinical trial, studied 47 consecutive patients from 1992 to 1998 with stable, neurologically intact burst fractures, comparing nonoperative and operative treatments. Patients in the nonoperative group seemed to do better compared with the operative group based on radiographic criteria, Short Form 36 (SF-36), and Oswestry questionnaires at an average follow-up of 44 months. The operative group showed more frequent complications and greater costs. Surgical treatment was heterogeneous and up to the discretion of the treating surgeon. In a recent publication consisting of 16- to 22-year follow-up of the same cohort, these authors presented long-term follow-up results. Of the original randomized 47 patients, 78% were included, and again, patients treated nonoperatively showed better outcomes in regard to pain and function. A later randomized controlled trial (RCT) of 34 subjects by Siebenga et al. found that stable burst fractures without neurologic deficits treated with operative intervention had better functional outcomes and higher return to preinjury work than the nonoperative group. These authors recommended that treatment should take into consideration the amount of deformity and patient preference with respect to complication risks.

Surgical Indications

Surgical treatment of thoracolumbar burst fractures is indicated in those patients with evidence of neurologic compromise or demonstration of instability. The TLICS and TL AOSIS can aid in surgical decision making. Additionally, there is some emerging evidence indicating surgical management of stable burst fractures without neurologic injury to promote early mobilization, avoidance of orthoses, and prevention of deformity. There are also some advantages of operative management for those who cannot tolerate brace wear due to polytrauma, skin lesions, and inability to wear orthosis due to body habitus. If operative intervention is chosen, the approach and technique should consider the fracture morphology, the need to decompress the neural elements, and the need to stabilize the injured tension-band construct. The goals of operative intervention are to (1) adequately stabilize the fracture, (2) correct kyphotic deformity, (3) reverse neurologic compression, and (4) allow for early mobilization.

Anterior, posterior, or combined anterior and posterior approaches may be used for burst fracture management. Fractures may be stabilized with or without fusion. Posterior instrumentation is generally short-segment fixation with or without the addition of screw fixation at the level of the fracture. Decompression can occur either indirectly by restoring the sagittal alignment or directly through corpectomy/laminectomy. In addition, minimally invasive surgery (MIS) is increasingly being utilized, which may change surgical indications in the future.

Advantages of an anterior approach include direct visual decompression to remove fracture comminution, torn disc, or disrupted ligaments within the canal. Anterior visualization may be particularly helpful in the subacute or chronic fracture setting, where callus may inhibit fragment mobilization. The ability to restore sagittal alignment and provide anterior column reconstruction potentially prevents the development of the kyphosis often seen with posterior treatment and is an important advantage of the anterior approach. Fewer levels may need to be fused through this approach. Anterior approaches in general are more invasive, are associated with greater blood loss and surgical time, and have less secure fixation than posterior approaches.

Posterior spinal instrumentation and fusion is an established and effective treatment for thoracolumbar burst fractures. The posterior approach is relatively safer compared with the anterior approach due to the reduced risk of pulmonary, visceral, and vascular injury. In cases with incompetent PLC, posterior instrumentation stabilizes the posterior elements. Disadvantages include potentially inadequate indirect reduction, mobilization of the dura if anterior compression is needed, failure of instrumentation, or inadequate stabilization in the setting of severe anterior vertebral body comminution or poor bone quality. In posterior instrumentation and fusion, the most frequent complications are implant failure and donor-site pain. Combined anterior and posterior approaches offer the advantages of both and may be necessary in the setting of severe comminution and neural injury.

Short-segment pedicle fixation is one of the most standard approaches to treat thoracolumbar burst fractures. These constructs consist of fixation at the levels directly cranial and caudal to the fractured vertebra. Due to modern instrumentation systems, levels can be preserved from fusion with shorter fixation segments. However, posterior-only instrumentation may fail due to the lack of anterior structural and mechanical stability at the fracture segment. The addition of a screw at the fractured vertebra can augment the short-segment posterior fixation and can protect the anterior column by increasing the stiffness of the construct. In most patients with burst fractures, the pedicles remain viable for screw fixation ( Fig. 33.2 ). Kanna et al. reviewed their series of 32 patients with severe, unstable thoracolumbar injuries with posterior fixation including the fractured vertebra over 2 years. There were no instances of implant failure in their series and minimal progression of kyphosis. Long-term follow-up outcomes remain to be seen.

The goal to reduce the morbidity of the surgical approach by reducing infection risk, blood loss, operative time, and iatrogenic injury leading to functional deficits has led to the emergence of MIS in the treatment of thoracolumbar burst fractures. Posterior percutaneous pedicle screw fixation, external spine fixators, kyphoplasty reduction and cement augmentation with or without pedicle screw constructs, and anterior endoscopic decompression and stabilizations have been developed as new techniques. These technologies may provide the ability for surgeons to reduce perioperative morbidities. Currently, there remains limited evidence to support routine use.

Posterior Instrumentation With or Without Fusion

There remains controversy regarding whether to preserve motion after posterior instrumentation rather than fusing the segments. Fusionless fixation has demonstrated similar long-term outcomes to those of fused fixation. Advantages of nonfusion include decreased intraoperative blood loss and operative time, preservation of segmental motion, and avoidance of donor-site morbidities. Typically, the hardware is removed 9 to 12 months after the initial operation. Fusion provides a lower risk of implant failure and improvement in radiographic parameters. In an RCT of stable burst fractures with a 5- to 7-year follow-up, patients were treated with or without fusion. There were no clinical or radiologic differences, and the nonfusion group had less operative time and blood loss. In a prospective RCT comparing fusion and nonfusion care, there was no difference in functional status, neurologic status, or pain between the two groups. Lan et al. performed a meta-analysis of the available English literature comparing fusion to nonfusion in treating thoracolumbar burst fractures and demonstrated satisfactory clinical and radiographic results in the nonfusion groups. There was no difference in hardware failure between the two groups.

Neurologic injury severity is related to the volumetric amount and time of compression. A systematic review of the literature regarding the timing of neurologic decompression strongly recommended, with a moderate level of evidence, consideration of early surgical decompression (within 24 hours) for any patient with spinal cord injury. Additionally, due to the implementation of modern resuscitation and stabilization techniques, there appears to be a gradual improvement in the prognosis for neurologic recovery.

Anterior Thoracolumbar Approach: Corpectomy and Instrumentation

Positioning

The patient is placed in the right lateral decubitus position for a left-sided approach in burst fractures at T12 or below ( Fig. 33.3A ). This avoids obstruction by the liver or having to mobilize the vena cava. More cranial thoracic fractures can be approached from the right or left side due to the ability to mobilize the aorta. If a thoracotomy is used for lower thoracic injuries, the anesthesiologist should intubate with a double-lumen tube to allow for deflation of the right lung to assist with the approach (see Fig. 33.3B ). The abdomen is allowed to fall forward, aiding the approach. Patients should be placed laterally, perpendicular to the ground, to maintain orientation of the vertebral body during the approach, corpectomy, and instrumentation. An axillary roll should be placed underneath the right axilla to prevent a brachial plexopathy. Elbows and knees are padded with gel rolls and pillows to avoid injuries caused by peripheral nerve compression. The left hip is flexed to decrease tension on the psoas muscle. The head and legs on the operating room table are lowered to elevate the fractured vertebral body and aid in the exposure. Additionally, a wrinkle-less, square bump made from bedsheets can be placed underneath the fracture level to assist in flexing the patient's surgical level. Before final instrumentation, the table should be leveled so as not to induce coronal deformity.

Instrumentation and Fusion

Reconstruction can be performed using an autograft or allograft structural graft or a metallic or synthetic cage. A structural autograft can be a tricorticate iliac crest or the resected rib. A Morcellized resected vertebral body or resected rib can be used for a graft within a cage. The cartilage on the superior and inferior endplates should be removed, but it is essential not to remove the bony structure to prevent later subsidence of the graft or cage. Alignment should be evaluated with fluoroscopy and reduction maneuvers performed. This can be achieved with manipulation of the spine from distraction within the corpectomy with the use of a laminar spreader or expandable cage ( Fig. 33.4A and B ). Also, external corporeal pressure can also be applied to correct deformity. The instrumentation system can be used to restore lordosis.

Internal fixation can augment the structural graft or cage (see Fig. 33.4C–F ). Transvertebral screws and rods or plates are placed posterio-laterally on the vertebral body to avoid irritation of the vessels and viscera. Screws are aimed laterally to the contralateral side. The screws are placed parallel to the endplates. Care must be taken to ensure that the screws do not violate the spinal canal. Screws should be precise in length to prevent iatrogenic visceral injury.

Posterior Thoracolumbar Approach: Corpectomy and Instrumentation

The posterior approach to thoracolumbar fractures is discussed later in this chapter. Patient positioning, anesthesia considerations, and instrumentation are similar when treating burst fractures. Posterior decompression can be performed using laminectomy and direct manipulation of retropulsed fragments. This is indicated in the presence of neurologic deficits in the lower lumbar spine where anterior decompression is more difficult. The retropulsed bone fragments can be difficult to mobilize, and it is best to perform a discectomy to create space. The burst fragments can then be pushed forward using a down-angled curette. Removal of the pedicle can give greater exposure and ease of decompression. If needed, subtotal corpectomy can also be done through a transpedicular posterior approach. Dural tears should be anticipated, and primary closure, if possible, is the best treatment. When laminectomy is performed, then fusion is mandatory, and a fusionless technique should be avoided. Further laminectomy increases instability. In these instances, longer instrumentation constructs at least two levels above and below should be considered.

Posterior reduction maneuvers are dependent on the direction of fracture instability. The preponderance of burst fractures is a kyphosing event and necessitates a combination of lordosis and compression to correct the deformity. Inducing lordosis can be achieved through a cantilever rod-reduction maneuver or in situ rod-bending techniques. When reducing the fracture, the rod loses some lordosis, so it is recommended to over-contour the rod before reduction. The superior articular process may block reduction and can be removed to aid in reduction. The rod should be inserted into the proximal screws bilaterally, and sequential reduction should be performed with the use of a persuader. Compression between pedicle screws after rod reduction may induce additional sagittal correction if needed.

Nonoperative Treatment Versus Operative Treatment for Thoracolumbar Fractures Without Neurologic Deficit

AOSpine type A3 or A4 bursts without neurologic injury and/or PLC injury are common fracture variants in the thoracolumbar spine. Throughout the world, there remains a wide range of “standard” treatment, which is highly region dependent. Treatment ranges from early mobilization without bracing to combined anterior/posterior surgical intervention. Surgical stabilization with or without decompression may result in earlier mobilization, decreased length of stay, and earlier return to work. These patients are at risk for surgical complications and revision surgery. Nonoperative management with symptomatic pain control and early mobilization in stable fractures would avoid the inherent risks of surgical intervention. A recent Cochrane review of RCTs of thoracolumbar burst fracture management was unable to conclude if there was a significant improvement in pain or functional outcomes between nonoperative and operative care.

Wood et al., in an RCT of 47 patients treated with either nonoperative management or surgery from 1992 to 1998, presented favorable outcomes for nonoperative treatment. At an average of 44 months, there were no clinical or radiographic differences. It should be noted that many of the surgical techniques used in this study would unlikely be used today. These authors again reported on the same cohort 16 to 22 years after treatment randomization. The surgical group had significantly worse functional outcome scores compared with the nonsurgical group. The average kyphosis between the two groups was not significantly different.

A multicenter, prospective RCT from the Netherlands of AO type A fractures without neurologic deficits demonstrated superior return to work in the operative group compared with the nonoperative group, 6.7 months versus 13.6 months, respectively. The surgical group underwent short-segment fixation with pedicle screws and fixed-angle dedicated fracture-reduction implants. Additionally, a multicenter study from German and Austrian trauma centers reported favorable outcomes and lower complications than previously reported for surgically treated patients. A meta-analysis in 2012 reviewed four trials consisting of 79 patients and compared the two treatment methods. Their findings showed improved kyphosis with the surgically treated but did not show improved function 4 years after injury. Additionally, there were increased complication rates with surgery.

In a prospective RCT, the authors compared nonoperative care with posterior instrumentation and showed that surgical intervention provided early pain control and kyphosis correction. Neurologic deteriorations did not occur in either group, and both showed progressive kyphosis over time. The surgical group had statistically significant improved pain scores at 1 and 3 months postoperatively but no difference at 6 months and beyond. In this study, a large percentage of patients who underwent surgical intervention were able to return to heavy labor compared with nonoperative patients.

Anterior Versus Posterior Fusion Outcomes

Anterior versus posterior approaches have been shown to be equivocal for radiographic and patient-reported functional outcomes. A systematic review determined that there were no relative advantages to the posterior, anterior, or combined approach for neurologic recovery. These authors concluded that the combined anterior and posterior approach appeared to maintain kyphosis correction. A meta-analysis, which reviewed seven RCTs including 331 patients, compared the anterior and posterior approaches and found no significantly superior deformity correction, neurologic recovery, or return to work between the two groups. In this study, there were no differences in complication rate. However, the anterior approach appeared to have increased operative times, blood loss, and costs compared with posterior approaches.

Lin et al. compared the two techniques in patients with neurologic injury in a prospective, randomized study. Sixty-four patients underwent either an anterior-approach fusion or a posterior approach that consisted of a subtotal corpectomy, decompression, and reconstruction of the spine. At a minimum of 2 years, all patients in the study went on to achieve a successful fusion, and there was no significant difference in motor-examination improvement between the two groups, although the posterior-approach group experienced fewer pulmonary complications, decreased rates of blood loss, fewer complications, and a shorter operative time.

A prospective multicenter registry performed in Germany compared the anterior and posterior approach versus the posterior approach in 733 patients. These authors reported no difference in neurologic recovery between the two groups. Correction of deformity was greater in the combined group. The single posterior approach showed superior functional results, including shortened return to work and reduced pain.

Loss of sagittal correction after a posterior-only approach may be due to a failure to reconstruct and maintain the anterior column. If using posterior-only instrumentation, high cantilever bending loads are placed on the pedicle screws, and increased kyphosis may occur over time. Sasso et al. reviewed sagittal alignment in 53 patients who sustained unstable thoracolumbar burst fractures. Forty patients underwent an anterior approach with strut placement, and 13 patients underwent short-segment posterior-only instrumentation. Although both groups initially achieved statistically significant kyphosis correction, the posterior-approach group demonstrated increased kyphosis deformity.

Flexion Injuries

Flexion injuries result from forceful flexion of the vertebral column where the force vector goes through the anterior to the vertebral body or through the vertebral body. The flexion force is usually associated with some sort of axial loading. According to the AO Classification system discussed earlier, fracture types that result from flexion injuries are compression type (AO type A), including compression fractures and bursting fractures; distractive lesions (AO type B), such as Chance and flexion-distraction injuries; and translation injuries (AO type C), such as fracture-dislocations.

Flexion-Distraction Injuries

Flexion-distraction injuries are from hyperflexion involving one or two spinal segments. In these types of injuries, the distraction force is the predominant force over the compression component. In the more common flexion-distraction injury, the center of spinal rotation is within the vertebral body. The posterior elements, including the pedicle, posterior longitudinal ligament, and posterior disc annulus, fail in tension. The anterior vertebral body, which is ventral to the center of rotation, is compressed, usually resulting in a compression fracture. Alternatively, with greater force, a burst fracture can occur. Thus the flexion-distraction injury has a tensile failure of the posterior and middle columns and compressive failure of the anterior column. Flexion-distraction injuries typically result from a seat-belt injury. This injury pattern is prototypical of the mechanism and has facilitated the clarification of the actual mechanism of these injuries as being caused by the spine rotating around a fixed fulcrum anteriorly (e.g., a lap belt) and sustaining a hyperflexion rotational moment ( Fig. 33.5 ).

The Chance fracture is a unique pattern because it is a purely bony distractive injury. The posterior spinal processes, pedicles, and vertebral body fracture and displace as they fail under tension as the spine rotates around an anterior fulcrum, whereas the anterior longitudinal ligament often remains intact. The force of this injury may also result in numerous severe intra-abdominal soft tissue injuries. In other types of injuries, an anterior column injury may occur through the disc, and the vertebral body may remain intact, whereas a posterior column injury may manifest as facet dislocations or posterior ligamentous disruptions. These purely discoligamentous injuries may be associated with bony fractures. These types of injuries are known as Chance variants. The vertebrae may realign after the initial injury, and the PLC injury may be missed in these patients unless an MRI is obtained. CT scans aid in revealing facet fracture or subluxation ( Fig. 33.6 ).

Due to the violent forces involved, flexion-distraction injuries are usually associated with severe intra-abdominal injuries, including trauma to the major blood vessels or hollow viscera. Owing to the potentially fatal nature of these injuries, a person with flexion-distraction injuries should be critically evaluated to rule out these injuries, and appropriate general surgical consults should be obtained.

Lateral Shear Injuries

Lateral shear injuries are rare and generally very unstable. They are often associated with other fracture patterns, including flexion-rotation and lateral compression, and thus create very complex injuries. These injuries often involve a combination of facet fractures, have oblique fractures through the vertebral body or a bursting component, can displace in any direction, and have a high association with concurrent complete neurologic injury. They are often associated with severe polytrauma, including to the chest and abdomen.

Flexion-Rotation Injuries

Flexion-rotation injuries occur after rotational forces are applied simultaneously with a flexion moment to the thoracolumbar spine, resulting in the complete disruption of the PLC and extension through the anterior disc and vertebral body. These injuries are often associated with facet fractures, are highly unstable, and can progress to a complete disassociation of the involved spinal segments. They carry a high likelihood of a complete neurologic injury. The hallmark of these injuries is lateral or rotational displacement and fracturing of the transverse processes.

Fracture-Dislocation Injuries

Any of the aforementioned forces of flexion, extension, lateral shear, distraction, and/or rotation, when severe enough, may lead to translation of the cephalad vertebra over the caudal vertebra, with dislocation of one or both facets. Vertebral translations can occur in the sagittal or coronal planes. These injuries are considered highly unstable, and the potential for neurologic injury is elevated. These fractures can be diagnosed on plain radiographs based on the translation of the vertebra and facet dislocation/subluxation. Axial CT scan imaging can also show the presence of an “empty facet sign” in the presence of bifacetal dislocations. Fracture-dislocations in neurologically intact patients need to be investigated emergently with MRI to assess the discoligamentous complex, and immediate surgical treatment is warranted.

In addition to the nature and direction of the force applied to the spine during a violent traumatic event, the strength of the bone and elasticity of the soft tissue are critical in determining the severity of the injury. Additional modifiers to the type of resultant fracture include the age of the patient, the level of fracture, the degree of comminution, the presence of osteoporosis and osteomalacia, the development of avascular necrosis, and damage to supporting structures such as the ribs and sternum. Preexisting deformities, including spondylolisthesis, kyphosis, and scoliosis, can also impact the fracture pattern and stability ( Fig. 33.7 ).

Treatment Considerations

Clinical and radiologic evaluation of the patient allows surgeons to determine the nature of the thoracolumbar injury. Surgeons are able to assess the neurologic status, morphology of the fracture, mechanism of injury, potential for instability, and presence of other injuries. The mechanism of injury can differentiate between low-energy or fragility fractures versus high-velocity injuries caused by falls from a height or MVAs. It is equally important to consider the overall general condition of the patient with respect to other skeletal and visceral injuries and the feasibility of operative treatment if it is considered.

After the initial assessment, the decision making for nonoperative versus operative treatment depends mainly on the presence or absence of a neurologic injury and the integrity of the PLC. In flexion-distraction injuries, MRI may be useful when the status of the PLC is indeterminate. On fat-suppressed T2 MRI, an increased signal transverses through the supraspinous and intraspinous ligaments or disruption of facet joints, indicating disruption of the PLC. In the presence of a PLC injury or neurologic deficits, the preferred treatment of choice is surgical stabilization with or without decompression. In the absence of a neurologic injury and with an unequivocally intact PLC, the choice of treatment is mainly nonoperative.