What Is the Optimal Treatment for Thoracolumbar Burst Fractures?

Published on 11/03/2015 by admin

Last modified 11/03/2015

Print this page

This article have been viewed 997 times

Chapter 6 What Is the Optimal Treatment for Thoracolumbar Burst Fractures?

MARCEL F. DVORAK, MD, FRCSC, CHARLES G. FISHER, MD, MHSc, FRCSC

Despite injury prevention initiatives and safer automobile designs, the incidence of thoracolumbar high-energy trauma remains significant.¹ Burst fractures of the thoracolumbar spine account for approximately 45% of all thoracolumbar trauma cases, and half of these patients remain neurologically intact after injury.² The 1990s and 2000s have brought significant technologic advancements, specifically the widespread use of pedicle screw fixation in the thoracic spine,³^–⁷ the design of stiffer and more rigid instrumentation,^3,^{8, 9} the ability to reconstruct the anterior spinal column with expandable cages¹⁰ and biologics,¹¹ and less invasive spinal surgical approaches.¹² The treatment of thoracolumbar trauma, however, and specifically that of burst fractures, continues to be one of the most controversial areas in spine trauma care despite the high incidence of these injuries and extensive published research.

A well-formulated question provides the foundation for a good systematic review, and this review investigates the variety of treatment options available for thoracolumbar burst fractures and tries to produce guidelines as to which treatment is most effective in producing predictable and safe outcomes for various types of injuries.

OPTIONS

It is important to define the population discussed here. Patients with burst fractures between T10-L2 inclusive, with or without neurologic deficit, who are a minimum age of 16 years are included. The term burst fracture refers to a fracture of the vertebral body with fracture lines that extend into the posterior vertebral body wall and result in a separation or widening of the pedicles.¹³ Burst fractures may be associated with varying degrees of disruption to the posterior vertebral elements, specifically the facet and laminar complex and the posterior ligamentous complex.¹⁴

Essentially, two major decisions need to be made by the treating physician: First, and more fundamentally, should the patient be treated with or without surgery? Second, if surgery is to be selected, what approach and technique should be used? Nonoperative care may involve the use of a thoracolumbar sacral orthosis (TLSO), body cast, hyperextension brace, or no orthosis at all, whereas operative treatment may involve anterior surgery alone, posterior surgery alone, or a combination of both. The posterior surgical fixation options include hook or wire constructs,^7,^{15, 16} short-segment pedicle screw fixation at one level above and one below the fractured vertebra,^3,^{4, 6, 7, 17} and long-segment fixation, characteristically two or three segments of fixation above and below the fracture.^7,^{18, 19} When the anterior column of the spine is surgically reconstructed, the vertebral body and discs may be approached indirectly through transpedicular bone^3,^{7, 20} or cement augmentation,^21,²² or by an indirect posterolateral approach. A direct anterior approach facilitates vertebral body resection, decompression of the anterior spinal canal, anterior reconstruction of the vertebra, and also anterior fixation with either plates or screw-rod constructs.^23,²⁴ Prosthetic devices (fixed and expanding cages), as well as autograft and allograft, are the most commonly used anterior vertebral reconstruction options.^10,²³^–²⁸

EVIDENCE

Operative versus Nonoperative Treatment

Five Level II studies directly compare operative with nonoperative care for thoracolumbar burst fractures (Table 6-1).^2,²⁹^–³² All of these studies include thoracolumbar burst fractures with normal neurology. Some of the fractures included would be described as unstable with significant kyphosis and some degree of posterior ligamentous disruption, though most would be described as stable.

TABLE 6-1 Thoracolumbar Fractures: Articles Comparing Operative and Nonoperative Treatment

Wood and colleagues ³¹ recruited 53 patients, 27 of whom were randomized to the nonoperative treatment arm, which contained two forms of nonoperative treatment, either a postural reduction and cast or a hyperextension custom-molded jacket TLSO worn for 12 to 16 weeks. Twenty-six patients were treated by a variety of surgical techniques, either a posterior screw/hook construct and fusion spanning between two to five levels or an anterior vertebrectomy, rib strut graft, and instrumentation. Patients’ outcome evaluation included the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), modified Roland Morris Disability Scale (RMDS) score, Oswestry Questionnaire, visual analogue pain scale (VAS), and a radiographic evaluation. The study does have significant limitations, including alack of standardization, multiple treatment options, outcome measures reported at varying intervals, lack of a priori determination of a primary outcome, no power calculations, and multiple comparisons without statistical adjustment.

A statistically significant difference between operative and nonoperative treatment was observed favoring nonoperative treatment, for physical function (P = 0.002) and role physical (P = 0.003). Wood and colleagues ³¹ report an average Roland Disability score of 8.2 for the operative group and 3.9 for the nonoperative group (P = 0.02). These authors also used the Oswestry questionnaire, reporting an average score of 20.75 and 10.66 for the surgical and nonsurgical groups, respectively. For both the Roland and Oswestry instruments, a lower score signifies better function.

In contradistinction with Wood and colleagues’³¹ study, Siebenga and investigators³⁰ included a homogeneously defined cohort and carefully standardized treatment of both the operative and nonoperative groups. By randomizing 34 patients to brace treatment and posterior short-segment fixation, Siebenga and investigators³⁰ showed a significant difference in pain (72 vs. 87 mm; P = 0.033), Rolland Morris Disability scores (8.9 vs. 3.1; P = 0.030), and return to work (38% vs. 85%; P = 0.018), each in favor of operative treatment. The methodology and uniformity of treatment applied to each group make this a strong study, whereas the lack of an a priori power calculation and an a priori description of a primary outcome prevent it from attaining Level I evidence status.

Shen and coauthors ²⁹ attempted to randomize patients to receive short-segment posterior instrumentation and fusion or nonoperative care using a hyperextension brace. Because of recruitment difficulties, some patients were not randomized, and as such, this study should be considered a prospective cohort study. Outcomes were measured by an independent assessor at 1, 3, 6, 12, 18, and 24 months. Also using a VAS, at 2-year follow-up, Shen and coauthors ²⁹ note the VAS to be 1.5 and 1.8 for the nonoperative and operative cohorts, respectively. For the 3- and 6-month follow-up, Shen and coauthors²⁹ show improved pain in the surgically compared with the brace-treated patients. The authors note a better Greenough low back outcome score in the surgically treated group for up to 6 months, but this effect was not observed with longer follow-up.

The systematic review by Thomas and investigators ³² was performed before the study by Siebenga and colleagues,³⁰ and concludes that there was no evidence of superiority of operative over nonoperative treatment for neurologically intact thoracolum-bar burst fractures. The final of the five Level II studies is a prospective comparative study by Domenicucci and coworkers,² which has multiple methodologic flaws, and although it favors surgery for patients with increased radiographic deformity (kyphosis over 20 degrees), issues of power and bias are significant.

Two Level III studies compare operative and nonoperative care (see Table 6-1).^33,³⁴ Rechtine and colleagues’³⁴ article is thought provoking in that it brings to mind the issues of costs of care, as well as patient preference, and shows that even in severely injured individuals, satisfactory outcomes may be obtained with or without surgery, however, with different treatment approaches, resource implications, costs, and risk.

In addition, a number of Level IV studies show satisfactory outcomes with nonoperative treatment of a variety of these thoracolumbar fractures.^33,³⁵^–⁴⁰ An example of one of these is an article by Mumford and coauthors,⁴⁰ who reviewed 41 of 47 patients treated with a variable period of bed rest (range, 7–68 days) followed by a custom-made TLSO for an average of 12 weeks. Inclusion criteria included burst fractures between T11-L5, and the data for each individual patient were reported. For patients treated without surgery, Mumford and coauthors⁴⁰ found that 50% of patients had little to no pain at final follow-up, as measured on a Likert Scale.

In summary, strong evidence has been reported to support satisfactory outcomes with both operative and nonoperative treatment. Two Level II studies ^29,³⁰ suggest improved outcomes with operative treatment, one of which shows improved outcomes only at 3 and 6 months, and not sustained out to longer follow-up.²⁹ Wood and colleagues’³¹ study has significant enough methodologic defects to make its conclusions nebulous.

Choice of Operative Approach

Specifically looking at the decision to operate from an anterior approach alone, posterior alone, or combined anterior and posterior approaches, two Level II studies ^28,⁴¹ and six Level III studies^5,^7,^35,⁴²^–⁴⁴ address this as their primary question (Table 6-2).

TABLE 6-2 Thoracolumbar Fractures: Articles Comparing Operative Approaches (Anterior and Posterior)

Wood and colleagues,²⁸ in an article that appears to be a subset of a previously reported randomized trial,³¹ randomly assigned 43 patients to either anterior partial vertebrectomy and Kaneda or Isola instrumentation or posterior Isola rod-hook stabilization. Unfortunately, the use of posterior rod-hook constructs is likely inferior to the more commonly used pedicle screw-rod systems currently used; thus, the high rate (11/18) of implant-related complications in the posterior surgery group is not that surprising. The generally good clinical outcomes in the anterior surgery group and the low complication rate make this a potentially reasonable option in the neurologically intact thoracolumbar burst fracture that requires surgical treatment.

Esses and coauthors’⁴¹ study strongly favors short-segment posterior fixation over an anterior fixation system (Kostuik–Harrington device) that tended to fail as frequently as the posterior short-segment fixator and required a greater degree of complexity and risk for its insertion.

Two Level III systematic reviews ^5,⁷ compare various surgical approaches and techniques. Though suffering from a lack of high-quality studies in his systematic review, Dickman and coworkers⁵ confidently state that segmental pedicle screw fixation of thoracolumbar fractures has a higher fusion rate than do rod-screw constructs or anterior fixation devices. Verlaan and investigators⁷ in a meta-analysis of 132 articles divided treatment into 5 categories, including posterior long- and short-segment instrumentation, a mixture of short- and long-segment instrumentation, anterior alone, and combined anterior and posterior fixation. Verlaan and investigators⁷ conclude that none of the five techniques reliably maintains alignment, and there is no compelling evidence of the superiority of one of these techniques over another.

Been and Bouma,⁴²

Buy Membership for Orthopaedics Category to continue reading. Learn more here