Cochrane in CORR^®: Surgical Versus Non-surgical Treatment for Thoracolumbar Burst Fractures Without Neurological Deficit

Importance of the Topic

Thoracolumbar burst fractures account for approximately 45% of all major thoracolumbar traumatic injuries [9]. They most commonly occur secondary to an axial compression mechanism and are characterized by failure of the anterior and middle spinal columns. Clinical features of thoracolumbar burst fractures include acute back pain and possible damage to the nerve roots or spinal cord, but more than 50% of these injuries may present without neurological deficit [9]. Characteristic radiographic findings include anterior wedging of the vertebral body, increased interpedicular distance, and narrowing of the spinal canal due to retropulsed bone.

The management of thoracolumbar burst fractures in patients without neurological deficits remains controversial [3]. Surgical stabilization and possible decompression may result in earlier mobilization, reduced time to hospital discharge, and faster return to work [9], but it may also expose patients to more-frequent early complications, an increased risk for subsequent revision surgery, and greater overall healthcare costs [9]. Nonoperative management including symptomatic pain control, early mobilization, and perhaps a brace may be an acceptable alternative in properly selected patients [1]. This Cochrane systematic review compared surgical versus non-surgical treatment for thoracolumbar burst fractures in patients without neurological deficits.

Upon Closer Inspection

The authors performed a systematic review and meta-analysis of two randomized or quasi-randomized controlled trials (n = 87) comparing surgical with nonsurgical treatment [1]. Quasi-randomized trials use nonrandom methods of allocation such as alternation to assign participants to the comparison group. They raise the possibility of selection bias because treatment allocation can be predicted before potential patients are enrolled [5]. Functional outcomes were reported according to the Roland and Morris Disability Questionnaire (RDQ), a well-validated, patient-reported outcome measure for back pain [6]. However, the RDQ does not address psychological or social dysfunction and is most useful when the disability level is only low to moderate [6].

Substantial clinical heterogeneity and a general lack of randomized trials limited the ability of the authors to provide valid treatment recommendations. Siebenga et al. compared surgical treatment with short-segment posterior only stabilization followed by a Jewett hyperextension orthosis [8], whereas Wood et al. [13] used posterior or anterior fusion and instrumentation. This meta-analysis was not able to address other controversial management issues such as the level and classification of thoracolumbar fractures, the optimal timing of surgery in spinal trauma, surgical approach, or the role of bone graft in addition to fusion.

Take-home Messages

There are not enough studies in this population to answer the question of whether surgical or nonsurgical treatment is better. In another meta-analysis with two additional trials not included in the present study, Gnanenthiran et al. [3] found no differences in pain, function, or return to work rates, with higher complications and costs associated with surgery. One of the trials was excluded from the Cochrane review because an intention-to-treat analysis was not reported, and the other because it was a nonrandomized prospective cohort study [4, 7].

Inherent barriers to randomized controlled trials in surgery include challenges with patient enrollment, randomization, blinding, and strong patient preferences [2].Trial innovations in recent years, however, have made conducting randomized controlled trials in surgical patients more practical [2].The Spine Patient Outcomes Research Trial is one such example, in which a preference trial was used to investigate operative versus nonoperative treatment for various spinal conditions [10–12]. Patients were given a choice of either randomization or enrolment in an observational arm. Although high crossover rates compromised the intention-to-treat analysis, these studies are celebrated as the highest quality of evidence in the absence of Level 1 studies. Efforts focusing on innovative methodological techniques may improve our ability to overcome some of the challenges inherent in surgical randomized controlled trials.

Appendix