Decision-making regarding operative versus non-operative treatment of patients with thoracolumbar burst fractures in the absence of neurological deficits is controversial, and evidence from trials is sparse. We present a systematic review and meta-analysis of randomised trials comparing operative treatment to non-operative treatment in the management of thoracolumbar burst fractures. With the assistance of a medical librarian, an electronic search of Medline Embase and Cochrane Central Register of Controlled trials was performed. Trials were included if they: were randomided, had radiologically confirmed thoracolumbar (T10-L3) burst fractures, had no neurological deficit, compared operative and non-operative management (regardless of modality used), and had participants aged 18 and over. We examined the following outcomes: pain, using a visual analogue scale (VAS), where 0=no pain and 100=worst pain; function, using the validated Roland Morris Disability Questionnaire (RMDQ); and Kyphosis (measured in degrees). Two randomised trials including 79 patients (41 operative vs. 38 non-operative) were identified. Both trials had similar quality, patient characteristics, outcome measures, rates of follow up, and times of follow up (mean=47 months). Individual patient data meta-analysis (a powerful method of meta-analysis) was performed, since data was made available by the authors. There were no between-group differences in sex, level of fracture, mechanism of injury, follow up rates or baseline pain, kyphosis and RMDQ scores, but there was a borderline difference in age (mean 44 years in operative group vs. 39 in non-operative group, p=0.046). At final follow up, there were no between group differences in VAS pain (25 in operative group vs. 22 non-operative, p=0.63), RMDQ scores (6.1 in operative group vs. 5.8 non-operative, p=0.85), or change in RMDQ scores from baseline (4.8 in operative group vs. 5.3 non-operative, p=0.70). But both kyphosis at final follow up (11 degrees vs. 16 degrees, p=0.009) and reduction in kyphosis from baseline (1.8 degrees vs. -3.3 degrees, p=0.003) were better in the operative group. Operative management of thoracolumbar burst fractures appears to improve kyphosis, but does not improve pain or function.
