Flexion-Distraction Injuries of the Thoraco-Lumbar Spine in Children: A Comparison of Operative and Non-Operative Management

Objectives: To compare clinical outcomes (fusion and loss of reduction) of patients treated operatively and nonoperatively and to critically appraise the utility of current classification systems for flexion-distraction injuries.

Summary of Background Data: Seat belts have remarkably reduced the risk of severe injury and death in children involved in motor-vehicle crashes; however, they also appear to influence patterns of spine fractures in children. MRI now more readily shows disruption of posterior elements and some flexion-distraction injuries may have previously been miss-classified as a more stable two column injury. Current classification systems have a widely disparate approach and no one system is routinely used.

Methods: Spine fractures at all levels presenting over a 19 year period were identified from a hospital trauma database which identified 306 patients having at least one spine fracture. >From these 20 cases of thoraco-lumbar and lumbar region flexion-distraction fractures were identified. Fractures were classified according to the systems of Gumley, Getzbein and Court-Brown, and Rumball and Jarvis. Patients were grouped by primary treatment: nonoperative versus operative. Sagittal Cobb angles were measured on radiographs taken at the time of fracture, initiation of treatment, and at farthest follow-up. Union rate as well as changes in angular deformity was monitored over time. Chi square analysis and Fischer's exact test were used for statistical analysis.

Results: Twenty flexion-distraction type fractures were identified representing almost 7% of all spine fractures seen at any level. Patients included 11 males and 9 females with an average age of 13.9 (range 6 - 17). Average length of follow-up was 5.4 years. Eighty-five percent of cases involved motor-vehicle crashes with seventy percent wearing some form of seatbelt. Other modes of injury included skiing, a fall and a motorcycle crash (one each). Four (20%) patients had a neurological injury associated. All patients initially treated surgically healed (100.0% union) with 62% having posterior fusion only and with an average loss of lordosis of 7.50 over time. Five of seven patients (71.4%) in the non-operative group developed non-union with instability on flexion extension views and with average loss of 33.2° of reduction. These five non-unions all went on to successful surgical fusion. Flexion-distraction fractures were 29.9 times more likely to heal successfully if having initial surgery (Odds ratio = 29.9; 95% C.I. 2.1 to 410.9). This was supported by both Chi square analysis and (p<0 .001) and Fisher's exact test (p<0.007). Classification systems used were found to either not take into consideration anterior column compression or burst, else too comprehensive be useful.

Conclusion: Non-operative management of paediatric flexion-distraction fractures leads to an unacceptably high rate of non-union. We recommend primary surgical management of all paediatric thoraco-lumbar and lumbar region flexion-distraction injuries.